C v D

This appeal is, as far as I am aware, the first time that this court has had the opportunity to consider the Bermuda Form which has emerged in the last 15 years or so, partly as a response to the problem of diminution in liability insurance capacity in the United States in the later part of the 20^th century. The striking feature of the form is that it requires the parties to arbitrate in London but provides for the proper law of the insurance contract to be the internal laws of New York. No doubt this represents a balancing of the conflicting interests of the insured on the one hand and liability insurers on the other. The authors of the standard work on the topic put the matter rather starkly when they say:-

The defendant (and appellant) was the liability insurer of the claimant New Jersey Company for 3 years between 1^st November 1997 and 1^st November 2000. The occurrence limit and the aggregate limit was US$100 million excess of $190 million. The policy was a claims made policy written on the Bermuda Form. The claimant was the named insured but the definition of the insured included any subsidiary, affiliate or associated company of the claimant as listed in the Schedule to the Policy. That list included 303 companies incorporated outside the United States of America so the policy offered world-wide cover. Various sections of the document deal with various matters under the head of “Insuring Agreements”. Section I deals with coverage but it is Section V with which this appeal is chiefly concerned. That is headed “Conditions” and there is then an alphabetical list of conditions to which the policy is subject e.g. (a) “Premium” (d) “Notice of Occurrence and Claim” (m) “Cancellation” and (n) “Currency”. Conditions (o) “Arbitration” and (q) “Governing Law and Arbitration” then provide as follows:-

During the policy period various claims were asserted against the claimant and a subsidiary with significant operations in Europe. The claimant paid damages and expenses in respect of these claims, considerably in excess of the policy limits, and made demand for payment under the policy which the defendant refused. On 2^nd May 2005, the claimant initiated arbitration against the defendant in London. The Tribunal’s terms of appointment dated 31^st August 2005 and signed by the parties and by the Tribunal included the following:-

Pursuant to Article V(q) of the Agreement, the law governing the insurance policy is the law of the State of New York, USA.

Pursuant to Article V(o) of the Agreement, the juridical seat of the arbitration is London, UK. Accordingly the law governing the arbitration itself [lex arbitri] is the English Arbitration Act 1996, as amended and supplemented, regardless of whether meetings and hearings take place elsewhere in the interest of saving costs or convenience.”

The defendant raised four defences to the claimant’s claim for indemnification. The first related to the scope of Endorsement number 5 to the policy; the second related to late notice; the third related to misrepresentations and/or non-disclosure prior to the inception of the insurance; and the fourth was a defence labelled as the “paediatric defence”. That defence consisted of the defendant’s allegation that the claimant had breached a purported duty of good faith and fair dealing under New York law and/or had violated public policy in relation to the alleged promotion by the claimant the use of its product by children.

By Procedural Order No. 3 dated 20^th February 2006, the Tribunal ordered that issues relating to the first three defences should be heard first and the “paediatric defence” should be deferred until later. The rationale for this was explained in the order since, if the claimant obtained an award which amounted to the full $100 million policy limit in relation to adult use, the paediatric use issue would no longer require determination. It was only if all the first three defences failed and the recoverable sum, without taking into account the paediatric use, was less than $100 million that the Tribunal would need to make any further determination. It has not yet been possible to determine whether the policy limit will be fully utilised in respect of adult use but it seems to be increasingly likely.

A hearing took place between 4^th and 12^th October 2006 to deal with the first three defences. Sixteen witnesses attended the hearing for cross-examination and there were extensive post hearing submissions. The Tribunal issued its Partial Award on 13^th March 2007, ruling that the claimant succeeded in full on its claim under the policy and that it was entitled to recover, dismissing each of the defendant’s first three defences and related claims for relief. The claimant was also awarded interest and costs. The Partial Award also provided that the paediatric defence would only be considered if the claimant could not establish that it had exhausted the policy limits, without including losses attributable to paediatric use. The parties were invited to seek to agree the quantum of the claims which the Tribunal had held were covered by the policy. It is agreed that this Partial Award is, in English law terms, final as to what it decides.

