This is an appeal in which a Part 20 defendant says that the judge was wrong not to stay the Part 20 proceedings against it and ought to have done so: both pursuant to a foreign jurisdiction clause and in the court’s inherent case management powers. As such the appeal attacks the exercise of the judge’s discretion.

It is a case of procedural and factual complexity, and to some extent mystery too, for at least four reasons. Its data makes it so. It stands at an interlocutory stage at which the court is reluctant to trespass on the decisions that properly belong to trial. There are proceedings abroad as well as here. And the claimant in both jurisdictions is not a party to the appeal in this court. For all or some of those reasons this court has to a material extent had to contend with conflicting submissions or assertions which it is not easy to evaluate.

The question is nevertheless whether behind these difficulties there lie some hard facts which either support or undermine the judge’s approach to the case.

The claimant in these proceedings is Konkola Copper Mines plc (“KCM”). It is incorporated in Zambia where it owns and operates the Nchanga copper mine. On 8 April 2001 an avalanche of rock and mud substantially damaged its mine, killing ten miners. Its claim is directed to recovering an indemnity for its loss under its insurance arrangements. For these purposes there is an important issue whether its mine suffered “collapse” or “landslip”.

At the relevant time KCM was a subsidiary within the Anglo American mining group, whose parent company, Anglo American plc, is incorporated in England with its head office in London. Through its Luxembourg subsidiary, ARH Limited SA (“ARH”), Anglo American purchased KCM in March 2000 but disposed of it in August 2002. ARH is also a claimant, as assignee of certain of KCM’s rights to insurance. I need not mention ARH again, since it will suffice to refer to KCM as the claimant.

There are potentially up to three layers of insurance. There are local insurers in Zambia (the “Zambian underwriters”). They insured KCM on a direct basis, which provided cover for named perils including collapse but not landslip. Their insurance incorporated a “Local Law and Jurisdiction Clause”.

A further layer of insurance was provided by Coromin Limited, an insurance company incorporated in Bermuda. At the time of loss Coromin was a captive insurer set up to insure or reinsure risks relating to companies within the Anglo American group. It is no longer controlled by Anglo American, and pursues its conduct in these proceedings independently. There are issues whether Coromin is an insurer of KCM directly (on an all risks basis) or only a reinsurer of the Zambian underwriters (on the named perils basis).

The third layer of insurance was provided by a group of European underwriters, led by Swiss Reinsurance Company (“Swiss Re”) and including Lloyd’s syndicates and companies domiciled in England, other European countries or Switzerland (the “Reinsurers”). There are issues as to the nature of the reinsurance provided by the Reinsurers. They say that they provided retrocession to Coromin on special terms (the “KCM wording”) which essentially mirrored the named perils cover provided by the Zambian underwriters and incorporated a Zambian law and jurisdiction clause. Coromin says that they included KCM in a global Anglo American group policy organised by Coromin which provided cover on all risks terms and incorporated an English law and jurisdiction clause.

The reinsurance provided by the Reinsurers was broked in London through Aon Limited, an English company.

The party structure of these proceedings, which were commenced on 27 May 2004, is as follows. KCM is claimant and sued Coromin and the Zambian underwriters: Coromin as a direct insurer on the all risks basis reflected in the global policy and the Zambian underwriters on their local insurance. Since that time KCM and the Zambian underwriters have agreed that their dispute should be heard in Zambia under Zambian law. Consent orders dated 10 and 17 March 2005 were made reflecting that agreement, which inter alia stayed these proceedings against the Zambian underwriters. Thus at the time of this appeal (and of the hearing below) Coromin was the single surviving defendant.

Coromin denied that it had entered into any all risks insurance of KCM but nevertheless sought to protect itself against KCM’s claim by commencing Part 20 proceedings against the Reinsurers. It served both its defence and Part 20 claim form on 24 September 2004. The Reinsurers say that they have reinsured only on a named perils basis and that KCM suffered a landslip and not a collapse. Coromin says that if it is liable to KCM directly at all, then it is reinsured by the Reinsurers on an all risks basis. The Reinsurers say that a Zambian law and jurisdiction clause applies and on that ground resist jurisdiction in England. Coromin relies on the English law and jurisdiction clause in its global policy.

Coromin has also more recently, and since the hearing and judgment below, joined Aon to its Part 20 proceedings, alleging that if, contrary to its primary argument that the Reinsurers have reinsured it on an all risks basis, the reinsurance is only on a named perils basis, then Aon is liable to it in negligence and breach of contract: since it had instructed Aon to obtain cover on the former basis and been told by Aon that that had been achieved.

It is necessary to say something further about KCM and the Zambian underwriters. KCM obtained leave to join them to its English proceedings as necessary or proper parties to its claim against Coromin. The Zambian underwriters’ response was to challenge jurisdiction on the ground of the Zambian law and jurisdiction clause in their contracts. They also denied that they had insured any more than 10% of the interest, saying that 90% had been insured by Coromin.

The consent orders to which I have referred above record the agreement by which KCM agreed to bring its claim against the Zambian underwriters in Zambia under Zambian law, and the Zambian underwriters agreed to accept, subject to the collapse or landslip coverage issue, potential liability as 100% direct insurers of KCM. The consent orders refer to the relevant insurance contract as being one for 12 months from 1 July 2000 to 30 June 2001 entered into on the same terms as a prior 3 month cover for the previous quarter 31 March to 30 June 2000.

The consent orders, staying KCM’s English proceedings against the Zambian underwriters, were made by Cresswell J and Andrew Smith J in the week immediately before the hearing of the Reinsurers’ application to stay Coromin’s Part 20 proceedings against them. When the judge on that hearing, Colman J, on 18 March 2005 learned of these consent orders, he was concerned, in the light of the submissions that had been addressed to him on behalf of Coromin and the Reinsurers, that the consent orders had made it impossible for him to canalise the litigation as a whole into a single set of proceedings, either in Zambia or in England, and thus raised the threat of inconsistent judgments in an intractable form. He therefore wrote a letter to all the parties or former parties in the proceedings, that is to say including KCM and the Zambian underwriters, neither of whom had been present at the hearing of 18 March, requesting them to attend at a further hearing at which he wished to consider whether the consent orders should be set aside as having been made ex improviso without the judges of those orders being told of the overall ramifications of such orders for the English proceedings as a whole.

In his letter Colman J said:

The judge heard the parties on 22 April 2005. The essence of what then happened is set out in the judgment under appeal, now reported at [2005] 2 Lloyd’s Rep 555:

There is no appeal from that part of the judge’s judgment, nor could there be any attempt to appeal it without providing for the presence in court of KCM and the Zambian underwriters.

It follows that the bifurcation of KCM’s claims between Zambia and England is a given in the present appeal.

That is an unusual state of affairs. It is prima facie abusive for a claimant to commence two separate actions on the same claim. KCM’s claims in Zambia and England, however, are not on the same claim, nor against the same parties, for in Zambia they rely on a claim against the Zambian underwriters (on a named perils basis) and in England they rely on a claim against Coromin (on an all risks basis). The claims are not, as I understand it, alternative although KCM of course does not seek more than an indemnity. There is nothing unusual about double insurance, and there are special provisions of the law, such as the right of co-insurers to claim a rateable contribution from one another in equity, to sort out the consequences. Insurance contracts often contain special clauses designed to short-circuit that process by providing that in the event of other insurance the insured can only recover a rateable proportion of his claim. Therefore, if an insured is covered by two insurances which each contain such a clause, he is compelled to sue both insurers to recover a full indemnity. It is possible that such provisions apply in the present case to KCM’s claims. Thus it seems that substantially the same “Other Insurance” clause was proposed in the case of both insurance contracts relied upon.

