In these proceedings, commenced on 30 December 2020, the Claimant Reinsurers (“Reinsurers”) contend that proceedings brought against them in South Africa by the Defendants, and commenced there on 23 November 2020, have been brought there in breach of an agreement or agreements between them that the Courts of England and Wales should have exclusive jurisdiction.

On 25 January 2021 Reinsurers applied for an interim anti-suit injunction. The Defendants’ South African solicitors were given notice but declined to attend, stating they intended to oppose the relief sought in due course. A hearing took place before Calver J on 2 February 2021 and in a judgment delivered on that day he granted an interim anti-suit injunction restraining the Defendants from pursuing the South African proceedings until the return date.

This is the return date hearing and is therefore the first occasion on which both sides’ arguments have been advanced. The question for me is whether I should continue, or vary, or decline to continue, the interim injunction made on 2 February 2021.

The resolution of those issues is intimately bound up with (though not necessarily determined by) the construction of various law and jurisdiction clauses. I raised with the parties the question of whether I was being invited to decide those points of construction at this stage. The Claimant Reinsurers invited me to decide those issues of construction, as though this hearing included a preliminary issue on the points of construction. The Defendants’ position was that although it was, they said, unnecessary to reach final decisions on the points of construction in order to resolve the application, they had no objection to the Court doing so. Neither party suggested that there might be further evidence or argument relevant to the points of construction beyond that which was available on this application. In those circumstances, and having heard full argument on the points of construction, it seems to me both appropriate and desirable to decide the construction issues. Accordingly this judgment first sets out the contractual background and facts, then determines the points of construction, and then addresses the question of what if any antisuit injunction is appropriate in the light of the findings as to construction.

The Fourth Defendant, (“ABSA Manx”) is a captive insurer incorporated in the Isle of Man. By two policies of insurance entered into for the years 2008/9 and 2009/10, it agreed to insure the First Defendant (“ABSA Group Limited”) together with its subsidiaries (including the Second and Third Defendants, (whom I shall call together

I shall call the contracts of insurance under which ABSA Manx insured the South African ABSA entities the Original Insurances, because that is how they are referred to in the reinsurance contracts. In the Original Insurances the relevant premiums (clause 4), indemnity limits (clause 6) and deductibles (clause 7) are all expressed in South African Rand (ZAR) and the Original Insurances contain a service of suit clause nominating a South African law firm. Clause 10 of the Schedule of the Original Insurances is headed “Wording” and incorporates all terms and conditions of the Primary Reinsurances. At clause 11 of the Schedule is a pay as paid clause which states:

At clause 12, the Original Insurances contain a “Direct Claims Payment Clause” which provides, in part, as follows:

At the time of the hearing the only such contract that was in evidence was a Primary Layer Original Insurance in which each claim was subject to a deductible of ZAR 30m and a limit of indemnity of ZAR 100m. The apparent absence of any contracts of insurance for sums exceeding ZAR 100m gave rise to a submission by Reinsurers that their contacts to reinsure sums above ZAR 100m could not give rise to liability because there was no underlying insurance to be reinsured. However, on the day after the hearing the Defendants sought to adduce a series of further very similar contracts recording insurance for first, second and third excess layers above ZAR 100m, and which correspond, in terms of sums insured, to the contracts of reinsurance which I shall describe in a moment. Having given the parties an opportunity to make submission on what course should be taken, I permitted those new documents to be adduced in evidence, and allowed the parties to make further written submissions in relation to them.

Although Reinsurers formally reserved their position as to the new documents the evidence now available suggests strongly that the South African ABSA entities and their captive insurer entered into a series of contracts of insurance for an original and

At the ex parte hearing Reinsurers had proceeded on the assumption that there would be corresponding insurances for the contracts of reinsurance they had written. Now that the hiatus caused by the Defendants’ temporary inability to locate the excess layer insurance contracts has been overcome, it appears that the Reinsurers were right to infer that such contracts existed, so the upshot is that both the ex parte hearing and this hearing proceed on the same basis, in terms of there being underlying insurances with the captive insurer, and back to back reinsurances with various of the Claimants.