In correspondence following the Partial Award, the defendant applied to the Tribunal to “correct” it, stating (inter alia) that the Tribunal’s findings constituted a “manifest disregard of New York law”, that the Partial Award fell outside the scope of the Convention of the Recognition and Enforcement of Foreign Arbitral Awards of 10^th June 1958 (the 1958 New York Convention) and as such was reviewable for error by any US Federal District Court having jurisdiction over the parties under the general federal venue statute. The defendant sought the Tribunal’s withdrawal of its findings as to the claimant’s duty to disclose, as to its expectation and intent and as to materiality, pending the outcome of the next phase of the hearing during which the Tribunal would hear evidence of the claimant’s promotion of paediatric use.

In further correspondence, the defendant intimated its intention to apply to a Federal Court applying US Federal Arbitration law governing the enforcement of arbitral awards, which was said to permit “vacatur” of an award where arbitrators have manifestly disregarded the law. It was in consequence of such intimation that the claimant sought and obtained an interim anti-suit injunction. The Tribunal made two “clerical” amendments but refused otherwise to amend the Partial Award, saying it had no power to do so.

The claimant’s case was that the defendant’s proposed challenge to the Partial Award in the United States (now particularised as a likely application to the courts of the Southern District of New York) was impermissible by reason of the agreement of the parties to London as the seat of arbitration and to the application of the English Arbitration Act of 1996 (“the 1996 Act”), by reason of the terms of the Arbitration Agreement and Agreement to Refer, and by reason of the court’s power of supervision over arbitrations held in this jurisdiction. The only permissible challenges to the Partial Award itself were those which can be made under the 1996 Act and the only permissible challenges to enforcement in other countries, which are parties to the 1958 Convention (as is the US), were those which arise under Article V of the Convention.

The defendant’s case was that, as New York law is the governing law of the insurance policy and that law entitles the defendant to a minimum standard of review of arbitration awards, when such arbitration takes place between US corporations in relationships without an important international element, the defendant cannot be deprived of exercising its right to such a review. Federal law, which is part of the internal law of New York, operated to require such a minimum standard of review, regardless of the terms agreed between the parties which might appear to restrict any ground of challenge. Although English law was the “curial law” of the arbitration, that did not exclude a challenge which reflected the parties’ express choice of New York law to govern their obligations under the policy. The defendant contended that the arbitrators had made fundamental errors of New York law in the Partial Award and that the court should proceed on the assumption that it had at least a seriously arguable case in that respect.

The claimant put its case on the basis of English law and maintained that, for the purposes of the court’s consideration of the final injunction now being sought, the law of New York was irrelevant. By contrast the defendant focused on the law of New York, seeking to establish that the Partial Award was a non-Convention Award under the terms of section 202 of the US Federal Arbitration Act (FAA), with the result that it was capable of challenge in New York. The defendant relied on English principles of conflicts of law to import New York law as the proper law of the arbitration agreement as well as the proper law of the underlying contract.

The judge was prepared to assume for the purpose of the argument before him, first that the defendant had a sufficiently arguable case that the Arbitration Tribunal has acted in manifest disregard of New York law and, secondly that it was arguable that the Partial Award was not a Convention Award for the purpose of the New York Convention, so that a wider challenge could be made to the Partial Award than permitted under Article V of that Convention. The claimant without accepting them is content for present purposes that those assumptions should be made and I will proceed on the same basis.

