BAILII [Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]

England and Wales Court of Appeal (Civil Division) Decisions


You are here: BAILII >> Databases >> England and Wales Court of Appeal (Civil Division) Decisions >> Liberty Mutual Insurance Company (UK) Ltd. & Anor v HSBC Bank Plc [2002] EWCA Civ 691 (16th May, 2002)
URL: http://www.bailii.org/ew/cases/EWCA/Civ/2002/691.html
Cite as: [2002] EWCA Civ 691

[New search] [View without highlighting] [Printable RTF version] [Help]


Liberty Mutual Insurance Company (UK) Ltd. & Anor v HSBC Bank Plc [2002] EWCA Civ 691 (16th May, 2002)

Neutral Citation Number: [2002] EWCA Civ 691
Case No: A3/2001/0450

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM CHANCERY DIVISION
(THE VICE-CHANCELLOR & Mr Justice PATTEN)

Royal Courts of Justice
Strand,
London, WC2A 2LL
16 May 2002

B e f o r e :

LORD JUSTICE JUDGE
LORD JUSTICE RIX
and
LADY JUSTICE ARDEN

____________________

Between:
LIBERTY MUTUAL INSURANCE COMPANY (UK) LTD & Another

Claimant/
Respondent
and –


HSBC BANK plc
Defendant/Appellant

____________________

(Transcript of the Handed Down Judgment of
Smith Bernal Reporting Limited, 190 Fleet Street
London EC4A 2AG
Tel No: 020 7421 4040, Fax No: 020 7831 8838
Official Shorthand Writers to the Court)

____________________

Gabriel Moss QC and Felicity Toube (instructed by Messrs DLA) for the Claimant/Respondent
Jonathan Sumption QC and Mark Arnold (instructed by Messrs Allen & Overy) for the Defendant/Appellant

____________________

HTML VERSION OF JUDGMENT
AS APPROVED BY THE COURT
____________________

Crown Copyright ©

    Lord Justice Rix:

  1. This is an appeal by HSBC Bank plc (the “bank”) against three judgments concerned with points of construction arising out of surety bonds in essentially the same form issued to it by two insurance companies, Liberty Mutual Insurance Company (UK) Limited (“Liberty”) and St Paul International Insurance Company Limited (“St Paul”, formerly Seaboard Surety Co (“Seaboard”)). The three judgments are those of the Vice-Chancellor and Patten J (two judgments). The appeals arise in two sets of proceedings in one f which the bank is the claimant and in the other of which Liberty is the claimant, but there is no need to distinguish between them, nor, for the most part, between Liberty and St Paul. I shall therefore refer save as may be necessary to both surety companies by the name of “Liberty”.
  2. The bonds were given to secure liabilities which the bank (formerly Midland Bank plc) entered into at the request of its customer, Ocean Marine Mutual Insurance Association Limited (“OMMIA”), a P&I Club incorporated in the Turks and Caicos Islands. As part of its business activities, OMMIA was regularly called upon to provide security in the form of bank guarantees to claimants who had arrested, or threatened to arrest, its members’ vessels in various parts of the world in respect of maritime claims such as those arising out of the carriage of goods or collisions. Such guarantees therefore provided security in place of the arrested vessel. A traditional form of such security is a bail bond or so-called admiralty bond, which is a security provided to the court itself: but nowadays a bank guarantee provided directly to the claimant by an acceptable financial institution is the more normal procedure. In OMMIA’s case, it would utilise its facilities with the bank to procure from local correspondent banks the provision of the necessary guarantees. In a few cases the bank provided the guarantees itself directly to the beneficiaries; but in the majority of cases it procured their provision from correspondent banks to whom it in turn granted a counter-indemnity. The practice was for OMMIA to name a local bank, which the bank could approve or disapprove, and a local firm of agents or lawyers to negotiate the actual wording of the guarantee. Such guarantees were limited to a specific sum and were “conditional” in the sense that they were payable only upon a judgment, arbitral award or agreement in settlement of the beneficiary’s claim. The bank’s counter-indemnity, however, took the form of an undertaking to pay, up to the stated maximum, on the correspondent bank’s first demand subject only to a statement that the correspondent bank was required to pay in accordance with the terms of its guarantee. Because the security provided to the claimant and the counter-indemnity provided by the bank can both be referred to as guarantees, I shall use the expression “admiralty bond” to refer to the former, and “counter-indemnity” to refer to the latter. I shall set out typical examples of each below.
  3. Liberty’s surety bonds were provided to support and secure the bank’s engagements, whether in the form of admiralty bonds (of which there were only a handful or so) or of counter-indemnities (of which there were some 140). Since these appeals raise points of construction upon such surety bonds, I shall set out their language without delay, utilising one provided in the case of a vessel called the Lady Lela; but they were all in substantially the same form, thus:
  4. “Bank Guarantee Number 102/113865/96 M/V LADY LELA
    Surety Guarantee Number…049-000-139
    WHEREAS OCEAN MARINE MUTUAL PROTECTION & INDEMNITY ASSOCIATION LIMITED has requested MIDLAND BANK PLC (hereinafter called ‘Bank’) to execute or procure the execution of a bond, undertaking or guarantee (hereinafter called “Guarantee”) as bail or security in connection with their lawful business.
    NOW THEREFORE LIBERTY MUTUAL INSURANCE COMPANY (UK) LIMITED (hereinafter called ‘Surety’) undertakes in the amount of TND 85,000.00 (EIGHTY FIVE THOUSAND TUNISIAN DINARS) to indemnify the Bank and hold it harmless from and against all liability, losses and damages which the Bank…may sustain or incur by reason of having executed or procured the execution of such Guarantee provided always that the loss sustained by the Bank arises from a claim that has been paid by the Bank strictly in accordance with the express terms of the Guarantee and that the aggregate liability of the Surety under this Agreement shall not exceed the sum of TND 85,000.00 (EIGHTY FIVE THOUSAND TUNISIAN DINARS) which is the value of the Guarantee.
    This Agreement is separate from any other security or right of indemnity taken in respect of the Guarantee from any person, including your above mentioned customer.
    This Agreement shall be a continuing agreement and binding on the Surety, its successors and assigns.
    This Agreement shall be governed by and construed in accordance with the Laws of England.”

    The preliminary issues

  5. Disputes arose between Liberty and the bank regarding the construction of the bonds. The bank had taken a fixed charge over all OMMIA’s book debts as a continuing security for all its liabilities to the bank, and had subsequently had to appoint a receiver. Liberty’s case was that, to the extent that it had discharged by payment any of its surety bonds given to the bank, it was entitled to be subrogated to the bank’s security under its fixed charge pari passu with the bank’s claims on OMMIA. The bank’s case was that the third paragraph of the bond (“This Agreement is separate…”) was a consensual exclusion or postponement of Liberty’s rights of subrogation. That dispute was decided by the Vice-Chancellor in favour of Liberty. I shall call this the “subrogation issue”.
  6. The second dispute was whether the proviso at the end of the second paragraph of the bond (“provided always that the loss…”) which requires a payment to have been made “strictly in accordance with the express terms of the Guarantee” refers to the terms of the admiralty bond given to the claimant beneficiaries or to the engagement given by the bank, whether admiralty bond or, as in nearly every case, counter-indemnity. In other words did “Guarantee” in the surety bond mean “admiralty bond” or “the bank’s engagement”? In theory it ought to make no difference, because the bank’s counter-indemnities can only be called on by the correspondent banks to the extent that the latter have to face a liability under their admiralty bonds: in theory, therefore, the bank is only exposed under its first demand counter-indemnities in the event of negligence or fraud on the part of its correspondents, of which there is no evidence. In practice, however, it is suggested by the bank that it may face difficulties in proving the condition of the proviso, if “Guarantee” does indeed refer to the admiralty bond and not to its own engagement. I shall call this the “Guarantee issue”.
  7. The third dispute relates to a small number of admiralty bonds where OMMIA’s instructions to the bank asked it to procure an admiralty bond with a validity of a year, but where the admiralty bond which emerged, although valid in the first instance for only a year, was automatically renewable for successive periods of twelve months. Liberty claims to be entitled to reject any liability under its bonds given in respect of such requests, at any rate beyond the first year, on the ground that its engagement relates only to the admiralty bond requested and not to the admiralty bond given in excess of OMMIA’s authority. If the bank’s submission on the second issue were to be correct, then this third issue would not arise, for the only question would be whether the bank had suffered loss by reason of its counter-indemnity. I shall call this the “one year issue”.
  8. The Guarantee issue was decided by Patten J in his first judgment in favour of Liberty, and the one year issue was decided by Patten J in his second judgment, also in favour of Liberty. Thus the bank is the appellant in this court on all three issues.
  9. These three disputes came before the courts in the form of preliminary issues. In all, five preliminary issues were ordered, of which those numbered (iii) and (iv) below are ancillary to the Guarantee issue and in this court have given rise to no additional debate. The preliminary issues are:
  10. The subrogation issue

    (i) Whether on the proper construction of the Bonds…[Liberty and St Paul] have waived in favour of [the bank] such rights as they might otherwise have to be subrogated to the rights of [Liberty and St Paul] under the charge dated 17 November 1997 executed by [OMMIA].

    The Guarantee issue

    (ii) What is the “Guarantee” within the meaning of Liberty’s standard form bond?

    (iii) Where the Guarantee has been issued by [the bank], how does [it] prove that liability attaches to Liberty under Liberty’s bond?

    (iv) Where the Guarantee has been issued by [the bank’s] correspondent bank, how does [the bank] prove that liability attaches to Liberty under Liberty’s bond?

    The one year issue

    (v) In cases where OMMIA stipulated that [the bank] should give or procure a Guarantee for one year, does Liberty have any potential liability under its bond in respect of loss arising from a claim made in respect of the Guarantee after the end of that period if:

    (a) OMMIA further stipulated that the Guarantee should adopt wording to be supplied by a named party, and the Guarantee did so, but such wording was inconsistent with the terms of OMMIA’s (first) stipulation; or
    (b) the provisions of the applicable system(s) of law governing the Guarantee provided for the extension of such period?

  11. Ultimately, issue (v)(b) was not pursued before Patten J, but a further issue was, namely whether, if the one year issue was decided against the bank, the bank could nevertheless say that Liberty was estopped from asserting that the relevant period of liability is only one year. The bank failed as well in relation to that submission, and the appeal on that point has not been pursued.
  12. This court is therefore concerned only with preliminary issues (i), (ii), (iii), (iv) and (v)(a); and has heard no oral argument as to (iii) and (iv).
  13. The Vice-Chancellor’s order on the subrogation issue, reflecting the language of preliminary issue (i), merely stated that Liberty and St Paul had not “waived” in favour of the bank its rights to be subrogated to the bank’s rights under its fixed charge.
  14. Patten J’s first order, dealing with the Guarantee issue and preliminary issues (ii), (iii) and (iv) above, stated, adopting this judgment’s terminology, as follows:
  15. as to (ii): the “Guarantee” is either the bank’s Guarantee issued directly to the claimant (in cases where the Guarantee was executed by the bank) or the correspondent bank Guarantee issued directly to the claimant (in cases where the Guarantee was procured by [the bank]) (in each case, the “admiralty bond”);

    as to (iii): the bank must prove that it has made payment strictly in accordance with the terms of the Guarantee by producing a copy of the admiralty bond, evidence of payment and whatever additional document (whether settlement agreement, arbitration award, court order or other document) as is specified in the admiralty bond as the pre-condition for liability;

    as to (iv): the bank must prove that it has made payment strictly in accordance with the terms of the Guarantee by showing that the conditions for liability under the admiralty bond have been satisfied, by proving the judgment, arbitral award or agreement (as the case may be) which triggered liability under the admiralty bond;
  16. Patten J’s second order, dealing with the one year issue, stated:
  17. as to (v): where OMMIA stipulated that the bank should give or procure a Guarantee for one year, Liberty does not have any potential liability under its bond in respect of loss arising from a claim made in respect of the Guarantee after the end of that period if OMMIA further stipulated that the Guarantee should adopt wording to be supplied by a named third party, and the Guarantee did so, but such wording was inconsistent with the terms of OMMIA’s (first) stipulation.