I turn then to the various contracts which describe themselves as reinsurance contracts, and which I shall therefore call for convenience the Reinsurance Contracts. There is a Primary Layer Reinsurance Contract, and then there are a series of three

I was shown a proposal form (in fact for the 09/10 year) signed by Absa Group Limited (D1) on behalf of the South African ABSA entities (as the “Proposer”) which seeks a limit of indemnity of ZAR 1 billion in total and in which the Proposer answers a series of question pertinent to the risks which are to be underwritten and which contains a note beneath the declaration that “if a policy is concluded it will be issued on a claims made basis ie to indemnify the Proposer for claims first made against it in the manner described in the policy…”. This proposal form (and, I infer, its 08/09 equivalent) was apparently used for all the layers of “reinsurance”.

In the primary layer of reinsurance there is a retention of ZAR 30m and a limit of ZAR 100m, matching the limit and deductible of the Primary Layer Original Insurance. The Schedule defines the Reassured as ABSA Manx and the “Original Assured” as the South African ABSA entities. It identifies an individual at a South African firm of solicitors as the person nominated to accept service of process. It states “Situation Worldwide”; and it contains a choice of law and jurisdiction clause which I shall set out later.

In the policy terms Section 2 is entitled Civil Liability Insurance and the opening words provide as follows:

The Special Conditions applying to section 2 contains the following provisions

The indemnity provided by this Section shall apply to final judgements against the Assured in the courts of any country in the world

Any legal proceedings commenced against the

There are General Conditions applicable to all sections of the policy. General Condition 5 makes clear that a third party claim is considered to be made when (broadly) a designated person at the Assured (not at Absa Manx) first receives a written demand or becomes aware that a demand might be made. General Condition 6 provides that in the event of any payment being made under the policy the Reinsurers shall become subrogated to the rights and remedies of the Assured. General Condition 13 is entitled “Interpretation: Service of Process: Jurisdiction” and begins with a choice of law clause in favour of English law. It then says

Clause 14 is entitled Direct Claims Payment Clause and provides as follows

Reinsurers accept the South African ABSA entities therefore have, in principle, the right to recover claims under the primary reinsurances directly from Reinsurers.

The Schedule contains a law and jurisdiction clause which is of course central to this dispute. I set out its terms below at paragraph 36.

The First Excess Reinsurance has a limit of ZAR 250m in excess of ZAR 100m. Its terms refer to the “Underlying Policy” which is the first layer reinsurance (not the Original Insurance). The terms provide that the policy is to indemnify the Assured (that is to say the South African ABSA entities, not ABSA Manx) inter alia for a claim or claims, as more fully defined in the Underlying Policy, first made against the Assured during the period of insurance, in excess of the amount being the subject of the Underlying Policy.

This excess layer corresponds to the first excess Original Insurance written by ABSA Manx.

Then there is Second Excess Reinsurance (for the next ZAR 650m), and Third Excess Reinsurance (for the next ZAR 500m), each corresponding with the corresponding excess layer Original Insurances between the South African ABSA entities and ABSA Manx.

In each of the Excess Reinsurances the Schedule contains a law and jurisdiction clause in the same form, which I set out below at paragraph 37.

The Defendants submit that the Original Insurances are essentially “fronting” insurance contracts with the real commercial relationship being between the Assureds (ie D1 to D3) and the Reinsurers, and they say that using a captive insurer in this way is common amongst large entities such as ABSA. They say that both the Original Insurances and the Primary and Excess Reinsurances allow the South African ABSA entities to “cut-through” by proceeding directly against the Reinsurers, effectively bypassing their captive insurer.

It seems to me that, although they are in form reinsurance contracts, the common intention of the parties was a commercial arrangement in which the South African ABSA entities could claim directly against the Reinsurers, subject to the various policy terms. The interposition of ABSA Manx as a captive insurer does not affect that, and the parties must have contemplated that any claims made under the suite of policies would always be directed at the Reinsurers in the first instance, bypassing ABSA Manx entirely, so that in practice the captive’s liability would only matter if for some reason the “reinsurance” did not respond to a direct claim.