The judge then proceeded to hold that the choice of England as the seat of the arbitration was determinative of the matter in as much as the parties had, by that agreement, expressly (or perhaps impliedly) agreed that any proceedings seeking to attack or set aside the Partial Award would only be those permitted by English law. That effectively meant that the Partial Award could only be attacked by reference to sections 67 and 68 of the 1996 Act (lack of jurisdiction and serious irregularity), the right of appeal under section 69 on points of law having been excluded by agreement. It was not therefore permissible for the defendant to bring any proceedings in New York or elsewhere to attack the Partial Award in any respect permitted by the law of that place e.g. for any supposed “manifest disregard” of the proper law of the contract. The judge also rejected arguments to the effect that the choice of the law of New York as the proper law of the contract amounted to an agreement that the law of England should not apply to proceedings post-award pursuant to section 4(5) of the 1996 Act and a further argument that the separate agreement to arbitrate contained in the Condition V(o) of the policy was itself governed by New York law so that proceedings could be instituted in New York. He then granted the claimant a final injunction.

The main submission of Mr Hirst QC for the defendant insurer was that the judge had been wrong to hold that the arbitration agreement itself was governed by English law merely because the seat of the arbitration was London. He argued that the arbitration agreement itself was silent as to its proper law but that its proper law should follow the proper law of the contract as a whole, namely New York law, rather than follow from the law of the seat of the arbitration namely England. The fact that the arbitration itself was governed by English procedural law did not mean that it followed that the arbitration agreement itself had to be governed by English law. The proper law of the arbitration agreement was that law with which the agreement had the most close and real connection; if the insurance policy was governed by New York law, the law with which the arbitration agreement had its closest and most real connection was the law of New York. It would then follow that, if New York law permitted a challenge for manifest disregard of the law, the court in England should not enjoin such a challenge.

Mr Eder QC contested all these points but submitted that they were all irrelevant because the judge was correct to decide that, once it was clear that England was the seat of the arbitration and that English law was, therefore, the “curial law” of the arbitration, it must follow that the parties intended only attacks which were permissible by English law and not attacks permitted by other laws, including those permitted either by the proper law of the underlying insurance contract or by the proper law of the arbitration agreement, if different.

I shall deal with Mr Hirst’s arguments in due course but, in my judgment, they fail to grapple with the central point at issue which is whether or not, by choosing London as the seat of the arbitration, the parties must be taken to have agreed that proceedings on the award should be only those permitted by English law. In my view they must be taken to have so agreed for the reasons given by the judge. The whole purpose of the balance achieved by the Bermuda Form (English arbitration but applying New York law to issues arising under the policy) is that judicial remedies in respect of the award should be those permitted by English law and only those so permitted. Mr Hirst could not say (and did not say) that English judicial remedies for lack of jurisdiction on procedural irregularities under sections 67 and 68 of the 1996 Act were not permitted; he was reduced to saying that New York judicial remedies were also permitted. That, however, would be a recipe for litigation and (what is worse) confusion which cannot have been intended by the parties. No doubt New York law has its own judicial remedies for want of jurisdiction and serious irregularity but it could scarcely be supposed that a party aggrieved by one part of an award could proceed in one jurisdiction and a party aggrieved by another part of an award could proceed in another jurisdiction. Similarly, in the case of a single complaint about an award, it could not be supposed that the aggrieved party could complain in one jurisdiction and the satisfied party be entitled to ask the other jurisdiction to declare its satisfaction with the award. There would be a serious risk of parties rushing to get the first judgment or of conflicting decisions which the parties cannot have contemplated.

It follows from this that a choice of seat for the arbitration must be a choice of forum for remedies seeking to attack the award. As the judge said in paragraph 27 of his judgment, as a matter of construction of the insurance contract with its reference to the English statutory law of arbitration, the parties incorporated the framework of the 1996 Act. He added that their agreement on the seat and the “curial law” necessarily meant that any challenges to any award had to be only those permitted by that Act. In so holding he was following the decisions of Colman J in A v B [2007] 1 Lloyds Rep 237 and A v B (No. 2) [2007] 1 Lloyds Rep 358 in the first of which that learned judge said (para. 111):-

Mr Hirst’s argument was that section 58 of the 1996 Act which provided for the finality of an arbitral award was not a mandatory provision of the Act and that there was a permissible “agreement to the contrary” contained in the arbitration clause itself which was governed by the law of the state of New York which permitted challenge for manifest disregard of the law.