In the case of the insurance with the Zambian underwriters the proposed clause was in the following terms:

I say that this clause was “proposed”, because it is Coromin’s contention that at the time of the incident at the mine the Zambian underwriters had not yet agreed the proposed policy wording of their 12 month contract. It is not clear to me whether Coromin’s evidence is that the wording for the named perils cover was never agreed (a form of what below is called the KCM wording) or whether it is rather that the contemplated changeover to all risks cover (to be based on what below is called the CCIP wording) was never agreed. As for KCM’s alleged all risks insurance with Coromin, the latter denies any such contract saying that that never matured beyond a draft proposal. However, the global all risks policy which Coromin in fact made with the Reinsurers for the Anglo American group and of which KCM says that it took the benefit at the primary level with Coromin as its direct insurer did contain an “Other Insurance” clause as follows:

The potential existence of the Other Insurance clauses means that ultimately it may not be possible to resolve KCM’s claims in either jurisdiction without finding a solution to both of them. Even if I put on one side matters of pure quantum, on the basis that the defendants in one jurisdiction (and their reinsurers) might be willing to accept the adjudication found in the other jurisdiction, and even if I also assume that, because Coromin accepts that it is the (90%) reinsurer of the Zambian underwriters on the named perils basis, both KCM and Coromin might otherwise be willing to compromise their dispute in England were KCM once to succeed in their Zambian claim, nevertheless Coromin is not able to do so when it takes its claim over against the Reinsurers into account. For if KCM is entitled to claim on both its alleged insurances, it can recover at best only half its indemnity under each. Therefore, Coromin could not recover from the Reinsurers more than half its liability without persuading the Reinsurers that it was liable only under its reinsurance of the Zambian underwriters.

In any event, if KCM were to fail in its Zambian claim (on the ground that the incident was not a collapse), it would still wish to press ahead with its all risks claim against Coromin in England. Such a claim avoids the factual difficulties, such as they may be, of establishing a collapse. On the other hand there may be legal difficulties in establishing its alleged direct claim, not only because Coromin denies that the insurance relied upon got any further than a draft proposal, but also inter alia because Coromin says that any direct insurance of a Zambian insured by other than local Zambian insurers is illegal under Zambian law. It is difficult to evaluate these issues, in part because the relevant documentation is not before the court. But it is also unnecessary to evaluate them because no one has suggested that, whatever be its ultimate merits, KCM’s all risks claim is not a claim in good standing. Jurisdiction against Coromin is not in question.

Be that as it may, KCM has made it clear that it wishes to press on with what it terms its primary claim against Coromin in England. Its particulars of claim identify that as its primary claim, it had made clear to the judge at the hearing of 22 April 2005 that it would prefer, if it felt it could, to have litigated everything in England, and in a letter dated 25 October 2005 addressed to Coromin’s solicitors to coincide with the appeal hearing, it reaffirmed its interest “in vigorously pursuing its primary all risks insurance claim” in the London proceedings against Coromin, despite in the meantime serving its statement of claim in the Zambian proceedings.

Evidence before Colman J suggested that it could take a year to reach trial in Zambia, and another 18 months to two years for any appeal. The matter may have been complicated since the appeal hearing, however, by the delivery of the Zambian underwriters’ defences. Four out of the five underwriters have declined to rest content with the consent orders in the English proceedings, for they have continued to plead that they are liable for a total of only 8% out of the 10% retained locally. Despite KCM’s pleading of the consent orders in its statement of claim, and the admission of them, these defendants allege that the “purported 100% cover of the risk by the defendants was a sham” in an attempt to circumvent the Zambia Insurance Act, and that they were blackmailed into consenting to that sham. It is not known at present what complications this may cause to the Zambian proceedings.

I have said above that the bifurcation of KCM’s claims against the Zambian underwriters and against Coromin is an unusual state of affairs. This is because the attitude of the English courts is, if possible, to avoid fragmentation of disputes between different jurisdictions where such fragmentation raises the twin dangers of waste of resources and of inconsistent decisions. The Brussels Convention on Jurisdiction and the Enforcement of Judgments in Civil and Commercial Matters 1968 (and now Council Regulation No. 44/2001 (“the Regulation”)) also approaches the risk of inconsistent decisions with the same dislike. However, the techniques of the English common law and of the Regulation are different. The common law ultimately relies on an exercise of discretion to reach what in each case seems to the court to be the right result. The Convention and Regulation state rules designed to avoid inconsistent decisions, but if those rules fail in a particular case to avoid that danger, there can be no fall-back on discretionary powers: see Erich Gasser GmbH v. MISAT Srl [2004] 1 Lloyd’s Rep 445, Owusu v. Jackson (Case C-281/02)[2005] QB 801.

Exclusive jurisdiction clauses in different contracts can in particular lead to such fragmentation of disputes: see, for instance, in the context of the Brussels Convention, Hough v. P&O Containers Ltd[1999] QB 834. Under the English common law, however, there remains a discretion, even in the case of an exclusive jurisdiction clause, to decline to enforce it if there are strong reasons for taking that course, such as to enable all aspects of an overall dispute to be canalised into a single set of proceedings: Donohue v. Armco[2001] UKHL 64, [2002] 1 Lloyd’s Rep 425. Even so, such canalisation is not always possible: see Credit Suisse First Boston (Europe) Ltd v. MLC (Bermuda) Ltd [1999] 1 Lloyd’s Rep 767, cited with approval by Lord Bingham of Cornhill in Donohue v. Armco (at para 28). Commenting on that case, Lord Bingham said (ibid):

So in the present case also it is impossible now to prevent the fragmentation of KCM’s claims.

In sum, so far as the basic claim of KCM in England is concerned, it stands for adjudication here. Any chance for staying it was lost following the decision of Colman J arising out of the hearing of all the parties to the proceedings on 22 April 2005. There has been no appeal from that decision. Therefore KCM’s claims will proceed in any event in two jurisdictions. The question that remains is whether Coromin’s Part 20 proceedings against its Reinsurers will be permitted to go forward to trial together with KCM’s English claim, as the judge ruled, or will be stayed, as the Reinsurers request on this appeal. In the latter case, Coromin would be exposed in this jurisdiction to a claim from its alleged insured without being entitled at the same time to proceed with its back to back claim against its Reinsurers. The essential question is whether this is something which is compelled by reason of a Zambian jurisdiction clause in the reinsurance or is otherwise in the interests of justice. The judge thought it was neither and therefore refused to stay the Part 20 claim.

Before I go on to examine the judge’s reasons for his decision and whether they are, as the Reinsurers submit, unreasonable and wrong, it is necessary to say something more about the nature of the dispute between Coromin and the Reinsurers.

As between Coromin and the Reinsurers the evidential position is far different from that between Coromin and KCM. In the latter case the court has very little other than the basic assertions of each side, from which it can be stated, rather baldly, that Coromin accepts that it is the 90% reinsurer of the Zambian underwriters for the 12 month period in question, on a named perils basis, but otherwise disputes KCM’s case. As between Coromin and the Reinsurers, however, the evidential and documentary material before the court is rich and complex, and it is the court’s task to select the minimum of what it needs to set the background to the current appeal.