    The background facts

  18. The Vice-Chancellor decided the subrogation issue essentially on the basis of the surety bond itself. The only other document in evidence was the fixed charge on book debts, annexed to a brief witness statement from the bank’s solicitor.
  19. For the purpose of the hearings before Patten J, however, there was further evidence, in the form of the bank’s facility letters, documents relating to the setting up of typical transactions, Liberty’s internal memorandum relating to the underwriting of its bonds, and a number of witness statements: from James Wade, the bank’s relationship manager responsible for its dealings with OMMIA; and from Graham Pitt, Liberty’s manager responsible for its relationship with OMMIA. Mr Wade and Mr Pitt were cross-examined at trial. It is agreed that for the purpose of these appeals, the evidence should be treated as generally available on all issues. Whereas such evidence is useful to throw light on the background and structure of the arrangements, it is not clear to me to what extent it assisted, or was admissible to assist, in the points of construction argued, although deployed at times by both sides in their submissions.
  20. The form of the surety bond was first developed by Seaboard. Nothing is known about the circumstances in which that bond was developed, save that its form was agreed by the bank to be acceptable to it. The earliest Seaboard bond goes back to December 1995. Seaboard/St Paul issued bonds to the bank from then until February 1997. In all there were only 21 such bonds and only 1 of these was issued after December 1996. The bank’s facility letter to OMMIA dated 4 November 1996 referred to such Seaboard bonds as being one form of acceptable security required by the bank for providing its engagement at OMMIA’s request. Such bonds, there referred to as an “insurance policy”, had to be “acceptable in form and substance to Midland in its sole discretion”. Mr Wade on behalf of the bank gave evidence to confirm that the Seaboard form was vetted by its internal lawyer. The Seaboard bonds appear to have been provided to OMMIA and assigned by OMMIA to the bank. Another acceptable form of security was a cash deposit in US dollars of not less than 25% of the face value of “each guarantee issued”. The bank also required an omnibus counter-indemnity from OMMIA “in respect of all past and future guarantees” given by the bank.
  21. Liberty took over from Seaboard/St Paul the provision of such bonds in the following circumstances. Late in 1996 Mr Pitt, Liberty’s senior underwriter based in London, began discussions with OMMIA which by July 1997 had matured into a proposal to his superiors in New York, set out in his memorandum dated 24 July 1997, for the provision to OMMIA of a facility of $25 million. The memorandum discussed the nature of a P&I Club’s business in arranging for admiralty bonds to be issued in support of its membership, and also the nature of such admiralty bonds themselves. Thus Patten J quoted most of the following extract:
  22. “Admiralty Bonds are basically “bail/court” bonds. When a vessel is involved in an accident, often the shipowner is required to post a bond in order to release his vessel.
    Admiralty bonds may be issued directly by the P&I clubs or by Guarantors (ie banks or surety companies)…
    The value of the bond is calculated on the maximum estimate of the value of the claim caused by the accident.
    The valuation is conducted by agents employed by OM and the claimant’s agents. A sum is agreed upon which is significantly higher than the settlement amount. Obviously, no liability is admitted by OM until a court decision has been made…
    Statistically, the average settlement of a claim is between 30-60% of the initial claimed sum…
    Claims can be settled via mutual consent, by arbitration or by a court award. On this basis all bonds are conditional – a sample bond issued to Midland is attached…
    The duration of the bond is a function of the length of determining the claim (bonds average 3 years however could possibly be as long as 7 years)…”

  23. The memorandum went on to speak of the “Administration of Admiralty Bonds” as follows:
  24. “In order for Midland to issue bonds, they require either a hold harmless from a surety company or 25% cash cover…
    ...it is perceived that we would be able to support a bank fronting on our behalf especially if the underlying obligation was similar to the obligation which a surety would issue.
    As these are conditional bonds, this would appear the most sensible route…
    Rating – the rate will be between 0.85-1% per annum…”

  25. The memorandum also annexed a form of counter-indemnity from OMMIA to Seaboard/St Paul, which it was proposed would similarly be provided to Liberty, if it went ahead.
  26. There is some confusion of thought apparent here, for instance the Seaboard bond enclosed was not an admiralty bond as that was elsewhere discussed, but a surety to support such an admiralty bond or even to support a counter-indemnity for such an admiralty bond. But the writer appears to have thought that because the ultimate, underlying obligation, reflected by the admiralty bond provided to the claimant, was conditional (upon agreement, judgment or award), therefore any supporting security would be similarly conditional and could be considered in pari materia. And in theory at any rate that was so.
  27. In due course Liberty agreed, by its letter dated 12 August 1997, to provide a “surety bond facility” to OMMIA. It was OMMIA which paid to Liberty the fee agreed of 1% per annum on the face value of the surety bonds provided. That was more advantageous to it than providing the bank with cash collateral of 25% of the value of the admiralty bond. The first surety bond provided by Liberty, slightly in advance of that letter, was dated 3 August 1997. Between then and 22 September 1998 Liberty provided some 142 bonds. The Liberty facility letter made it clear that Liberty was not, however, committed to the provision of its surety bonds: that would depend inter alia on underwriting criteria such as the financial strength of OMMIA itself.
  28. On 29 October 1997 the bank had issued a further facility letter to OMMIA. This continued its $43 million facility for its engagements, subject to an omnibus counter-indemnity from OMMIA and security in the form of either a Seaboard surety bond or a Liberty surety bond or a 25% cash deposit. In addition it required a fixed charge on book debts, but only in respect of a separate $12 million standby letter of credit facility. The fixed charge was executed on 17 November 1997.
  29. Two particular transactions were focused on below to demonstrate and illustrate how in an individual case Liberty’s surety bond was generated. These two transactions concerned the Lady Lela and the Skimmer, the latter being an illustration of a “one year” admiralty bond requested by OMMIA.
  30. The Lady Lela. The surety bond provided in this case related to an admiralty bond which had been procured by the bank at OMMIA’s request back in May 1996. The counter-indemnity given by the bank to its correspondent bank had presumably been secured in the meantime by a 25% cash deposit.
  31. On 1 May 1996 OMMIA issued instructions to the bank in respect of a “New Guarantee” re the Lady Lela. The instructions, which were in typical form (subject to the “one year” exception), began by referring to the incident which had given rise to the claim and to the amount of the guarantee and contained paragraphs detailing the “Beneficiary”, the “Issuing Bank” (ie the correspondent bank), the party “Wording Supplied by” and the “Expiry Date”, here “Until Further Notice”. It is clear from such instructions that the “New Guarantee” is a reference to an admiralty bond. There is no express or implied reference to the bank’s counter-indemnity.
  32. On the same day the bank telexed its correspondent bank as follows:
  33. “At request [OMMIA] and under our full responsibility please consult with [the party by whom the wording was to be supplied]…immediately and in accordance with their instructions and using the text they provide arrange for the issue and hold for collection by above of guarantee for [the amount] favour [the beneficiary] valid for period required relating m/v Lady Lela – alleged damage to cargo at Sousse on 8 April 1996.
    In consideration of your so doing we issue our unconditional counterindemnity no 102/11386596 your favour and undertake to pay to you on your first demand by mail or by tested telex/swift despite any contestation by our principals any sum or sums that you state are required to pay in accordance with terms of your guarantee.
    Our counterindemnity is valid until we receive your notice we released from all liability towards you hereunder.
    Our liability limited to amount not exceeding [the amount]…
    No mail confirmation follows. Please forward two translated copies your guarantee quoting our reference 102/11386596.”
  34. On 7 May 1996 the correspondent bank sent an (untranslated) copy of the Lady Lela admiralty bond. Even in French one can see in its final paragraph the reference to the need for either agreement, arbitration or judgment to dispose of the claim thus guaranteed. The amount of the admiralty bond was subsequently increased, again at the request of OMMIA, but nothing turns on that.
  35. In due course on 29 September 1997 OMMIA sent to Liberty a copy of the bank’s instructions to the correspondent bank, and on 30 September Liberty issued its surety bond (see para 3 above). The “Bank Guarantee Number” cited at the top of the surety bond was the bank’s reference number for its counter-indemnity. As to that, Patten J found as follows (para 17 of his first judgment):
  36. “…this remained a feature of all bonds issued by Liberty throughout the relevant period. Mr Pitt’s evidence was that the Bank Guarantee Number was simply a convenient method of identifying the particular bond by reference to the claim in question. Liberty could have used OMMIA’s reference number but chose instead to follow the practice used in the Seaboard bonds of quoting the Bank’s guarantee number. It seems to me that the practice of an identification number linking the bonds with the Bank’s guarantee number is an obviously convenient one and is consistent with the bond being given as additional security for the engagements facility. But beyond that it offers little real assistance in the question of construction which I have to resolve."

  37. The Skimmer. This example illustrates a number of differences from the case of the Lady Lela. Since it was dealing with a current casualty and claim, OMMIA sent contemporaneously to Liberty a copy of its instructions to the bank. It also sent to Liberty the bank’s instructions to its correspondent bank, as soon as it had received a copy of them from the bank. Secondly, OMMIA’s instructions were for an admiralty bond whose “Expiry Date” was given as “One Year”: that was included in the bank’s instructions to its correspondent bank, but even so the bank’s counter-indemnity was worded again to be “valid until we receive your notice we released…” Thirdly, the undertaking in the counter-indemnity was worded “to reimburse you on your first demand…any sum or sums that you are required to pay”, whereas in the case of the Lady Lela the undertaking was “to pay to you on your first demand…any sum or sums you state that you are required to pay”. Fourthly, Liberty provided its surety bond, as requested by OMMIA, dated the same date as the bank’s instructions to the correspondent bank, 19 November 1997, even though the admiralty bond was not executed until 23 November and not sent on by the bank to OMMIA until 15 December. In doing so, the bank referred to the “Guarantee” provided by its correspondent and to its own “counterindemnity”. Fifthly, the admiralty bond dealt with the question of its expiry date in the following fashion:
  38. “This guarantee shall be valid for a period of twelve months from the date of its issue and shall be automatically renewable for further consecutive periods of twelve months each and will automatically expire upon the fulfilment of our undertaking hereunder during the first or any period of twelve months.”