The fact, however, that the 1996 Act allows parties to contract out of its non-mandatory provisions does not mean that the proper law of a contract to refer disputes to arbitration can constitute an “agreement to the contrary” and thus import a method of challenge to the award not permitted by the seat of the arbitration. For example section 49 of the 1996 Act gives an arbitration tribunal power to award interest. That provision is one of the non-mandatory provisions of the Act. It was argued in Lesotho Highlands Development Authority v Impregilo SpA [2006] 1 AC 221 that, if the proper law of the underlying contract did not permit an award of interest, the choice of that proper law amounted to an agreement to the contrary so as to preclude the Tribunal from awarding interest. Lord Steyn (with whom the majority of the House agreed) pointed out (para. 37) that by reason of section 5 of the Act only an agreement in writing as defined by the 1996 Act could qualify as an “agreement to the contrary” and that a choice of proper law clause was not such an agreement. That is reinforced by the terms of section 4(5) of the Act which refers not to a choice of law clause generally but to a choice of law as “the applicable law in respect of a matter provided for by a non-mandatory provision of this part” of the Act. In other words there has to be a choice of law with regard to the specific provision of the Act which the parties agree is not to apply.

Even if therefore the first plank of Mr Hirst’s argument (that the arbitration clause itself was governed by the law of New York) were to be correct, it would not qualify as an “agreement to the contrary” in the 1996 Act. Still less would it entitle the defendant to mount a challenge to the award in a country other than the seat of the arbitration.

It is therefore unnecessary to engage with Mr Hirst’s first argument that the arbitration agreement is governed by New York law. But since the point was fully argued, I will express my view upon it.

It is necessary to distinguish between the proper law of the underlying insurance contract which is, by agreement, the internal law of New York and the arbitration agreement which is, by virtue of section 7 of the 1996 Act, as well as by virtue of common law, a separable and separate agreement, see Harbour Assurance Co. (UK) Ltd v Kansa General International Insurance Co Ltd [1993] QB 701 before the Act and Premium Nafta Products Ltd v Fili Shipping Co Ltd [2007] UKHL 40 after the Act. There is also the law of the seat of the arbitration, namely English law, which will be relevant. The question then arises whether, if there is no express law of the arbitration agreement, the law with which that agreement has its closest and most real connection is the law of the underlying contract or the law of the seat of arbitration. It seems to me that if (contrary to what I have said above) this is a relevant question, the answer is more likely to be the law of the seat of arbitration than the law of the underlying contract.

In the days before the separability of the arbitration agreement was fully apparent it was often said that if a contract chose a place of arbitration, the law of that place was the proper law of the contract on the principle of ‘Qui elegit judicem elegit jus’ see Tzortzis v Monark Line [1968] 1 WLR 406. (No doubt it would conversely have been said that, if a contract had an express choice of law clause, the law of the arbitration agreement would have been the same as the proper law of the contract). This convenient but stark proposition was departed from by the House of Lords in Compagnie Tunisienne De Navigation S.A. v Compagnie D’Armement Maritime S.A [1971] AC 572 in which it was pointed out that the inquiry must always be to discover the law with which the contract has the closest and most real connection. It was there decided that the mere fact that arbitration was to be in London did not mean that what was in reality a French contract of affreightment had to be governed by English rather than French law. It did not matter at all that English arbitrators would have to apply French law. In these circumstances it cannot be automatic that if the relevant inquiry is the converse inquiry (namely to discover the proper law of an arbitration agreement) the answer to that inquiry is to be the proper law of the agreement. The inquiry is, as I have said, to discover the law with which the agreement to arbitrate has the closest and most real connection.