When in early 2000 KCM was about to join the Anglo American group, Coromin, as captive insurer, sought to bring KCM within its global insurance and reinsurance policies. Such cover was, Coromin says, reinsurance driven. In the then current 1999/2000 year Coromin had already arranged a Global Master All Risks Policy which incepted for the 12 months commencing on 1 July 1999 and which provided insurance or reinsurance by Coromin for all Anglo American subsidiaries and/or their local insurers. This was backed by a reinsurance agreement (the “1999/2000 Reinsurance”) provided by Swiss Re and other reinsurers. The wording attached to the 1999/2000 Reinsurance was said to be the same as that of the original policy (except where specifically modified) and was described as the Commercial Combined Insurance Policy (“CCIP”). It was that wording which at clause 4.3 contained the “Other Insurance” clause cited above, and at clause 4.7 contained an English law and jurisdiction clause. A further clause of the CCIP wording was a “Difference in Conditions” clause (“DIC”) in the following terms:

Thus the effect of the DIC clause is to extend the coverage of an original assured to the full width of the all risks cover, despite any limitation of local or primary insurance. In effect, therefore, reinsurers under the CCIP wording become primary insurers to the extent of any difference in conditions.

Coromin sought to attach KCM, as a new Anglo American subsidiary, to its insurance and reinsurance arrangements for the last quarter of the 1999/2000 policy year. However, it could not interest Swiss Re, as lead reinsurer, to do so by a straightforward endorsement to the 1999/2000 Reinsurance, since Swiss Re was unwilling to insure KCM on an all risks basis. A special endorsement was therefore agreed by which KCM was reinsured by the 1999/2000 reinsurers on a named perils basis, with increased excesses and special limits and sub-limits. For this purpose special wording, the “KCM wording” was developed. The KCM wording in many respects followed the form of the CCIP wording, but for present purposes crucially (i) only covered named perils; (ii) did not include the DIC clause; and (iii) included a Zambian law and jurisdiction clause. Moreover, the cession was only of 90%. The relevant endorsements appear to be those scratched and dated 28 March and 25 May 2005. A fax of 25 May 2005 also refers to the agreement of the KCM wording. (The underlying primary insurance provided by the Zambian underwriters for this same three month period is similarly evidenced by a cover note scratched by the lead local insurer on 28 March 2005.) An additional premium was agreed.

The three month insurance was only a stop-gap pending the new broking of Coromin’s global insurance and reinsurance arrangements for the next period commencing 1 July 2000. It is at this point that Coromin’s and the Reinsurers’ cases part company.

Coromin’s primary case is that it instructed Aon and that Swiss Re was fully informed about its intention that KCM be brought fully within the renewal of its global all risks arrangements and that the three month endorsement relating to the 1999/2000 year was merely a temporising expedient. Thus Anglo American developed, with Coromin and Aon, a booklet entitled “Global Underwriting Information Renewal Year 2000” (the “2000 Underwriting Submission”). The booklet explained that Anglo American’s global underwriting for 2000 and forwards needed to be looked at again because of the important merger with Minorco SA in May 1999. It seems that although the KCM acquisition was not part of that merger, nevertheless it was to be swept up in the new arrangements. All Anglo American’s operating subsidiaries, including KCM, were listed in a section called “Risk profile information”. In due course Swiss Re was to scratch a slip stating that it had seen the 2000 Underwriting Submission.

On 19 June 2000 Aon faxed Swiss Re a draft slip of the same date for a 36 month global reinsurance. The slip refers to “TYPE” as “All Risks of Physical Loss or Damage Reinsurance”, and to “FORM” as “Following original – Slip Policy NMA 1779a – with original Manuscript Global wording as expiring…”, a reference it is said to the 1999/2000 Reinsurance and its CCIP wording. Annual premium was to be $4 million, and the sum insured was $500 million. “INFORMATION” stated “2000/2001 Underwriting Submission seen by Reinsuring Underwriters and retained at Aon”. The covering fax concluded:

Among the attachments was one relating to KCM giving relevant financial details about it, and containing a line stating “Sublimits tba”.

Whether that slip was scratched or not, a subsequent revised slip (also dated 19 June 2000 but reformulated on 8 July) was scratched by Swiss Re’s Mr Cochrane on 13 July 2000 for 26% and in due course by the other Reinsurers for their respective percentages. That slip was in what is for present purposes an identical form, save that it was for only 24 months. Coromin submits that the scratching of that slip bound the Reinsurers to reinsure KCM on all risks terms subject to the CCIP wording, including the DIC clause and the English jurisdiction clause.

There were two subsequent endorsements concerning KCM in particular. The first was dated 8 January 2001 “Attaching to and forming part of slip dated 19^th June 2000” and agreed certain sub-limits in respect of refurbishment, course of construction and minor works. The second was dated 5 February and scratched 12 February 2001 “Further to agreement dated 8 January 2001” and agreed a detailed list of KCM sub-limits “as attached”. Attached was a two page document, each page also scratched. The first reproduced the financial information originally contained in the KCM profile attached to the 19 June 2000 slip, but amended to contain the additional line: “Coverage All Risks of Physical Loss or Damage”. The second contained the sub-limits and an additional paragraph on Business Interruption. Coromin says that these endorsements are entirely consistent with KCM being part of the all risk global policy. To the extent that previously sub-limits had yet to be agreed, there were none.

The Reinsurers’ interpretation of these documents is different. Just as the previous year’s reinsurers, led by Swiss Re, had carved out a special position for KCM for the three month period, albeit as an endorsement to the 1999/2000 Reinsurance, so the same had happened for the purposes of the renewal for 2000. So far as KCM was concerned, the “as expiring” renewal was based on the KCM wording, viz named perils cover only, no DIC clause, and Zambian jurisdiction. It could not be supposed that Swiss Re, who as recently as 25 May 2000 had agreed on the KCM wording, could have intended to renew cover for KCM on an all risks basis when it scratched the renewal slip on 13 July 2000, less than two months later. Just as KCM’s insurance at the local primary level and at the Coromin reinsurance level was treated differently, namely as named perils cover with a local law and jurisdiction clause, so the same was true at the ultimate reinsurance level. That after all was Coromin’s primary case in response to KCM’s claim in England. Mr Cochrane had never seen the 2000 Underwriting Submission, whatever his scratch might otherwise suggest. In any event, the sub-limits dealt with by the January and February endorsements only made sense on the basis that the KCM wording applied. As for the reference to “All Risks of Physical Loss and Damage” in the attachment to the February endorsement, that was equally consistent, at any rate as a form of shorthand, with the KCM wording since the three month slip referred to “Assets All Risks” but continued “limited to specific insured perils”. In any event, if it had a different meaning such as that relied on by Coromin, that was a change that unfairly had not been pointed out to Mr Cochrane.

The judge was faced by three arguments. The Reinsurers submitted (i) that their reinsurance contract was governed by an exclusive Zambian jurisdiction clause and so on that ground the Part 20 proceedings should be stayed in favour of Zambian proceedings; (ii) that in any event there should be a case management stay of the Part 20 proceedings until the issues between KCM and the Zambian underwriters in Zambia and between KCM and Coromin in England had been determined; and (iii) that alternatively there should be a stay of both the Part 20 proceedings and the hearing of the Reinsurers’ application until those issues raised by KCM had been determined. The first submission argued for a permanent stay; the second and third submissions argued for a temporary stay on the ground that the Part 20 proceedings were premature and might never need determining.

The judge rejected all three submissions. As for the first, he applied Waller LJ’s gloss in Canada Trust Co v. Stolzenberg(No2) [1998] 1 WLR 547 of the “good arguable case” test (ex Seaconsar Far East Ltd v. Bank Markazi[1994] 1 AC 438) in the context of a choice for founding jurisdiction as requiring that “one side has a much better argument on the material available” (at 555). He applied that test to the mirror context of declining jurisdiction, and found that the Reinsurers had failed to meet it. He said (at para 68):

That finding ended the Reinsurers’ attempt at obtaining a permanent stay of the Part 20 proceedings on jurisdictional grounds.