  39. That admiralty bond was rejected by neither the bank nor OMMIA as failing to comply with their respective instructions.
  40. It seems that, as before in the case of the Seaboard bonds, the Liberty bonds were sent by Liberty to OMMIA, who after all had requested and was paying for them, and then sent on by OMMIA to the bank.
  41. When in due course claims under the admiralty bonds were settled, and the correspondent banks began to make demands under the counter-indemnities, the bank did not look to Liberty for recoupment, but debited OMMIA under a running account with it. Indeed, Patten J found that this was the intention of the parties, stating (at para 10 of his first judgment) that –
  42. “The Liberty bond, unlike the Admiralty Bond and (when the Admiralty Bond was provided by a correspondent bank) the counter-indemnity given by Midland to the correspondent bank, was not intended by the parties to be part of the mechanics of payment. A demand under the Admiralty Bond would lead to payment by Midland either directly or to the correspondent bank. The sums involved would be debited to OMMIA’s overdraft facility and discharged as part of these continuing banking arrangements. No resort would ever be made to the Liberty bond unless OMMIA was unable to discharge its own indebtedness to Midland. The Liberty bond was therefore unlikely to be called upon except in the event of OMMIA’s insolvency.”

  43. Unfortunately, OMMIA did become insolvent. Before then, the bank had extended its fixed charge to secure all the facilities it provided to OMMIA, including the $43 million facility under which its relevant engagements occurred. This extension of the fixed charge occurred under the terms of the bank’s facility letter to OMMIA dated 22 September 1998. It was only from this time that the bank’s fixed charge security became relevant to any rights of subrogation which Liberty might acquire by reason of payment under its surety bonds. Then on 4 March 1999 OMMIA went into provisional liquidation and on 2 July 1999 the bank appointed receivers under its fixed charge.
  44. Thereafter the bank has looked to Liberty to recoup its losses under its engagements. For a period those claims were met. Latterly, however, Liberty took the points that the bank was not entitled to be paid because (a) it had not proved that its own payment was “strictly in accordance with the express terms” of the admiralty bonds, and (b) in the case of OMMIA’s “one year” instructions, its payment fell outside the terms of the admiralty bond requested by OMMIA. These points are illustrated by Liberty’s reactions to the bank’s claims under the Lady Lela and Skimmer surety bonds themselves, which remain unpaid.
  45. Thus in the case of the Lady Lela, on 9 June 2000 the bank’s correspondent bank made demand on the bank under its counter-indemnity for the maximum amount and was paid. On 22 June 2000 the bank wrote to Liberty enclosing a copy of its counter-indemnity and of the correspondent bank’s demand and confirming its own payment. It referred to its own counter-indemnity as its “Engagement” and glossed the words of the surety bond by stating that it had paid the demand made on it “strictly in accordance with the express terms of the Engagement”. Liberty refused to pay. In the case of the Skimmer, the correspondent bank made its demands in August/September 2000 and the bank paid and claimed on Liberty in similar terms. However, Liberty had already written by its solicitors to the bank on 10 May 2000 noting OMMIA’s “one year” instruction, stating that the admiralty bond had therefore expired and requesting the return of its own surety bond for cancellation.
  46. I can now turn to the three issues which have been argued on these appeals.
  47. The subrogation issue

  48. The critical paragraph is the third in the surety bonds, viz
  49. “This Agreement is separate from any other security or right of indemnity taken in respect of the Guarantee from any person, including your above mentioned customer.”

  50. On behalf of the bank, Mr Sumption QC has submitted that this paragraph, with its concept of separateness, is a derogation on Liberty’s rights of subrogation. The effect of the clause, he submits, is to postpone but not to exclude such rights of subrogation where, as in the case of the bank’s fixed charge, there has, since 22 September 1998, been a security available to the bank which covers both the bank’s claims to be indemnified by OMMIA against the bank’s losses arising out its engagements in respect of the admiralty bonds and the bank’s other claims against OMMIA. He calls such a security a “partially overlapping security”, and distinguishes it from a security which is either totally confined to claims themselves covered by the surety bonds (which he terms a “totally overlapping” security) or on the other hand exclusively distinct from or completely unconnected with such claims. In the latter case, the distinct security in any event is unavailable to and thus does not assist the subrogated surety. In the former case he accepts that the surety is indeed entitled to be subrogated to the security and that the clause in question does not affect such a case either. In the case, however, of a partially overlapping security the creditor’s concern is to use his security first to obtain a complete indemnity for his other claims against the debtor, ie for those for which he has no indemnity from the surety, and only thereafter to permit the subrogated surety to utilise his, the creditor’s, security. Only in such a way is his security kept “separate” from his rights under the surety bond. Only in such a way does a clause which appears designed to benefit the creditor give to him rights which are not already enshrined in what the law would otherwise accord him even in the absence of the clause.
  51. It is important to note that Mr Sumption argues only for the clause having the effect of a postponement of Liberty’s rights of subrogation. He does not say that it excludes such rights. This was an important clarification of the bank’s case. No one has suggested that such a clarification was not open to the bank, or that the bank was not to be entitled to make that argument. It is a pure matter of construction. It is nevertheless relevant to trace for a moment the history of the bank’s case in this respect. In its pleaded defence it alleged that the effect of the clause was that Liberty and St Paul had “waived such rights of subrogation as they might otherwise have had upon payment under the Bonds” (para 8(1)). Nothing was said of an alternative case of postponement. The concept of waiver was therefore picked up both in the language of the preliminary issue ordered (see para 8 above) and in the Vice-Chancellor’s declaration incorporated in his order following his judgment (see para 11 above). It appears from that judgment, however, that in the argument before the Vice-Chancellor concepts both of exclusion (waiver) and of postponement were argued, and that is to some extent indicated also by the terms of the bank’s skeleton argument below, which submits that the benefit intended by the clause “must have been the right to enjoy the security afforded by both the bond and the other securities until the relevant liabilities of [OMMIA] had been discharged in full”. It is common ground therefore that the concept of postponement was in play before the Vice-Chancellor. In its skeleton argument on appeal, moreover, both concepts, of total exclusion and of postponement, were expressly advanced. Indeed, it is there stated that whereas before the Vice-Chancellor the bank’s argument was that the clause constituted a postponement rather than a complete waiver, the bank proposed to address this court on the basis that the clause “constituted either a postponement or a complete waiver” (at para 16). The primary argument therefore was one of complete waiver, that any other security or the proceeds of its enforcement were not to be brought into account or shared at all; and the alternative argument was that of postponement. It was only on the second day of this appeal, during the submissions of Mr Moss QC on behalf of Liberty, that Mr Sumption made it plain that he was not suggesting that the clause effected a total exclusion, only a postponement (Day 2.55). This short history of the windings of the bank’s submissions is undertaken only to indicate that even on the bank’s side there has been some uncertainty as to the effect of the clause.
  52. On behalf of Liberty Mr Moss has, on the other hand, submitted that rights of subrogation are inherent in the justice of the situation between a surety and a principal creditor. It would therefore take clear words to remove or limit such rights by contract, even if, as he accepted, that was possible in theory. Nevertheless, no case had been cited and he knew of no case where a postponement had been achieved by language of this nature. The closest parallel to the present language were clauses which speak of a surety being “in addition to” other security, but authorities on such clauses were to the effect that they did not affect rights of subrogation. Banking clauses which were expressly concerned with postponement of rights of subrogation in the event of an overlapping security were well known and were a recognised part of the bank’s own armoury of clauses; but such clauses were only used to protect a creditor from an “insider” such as a director who guarantees the debts of his own company from utilising via the concept of subrogation the bank’s own security over his company’s assets. Absent such situations, it would be most unfair and uncommercial for a bank protected by a guarantee to postpone the surety’s rights of subrogation until the bank had obtained a complete indemnity in respect of all other debts owed to it, for in such a case not only would a bank’s right to be subrogated to the creditor’s security be postponed, so that the security was probably rendered worthless, but its unsecured rights of subrogation would also be overwhelmed. As for the meaning of the clause and its concept of separateness, it was not incumbent on Liberty to solve definitively the riddle of its meaning, but in any event there were many arguments and potential defences available to a surety against a principal creditor which the concept of the autonomy of the guarantee and its separateness from any other right of the principal creditor would greatly assist the principal creditor to eradicate: and it made no difference that the existing law would be in favour of the principal creditor in many of these situations even in the absence of such a clause.
  53. The Vice-Chancellor found in favour of Liberty’s submissions, as he said “not without hesitation” and he gave the bank permission to appeal. In his judgment he stressed four reasons in particular for his decision (at paras 15/16). First, he accepted that “clear words are required to exclude, reduce or postpone” the right of subrogation. Secondly, he regarded the clause as more apt to deal with the position before discharge of the surety bond than after and in consequence of its payment in full. Thirdly, he regarded postponement as contrary to commercial reality, as its practical effect was “to extend liability under the bond”. And fourthly, he regarded the words “in respect of the Guarantee” as an important qualification, to be understood in the sense of excluding (as it were by interpolation of the word “only”) a security to the extent that it was in any event separate because it was given in respect of a different liability.
  54. I have given anxious consideration to the competing submissions of the parties, but in the end I have concluded, even firmly concluded, that Liberty’s case is to be preferred, inter alia for the reasons given by the Vice-Chancellor. But before I seek to express in my own words my reasons for that decision, it is necessary to say something both about the nature of subrogation and about some leading modern authorities on construction which have been referred to.
  55. Subrogation. There was no dispute about the applicable principles of subrogation. It is a remedy, rather than a cause of action, founded in equity and natural justice not in contract, designed to prevent unjust enrichment. A useful statement is that to be found in GOFF & JONES, The Law of Restitution, 5th ed, 1998, at 133/4:
  56. “A surety who pays off the debt owed by the principal debtor is subrogated to any securities given by the debtor as security for the debt. The surety’s “right to have those securities transferred to him”, and his right to seek contribution from a co-surety, are said to be based on “a principle of natural justice.” In Craythorne v. Swinburne [(1807) 14 Ves 160], Sir Samuel Romilly described why equity intervened in the following words [at 162, as counsel, arguendo], which gained the approval of Lord Eldon [at 169]. He said:
    “…a surety will be entitled to every remedy, which the creditor has against the principal debtor; to enforce every security and all means of payment; to stand in the place of the creditor; not even through the medium of contract, but even by means of securities, entered into without the knowledge of the surety; having a right to have those securities transferred to him; though there was no stipulation for that; and to avail himself of all those securities against the debtor. This right of the surety also stands, not upon contract, but upon a principle of natural justice: the same principle, upon which one surety is entitled to contribution from another.”
    Subrogation is not then based on contract, for the surety “seldom if ever stipulated for the benefit of the security which the principal debtor has given” [Yonge v. Reynell (1852) 9 Hare 809, 818/819, per Turner V-C]. Its basis is natural justice; it is against conscience for the debtor to regain the securities from the creditor on the discharge of the debt by the surety, because it is the debtor’s obligation to indemnify the surety against any loss he incurs.”