The matter is not entirely free from authority. In Black-Clawson v Papierwerke [1981] 2 Lloyds Rep 446, 483 Mustill J set out the three potentially relevant laws, namely (i) the law governing the substantive agreement; (ii) the law governing the agreement to arbitrate and the performance of that agreement; and (iii) the law of the place where the reference is conducted (the lex fori). He then said:-

Mr Hirst submitted that, by the time Lord Mustill came to give judgment in Channel Tunnel Group Ltd v Balfour Beatty Construction Ltd [1993] AC 334, he had changed his mind on this question. In the course of deciding whether section 12(6)(h) of the Arbitration Act 1950 (relating to the power to grant interim injunctions) applied to arbitrations with their seat outside England, Lord Mustill said (357A – 358A):-

It does not seem to me that Lord Mustill is in fact saying any such thing. He is merely saying that although it is exceptional for the proper law of the underlying contract to be different from the proper law of the arbitration agreement, it is less exceptional (or more common) for the proper law of that underlying contract to be different from the lex fori or curial law namely the seat of the arbitration. He is not expressing any view on the frequency or otherwise of the law of the arbitration agreement differing from the law of the seat of the arbitration. One is therefore just left with his dictum in Black-Clawson (with which I would respectfully agree) that it would be rare for the law of the (separable) arbitration agreement to be different from the law of the seat of the arbitration. The reason is that an agreement to arbitrate will normally have a closer and more real connection with the place where the parties have chosen to arbitrate than with the place of the law of the underlying contract in cases where the parties have deliberately chosen to arbitrate in one place disputes which have arisen under a contract governed by the law of another place.

Mr Hirst might say that Miller v Whitworth is a good example of a case where the agreement to arbitrate took its colour from the proper law of the underlying contract rather than from the fact that the arbitration took place in Scotland. But (1) although the arbitration did in fact take place in Scotland it was not a requirement of the contract that it should – it was never decided that the arbitration was contractually required to take place in Scotland; and (2) the case, if it is support for Mr Hirst’s submissions on this part of the case, becomes a strong authority against him on the primary part of the case. That is because the House of Lords determined that it was the law of the seat of the arbitration that decided whether it was possible to require the arbitrator to state a special case. Since the seat of the arbitration was Scotland, there was no such power. The fact the arbitration agreement was itself governed by English law did not at all mean that an English law remedy was available, if it was not available at the seat of the arbitration. The decision was that the losing party in an arbitration whose seat was in Scotland could not go to the English courts and ask for an English remedy. Likewise the insurers who have lost an arbitration in England cannot go to another state (e.g. New York) and invoke the local remedies in that state.

As the judge observed in paragraph 45 of his judgment, these are only general considerations; much more forceful in the present case are the positive indications in the arbitration agreement itself which point to English law governing the agreement. Moreover as the judge points out in paragraph 47, the provision that the arbitral decision shall be final and binding and

For all these reasons Mr Hirst’s first argument that the proper law of the arbitration agreement was New York law rather than English law cannot get off the ground and the only remaining questions relate to remedy and costs.

The judge granted an anti-suit injunction preventing the defendant insurers from initiating proceedings on the Partial Award in New York and also preventing them from relying on law of the New York in any application to enforce the Partial Award. Mr Hirst reminded us of the caution that the English court always exercises in relation to such injunctions by reason of the possibility that they may be thought to interfere with decisions or potential decisions of a foreign court. Having every regard to that caution, it nevertheless seems to me that the judge was right not only to grant a final injunction but to frame it in the way in which he did. It is only by doing so that the parties’ legitimate expectations in relation to the Bermuda Form can be respected and enforced. I have already said that the form constitutes a balancing of the opposing interests of the insured and their insurers. If either party was permitted to challenge an award in a manner intended to be excluded by the form, that balance would be fatally compromised. This is just as much in the interest of insurers as well as of the insured. This particular case is one in which it is the insured who seeks injunctive relief but tomorrow it may be the insurer in whose interest it is to uphold the intentions of the parties as expressed in the Bermuda Form. The form of relief is, in any event, a matter for the judge’s discretion with which this court will not lightly interfere. Since the insurers have indicated that they will seek relief unless they are restrained, the judge’s exercise of his discretion is, in my judgment, unassailable.