Finally, there is a policy of Aggregate Retention Reinsurance (“ARR”) in relation to the retention of the first ZAR 30m of each claim. It would in my view be inaccurate to regard the ARR and the First Excess Reinsurance as somehow sandwiching the Primary Reinsurance, if by that was meant there was a direct analogy between the role of all three “reinsurance” policies. The ARR is not a policy which allows cut-through claims for the first ZAR 30m of each claim. Rather, it provides for a more complex scheme under which ABSA Manx can claim if the aggregate of retentions exceed various thresholds.

The ARR also contained law and jurisdiction provisions, which I set out below at paragraphs 38 to 40.

In terms of which Reinsurers were a party to which reinsurance contracts: only C1 was party to the ARR policy; each of C1 to C7 subscribed for the primary reinsurance plus at least one layer of excess reinsurance (but not necessarily adjacent layers), and they almost invariably signed each of the policies to which they did subscribe on the same date. In contrast, none of C8 to C12 subscribed to the primary layer reinsurance, although each subscribed to one or more of the excess layers, again generally signing each excess layer policy to which they did subscribe on the same date or adjacent days.

All of the reinsurance contracts were made in London.

The underlying dispute, in respect of which the Defendants now seek indemnity from the Reinsurers, arose as follows. ABSA Bank was trustee or custodian of a cash management fund which formed part of a collective investment scheme called (latterly) the Ayanda Collective Investment Scheme. During the period when ABSA Bank was trustee or custodian, the portfolio manager of the fund allegedly invested monies in illiquid promissory notes. ABSA Bank has been sued by investors in that Scheme, on the basis that the fund, which has now been wound up, was grossly overinvested in illiquid securities and the investors have suffered loss for which ABSA Bank is liable. The South African ABSA entities say that they have settled these claims, and they now seek indemnity from Reinsurers in respect of those settlements.

In the South African proceedings the South African ABSA entities claim against the relevant Reinsurers under the Reinsurances. Their primary claim is on the 09/10 policies but they claim on the 08/09 policies in the alternative. The total claim is ZAR 467m plus ZAR 60m costs, so the claim is for ZAR 100m under the Primary Reinsurance, ZAR 250m under the First Excess layer and 117m plus costs under the Second Excess layer. No claim is brought in the South African proceedings under the ARR.

Reinsurers’ essential answer to those claims (wherever they might be made), and their basis for seeking a declaration of non-liability in these proceedings, is

that the Reinsurers are not liable to indemnify the defendants, under the 2009/2010 reinsurances, as alleged in the South African proceedings, nor under the ARR, by reason of the operation of general exclusion 1(c), which excludes cover for losses or claims where a designated person was aware of a circumstance or occurrence which would cause a reasonable person to assume that a third-party claim covered by the reinsurances could be made, and if they were not disclosed to the Reinsurers at inception; and

that the Reinsurers are not liable to indemnify the defendants, or any of them, under the 2008/2009 reinsurances, as alleged in the South African proceedings, nor under the ARR, because the underlying claims were not first made and are not deemed to have been first made during the policy year of the 2008/2009 reinsurances or the 2008/2009 ARR.

I can now set out the law and jurisdiction clauses (“L&J clauses”) from the various contracts. There are three distinct wordings: the wording in the Primary Layer Reinsurance, the wording in the Excess Layer Reinsurances (all of which are the same), and the wording in the ARR policy.

In the Primary Layer Reinsurance, next to the marginal heading “Choice of Law and Jurisdiction”, the following wording appears. In the original the wording is a single continuous paragraph, but for ease of reference I have added line breaks between the sentences and numbered them:

Any dispute concerning the interpretation of the terms, conditions, limitations and/or exclusions contained in this policy is understood and agreed by both the Reinsured and the Reinsurers to be subject to England Wales Law.

Each party agrees to submit to a worldwide jurisdiction and to comply with all requirements necessary to give such court jurisdiction.

In respect of claims brought against the insured and indemnified under this policy, as more fully described herein, the choice of law applicable is Worldwide and the choice of jurisdiction is Worldwide.

The equivalent wording in each of the Excess Layer Reinsurances, again with added numbering, is as follows:

Any dispute concerning the interpretation of the terms, conditions, limitations and/or exclusions contained in this policy is understood and agreed by both the insured and the insurers to be subject to England and Wales.

Each party agrees to submit to the jurisdiction of England and Wales to comply with all requirements necessary to give such court jurisdiction.