  57. Thus, even if the securities were given after the contract of guarantee was made, and even if they are deemed to be satisfied by the payment of the debt (see Mercantile Law Amendment Act 1856, section 5), the surety is entitled to the benefit of them.
  58. An insurer as well as a surety is entitled to the rights of subrogation when once he has paid the assured. In Lord Napier and Ettrick v. Hunter [1993] AC 713 at 738B Lord Templeman described the principles of subrogation as follows:
  59. “The principles which dictated the decisions of our ancestors and inspired their references to the equitable obligations of an insured person towards an insurer entitled to subrogation are discernible and immutable. They establish that such an insurer has an enforceable equitable interest in the damages payable by the wrongdoer. The insured person is guilty of unconscionable conduct if he does not provide for the insurer to be recouped out of the damages awarded against the wrongdoer. Equity will not allow the insured person to insist on his legal rights…”

  60. Thus subrogation provides to the surety or the insurer an equitable or proprietary interest in the creditor’s rights against the debtor. As Millett LJ said in Boscawen v. Bajwa [1996] 1 WLR 328 at 335C/D (in a judgment approved by Lord Hoffmann in Banque Financiere de la Cité v. Parc (Battersea) Ltd [1999] AC 221 at 233F) –
  61. “It is available in a wide variety of different factual situations in which it is required to reverse the defendant’s unjust enrichment. Equity lawyers speak of a right of subrogation, or of an equity of subrogation, but this merely reflects the fact that it is not a remedy which the court has a general discretion to impose whenever it thinks it just to do so. The equity arises from the conduct of the parties on well settled principles and in defined circumstances which make it unconscionable for the defendant to deny the proprietary interest claimed by the plaintiff. A constructive trust arises in the same way. Once the equity is established the court satisfies it by declaring that the property in question is subject to a charge by way of subrogation in the one case or a constructive trust in the other.”

  62. It is also common ground that in English (and Commonwealth law) a surety is entitled to be subrogated pari passu to what Mr Sumption has called a partially overlapping security. The authority for that goes back at least to Goodwin v. Gray (1874) 22 WR 312. As Mr Richard Sykes QC, sitting as a deputy high court judge, said in Re Butlers Wharf Ltd [1995] 2 BCLC 43 at 50a/f:
  63. “…I now turn to the question mainly argued before me, which is this: is a surety for a part of a larger debt who has discharged the full amount for which he is surety entitled to share rateably with the principal debtor in any security which has been given for the whole debt?
    Three principles have been established beyond question:
    (1) The surety, in the situation described above, is entitled both to an indemnity claim against the principal debtor and to all securities which the principal debtor gave to the principal creditor in respect of and limited to the guaranteed part of the debt…
    (2) The surety is neither entitled to amounts paid voluntarily by the debtor in respect of the part of the debt not guaranteed by the surety nor to security given only for the part of the debt not so guaranteed…
    (3) The surety is entitled, upon the insolvency of the principal debtor, to share rateably with the principal creditor in amounts paid by way of dividend in respect of the whole debt…
    Each of these principles is of course subject to express terms of the guarantee to the contrary effect.
    All the textbooks which deal with the subject regard a fourth principle as established, namely that any security given by the principal debtor for the whole debt must be shared between the principal creditor (the remainder of whose debt remains outstanding, of course) and the surety; Goodwin v Gray (1874) 22 WR 312 is the authority most often cited for this principle.”

  64. At 51h Mr Sykes acknowledged that the law in the United States appears to be different. This is reflected in the US Restatement of the Law of Security, 1941, at 387 (“The surety in such cases is subrogated only after every duty protected by the security has been fully satisfied”) and in the US Restatement of the Law of Suretyship and Guarantee, 1996, at 115 (“Third, when collateral for the underlying obligation also secures other obligations of the principal obligor to the obligee, the secondary obligor is not entitled to subrogation to the obligee’s rights to the principal obligor’s collateral until all obligations secured by that collateral are satisfied or discharged”): but that merely indicates that another view about such matters is possible. It is common ground that the law in England, and as it would seem (see Re Butlers Wharf at 52a) in Ireland, Canada, Australia and New Zealand and probably other Commonwealth countries, is as found by Mr Sykes.
  65. Construction. It was also common ground that the normal incidents of rights of subrogation may be excluded or varied by contract. Mr Moss submits that clear words are needed to exclude or limit the rights otherwise assured by law. He cites in support of that submission Trafalgar House Construction (Regions) Ltd v. General Surety & Guarantee Co Ltd [1996] AC 199 at 208C per Lord Jauncey of Tullichettle (with whose speech their other Lordships agreed) where he said –
  66. “There is no doubt that in a contract of guarantee parties may, if so minded, exclude any one or more of the normal incidents of suretyship. However, if they choose to do so clear and unambiguous language must be used to displace the normal legal consequences of the contract…”
  67. Mr Sumption disputed the need for clear words: but in truth, the point is a general and well recognised one. Lord Jauncey was speaking in the specific context of contracts of guarantee, but the same point has been classically made by Lord Diplock in the context of rights of set-off in Gilbert Ash (Northern) Ltd v. Modern Engineering (Bristol) Ltd [1974] AC 689 at 717H.
  68. Mr Sumption also submitted that rights of subrogation may be excluded by implication, let alone express language clear and unequivocal or otherwise. I would allow that such implication is always possible, for where the implication is, as it must be, necessary, it is not to be denied. Mr Sumption’s sole example of such an implication is Morris v. Ford Motor Co Ltd [1973] QB 792. There a firm of cleaners contracted to Ford agreed to indemnify Ford against any loss or claim arising out of the performance of their contract. One of the cleaners’ employees was injured by the negligence of a Ford employee. The cleaners’ employee sued Ford, the cleaners indemnified Ford pursuant to their contract and sought to be subrogated to Ford’s common law right (see Lister v. Romford Ice and Cold Storage Co Ltd [1957] AC 555) to recover against its own employee. A majority of this court held that the cleaners’ claim failed, but varied in their reasoning. Lord Denning MR, having pointed out that the Lister decision had been set aside in the case of fellow-employees by a “gentleman’s agreement” reached by the members of the British Insurance Association, and that that agreement did not cover the instant case because it did not involve fellow-employees, adverted to Ford’s policy of not pursuing its rights against employees for fear of causing serious industrial repercussions. He went on to hold that equity did not compel Ford to lend its name to the cleaners to sue its employee, since “To make the servant personally liable would not only lead to a strike. It would be positively unjust” (at 801F). His secondary ground, if subrogation was to be looked at as a matter of implied contract rather than as founded on principles of equity (modern doctrine would approve Lord Denning’s preference for the principles of equity, see Lord Hoffmann in Banque Financiere at 234C), was to say that no implication requiring subrogation was to be made in the circumstances. Only James LJ (at 815C) put the matter in terms of an implied term whereby subrogation was excluded. He continued:
  69. “The implied term springs from the nature and terms of the contract between these parties. Their agreement was operative in an industrial setting in which subrogation of the third party to the rights and remedies of the defendants against their employees would be unacceptable and unrealistic.”

  70. Stamp LJ dissented, but it is clear from a passage at the end of his judgment at 807D/F that he might have had sympathy for Lord Denning’s primary ground had it not been for the fact that the point had not been covered in the pleadings, nor taken by the appellant, nor had the respondent any opportunity of dealing with it.
  71. In my judgment this is an unhelpful authority for Mr Sumption’s purposes. In truth it does not proceed so much by construing the contract as by reaching a conclusion based on the equity of the situation, where the alternative conclusion was described as unjust or unacceptable and unrealistic. It is also analogous to the case of the “insider”, where the contract will simply not work in any business sense if the right of subrogation survives. A clear example of such a case in the insurance setting is National Oilwell (UK) Ltd v. Davy Offshore Ltd [1993] 2 Lloyd’s Rep 582, where the “insider” was a co-assured (see headnote 13). In the present case, however, it is common ground that subject to contrary agreement equity requires Liberty, to the extent of any surety bond discharged by payment, to be entitled to share rateably in the bank’s fixed charge. There is nothing in the general context of the relations between the parties (subject of course to the issue of construction) to support the bank’s claim for preference in the exercise of its fixed charge over Liberty’s claim to share rateably in it, and the sole question is what the disputed clause means.
  72. Reference has been made to modern cases on construction such as Mannai Investment Co Ltd v. Eagle Star Life Assurance Co Ltd [1997] AC 749 especially per Lord Steyn at 964E/H, Investors Compensation Scheme Ltd v. West Bromwich Building Society [1998] 1 WLR 896 especially per Lord Hoffmann at 912H/913E, and Bank of Credit and Commerce International SA v. Ali [2001] 2 WLR 735 especially at para 8 per Lord Bingham of Cornhill. The principles are well known. Against the background of the admissible matrix of facts known to or at least reasonably available to the parties, the meaning sought is that which the language in question would convey to the reasonable man. In that context the language used is to be given its natural and ordinary meaning, unless the reasonable man would conclude that something has gone wrong in expressing the parties’ intentions.
  73. It was not suggested, however, that in the case of the present subrogation issue anything had gone wrong along the lines of the mistake in ICS v. West Bromwich. And, whereas both parties had submissions and claims to make about the commercial context and possible purposes of the clause in question, neither could properly say that the solution of the other was absurd or uncommercial, or in the well known and oft repeated words of Lord Diplock in Antaios Compania Naviera SA v. Salen Rederierna AB [1985] AC 191 at 201 “flouts business commonsense”. After all, on Liberty’s construction the general rule of subrogation would prevail, while on the bank’s construction a priority favourable to the bank but perfectly comprehensible in a business setting was proposed.
  74. It was suggested nevertheless that such modern authority is hostile to a maxim of construction requiring “clear words” in certain contexts if the prevailing rule is to be excluded or limited, such as Lord Jauncey’s statement in the context of contracts of guarantee and Lord Diplock’s statement in the context of set-off. I do not agree. It is true that modern authority (and I have in mind cases other than those referred to above) has moved away from technical or hostile attitudes to exclusion clauses. Even so, situations where “clear words” are required remain. A central example is the exclusion of negligence: see Ailsa Craig Fishing Co Ltd v. Malvern Fishing Co Ltd [1983] 1 WLR 964. This is because the reasonable man does not expect the parties to intend to exclude negligence unless the contract clearly says so. So also the reasonable man does not expect fundamental principles of law, equity and justice, such as rights of set-off or of subrogation to be excluded unless the contract clearly says so. A recent application of Lord Diplock’s requirement of clear words in relation to remedies for breach of contract can be found in Stocznia Gdanska SA v. Latvian Shipping Co [1998] 1 WLR 574 at 585C per Lord Goff. Even in BCCI v. Ali itself Lord Bingham expressed himself in similar terms in the context of the problem of construction relevant to that case (I do not say to this), when he said (at para 10):
  75. “But a long and in my view salutary line of authority shows that, in the absence of clear language, the court will be very slow to infer that a party intended to surrender rights and claims of which he was unaware and could not have been aware.”