There has been much recent debate at first instance whether, in a case where a party has acted in breach of a jurisdiction clause, an arbitration clause or an anti-suit clause, it is proper to make an order for costs on an indemnity basis rather than the standard basis. In Kyrgyz Mobil v Fellowes International [2005] EWHC 1314 (Comm.) Cooke J awarded indemnity costs against a defendant who had started proceedings in the courts of Kyrgyzstan in breach of an arbitration agreement saying that that was the general approach of the Commercial Court. In A v B (No. 2) [2007] 1 Lloyds Rep 358 Colman J said he had not come across any such practice but decided that, where a claimant had wrongly invoked the jurisdiction of the English court to make claims which were covered by an arbitration agreement or were subject to the jurisdiction of the Swiss courts (the seat of the arbitration being Switzerland), the successful defendant was entitled to be paid costs on an indemnity basis. In National Westminster Bank v Rabobank Nederland RV [2007] EWHC 1742 (Comm.) Colman J decided that damages for breach of a jurisdiction clause and anti-suit clause should include the costs of the wrongful proceedings assessed by reference to the indemnity basis rather than the standard basis. In the present case Cooke J followed what was, at any rate, his practice and the question is now whether it is right that the insured should have received the costs of his successfully proceeding for an anti-suit injunction on an indemnity basis.

This appeal is not, in my judgment, a suitable vehicle for coming to any definitive decision on the proper approach for awarding costs in such cases. That is partly because we only received short oral argument on the question (the detailed and intricate judgment of Colman J only coming to our attention after the oral argument had concluded) but mainly because, in all the cases to which our attention has been drawn, there had been conduct which was found to be in breach of a relevant exclusive jurisdiction or arbitration clause.

This case is, however, different. Although costs are normally a matter for the judge’s discretion, questions of principle do arise. In the first place neither the judge (nor I) have held that there is an exclusive jurisdiction clause in fact; I have merely agreed with Colman J in A v B (No. 1) and Cooke J in the present case that agreement on the seat of the arbitration is “analogous to an exclusive jurisdiction clause”. Secondly, there has been no instigation of proceedings which can be categorised as an actual breach of the agreement that the seat of the arbitration should be London, England. Insurers have very properly not (as far as I am aware) actually started proceedings in New York, but have merely intimated that they will start such proceedings, if they are not restrained from doing so. Thirdly, the issues which have been raised on the Bermuda Form are novel issues on which up to now conflicting views could legitimately be held. For these reasons in this particular case I do not think an order for indemnity costs was appropriate. The wider issue will have to await further argument on another day.

We held the oral hearing in private at the joint request of the parties. It is not the practice of this court to sit in private on arbitration (or indeed any other) appeals unless there is a special reason to do so. That is the case even though parties to arbitrations can legitimately expect that arbitrations are themselves confidential to the parties. Having heard the appeal, I am satisfied that there was no good reason to have sat in private at any rate if anonymisation of the parties was continued for the hearing. I would therefore now rescind the order which the court originally made for a private hearing but continue the order for anonymity for the moment.

As regards anonymisation, we received some short submissions from Mr Eder on instructions setting out why anonymity was desirable. I was not convinced by those submissions that anonymity should continue to be preserved, particularly because there was no evidential basis for the submissions which he, on instructions, was able to make. I would therefore propose that, unless the claimant wishes to make further submissions based on some form of actual evidence within the next 14 days, the order for anonymity should also cease to have effect after that 14 day period.

I would further wish to emphasise that any future application for privacy or anonymity (for arbitration appeals) should be supported by written evidence at the time such application is made in the form of a statement from someone at managerial level explaining the need for privacy or anonymity.

As it is, I would dismiss this appeal, save as to costs.

I agree.

I also agree.