The judge went on, however, in an obiter part of his judgment to consider what might be the position if he had concluded to the contrary that the Reinsurers had made out a good arguable case, that is to say at the provisional stage had succeeded in making out a much better argument in favour of the incorporation of the KCM wording including the Zambian jurisdiction clause. He concluded, nevertheless, that as a matter of his discretion there were here strong reasons (see Donohue v. Armco) for declining to give effect to the Zambian jurisdiction clause (at paras 102/112), albeit he found it to be an exclusive jurisdiction clause (see paras 69/73). Along the route to his conclusion, he considered the question of whether, in the light of the European Court of Justice decision in Owusu v. Jackson, it was open to the court to decline in its discretion to enforce the Zambian jurisdiction clause in circumstances where, had he been considering an exclusive jurisdiction clause in favour of the courts of one of the state parties to the Regulation, he would have been bound absolutely to give effect to it under what is now article 23 of the Regulation. He concluded that Owusu v. Jackson did not apply to exclusive jurisdiction clauses in favour of the courts of a state not party to the Regulation and that therefore he retained the Donohue v. Armco discretion to decline for strong reasons to give effect to an exclusive jurisdiction clause (at paras 74/101).

There is no appeal by the Reinsurers from that last ruling. They do, however, appeal his decisions that (a) there was no good arguable case in favour of the Zambian jurisdiction clause and (b) that as a matter of discretion he would not have given effect to such a clause in any event. On the other side, Coromin dispute by way of respondent’s notice his decision that the Zambian jurisdiction clause was an exclusive clause.

As for the judge’s reasons for exercising his discretion, they can be expressed as follows. First, if a stay were granted, Coromin would be unable to join Aon to its Part 20 proceedings, even though it needed to do so before a limitation period expired. Secondly, neither the Reinsurers nor Aon would be bound by the outcome of KCM’s claim against the Zambian underwriters in Zambia, unless they were joined in those proceedings, something which was uncertain and lay in the future. Thirdly, there was no basis for preventing KCM from pursuing its English claim against Coromin and that trial could be completed in 2006. Fourthly, since KCM’s claim, in one of its alternatives, involved a claim under the DIC clause, it would be necessary to decide in England (as well as in Zambia) whether the incident was a collapse or a landslip: but that might be done as easily in England as in Zambia without substantial costs implications. Fifthly, the probability was that both primary claim and the Part 20 proceedings (against the Reinsurers and alternatively against Aon) would be completed in London in 2006 before a trial would be completed in Zambia. In those circumstances, it was unlikely that the claim in Zambia would be pursued and the prospects were that all parties, including the Zambian underwriters themselves, would settle all outstanding issues on the basis of the English court’s judgment. In that way the risk of conflicting decisions would be avoided.

The judge therefore concluded as follows (at 112):

The judge therefore decided the question of a permanent stay on jurisdictional grounds against the Reinsurers both as a matter of principle and as a matter of discretion. The judge then turned, albeit very briefly, to the Reinsurers’ two fall-back positions that a temporary stay should be granted for case management reasons and rejected them for essentially the same reasons. He said (at para 114):

Subsequently, a window for trial was fixed for June 2006, although there are now some arguable grounds for thinking that that timetable may be too tight.

On this appeal Mr Joe Smouha QC on behalf of the Reinsurers placed primary emphasis in his submissions on what had been their alternative (second and third) arguments below, namely for a temporary, case management stay of the Part 20 proceedings against them. In effect, since their third argument, for a stay of the application itself (as well as of the Part 20 proceedings) had been consumed in the process, the weight of his appeal was put behind the submission that the Part 20 claim against the Reinsurers should await the outcome of KCM’s claims against the Zambian underwriters and/or Coromin, and in particular the former.

A difficulty about a stay pending the outcome of KCM’s claim against Coromin in England is that that would leave Coromin exposed were that claim to succeed in circumstances where Coromin would say that its reinsurance contract with the Reinsurers was intended to provide an indemnity. It would also leave it exposed to the possibility of inconsistent judgments (on KCM’s claim and Coromin’s Part 20 claim respectively) unless the Reinsurers would agree to be bound by any decision in the former. They did not say that they agreed to be bound, while asserting that in practice they would be. However, the 2000 reinsurance contract, whether in the form of the CCIP wording or the KCM wording, does not contain a follow the settlements clause; although it does contain a claims cooperation clause. It would therefore seem to be the Reinsurers’ intention to seek to control Coromin’s defence, while at the same time standing formally aloof from the proceedings by way of the stay applied for. Of course, were KCM’s claim against Coromin to fail, then it would be true that there would be no need for any Part 20 proceedings.

It was perhaps for these or some such reasons that Mr Smouha emphasised a temporary stay pending KCM’s claim in Zambia. The essence of his submission was that until it was known whether or not the incident at the mine was or was not a collapse, everything else was premature. That issue could only be decided, or at any rate was most easily and efficiently decided, under Zambian law in the Zambian courts. That was, factually speaking, the big issue: in terms of investigations, expert witnesses, cost, time, and the potential wastefulness of duplicative investigation. It was also the key issue to the whole litigation. A decision on that issue would be effectively, if not formally, binding on all parties, none of whom would have any wish to relitigate it in London. It did not arise in the London litigation between KCM and Coromin, only potentially albeit indirectly in the Part 20 proceedings (since the Reinsurers were saying that their liability was only on a named perils basis). In any event, Coromin, the Reinsurers and even Aon could be joined, if necessary, to the Zambian proceedings.

Thus the only risk of inconsistent judgments was between a decision on the collapse issue in Zambia and in the Part 20 proceedings. But that risk could be efficiently managed and practically eliminated by staying those proceedings in London pending a decision in Zambia. There was no reason to think that a trial in Zambia would take any longer to reach and conclude than the trial in London: on the contrary, if the litigation here was to proceed in all its aspects.

Turning to the reasons given by the judge for his decision, Mr Smouha submitted that he had given no effective separate consideration to a temporary, case management, stay. As for his reasons for rejecting a jurisdictional stay: the first, the need to join Aon, had already been achieved and could always have been permitted pending a stay; the second, the need to bind other parties to a Zambian court decision, could be achieved formally by joining them to the Zambian proceedings and would in any event be achieved in practice; the third, the inability to stay KCM’s English claim against Coromin, was not a vital point since that claim had no overlap with the Zambian proceedings and was in any event a necessary precondition to the Part 20 proceedings; the fourth, that the collapse issue had to be decided in England as part of KCM’s alternative DIC claim, was a mistake which had entered into the judge’s third reason as well – neither KCM nor Coromin were asserting that there was no collapse, it was simply that a (Zambian) decision that the loss was not within the named perils cover was a precondition for triggering the DIC claim; the fifth, as to the relative timings of litigation in Zambia and England, was unjustified on the evidence and was in the process of being shown to be an error.

In sum, in circumstances where KCM claimed on a named perils basis, where the Zambian underwriters accepted (under the consent orders) that they had insured 100% on that basis, where Coromin accepted that it had reinsured the Zambian underwriters for 90% on that basis (so much so that the Zambian underwriters saw no need to resort to litigation against Coromin), and where the Reinsurers accepted that they had reinsured Coromin on that basis, the “collapse” issue was the key issue in the case. It was an entirely Zambian centric issue, and, albeit KCM’s all risks claim was some $2.6 million greater than its $47.9 million named perils claim, success for KCM on the Zambian claim would be the practical solution to the litigation as a whole. The Part 20 proceedings were premature and should be stayed.