  76. As the Vice-Chancellor said in this case (at para 15 of his judgment) –
  77. “The nature of the right and the circumstances in which it arises are such that an intention to exclude, reduce or postpone it is not lightly to be attributed to the parties.”
  78. I am now in a position to give my reasons for preferring the submissions of Liberty on the subrogation issue. I would seek to put the matter in the following way.
  79. (1) I agree with the Vice-Chancellor that clear words (or the necessity of supplying those words by implication) are required if rights of subrogation are to be removed.
  80. (2) I address myself to the language of the clause itself, and note first, that subrogation is not expressly mentioned, nor is the concept of postponement; and secondly, that the focus of the clause is the separateness of the surety bond (“This Agreement”) from “any other security or right of indemnity” taken “in respect of the Guarantee”. Three matters are emphasised in that language: one is that the clause chooses to underline the separateness of the surety bond, not the separateness of any other security or indemnity; the second is that, whether or not the word “only” is to be interpolated before the phrase “in respect of the Guarantee”, that phrase is hardly conducive to a meaning the focus of which, on the bank’s case, is not rights in respect of the Guarantee but rights which go beyond rights in respect of the Guarantee (I shall revert to this matter below); and the third is that the earlier phrase “any other security or right of indemnity” extends beyond a security (whatever that means) in the sense of a secured interest and into merely personal rights. Thus Mr Sumption specifically accepted that it embraced the right of the bank, for instance under its omnibus counter-indemnity, to be indemnified by OMMIA against any loss incurred by the bank by reason of carrying out OMMIA’s instructions.
  81. (3) Mr Sumption also accepts that the clause does not exclude Liberty’s rights of subrogation, but merely postpones them. It is necessary to consider the ramifications of this approach, not for the fixed charge which is understandably the focus of debate, but for just such a personal “right of indemnity” as the omnibus counter-indemnity owed by OMMIA. The suggestion is that such a right of indemnity is not excluded but postponed. This would mean that in a situation where the bank looked for an indemnity from Liberty under the surety bond before OMMIA’s insolvency, as under the bond it would of course be entitled to do, for instance in circumstances where the bank was unwilling to give OMMIA any further credit, Liberty would be postponed in its ability to exercise rights of subrogation against OMMIA until the bank had satisfied all its other claims against OMMIA under the omnibus counter-indemnity. The same would occur in any insolvency situation, where, even in the absence of a fixed charge (and it will be recalled the fixed charge only comes on the scene in any relevant sense, as a “partially overlapping” security, from 22 September 1998), Liberty’s right to share rateably in the bank’s dividend in the liquidation would be postponed until the bank had a complete indemnity, even though the bank’s unsecured claim would I suppose be increased by reason of Liberty’s subrogated claim. These are odd results. If they have been agreed, so be it; but it is not easy to see how they are to be found in the clause’s brief wording. It is true that in practice the parties did not contemplate that the surety bond would be part of the mechanics of payment. But there is nothing in the bond to prevent it being so, even though Mr Sumption was prepared on occasions to suggest that the reference to the need for the bank to show a “loss sustained by the Bank from a claim that has been paid” would have prevented an immediate claim under the surety bond before it had tried and failed to secure payment from OMMIA. In my judgment, however, that is not correct, as Mr Sumption recognised on other occasions when he accepted as almost axiomatically true that the bank was entitled, once it had paid up on any of its engagements by way of admiralty bond or counter-indemnity, to look to its rights under the surety bond even before seeking repayment by OMMIA. Its loss, after all, was crystallised by its payment, not by its failure to recover it from OMMIA.
  82. (4) Mr Sumption submits that the concept of postponement is inherent in the concept of the separateness of the surety bond, or, as he would prefer to present it, in the concept of the separateness of the other security. To my mind, however, the concept of postponement is a complicated one to derive from a concept of separation, for the very reason that it does not require complete separation. I can understand deriving exclusion of the right to subrogate from a concept of the separateness of some other security: in effect, the submission would be that the creditor is entitled to keep what is separate entirely for himself, and that would be inimical to subrogation, which involves a form of equitable transfer of the security to the subrogated surety. However, the concept of postponement does not insist on that complete separation: it allows the surety to share the security, but only after the creditor has satisfied himself completely in respect of other debts. Moreover, it is a necessary and inherent element of the concept of postponement that it divides the partially overlapping security into independent parts: that reflecting the guaranteed debt and that reflecting all other debts. Thus the creditor is first allowed to secure for himself a complete indemnity in respect of all other debts owed to him, and only then is required to permit the transfer to the surety of the right to secure the guaranteed debt. Any blurring of that division is said to be inconsistent with the doctrine of separation. All of that, however, as it seems to me, is a great deal to derive from the simple language of the clause, especially from a clause which says nothing about partially overlapping securities or rights (despite Mr Sumption’s submission that these are its sole subject-matter) but on the contrary speaks of other securities and rights “in respect of the Guarantee”: thus on one view dealing with securities and rights which are not concerned with other debts, or at the very least, if partially overlapping securities and rights do come within its reach, emphasising that part of them which relates to the guaranteed debt as distinct from other debts. And so, at one and the same time, the concept of postponement emphasises the separateness of other securities only in so far as they are concerned with other debts and not the guaranteed debt, and allows subrogation (albeit postponed subrogation) and thus the surety’s participation in the “separate” security in respect of the guaranteed debt. To my mind this is not a process of construction of a document, but a remaking of it. It is not so much seeking to find what can be derived from the language of a document, but how much can be poured into it.
  83. (5) I am confirmed in these thoughts by the following considerations deriving from the business setting of the surety bonds. The parties involved are financial institutions who are or at least must be regarded as being entirely familiar with the concepts which are the focus of debate. The bank was entitled to be wholly satisfied with the wording and substance of the bonds, and did in fact vet them in their legal department. In these circumstances if, as the bank’s thesis goes, it was concerned via the concept of postponement to mitigate the surety’s rights of subrogation in the case of partially overlapping securities, it seems surprising that it did not adopt for use a clause which it had to hand, such as that which can be found set out in the then contemporaneous edition of ROWLATT on The Law of Principal and Surety, 1982, 4th ed, at 247 (incorporating some appropriate modification of such language as “unless and until the whole of such principal moneys and interest shall have first been completely discharged and satisfied”). Mr Moss, who has considerable experience in this field, has told the court that the use of the postponement technique outside the field of the “insider” surety, such as the director of a company who guarantees to a bank his company’s borrowings from that bank but may also have made loans to and taken security from his company, is unknown to him; and that even in such a situation the postponement only operates in relation to the debt owed to the director and not to unrelated debts. Certainly, no case to gainsay that information has been presented to the court, and Mr Sumption, instructed as he is by the bank, has not disputed it. (Mr Sumption can point to a passing reference by Oliver LJ in an obiter part of his judgment in Barclays Bank Ltd v. TOSG Trust Fund Ltd [1984] AC 626 at 644B to it being normal in bank guarantees to exclude the surety’s right to prove in priority to the creditor, but it is uncertain to what that is a reference, unless to the problem of the “insider”.) Moreover, in March 1995, just before the bank was vetting the Seaboard bond which is the exemplar of all the bonds in question (it will be recalled that the earliest Seaboard bond is December 1995), the bank was itself a party to the judgment of Mr Sykes QC in Re Butlers Wharf, a case where the bank argued for a postponement solution, but failed in that argument, in respect of another of its clauses. The clause there in question said that “This security is in addition to…” any other security, which before the Vice-Chancellor the bank argued had the same meaning in substance as the instant clause’s language of “separate from” any other security. Mr Sykes actually pointed out (at 55i) that
  84. “Midland’s standard form of guarantee…also contained a clause which unequivocally postponed the surety’s right of proof and subrogation until payment of the whole amount of the debt…”

    In all these circumstances I see even less reason to think that the meaning conveyed to the reasonable man (here the reasonable banker or surety bond provider) by the clause before this court would be that contended for by the bank.

  85. (6) If therefore, as I think, the postponement of Liberty’s right of subrogation is very far from being the obvious focus of this clause, the question therefore arises as to what its meaning is. As to that, the Vice-Chancellor said this (at para 15 of his judgment):
  86. “I accept that it is not clear what the draftsman had in mind when inserting the provision in question. But I do not think that such uncertainty can be converted into the clear words which, in my view, are required to give rise to the exclusion or postponement for which the Bank contends.”
  87. In giving permission to appeal the Vice-Chancellor expressly excluded a ground whereby the bank wished to know what, in law, the ambit of the relevant provision is. He said “It is not the function of the court to give general guidance on matters which are unnecessary to the decision of the case.” On this appeal, however, there have not unnaturally been submissions as to what the clause might mean, if not what the bank says that it means. Mr Moss has been at pains to make a number of suggestions, and Mr Sumption to undermine them with arguments to the effect that they provide nothing more than the law otherwise provides, ie that such solutions would be mere surplusage, or that they are unrealistic and uncommercial compared with the hard realistic problem of the partially overlapping security posited by the bank. I respectfully agree with the Vice-Chancellor to this extent, that if it is unnecessary to the decision in this case to decide affirmatively what the clause means, then it is undesirable and indeed dangerous to attempt to do so: experience shows that it is against the background of a particular issue that points of construction are best dealt with. But I also agree that if the clause could be shown to be bereft of any sensible meaning if it did not mean what the bank contends it means, then that of course would be a considerable weapon in favour of its contention. In my view, therefore, the question best asked is whether the clause could sensibly have some meaning other than that contended for by the bank, but not whether it does definitively have some such other meaning.
  88. Mr Moss’s suggestions focused on the clause being intended to protect the bank, for he conceded that it was intended for its benefit, against arguments which, whether or not they were or might prove to be good arguments in law, were best knocked on the head by a simple clause. One example of this technique was that, in his submission, the clause emphasised that the rights afforded by the surety bond were separate from and thus could be exercised separately from any other security or right in respect of the guaranteed debt. Mr Sumption responded by pointing out that it was undisputed law that the bond could be used so, for the law did not require a guaranteed creditor to exhaust any other remedy before relying on his guarantee. In my judgment, however, there is force in Mr Moss’s suggestion, particularly in circumstances where the parties in fact contemplated that the bond would not be used as part of the mechanics of payment. Indeed, Mr Sumption had at times seemed to suggest that if OMMIA were in a position to pay, then the bank could not be said to have suffered a loss. Moreover, Mr Moss was able to point out that in the 1982 edition of ROWLATT, ie the edition contemporaneous with the vetting of the Seaboard bond, there is a discussion at 132/3 of what is there considered to be a valid argument that a surety has an equity to stay a creditor attempting unfairly to place the whole burden of a debt upon the surety at any rate in circumstances where the debtor is solvent or a security is available which could easily be realised to pay the debt. Therefore I do not consider Mr Moss’s suggestion to be fanciful or unrealistic.
  89. Mr Moss’s other suggestions as to why the separateness of the surety bond was emphasised were as follows: to prevent an argument that the surety bond is discharged by a debit to OMMIA’s running account with the bank, which was certainly a realistic problem on the facts; to deal with the problem where a surety bond is taken as part of a package of securities, a problem discussed in Byblos Bank SAL v. Al-Khudhairy [1987] BCLC 232; to deal with other security being given in relation to the same debt (ie in Mr Sumption’s terms, totally overlapping security) so as to emphasise the independence of each; to protect the bank against arguments that the surety is released by the release of the debtor, or by the debtor being given time to pay, or by a composition with the debtor (see ROWLATT at 162ff), or by the release of a co-surety (ibid at 184 and 188/190). In my judgment these are again realistic suggestions. There is no other provision in the surety bond dealing with problems of release, which understandably form a normal part of an informed creditor’s concerns. I am not convinced by Mr Sumption’s submissions that concerns such as these are all subsumed by underlying black letter law and/or unrealistic. I am satisfied that, whether or not judicial decision would prove that the clause thus understood would add nothing to the existing law, it is entirely realistic to desire such a clause in an instrument such as the bond. However ironic the observation may strike the bank in the light of the present litigation, it makes sense to try to make a bond which is looked to as the equivalent of cash as fool-proof as reasonably possible. So I agree with the Vice-Chancellor that the situation being considered by the clause can realistically be viewed as going to the value of the surety bond before it has been discharged by payment (“This Agreement is separate…”), rather than as restricting the rights of subrogation which would only arise after such discharge.
  90. I would therefore conclude that the clause has realistic potential value if given a meaning which stresses the independence of the surety bond from other securities or rights given in respect of the same debt. Indeed the advantage of such an interpretation is that it better accords with the language of the clause. It explains why the focus is on the separateness of the surety bond rather than on the separateness of the other security or right; and it accords better with the qualification “in respect of the Guarantee”. Whereas, for the reasons given above, I find that the concept of postponement fits ill with the language of the clause. In the Vice-Chancellor’s words (at paragraph 15) –
  91. “Had it been intended to exclude, reduce or postpone the right of subrogation which would otherwise arise a less oblique approach is to be expected.”
  92. (7) Finally, I would refer to the authorities considered by the Vice-Chancellor on the type of clause which refers to the surety bond being “in addition to” other securities and rights (see paras 8/13 of his judgment). The Vice-Chancellor did not give permission to appeal in respect of his conclusions about those authorities, viz Re Sass [1896] QB 12, Barclays Bank v. TOSG [1984] AC 626 and Re Butlers Wharf [1995] 2 BCLC 43. It suffices to say that I agree with what the Vice-Chancellor had to say about them.
  93. For all these reasons, I would conclude that on the subrogation issue the bank’s appeal fails. Mr Moss also had submissions to make as to the effect on the receivership should Liberty’s case on this issue prevail. Those submissions affect issues within this litigation which lie outside the preliminary issues which have been ordered, and therefore I have not thought it right to afford them consideration at this stage.
  94. The Guarantee issue