On behalf of Coromin, Mr Steven Berry QC had three main themes. The first was that it is very difficult to upset on appeal the commercial judge’s exercise of his discretion in such a matter, a fortiori in the context of a decision of case management. The leading authority in such a context is Reichold Norway ASA v. Goldman Sachs International [2000] 1 WLR 173, where Lord Bingham CJ said that “stays are only granted in cases of this kind in rare and compelling circumstances” (at 186). The second was that in any event the judge was right to exercise his discretion in the way he did for the reasons that he did. He had taken all relevant matters into account, in particular the (assumed good arguable case for the) application of the Zambian law and jurisdiction clause and the evidential centre of gravity of the “collapse” issue (at para 112). If his decision on discretion was good in the context of the application for a jurisdictional stay, it was all the more unassailable on the hypothesis of a case management stay unassisted by the jurisdictional arguments.

His third main strand was to underline Coromin’s preferred way of looking at the litigation as a whole: namely, not through the lens of KCM’s claim against the Zambian underwriters on a named perils basis, but through the lens of KCM’s claim against Coromin on the all risks basis. That after all was the English claim. It relied on a contract made or negotiated in England, incorporating ex hypothesi an English law and jurisdiction clause, with back to back reinsurance from the Reinsurers on the same basis. Why should such a claim or the Part 20 claim over for an indemnity from the Reinsurers be made subservient to the Zambian claim? In any event the issues were interwoven and could not be disentangled. And if the contingent claim against Aon was to be progressed, it were better that it should be done before any more time had been allowed to lapse since the events of 2000.

The first matter which has to be decided is whether to give permission to appeal to the Reinsurers on their case management challenge (their grounds five and six). Waller LJ, who gave permission to appeal on their jurisdictional challenge (grounds one to four), adjourned the application on grounds five and six for the full court. Because the two forms of challenge are not entirely distinct from one another and also because this court has heard very extensive argument on all the grounds, I would extend permission to appeal to grounds five and six and treat the Reinsurers’ submissions on those grounds as part of their appeal.

If the case management challenge is looked at strictly on its own terms, then it seems to me that it would be wrong to interfere with the judge’s discretion. This is in the first place because the case management challenge is made on the basis that the jurisdictional challenge to the Part 20 proceedings has failed. In any event the KCM claim in England is properly based here jurisdictionally, is a case in good standing (see above), and it is a claim based on an English law and jurisdiction contract at least in part made or negotiated here through Aon. The same can of course be said about Coromin’s Part 20 claim against the Reinsurers. The most that can be said on behalf of the Reinsurers is that, albeit as an endorsement to a global reinsurance contract itself made and negotiated in England through Aon and governed by an English law and jurisdiction clause, it is asserted that the extension of that contract to KCM was separately negotiated on KCM wording incorporating a Zambian law and jurisdiction clause. For these purposes there is no need for a choice between the contentions in the sense of finding that one party’s version is to be characterised as a “much better argument” than the other. Both KCM’s and Coromin’s claims are jurisdictionally soundly based here. There are competing submissions as to both claims with Coromin of necessity having to face both ways. It is undesirable for the court at an interim stage to say anything about the competing value of those claims, unless either it is forced to take a provisional view (as it may have to do for the purpose of a jurisdictional challenge, see Canada Trust) or unless one or other party undertakes the severe burden of seeking to show that its opponent’s version has no real prospect of success. Thus it may be that both versions, considered separately, constitute good arguable cases. Where a choice between such versions does not have to be made (and should not be made) it is perfectly possible to think of each version as being a good arguable case. I would so regard the respective submissions set out above.

Therefore the background to the Reinsurers’ application for a case management stay is first the hypothesis that their jurisdictional challenge has failed and secondly or at least that there is no reason to prefer their analysis of the contractual context to that of KCM and/or Coromin.

Secondly, it is common ground that Reichhold lays down the relevant test. Thus it is accepted that a case management stay is possible, but also that it requires rare and compelling circumstances. In that case this court upheld the commercial judge’s decision to order a case management stay of an English action pending the determination of related arbitration proceedings. However, the claimants in both the action and the arbitration were the same parties making essentially the same claims: in the arbitration they claimed against the Norwegian seller of a company under a Norwegian law contract in respect of that seller’s contractual warranties; and in the action they claimed against the seller’s agents, Goldman Sachs, in tort for negligent misstatement. The judge described the two sets of proceedings as concurrent with a significant degree of overlap. Reichhold, the claimant in each, provided the judge with no reasons for the practical advantages of pursuing the action in advance of the arbitration nor put forward any prejudice which it would suffer if there was a brief delay which could be adequately compensated in interest.

That case may be compared with the present, where Coromin is not making concurrent claims but responding to a claim by seeking to pass it on to a reinsurer; where, if Coromin were to be found liable for KCM’s all risks claim, it could be severely prejudiced if it could not in the same trial be entitled to seek to pass it on to the Reinsurers (or to Aon); and where if KCM succeeded in Zambia in proving a collapse within the named perils cover but the Reinsurers wished to dispute that result, Coromin could be similarly prejudiced. In its defence to the Part 20 proceedings, the Reinsurers do dispute any liability by Coromin to KCM: they say that there was no direct all risks insurance by Coromin of KCM; that any such direct insurance would have been illegal, and that in any event the incident at the mine was not a collapse.

Thirdly, the fact is that it would therefore be unfair to Coromin to stay its Part 20 claim, even temporarily, and thus leave it exposed to KCM’s claim: unless either (1) it was possible to stay KCM’s English claim, as well as Coromin’s Part 20 claim, pending the resolution of KCM’s Zambian claim, or (2) there was good reason to believe that KCM’s real interest was in its Zambian claim and that it was itself uninterested in carrying forward its English claim pending the resolution of its Zambian claim. However, the opportunity to stay KCM’s English claim was lost when the judge decided to allow the consent orders to lie undisturbed and was unimpressed by the Reinsurers’ invitation to stay the proceedings as a whole – from which there is no appeal. And there is no evidence that KCM’s primary interest is in its Zambian claim – on the contrary, as stated above, it asserts that its interest is in carrying forward its “primary claim” in these proceedings and that is what has been happening in the commercial court pending this appeal.

Fourthly, for the same reasons it would be wrong to look at the litigation by concentrating simply on the Zambian claim, as much as it would be wrong to concentrate simply on the English proceedings. It simply is not feasible at this stage to anticipate the course of the Zambian proceedings, and therefore it is not really profitable to speculate on certain contingencies. It is of course possible that the Zambian claim (a) might be prosecuted speedily; (b) might result in a victory on the “collapse” issue for KCM; so that (c) the litigation as a whole might at that stage be settled. Equally, however, it is possible that (a) the Zambian claim might run into difficulties, eg because of the allegation that any Zambian insurance beyond that of 10% is a sham, or because of the need for an appeal; (b) the claim might be lost on the “collapse” issue, thereby promoting the need for KCM to develop its all risks claim; and/or (c) KCM will in any event wish to push ahead its (larger) all risks claim in England. Similarly it is possible that the other parties, such as Coromin, the Reinsurers, and/or Aon might be joined to the Zambian proceedings: but that is mere speculation, for KCM has shown no sign of wishing to take any steps to do so, nor has Coromin shown any signs of initiating litigation in Zambia against KCM, nor has any evidence been laid before the court of how the foreign parties could be joined to the Zambian proceedings.

As for the English proceedings, although I have already stated that it would be wrong to consider them in isolation from the overall situation, the judge did not make that error. But he was at least entitled to consider the ramifications of those proceedings, and his instincts as to how they might progress would at any rate be better founded than in the case of the Zambian proceedings. Thus he was, in my judgment, entitled to think that they could be efficiently processed; that it would be unfair to leave Coromin exposed without being entitled to progress its Part 20 claim against both the Reinsurers and Aon; that the validity of the Zambian claim under the named perils insurance would arise as an issue, whether that be in the course of KCM’s claim against Coromin (eg, as I would suggest, because of the “other insurance” clause) or (as the Reinsurers themselves accept) because of their defence that KCM has no primary claim of any kind to make, a defence that Coromin would be forced to make in its defence as well; and that without permitting the Part 20 proceedings to progress together with KCM’s all risks claim there would be the danger of inconsistent decisions within the English proceedings on top of the danger of inconsistent decisions between the English and Zambian proceedings. This court can do nothing about the latter, but I share the judge’s concern not to take a step (that of staying the Part 20 proceedings) which would aggravate the risk of the former.

Moreover, so far as Aon is concerned, although the risk that a claim against it would be time barred has now been overcome, there is no application before the court which would permit us, in the absence of Aon, to stay Coromin’s claim against it. It would not, however, be fair or efficient to allow the Part 20 claim against Aon to go forward in the absence of the Part 20 claim against the Reinsurers. These two claims have to be heard at the same time, otherwise neither claim (or defence) can be fairly conducted: they are closely intertwined. Thus they must either proceed together, or be stayed together. Even though a stay of the proceedings against the Reinsurers might be a prelude to a stay against Aon, that cannot be a reason for taking a critical decision affecting Aon in its absence. In any event I agree with Coromin’s submission that delay in the advancement of that claim is undesirable.

In sum, for all these reasons, which in substance duplicate (or add to) the reasons given by Colman J, I would agree with him that a case management stay is not in the interests of justice. In any event, the question is not whether this court agrees with the judge, but whether his discretion should be interfered with. I can see no basis upon which it would be right to do so. I would therefore reject the Reinsurers’ appeal on the grounds that there should be a temporary case management stay.

For the same or much the same reasons I can see no basis upon which it would be right for this court to interfere with the exercise of the judge’s discretion not to enforce a Zambian exclusive jurisdiction clause even on the hypothesis that the Reinsurers had established a much stronger argument in favour of the applicability of such a clause to Coromin’s Part 20 claim. The critical factor here, which brings the other reasons in its train, is that the KCM claim is established in this country as a claim in good standing, based on an alleged insurance contract containing an English law and jurisdiction clause. In these circumstances, the potential bifurcation of KCM’s claims in two separate jurisdictions, with its attendant dangers, is, as I have said above, a given. The problem for the English court is how best to overcome those dangers, or, to put it another way, how to reach the least bad of the remaining options. At least in this country it has been shown to be possible to assemble as litigants all those involved – except of course the Zambian underwriters. Nevertheless, the issue which affects those underwriters, the “collapse” issue, will arise in England at the very least, as Mr Smouha himself accepts, on the Reinsurers’ defence. The judge was entitled to think that in such a situation the jurisdiction in which by a process of litigation, assisted, I might add, by a process of negotiation, the parties might find a solution to their problems as a whole would or could reasonably be thought to be England. As he said (at para 111):

There is one other factor I would stress in this context, even on the present assumption that the Reinsurers are entitled to say that they have a much stronger argument in favour of a Zambian jurisdiction clause. The position is not only that KCM is bringing a claim here against Coromin on the basis of an all risks policy including an English law and jurisdiction clause, but also that, subject to the court’s power to enforce the Reinsurers’ Zambian jurisdiction clause (at best a merely provisional finding at this stage), jurisdiction in England over each of the Reinsurers has been established under the provisions of the Regulation or (in the case of the Swiss Reinsurers such as Swiss Re itself) the Lugano Convention. Such jurisdiction is mandatory under article 2 itself in the case of the English reinsurers, or has been established under article 6 (as an alternative to suit in their own domiciles) in the case of other EU or Swiss reinsurers. Thus jurisdiction over all the Reinsurers is established quite irrespective of any reliance on an English jurisdiction clause under article 23 of the Regulation (or article 17 of the Lugano Convention). The judge has held, in the course of his consideration of Owusu, that article 23 (article 17) does not apply to jurisdiction clauses in favour of the courts of states not within the EU (or Lugano Convention states). Therefore the effect of a provisional preference for a Zambian jurisdiction clause is a matter entirely for the English common law and not for the Regulation (or Lugano Convention). There is no appeal from that decision. The question therefore of whether there are strong reasons for not enforcing a provisionally found Zambian jurisdiction clause has to be determined against that background. There are two aspects of this point. Jurisdiction in England has been established (subject to it being declined on the basis of a foreign jurisdiction clause); and that basis for declining jurisdiction is not itself an established fact (as it typically is, see, for example The El Amria [1981] 2 Lloyd’s Rep 119), but a merely provisional preference. The court has not been cited any case in which the question of whether there are strong reasons for not enforcing an exclusive jurisdiction clause has been determined against such a background. I will return to this matter below in considering whether there is a good arguable case or much the better argument in favour of the Zambian clause.

In my judgment, this background is significant. The strength of a strong reason for not enforcing a foreign jurisdiction clause can only ultimately be evaluated against the background of the strength of the argument in favour of that clause (or the strength of the argument in favour of English jurisdiction). Thus even on the assumption of a provisional finding in favour of the incorporation of a Zambian (rather than an English) jurisdiction clause, the judge was entitled to have regard both to the otherwise clearly established jurisdiction for Coromin’s Part 20 claim and to the clear dangers for Coromin of facing KCM’s claim without an opportunity to pass it on to its Reinsurers.

In sum, in the light of the complexities of the issues and the manifold possibilities which the future might bring, and given the assumption for present purposes of a finding of a much better argument in favour of a Zambian jurisdiction clause, it is easy to contemplate that judges might differ over the solution to the present conundrum. However, the question is whether Colman J has gone wrong in principle, or come to an unreasonable solution, or has misled himself by either omitting an important consideration or taking materially into account something irrelevant or wrong. I am not satisfied that he has. In a situation where the least bad solution is being sought, I do not think that the judge has erred in the exercise of his discretion. On the contrary, I think he was well entitled to consider in his discretion that a permanent stay on jurisdictional grounds would carry with it a great danger of injustice to Coromin, and that the interests of justice required, to quote again from his para 112, that “conflicting decisions on key issues should be avoided” and that counter-arguments were outweighed by the desirability of having the “issues between KCM, Coromin, the Reinsurers and Aon all decided by one tribunal and in proceedings in which they are all entitled to participate”.

There remains the formal basis of the Reinsurers’ jurisdictional challenge, logically the first point but relegated by Mr Smouha to subsidiary status. This challenge has gone through some reformulation, although the essence of it is that the Reinsurers’ case in favour of the KCM wording with its Zambian jurisdiction clause is to be preferred to Coromin’s case in favour of the global all risk policy’s CCIP wording with its English jurisdiction clause.