  95. The essential problem here is as to the meaning of “Guarantee” in the surety bond. Does that expression refer to the admiralty bond, as Liberty contends and Patten J held, or does it refer to the bank’s engagement, whether that is an admiralty bond or a counter-indemnity, as the bank contends?
  96. The word “Guarantee” occurs in all seven times in the bond. Its first two appearances are right at the top, prior to the text of the bond, where the word appears in connection with the reference numbers given to each individual bond, one for the bank and one for the surety. Where, as in most cases, the bank’s engagement took the form of a counter-indemnity, the bank’s reference therefore refers to that. Patten J found, however, that this was a mere matter of convenience and administration and did not affect the construction of the bond (see para 28 above). Mr Sumption relied on the first appearance (“Bank Guarantee Number”), but recognised that it did not take him very far. Its third appearance is in the first paragraph of the bond: “to execute or procure the execution of a bond, undertaking or guarantee (hereinafter called ‘Guarantee’) as bail or security...”. Since the word is here defined, it seems to me that this may well be its critical appearance, and certainly this passage of the bond was carefully addressed by Mr Sumption and Mr Moss. The word then appears three times in the second paragraph of the bond, first in the undertaking itself, obviously another important appearance (“by reason of having executed or procured the execution of such Guarantee”), next in the first condition of the proviso (“paid by the Bank strictly in accordance with the express terms of the Guarantee”) which is where its meaning matters for the purpose of the current issue, and thirdly in the second condition of the proviso dealing with aggregate liability (“which is the value of the Guarantee”). The parties made no submission in respect of that last appearance, no doubt because it will work equally in respect of admiralty bond or counter-indemnity. The final reference is in the third paragraph dealing with the separateness of the bond, in the phrase “any other security or right of indemnity taken in respect of the Guarantee”. The parties again made no submission in respect of the word’s appearance in that paragraph, until asked to do so by the court following the close of the oral hearing. The parties responded in writing.
  97. Mr Sumption submits that the critical appearance is the fifth, ie the first within the proviso, since the words “paid by the Bank strictly in accordance with the express terms of the Guarantee” can, he says, only refer to a payment under the bank’s engagement: it is only such a payment that can be said to be paid “by the Bank”. He therefore submits as to the first paragraph containing the word’s definition that the only question is whether the bank’s engagement, be it an admiralty bond or a counter-indemnity, falls within the words “to execute or procure the execution of a bond, undertaking or guarantee…as bail or security”. He answers that question by saying that it does, for even a counter-indemnity is an “undertaking or guarantee” executed “as…security”. He stresses the alternatives permitted within that language. Therefore any loss arising from a counter-indemnity is incurred “by reason of having executed…such Guarantee” within the meaning of the undertaking in the second paragraph. The seventh appearance is similarly a reference to the bank’s engagement, since the bank does not need to take any other security or right of indemnity in respect of an admiralty bond which it has not executed itself. Mr Sumption supported these submissions on the language of the document by reference to the background facts. He stressed that (save in one anomalous case) Liberty had never received the admiralty bonds, only the bank’s instructions to its correspondent banks, containing as they did the terms of the bank’s counter-indemnities. In that context it was simply mistaken to view the “Guarantee” as referring to the admiralty bond. Payment against a first demand bank undertaking was totally different from the nature of the liability undertaken under the admiralty bonds, and it would in practice be impossible for the bank to obtain evidence from its correspondents as to strict compliance with those bonds. Against that background, the reasonable man would conclude that “Guarantee” meant the bank’s engagement and nothing else.
  98. Mr Moss, however, submits that the critical appearance of the word “Guarantee” is the third, that within the opening paragraph of the bond, where the definition of the word is set out. The word “bail” points towards an admiralty bond, even if that is an incorrect description of a bond given other than to a court. Moreover, only an admiralty bond can be spoken of in terms of “execute or procure the execution of”, for the bank’s engagement, whether admiralty bond or counter-indemnity, is always executed by the bank. It never procures the execution of its own engagement. Therefore, unless the Guarantee is the admiralty bond, all reference to “or procure the execution of” is redundant. The same is true of the word’s fourth appearance, in the undertaking, where the language of the definition paragraph is tracked and thus similar language “having executed or procured the execution of such Guarantee” can be found. All other appearances are neutral. As for the fifth appearance (“paid by the Bank” etc), it is natural and acceptable to speak of the bank’s payment of its counter-indemnity as amounting to a payment in accordance with the admiralty bond and even payment of the ultimate claimant’s claim under the admiralty bond can similarly be spoken of as paid by or through the medium of the bank. As for the seventh appearance, the fixed charge is appropriately referred to as taken “in respect of” the admiralty bond, even where the bank has given its own counter-indemnity. As for the background facts, they did not assist. There was no evidence as to the circumstances in which the original Seaboard bond was vetted and settled. If, nevertheless, background facts were to be taken into account at all, the evidence was that Liberty regarded its liability under its bonds as reflecting the “conditional” nature of the admiralty bond, rather than dependent on whatever terms the bank may have agreed with its correspondents. There was no evidence before the court as to the difficulties of obtaining material from correspondent banks with which to demonstrate strict compliance with the admiralty bonds.
  99. Patten J in essence accepted Liberty’s submissions on this issue. He regarded the language of the first paragraph of the bond, tracked in the wording of the undertaking in the second paragraph, as the most important. In that connection he was impressed by the references in the first paragraph to “procure the execution of” and “bail”. As for the former, he said (paragraph 25 of his first judgment):
  100. “The Guarantee was to be given either as bail or security. The reference to bail seems to me to be a clear pointer towards the Admiralty Bond being the instrument described. I do not see how that description could ever properly be applied to one of the counter-indemnities. Similarly the use of the word “Guarantee” to describe the Admiralty Bond makes the word “procure” a real and accurate part of the description. On the Bank’s argument it is redundant.”

  101. As for the proviso, Patten J said this (in para 32):
  102. “As a secondary argument, initially encouraged by me, Mr Adkins [counsel for the bank] contended that the use of the word “paid” in the proviso was another pointer in favour of “Guarantee” meaning engagement. But this is I think only a shorthand method of describing a liability which has arisen either directly or indirectly as a result of the Bank having executed or procured the execution of the “Guarantee”. I do not consider that the word “paid” is sufficient in itself to dictate or alter the meaning of the word “Guarantee” derived from an examination of the earlier parts of the Liberty bond.”

  103. As for the background facts relied on by the bank, Patten J was dismissive. On the contrary, he was prepared to take account of the fact that a 1% premium was a limited reward for taking a risk for which the bank had required a 25% retention and reflected Liberty’s understanding that the risk was based on the conditional nature of the admiralty bond. He said:
  104. “31…I am not persuaded that the factors [the bank] relies upon are sufficient to justify me in departing from what I regard as the most likely and natural meaning of the language used in the bonds themselves. There has been no evidence from the Bank that it has or would have encountered any difficulties in obtaining copies of relevant awards or judgments from its various correspondents. Nor is there any evidence to show that the counter-indemnities could not have included some form of requirement to produce the necessary documents. Indeed the extract from the ICC Uniform Rules for Demand Guarantees referred to in the latest edition of Paget’s Law of Banking which was cited to me indicates that a demand guarantee (properly so called) may include a requirement for the provision of documents. Liberty’s apparent lack of concern to see the Admiralty Bonds that were issued is a curiosity but I accept Mr Pitt’s evidence that he was aware from the due diligence which he undertook that liability under the bonds was always conditional upon one or more of the following events, viz an agreement, a court judgment or an arbitral award; and that it was not considered necessary in the calculation of the risk to inspect each Admiralty Bond as it came to be issued…
    “33. It therefore comes to this: there is nothing in the transactional background or other admissible evidence which I believe justifies me in departing from the relatively clear language of the Liberty bond itself. This is not a case of obvious mistake or where the construction contended for by either party produces absurd results. On the contrary both parties have been able to produce what they say are sound commercial reasons for the interpretations which they respectively advance. But the court is not entitled in my judgment to speculate in these circumstances. The only safe and proper course is to be guided by the ordinary meaning of the language which the parties have themselves used."