The reformulation has two aspects. The first is as to who bears the burden of proof. Before Colman J and in their original skeleton argument for the appeal, the Reinsurers accepted that if their application for jurisdiction to be set aside was to succeed, it was they who bore the burden of proof (or persuasion) that their case was to be preferred. They now wish to submit that the burden remains on Coromin as the party seeking jurisdiction in England. The second is as to the nature or standard of that burden, whoever bears it. They complained in their skeleton argument that the judge had adopted the wrong test: “Reinsurers do not have to establish that they have a much better argument than Coromin, merely that on the balance of probabilities it is more likely that the KCM wording applied.” In other words they suggested that the Canada Trust test adopted by the judge was an even more severe test (“more onerous”) than that of the normal standard of proof in civil proceedings, that of the balance of probabilities. They subsequently abandoned that position, however, in my judgment rightly so, and submitted that the judge had erred in his application of the Canada Trust test, not only by reversing the burden of proof but also by finding merely that Coromin’s case “appears at this stage to be stronger” than the Reinsurers’ case: whereas in Canada Trust Waller LJ had spoken of the need for “a much better argument”.

Since I would uphold the judge’s solution to the Reinsurers’ jurisdictional challenge even on the basis that they have much the better of the argument as to the applicable wording and jurisdiction clauses, there are two reasons why I prefer to be less than full in debating the parties’ respective contractual submissions under this heading. One is that this topic is not necessary to the decision on this appeal. But the other is that at this interim stage it is desirable to say as little as possible about what is a crucial issue between Coromin and the Reinsurers, namely whether Coromin was reinsured on the named perils basis or on the all risks basis. This is because, as sometimes occurs, a jurisdictional issue involves an argument which goes to the heart of the ultimate merits of the case.

The problem in such a situation is to find a test which answers the jurisdictional issue without trespassing further than is necessary on the merits. It was for this purpose that a test which was intended to be less onerous than the ordinary civil standard of proof on the balance of probabilities was adopted. Waller LJ spoke of this problem in his judgment in Canada Trust (at 555/556) in a lengthy passage cited by the judge (at para 56 of his judgment), from which I would extract the following:

Waller LJ then went on to speak of the flexibility of the “good arguable case” test. Thus he contrasted the case where the jurisdictional factor, viz breach of contract within the jurisdiction, could raise an issue whether there was a contract at all, with the different case, where the jurisdictional factor, viz domicile of the defendant, does not enter into the merits of the parties’ dispute and so will not arise at trial at all. In the former case the court, although needing to be satisfied as to the existence of the contract, will not wish to give “even the appearance of pre-trying the central issue” while scrutinising “most jealously” the factor which provides jurisdiction, whereas in the latter case that scrutiny will be affected only by the limitations of an interlocutory process rather than by any inhibitions as to trespassing on to the merits of an ultimate trial.

Canada Trust itself was of the latter variety, for it involved an issue as to domicile.

The present case, however, while closer to the former variety, is either an extreme form of it or, as I would prefer to think, a third variety where the jurisdictional factor is part and parcel of the ultimate merits of trial. On the structure of the argument presented before Colman J and in the parties’ original skeleton arguments for appeal, the issue, Zambian or English jurisdiction clauses, was simply one aspect of the broader issue: KCM wording (named perils cover) or CCIP wording (all risks cover). But that is the issue at trial between these parties (and the “collapse” issue only becomes relevant if the reinsurance contract is on the KCM wording). In such a case it seems to me that Waller LJ’s warnings about avoiding even the appearance of pre-trying the central issue move to centre stage.

What is to be done in such a case? One possibility is that the Canada Trust gloss (“a much better argument”) of the “good arguable case” test is not really appropriate in such circumstances. It is quite possible, especially at the opening, jurisdictional, stages of a dispute for both sides to have a good arguable case as to the central merits of a dispute. I would regard the present case as being a good example of just such a dispute. It is after all familiar enough, even at the end of a trial, for a judge to think that the merits of the opposing parties are fairly evenly balanced. If a court, in applying the good arguable case test, as well as taking account of the opposing arguments as it has always done, in addition had to decide and rule as to which side had the “much better argument”, I fear that, however much the judge couched his reasoning in terms of the provisional nature of his decision on the material available, he would inevitably be drawn into a trial of the merits. This is plainly undesirable. Is it necessary? I am doubtful that it is, especially in circumstances where, as was generally the case under the old RSC Order 11 or is now the position under CPR 6.20, the power to give leave to serve out of the jurisdiction is at the end of the day a discretionary one. However important the proper disposition of a jurisdictional challenge is, it is not something which should be allowed to subvert the merits of a potential trial.

In this context, moreover, it is to be noted how special the jurisdictional factor of domicile (the immediate factor in Canada Trust) is in the case of the Regulation. Not only is that a factor which, generally speaking, is irrelevant to the merits at trial, but it is also, under the Brussels and Lugano Conventions and now under the Regulation, a factor which, subject to the exceptions contained in those instruments, leads to mandatory consequences.

As it is, however, we are not in this case even concerned with the establishment of jurisdiction in England. It is to be recalled that all that Waller LJ said was in that context. As I have observed above, jurisdiction over the Reinsurers in England has been established, and the question is whether nevertheless that jurisdiction is to be set aside on the ground that there is a better claim to exclusive jurisdiction elsewhere. In the typical case outside the Regulation the exclusive jurisdiction clause is established and the only question is as to the exercise of discretion as to its enforcement in circumstances where the principle is that in general the clause will be enforced subject to strong reasons to the contrary. What then is to be the test where there is a dispute as to the existence of the jurisdiction clause relied upon? The judge was content to assume that for these purposes the Canada Trust gloss of the good arguable case test applied as well (at para 57: “Thus in both cases there is a pre-trial investigation ancillary to the main action as to which the court should avoid any final determination of any issue of fact or law going to the substantive rights of the parties which will have to be conclusively determined at the trial”). He therefore asked himself whether there was a better case for the Zambian clause or for the English clause. The parties on appeal have squared up to this analysis, with the Reinsurers submitting that Coromin should be regarded as failing the test, since (a) it bears the burden of proof, and (b) had failed to show, even on the reasoning of the judge, that it had “a much better argument” (emphasis added).

It seems to me that the issue of the right test to apply to this question is a difficult one. On the one hand, one would wish to avoid a test which sets up a trial on the merits, even if only a provisional one. On the other hand, where there is established Regulation or Convention jurisdiction subject to a jurisdiction clause, the choice is not merely between two competing jurisdiction clauses, but between the established jurisdiction and a jurisdiction clause derogating from that jurisdiction. Since an English jurisdiction clause does not here derogate from the established jurisdiction, the real competition in this case is between England as the established jurisdiction and the Zambian clause relied on by the Reinsurers. Moreover, there may be a difference between a jurisdiction established by mandatory international rules as under the Regulation and a jurisdiction merely accepted by the English courts as part of their national rules, including their long-arm statute, and always subject to their overriding and inherent discretion. Where the choice arises under the Regulation itself and article 23 is in question – which Colman J has held is not the case here in respect of the Zambian clause since that is not a clause giving jurisdiction to “a court or the courts of a Member State” (article 23.1) – article 23 itself lays down the conditions for the clause in question, eg that it should be “in or evidenced in writing” (article 23.1(a)); and ECJ jurisprudence has held that it is necessary for the applicability of the jurisdiction clause to be without doubt, in the sense that it should be established that the parties’ consent has been clearly and precisely demonstrated: see Etsasis Salotti di Colzano Aimo e Gianmario Colzani v. Ruwa Polstereinaschinen GmbH [1976] ECR 1831 and Partenreederei MS Tilly Russ v. Haven & Vervoebedrijf Nova NV [1985] QB 931. By analogy, therefore, there is an argument for saying that the test for the applicability of the clause in this case should be a stringent one.

I am, however, reluctant to decide this issue in a case where a decision is not necessary. No authorities have been cited which determine it, although, as will appear below, it appears to have been assumed that the Canada Trust test applies. In this context, the judge’s analysis was more subtle than the parties have given him credit for. He said (at para 57):

Thus it is possible that, given the flexibility of the “good arguable case” test, the answer could simply be that the applicant should make out a case which is sufficient in the circumstances to render it just to derogate from the established jurisdiction, but which still remains short of proof on the balance of probabilities. If, therefore, in terms of a provisional argument at an interim stage which properly stops short of a trial mode, the applicant fails to make out such a case, then, because he bears the burden of proof, he fails.

In such circumstances, the question of burden of proof may be a significant one. Who, therefore, for such purposes is the “applicant” and who bears the burden of proof?

On this question, Mr Smouha relied on a decision of Mr Richard Siberry QC sitting as a deputy high court judge in Carnoustie Universal SA v. International Transport Workers’ Federation[2002] EWHC 1624 (Comm), [2002] 2 All ER 657. There a settlement agreement of a Finnish union’s strike action provided for the jurisdiction of the Finnish courts. The claimant shipowners commenced two actions in England, alleging that the settlement agreement was void or voidable for duress, and sought to join the union to the first and issued directly against the union in respect of the second. Another defendant to both actions was domiciled in England. The shipowners therefore relied on article 6(1) of the Brussels Convention to establish jurisdiction against the union. Almost simultaneously with issue of the second action, the union commenced its own action in Finland against the shipowners. The union challenged jurisdiction, relying on the Finnish jurisdiction clause and an argument that the Finnish court was first seised. The shipowners argued that (a) service under their action preceded service under the Finnish action so that the English court was first seised; and (b) the Finnish jurisdiction clause did not encompass the issue under their English action that the settlement agreement was void or voidable for duress. Jurisdiction in England was upheld and the union’s challenge failed.

The jurisdiction issues therefore did not encompass the ultimate merits of the dispute, as in the present case. It was “common ground that the applicable standard of proof was that of a good arguable case, and that this required less than proof on the balance of probabilities” (at para 38). However, the question of burden of proof was debated. Each party asserted that the burden of proof was borne by it, on the paradoxical basis that the party who bore a burden which was less than that of the balance of probabilities had an advantage in a close fight. The deputy judge held that the burden of proof always rested on the party seeking jurisdiction in England. He said (at para 45):

We were also referred to two other cases. In Provimi Ltd v. Roche ProductsLtd[2003] EWHC 961 (Comm), [2003] 2 All ER 683 (Comm) a claim was brought against a number of defendants domiciled in England and other EU or Lugano Convention states, relying on articles 5(3) and 6(1). French, German and Swiss jurisdiction clauses were relied on, but, as in Carnoustie, the judge, Aikens J, held that the scope of these clauses did not extend to the claims in dispute and jurisdiction in England was upheld. Once again, the matters in issue did not relate to the essential merits of the dispute for trial. Aikens J said (at para 55):

He also went on to adopt the Canada Trust test as the standard of proof, but again that would seem to have been common ground.

In Bank of Tokyo-Mitsubishi Ltd v. Baskan Gida Sanayi ve Pazarlama AS[2004] EWHC 945 (Ch), [2004] 2 Lloyd’s Rep 395 Lawrence Collins J was again dealing with the question (inter alia) of whether a foreign (Italian) jurisdiction clause displaced jurisdiction in England. While referring to Canada Trust, he rightly, as it seems to me, said (at para 193) that –

He continued (at para 194):

I would seek to sum up these authorities, which are all at first instance, in this way: that where an established Regulation (or Convention) jurisdiction in England is challenged under article 23 (or article 17), (1) there are conflicting views as to where the burden of proof lies (there is a decision in Carnoustie that the burden remains on the claimant, a decision in Knauf (at first instance) that it is on the defendant, and a view in Bank of Tokyo-Mitsubishi also to the latter effect); (2) that the standard of proof has not been settled, but that there is a general tendency to apply the good arguable case test in a form which is more or less consistent with the Canada Trust gloss, but that question was expressly reserved in this court in Knauf; and (3) that no case cited to us has dealt specifically with either aspect of the present case which is of particular interest here, namely (a) a situation where the foreign jurisdiction clause is not within article 23 (17), and (b) the jurisdiction clause issue goes to the heart of the ultimate merits at trial.

As for the difference of opinion at first instance on burden of proof, I would hazard the opinion, without seeking to decide the issue, that the views of David Steel J and Lawrence Collins J are to be preferred. It seems to me to be counter-intuitive to think that, where a statutory jurisdiction has been established but an exceptional jurisdiction elsewhere is put forward based on a contract which must be clearly shown to have the assent of both parties, it remains the burden of the claimant to prove a negative rather than that of the applicant who challenges the established jurisdiction to prove that he is entitled to rely on the clause in question. After all, article 23 comes in a section of the Regulation (section 7) called “Prorogation of Jurisdiction”.

If I am right about this, I think it is a potentially important point in a problem such as that faced in this case. Judges often say that burden of proof is of little assistance by the end of a trial once all the evidence is in hand. Whether that is so or not, it seems to me that at a pre-trial stage, where jurisdiction is in issue, the position may be otherwise. Ex hypothesi there has been no trial and all the evidence is not in hand. If the test remains some form of the flexible good arguable case test, it will remain necessary, in considering a jurisdictional issue which goes to the heart of the merits at trial, to be sensitive to the need to avoid pre-judgment. In this case, therefore, I would prefer to say, on the documents, evidence and arguments that have been shown to the court, that both parties had made out a good arguable case, without saying which was much the better of the two. Or to put it another way, that the Reinsurers had not shown that their case in favour of the KCM wording was sufficiently good either to be preferred to Coromin’s case in favour of the CCIP wording or to displace the otherwise established jurisdiction in England. On that basis the Reinsurers, who bear the burden of giving effect to their clause, would fail in their jurisdictional challenge.

Finally, there is an additional issue introduced by Coromin on appeal to the effect that the slip or slips which constitute the 2000 reinsurance contract do not expressly incorporate a law and jurisdiction clause, but merely proceed by way of a general incorporation of expiring wording. The issue is whether that expiring wording is the KCM or the CCIP wording. Each of course contains its respective Zambian or English law and jurisdiction clause, but the slips do not expressly state that such clauses are incorporated. Mr Berry relies on a general principle of construction whereby ancillary clauses such as arbitration and jurisdiction clauses are not incorporated without express reference. On that basis, the Reinsurers’ argument for a permanent stay on jurisdictional grounds does not get off the ground. It is unnecessary, however, to rule on this new point.

It is also unnecessary to consider Mr Berry’s submission that the Zambian clause is in any event a non-exclusive one, a submission considered and rejected by the judge.

The Reinsurers’ application for a stay of Coromin’s Part 20 claim, whether on a permanent basis premised on a jurisdictional challenge in reliance on a Zambian jurisdiction clause, or on a temporary case management basis premised on the solution of a primary determination of the “collapse” issue in Zambia raised by Coromin’s claim there against the Zambian underwriters, fails on this appeal as a matter of discretion. It is unnecessary to decide whether the Reinsurers’ case in favour of a Zambian clause or Coromin’s case in favour of an English clause has much the better of the argument, even if that is the right test.

I agree.

I agree that, in addition to the permission granted on grounds one to four by Waller LJ, permission to appeal should be granted on grounds five and six as proposed by Rix LJ. I also agree that the Reinsurers’ appeal should be dismissed for the reasons given by him. As to the interesting jurisdictional questions considered but not decided by Rix LJ in the latter part of his judgment, his reasoning seems to me to be persuasive but, like him, I would prefer not to reach a final conclusion upon them until they arise for decision.