  105. I agree with the substance of Patten J’s reasons, and it is only in places that I would place a different emphasis on the matters in issue. I can therefore be briefer than I might otherwise be in setting out my reasons in my own words. Thus (1) it is correct to begin with the language of the document. (2) The first paragraph of the surety bond used the word “Guarantee” for the purpose of defining it. Only an admiralty bond is capable of being either executed or procured by the bank. The question is not, as Mr Sumption presents it, whether the bank’s engagement can in every case be fitted within one or other of the words used, but what the words which define “Guarantee” point to. If they point to an admiralty bond, then a counter-indemnity will not be within the meaning of the word. (3) The reference to “bail” is another pointer in this connection. It is true that the word is a misnomer for a bond given otherwise than to the court: but it is, like the admiralty bond, an instrument which procures the release of a vessel from arrest against the provision of alternative security. The bank’s counter-indemnity is none of this. The words “as bail” are given a prominence which they do not deserve unless “Guarantee” refers to the admiralty bond. (4) The definition of the first paragraph is tracked in the language of the undertaking. (5) The words in the proviso “paid by the bank” etc are wholly appropriate where, of course, the surety bond is given to secure the bank’s loss; without payment by the bank the surety bond is not engaged. The measure of the bank’s loss, however, depends on what is payable to the claimant, and therefore recoupable from the bank, in accordance with the terms of the admiralty bond. Even if the bank’s counter-indemnity is payable on evidence or in response to a demand which is different from the evidence or claim which triggers payment under the admiralty bond, the intention of the whole transaction of course is that it is the liability under the admiralty bond which is the measure of everyone’s liability. As the bank’s counter-indemnity in the Lady Lela said: “any sum or sums which you state you are required to pay in accordance with terms of your guarantee”; and as the Skimmer counter-indemnity says: “any sum or sums that you are required to pay in accordance with terms your guarantee”.
  106. (6) While it is true that in theory the correspondent bank may be negligent or even fraudulent in the terms of their demand, there is no evidence that in practice such errors or deceit occur or that the bank is not well capable of looking after itself in the policing of such engagements. There is no evidence that the possibility of the counter-indemnities not working in the way in which all parties intended that they would work was within the realistic contemplation of these parties, or would be within the realistic contemplation of reasonable men placed in the parties’ business situation. If such evidence was to hand, the bank, which is one of the world’s major financial institutions, ought to be in the best possible position of presenting it to the court. In these circumstances it seems to me to be idle to insist on a construction which depends on theoretical evidential differences between liability under an admiralty bond and under a counter-indemnity for that liability.
  107. (7) If the parties had really intended to make the bank’s payment in accordance with the strict terms of their own engagements the essence of the surety bond, that document could easily have been worded accordingly. Either the definition of “Guarantee” could have been changed, or the simple exchange of “its engagement” for the fifth (and possibly sixth) appearance of the word “Guarantee” would have sufficed. (8) The references at the top of the document do not really assist. They occur before the word “Guarantee” is defined. In any event, the second appearance of “Guarantee” in the reference to “Surety Guarantee Number” shows that the word “Guarantee” is here being used in a sense different from the body of the text: the surety’s reference number is a reference to the surety bond not to the bank’s engagement. (9) The words “in respect of the Guarantee” in the third paragraph do not assist because the other security or right of indemnity there referred to is designed to cover the bank’s loss “in respect of” what is regarded as the claimant’s valid claim on his admiralty bond passed down to the bank through the correspondent bank.
  108. (10) References to the background facts are in my judgment on the whole unhelpful. Nothing is known of the background to the original settling of the Seaboard bond. There is no evidence that any of the parties, let alone all the parties, contemplated or even ought reasonably to have contemplated that there was a real danger that the bank’s liability under its counter-indemnities would be greater than a liability under the admiralty bonds themselves. If, nevertheless, that danger was regarded as a real one, it would give point to the words “strictly in accordance with the Guarantee” and especially to the word “strictly”. Although the terms of Liberty’s internal memorandum of July 1997, with its emphasis on the conditional nature of admiralty bonds, is suggestive, I doubt that it is admissible against the bank for the purpose of an objective evaluation of the surety bond’s matrix. All that can be said is that it is reasonable to assume that all the parties concerned were well aware of the so-called conditional nature of admiralty bonds. No inference can be drawn from the level of the 1% rate set by Liberty. Whether or not that is a modest rate, and would be known by all parties to be so, it is impossible to say whether or not the rate takes account of a real risk that the bank’s liabilities would prove greater than liabilities under the admiralty bonds. Seeing that the admiralty bonds were not sent to Liberty even in those cases where they were executed by the bank itself, it is difficult to read very much into the fact that the bank’s instructions to its correspondent banks were sent. What perhaps is as suggestive as anything, but I would not seek to read any more into it than I am inclined to do for the background facts in general, is that OMMIA’s instructions to the bank, which it sent in copy to Liberty when requesting the bank to procure the execution of a surety bond, make request of a “Bank Guarantee” specifying a correspondent bank as the issuing bank, but say nothing at all regarding the provision of a counter-indemnity. (11) Both constructions make commercial sense. Neither is absurd. Nor is it a case where it is clear that some obvious mistake has been made. Indeed, I do not think it has even been suggested that that is a possibility.
  109. In these circumstances it is my judgment that the meaning which a reasonable man would understand as intended by the word “Guarantee” in the proviso is a reference to the admiralty bond, not the bank’s counter-indemnity.
  110. The question then arises as to the evidence which the bank must produce to prove that liability attaches to Liberty under its bond. These matters have not been separately debated at the hearing before this court, and therefore, given my view on the Guarantee issue, must in my judgment lie as stated in Patten J’s order. I would merely for myself sound the note of caution that it seems to me that ultimately it is up to a trial judge to find whether the evidence put before him is or is not sufficient to satisfy him on the balance of probabilities that the bank has engaged the liability of Liberty under any of its bonds. Where a contract stipulates the evidence that must be provided, then that becomes the condition precedent for liability. Where, however, a contract does not stipulate the evidence that must be provided, it is a matter for the court to determine what evidence is sufficient to satisfy it that any factual condition precedent has been fulfilled in order to engage liability.
  111. The one year issue

  112. This is a nicely poised issue, whose solution has been made more difficult by the form in which the preliminary issue has come forward.
  113. The preliminary issue (see para 8 above) assumed in its very wording that the relevant admiralty bonds delivered by the correspondent banks providing for an initial validity of one year automatically renewable for successive periods of one year were in a form inconsistent with the instructions given by OMMIA to the bank to procure an admiralty bond for a period of one year. The issue also assumed in its wording that this inconsistency remained despite OMMIA’s further instruction to the bank to procure an admiralty bond whose wording was to be supplied by a named third party. This is because the issue raised the question whether Liberty had any liability under its surety bond in such a case after the end of the initial year –
  114. “if…OMMIA further stipulated that the [admiralty bond] should adopt wording to be supplied by a named third party, and the [admiralty bond] did so, but such wording was inconsistent with the terms of OMMIA’s (first) stipulation? [emphasis added]

  115. On that narrow issue the question ultimately resolved itself, in the judgment of Patten J and again on the submissions of counsel in this court, to whether Liberty’s liability under its surety bonds was premised on the terms of the admiralty bond requested by OMMIA or on the admiralty bond given? To that question, Patten J returned the answer, that Liberty’s liability depended on the admiralty bond requested. He said (at para 8 of his second judgment):
  116. “The liability which the Liberty Bond secures is that of the Bank arising from it having procured the execution of “such Guarantee”. That must be a reference to the Guarantee which in the case of the Skimmer or any other particular claim OMMIA had requested the Bank to procure. It is what the recital to the Liberty Bond says. The fact that it refers to the Guarantee in descriptive terms which involve the use of the indefinite rather than the definite article is to my mind completely irrelevant. When OMMIA requests the Bank to provide a Guarantee that then becomes the Guarantee for purposes of the Liberty Bond. Therefore in cases such as the Skimmer where the Guarantee was intended to endure for only a year liability under the Liberty Bond does not without more extend beyond that period.”

  117. In that final sentence the judge’s words “where the Guarantee was intended to endure for only a year” reflected the concession built into the preliminary issue that the admiralty bond given was inconsistent with the admiralty bond requested. I say “concession”, but I do not know why or in what circumstances the preliminary issue was drafted in the way it was. For all I know it reflects the bank’s concern to allow it room to argue that, because it faithfully translated OMMIA’s request for a one year admiralty bond in its own instructions to the correspondent banks, therefore it was itself entitled to argue as against those banks that it had no liability after the first year had passed. I merely note that in the case of the Skimmer itself, it did not take that point: and on the history of the matter (see para 35 above), it appears to have chosen not to take the point even after it was exposed to it by Liberty itself. But of course that may have happened without insight.
  118. On that narrow issue the submissions in this court have tracked the argument recorded in the judgment below. Mr Sumption for the bank has emphasised that what the bank sought in the surety bond was security for any loss which it might suffer by carrying out OMMIA’s request. If that request had been faithfully translated into its instructions to the correspondent bank, it was not its fault if the correspondent bank had erred. It remained exposed to loss on its counter-indemnity. It was for that loss that Liberty secured it. The difficulty with that argument, however, apart from the essential issue of whether the surety bond responds to the admiralty bond requested or the admiralty bond given, is that it begs the question of whether the bank is itself truly liable on a counter-indemnity given in respect of an admiralty bond given in terms beyond the bank’s own instructions.
  119. There was some discussion about the effect of an admiralty bond drawn up in terms in other respects outside those requested by OMMIA, for instance an admiralty bond which names the wrong ship, or which provides for a maximum amount greater than that stipulated but where the ultimate claim is within the smaller maximum stipulated. I did not find these speculations helpful. The example of the wrong ship would presumably be solved according to the limits of the doctrine whereby an erroneous description does no harm (falso demonstatio non nocet). The example of the larger maximum itself begs the question at issue. If the surety bond covers the bank’s liability in accordance with the terms of the admiralty bond given, then there would be no problem. If, however, the surety bond covered liability only in accordance with the terms of the bond requested, then there would be no problem if the bank’s liability were limited to the maximum stipulated by OMMIA. Perhaps the most that could be said is that the example of the surety bond issued for more than the maximum stipulated by OMMIA raises in a strong form the consequences of the narrow issue presently under discussion. It would be surprising if Liberty (and OMMIA) were liable for more than OMMIA had stipulated just because the admiralty bond produced by the correspondent bank was drawn up with a maximum greater than that stipulated.
  120. On the narrow issue raised by the preliminary issue stated, I would be content to say that I agree with Patten J, for the reasons given by him. I would add that this conclusion, derived as a matter of the language of the surety bond, and making good sense from the starting point of an acknowledged inconsistency between instructions and outcome, also has the advantage of reflecting the practice between the parties relating to the issue of the surety bond: for it was produced on the basis of OMMIA’s and the bank’s instructions, and even before the admiralty bond was yet in place.
  121. However, at the hearing of the appeal the court was concerned that the one year issue had been too narrowly focused. I am not sure, however, that the court was fully alive to the implication of the terms in which the preliminary issue had been drawn. Already before Patten J there was an attempt by the bank’s counsel to widen the issue, but it was dismissed by the judge (at para 8 of his second judgment) in these terms:
  122. “I should mention for completeness that Mr Adkins relied in support of the Bank’s contentions upon the reference in OMMIA’s instructions to the wording of the Guarantee being supplied by a nominated third party. In my judgment that was simply intended to allow the third party to identify the claim and the basis upon which it should be resolved. It did not delegate to that third party the ability to require the correspondent bank to issue a Guarantee for a period longer than that permitted under OMMIA’s express instructions.”
  123. On the appeal hearing, however, the court encouraged the parties to revert to a consideration of whether a renewable bond was indeed outside and inconsistent with OMMIA’s instructions. Mr Moss fairly submitted that he had not come prepared to deal with the wider proposition. It was arranged that the parties would address the matter in written submissions following the hearing, which they did, beginning, in accordance with the court’s invitation, with submissions from Liberty to which the bank then responded.
  124. In those written submissions on behalf of Liberty, Mr Moss contended that a one year bond with an automatic renewal clause could not be said to be within any express or implied authority of OMMIA’s request. It was after all in substance the equivalent of a bond without any limit of time. There was no evidence that the renewal was even accompanied by any element of supervision: and at most any such supervision would be a “grandified record keeping process” to pick up Mr Sumption’s own language.
  125. Mr Sumption on the other hand responded to the effect that an admiralty bond had to be in an acceptable form to a claimant, for otherwise its function, which was to obtain the release of a vessel from arrest or the threat of arrest, could not be achieved. In circumstances where disputes could often take two or three or even more years to resolve (as Liberty acknowledged in Mr Pitt’s July 1997 memorandum), a one year bond renewable on request or even automatically renewable was within the scope of OMMIA’s instructions in the light of the fact that OMMIA left it open to the nominated party to provide suitable wording for the admiralty bond. Although Liberty could point to an example within the papers of a one year admiralty bond without any provision for renewal (in the case of the Gemerela), that was the exception and it was of course always open to a claimant to accept a one year bond where he was confident that the dispute would be settled within that period.
  126. It is less than satisfactory for this point to be dealt with in this rather informal way, without any real investigation into the circumstances and practice of the provision of one year bonds of this type. For that reason I have finally come to the conclusion that the proper course is to confine the present appeal to the stated issue, on the basis of its assumption that the admiralty bond provided was inconsistent with OMMIA’s instructions, but to permit the wider ramifications of the point to be developed, if the parties so wish, at trial of the outstanding pleaded issues. It seems to me that the question whether a renewable one year admiralty bond is within OMMIA’s (and thus the bank’s) instructions potentially goes wider than merely asking whether the clause permitting renewal was within the implied authority allowed to the draftsman of the admiralty bond. Another question, or at any rate another aspect of the same question, might be whether parties such as the bank and the correspondent banks who act in good faith upon a reasonable interpretation of instructions which might be said to be ambiguous are entitled to be regarded as fulfilling their instructions: see CHITTY on Contracts, 1999, 28th ed, Vol II, at para 32-044. A third question in this context is the role and significance of the fact that OMMIA and the bank appear to have as a matter of practice accepted the admiralty bonds in issue as compliant with their respective instructions. Whether that is put as a matter of retrospective ratification, or as a course of practice enlarging the scope of the instructions, it also appears to be relevant as helping to define the nature and width of the instructions, at any rate after the first of such bonds had been accepted.
  127. I regret leaving the parties with a loose end of this kind; but I comfort myself with the thoughts that I have answered the question which the preliminary issue posed and that to go further risks unfairness and error.
  128. Summary

  129. In conclusion, I would dismiss all three appeals from the judgments of the Vice-Chancellor and Patten J.
  130. Lady Justice Arden:

  131. I gratefully adopt Rix LJ’s careful analysis of the facts and issues in this case.
  132. The Subrogation Issue

  133. The critical paragraph in the Liberty bond, to which I shall refer as “the Clause”, reads:-
  134. “This Agreement is separate from any other security or right of indemnity taken in respect of the Guarantee from any person, including your above named customer.”
  135. In my judgment the approach of the Vice-Chancellor to the construction of the Clause was correct and I can state my additional reasons shortly.
  136. The Vice-Chancellor expressed his reasons thus:-
  137. “15. I prefer, not without hesitation, the arguments for the claimants. First, I accept that clear words are required to exclude, reduce or postpone the right of subrogation. The nature of the right and the circumstances in which it arises are such that an intention to exclude, reduce or postpone it is not lightly to be attributed to the parties. Second, the requirement to keep the bond separate from any other security appears to be apt to cover the position before it has been discharged but not to rights arising after and in consequence of its payment in full. I accept that it is not clear what the draftsman had in mind when inserting the provision in question. But I do not think that such uncertainty can be converted into the clear words which, in my view, are required to give rise to the exclusion or postponement for which the Bank contends. Had it been intended to exclude, reduce or postpone the right of subrogation which would otherwise arise a less oblique approach is to be expected. Third commercial reality supports the case of the claimants. It is true, as the Bank contends, that the issuer of the bond will not know whether the Bank holds other security for the same liability but the possibility that it may is envisaged by the provision on which the Bank relies. That security may cover not only the liability under the guarantee but, as in this case, other liabilities, present or future, for which the issuer of the bond is not liable. If on payment in full of the liability under the bond the exercise of the right of subrogation is to be postponed until the other liabilities have been discharged as well the practical effect is to extend the liability under the bond.
    16. The third consideration leads to the fourth. Even if it be accepted that the effect of the provision relied on by the Bank is to exclude a right of subrogation to “the other security” to which the provision refers why should that requirement extend to such security if and in so far as it secures different liabilities. In the case of an all moneys charge, as in this case, both the creditor and debtor could postpone the right of subrogation of the issuer of the bond for ever by the creation of further liabilities. This, as I understood it, was the short answer to the Bank’s contentions advanced by counsel for the claimants in his oral argument. It would involve interpolating the word “only” either before or after the words “taken in respect of the guarantee”. Counsel for the Bank sought to refute the argument on the ground that insofar as the second security covered other liabilities then it was separate anyway. So, he submitted, the requirement for separation applied to the extent to which the two securities overlapped.
    17. In my view the Bank’s argument confirms, rather than refutes, the argument advanced by the claimants as the short answer. Accepting that insofar as a charge for different liability is separate from the bond anyway then to make the interpolation to which I have referred does reflect what the parties must have intended. If such an interpolation is made then the provision applies if and to the extent only that the second charge is security for the same liability. But in that event, in payment in full of the liability under the bond, the bank will have received 100% of that liability. Thereafter there is no ground for denying the right of subrogation to the issuer of the bond to enable him to share with the Bank the benefit of the security held by the Bank to secure the other liabilities. In short is appears to me that the argument for the Bank proves too much.”
  138. Subject to the effect of the Clause, Liberty, as a surety, is entitled on payment in full of its bond “to have assigned to [it], or to a trustee for [it], every judgment, speciality or other security which shall be held by the [Bank] in respect of the [bond] whether such judgment, speciality or other security shall or shall not be deemed at law to have been satisfied by the payment of the [bond]” (Mercantile Law Amendment Act 1856, section 5). Where the Bank’s security also secures other debts, apart from the debt guaranteed, Liberty is entitled, on payment in full of its bond and subject again to the effect of the Clause) to a rateable share of the Bank’s security: Goodwin v Gray (1874) 22 WR 312. I accept that the parties to the bond could vary or exclude these rights, or subordinate the rights of either party to those of the other (see Re T.H.Knitwear (Wholesale) Ltd [1987] 1 WLR 371 at 376 per Sir Nicholas Browne-Wilkinson VC). However, because the arrangement to postpone Liberty’s right would be unusual and asymmetric, in my judgment, the Vice-Chancellor was right to say (as he did in his first reason) that clear wording was required. A comparison may be made between this case and BCCI v Ali [2002] AC 251 where it was contended that a general release of claims on its true interpretation effected the release of a claim for stigma damages of which the parties were not aware and could not have been aware because it had not yet been enunciated by the courts. Lord Bingham, with whom Lord Browne-Wilkinson agreed, and Lord Clyde all held that for the release to have this effect, clear or specific language was necessary (see paragraphs 8 to 10, 17, 85, 86 of their speeches).
  139. The wording of the Clause is far from clear. Moreover, there is neither any significant factual background as to the parties’ objectives in agreeing the Clause (cf Investors’ Compensation Scheme v West Bromwich Building Society [1998] 1 WLR 896) nor any evidence as to any market practice (cf Charter Re-insurance v Fagan [1997] AC 313). I accept that words can have different tiers of meaning but in the absence of context the difficulties of putting a gloss on the words of the Clause are overwhelming.
  140. Mr Jonathan Sumption QC, for the Bank, makes the point that the weakness of Liberty’s case is that Liberty is unable to give any cogent or consistent meaning to that phrase, but in the particular circumstances of this case I would regard that point as only a forensic point. It may be – but Mr Sumption has not taken this line - that what the draftsman intended was that the proceeds of the bond should be kept separate (in a securities realisation account) and not applied in reducing the amount due from OMMIA until the Bank chose to apply those proceeds in this way (see The Law and Practice of Banking, J Milnes Holden, 8 edition, 1993, Vol.2 paras. 19 – 19, 2l – 11). If the Clause had been so formulated it might well have had the effect of postponing Liberty’s right of subrogation. The parties could also have achieved postponement of Liberty’s right of subrogation in other ways, such as by subordination though this may have involved the commercially inconvenient result of creating registrable charges over book debts. For my own part I am not satisfied that, as suggested by Mr Gabriel Moss QC, for the respondent, the Clause effectively disapplies any duty of the Bank as principal creditor to marshall its securities. I am forced to the conclusion that if the draftsman intended to create a stipulation which would alter Liberty’s right of subrogation, he failed to achieve that objective. In this situation I agree with Rix LJ that the court does not have to find a meaning for the Clause in order to reject those put forward by the parties. I would regard this conclusion as an exceptional and unusual one. But it is not impossible. For instance, the draftsman may have proceeded on a mistaken view, which may have been widely held among banking law practitioners, as to the effect of Re Sass [1896] 2 QB 12. We have not been asked to differ from the Vice-Chancellor’s conclusions on that case or on the observations of Lord Oliver in Barclays Bank Ltd v T.O.S.G.Trust Fund Ltd [1984] 1 AC 626 at 643 to 644.
  141. I agree with the Vice-Chancellor’s second reason. The wording is more apt to pre-empt any argument that by taking the bond the Bank waived some other secured or unsecured right in respect of the bond. I would further add that the conclusion that the Clause is more appropriate to cover the situation before discharge is consistent with the technical analysis of subrogation. When the surety pays the amount owed by the debtor, the latter’s debt is in law discharged (see Copis v Middleton (1823) T & R 224, 229). It has to be reinstated by means of subrogation. This is reflected in the wording of section 5 of the Mercantile Law Amendment Act 1856, which I have set out above. As payment by a surety of the guaranteed debt discharges the principal debtor’s liability, that from which the Agreement has under the Clause to be kept apart or separate, that is the guaranteed debt, ceases on discharge to exist. The obligation to keep the Agreement “separate” is not apt to describe an obligation not to seek to reinstate by subrogation a right against OMMIA which ex hypothesi could not be brought into existence if the Bank’s construction is correct.
  142. In any event, the Clause has no obvious effect on Liberty’s own right to sue OMMIA, or to prove in its liquidation and so the Clause would only be partially effective at best even on Mr Sumption’s construction.
  143. I agree with the Vice-Chancellor’s third reason, which deals with the practical, or economic, effect of the Bank’s interpretation of the Clause.
  144. The only reason which the Vice-Chancellor gives with which respectfully I do not agree is his fourth reason. Mr Sumption does not submit that the word “only” should be inserted, and indeed to do so would be inconsistent with the Bank’s case on the subrogation issue. However, this reason was not essential to the Vice-Chancellor’s decision.
  145. For these reasons, and those given by the Vice-Chancellor, I agree with Rix LJ that the appeal on the subrogation issue must be dismissed.
  146. The Guarantee Issue

  147. I agree with the conclusion of Rix LJ for the reasons given by him.
  148. The One Year Issue

  149. I also agree with Rix LJ on this issue for the reasons which he has given.
  150. Accordingly, I too would dismiss this appeal.
  151. Lord Justice Judge:

  152. I have read the judgments of Rix and Arden LJJ.
  153. I have reminded myself that we have been considering and deciding the meaning of a short contractual document, the whole of which has been typed and is contained on a single page. In argument before us, although both sides invited us to eschew technical niceties, the skeleton arguments designed to assist us to discover the true meaning of this single page, or more accurately, some, but by no means all the words on this page, extended to very nearly 100 typed pages of analysis and argument. I have re-read them, with admiration, and reflected on the record of the oral submissions during the hearing before us.
  154. Having done do, I shall content myself with recording that I agree that both the appeal from the Vice-Chancellor, and the appeals from Patten J should be dismissed for the reasons given by Rix and Arden LJJ. The areas of minor disagreement between their four respective judgments are not material to the outcome of either appeal, and I do not propose to comment on any point of controversy unless it is fundamental to the decision. As there is none, I have nothing to add.
  155. Order:
  156. Appeals dismissed
  157. HSBC shall pay the costs of Liberty of and occasioned by the Appeals to be assessed on a detailed assessment if not agreed.
  158. HSBC do pay the costs of St Paul of and occasioned by Appeal 1 to be assessed on a detailed assessment if not agreed.
  159. (Order does not form part of the approved judgment)


© 2002 Crown Copyright


BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/ew/cases/EWCA/Civ/2002/691.html