In respect of claims brought against the insured and indemnified under this policy, as more fully described herein, the choice of law applicable is Worldwide and the choice of jurisdiction is Worldwide.

In the ARR reinsurance policy there are two potentially relevant clauses.

In the Schedule, next to the heading Choice of Law and Jurisdiction this wording appears (again I have added numbering):

Any dispute concerning the interpretation of the terms conditions limitations and/or exclusions contained in this policy is understood and agreed by both the (re) insured and the (re)insurers to be subject to the law of the courts of England and Wales.

Each party agrees to submit to the jurisdiction of Worldwide to comply with all requirements necessary to give such court jurisdiction.

But a supplemental clause, identified in the Schedule as a “Policy Interpretation Jurisdiction and Service of Suit clause”, provides as follows:

The parties agreed that each reinsurance contract contained a clause choosing the law of England and Wales.

The parties’ competing submissions on the interpretation of the jurisdiction provisions can be summarised at a high level as follows.

For the Reinsurers, Mr MacDonald Eggers QC submitted that

although the Primary Layer Reinsurance Contract does not contain an express exclusive jurisdiction clause, on the true construction of the contract, or by way of an implied term, to be implied by reason of obviousness and/or necessity for business efficacy, where a claim has impacted or would impact the Excess Reinsurances and/or the Aggregate Retention Reinsurances, the Reinsurers and the Defendants are obliged to submit to and to submit any dispute arising under or connection with the Primary Reinsurances to the exclusive jurisdiction of the Courts of England and Wales;

Sentence 2 in the Excess Layer L&J clause confers exclusive jurisdiction on the courts of England and Wales;

the ARR policy confers exclusive jurisdiction on the Courts of England and Wales.

For the Defendants Mr Lynch QC submitted that

the Primary Layer Reinsurance does not contain an exclusive jurisdiction clause and it is not subject to the implied term contended for by Reinsurers;

Sentence 2 in the Excess Layer Reinsurances confers jurisdiction on the Courts of England and Wales but is not an exclusive jurisdiction clause: it prevents the parties from objecting to the jurisdiction of the English Court but does not compel them to sue there.

the ARR policy confers exclusive jurisdiction on the Courts of England and Wales.

The parties, correctly in my view, proceeded on the basis that the three categories of reinsurance contract (ARR, primary layer, and excess layers) fell to be construed together, even though some claimants were a party only to the excess layer policies and even though only C1 was a party to the ARR.

I begin with the ARR, which in my view has only passing relevance. Both parties accept that law and jurisdiction is governed by the bespoke clause set out at paragraph 40 above and that, by that clause, C1 and ABSA Manx agreed that the courts of England and Wales should have exclusive jurisdiction. Mr Lynch points out that it is not at all surprising that an Isle of Man company might agree with predominantly English reinsurers that any claim relating to the aggregate of retentions should be litigated in, and only in, England & Wales. The ARR demonstrates that C1 had available to it, to use if it wished to, wording which unequivocally achieves exclusive jurisdiction.

Turning then to the Primary and Excess Reinsurances, for understandable forensic reasons Mr MacDonald Eggers wanted to start with the excess layers, submit that they contained an exclusive jurisdiction clause, and then address the primary layer, whereas Mr Lynch wanted first of all to focus on the primary layer and then consider the excess layers. There are obvious dangers with either sequential approach, since the contracts need to be construed together, but I think I can safely proceed by considering first of all the primary layer, and construing it on a working assumption (which favours Reinsurers) that the excess layers contain an exclusive jurisdiction clause.

The critical part of the Primary Layer Law and Jurisdiction clause is the second of these three sentences:

Any dispute concerning the interpretation of the terms, conditions, limitations and/or exclusions contained in this policy is understood and agreed by both the Reinsured and the Reinsurers to be subject to England Wales Law.

Each party agrees to submit to a worldwide jurisdiction and to comply with all requirements necessary to give such court jurisdiction.

In respect of claims brought against the insured and indemnified under this policy, as more fully described herein, the choice of law applicable is Worldwide and the choice of jurisdiction is Worldwide.

Although the parties were, of course, agreed that Sentence (2) was not an exclusive jurisdiction clause, neither found it easy to articulate exactly what its intended effect was. Does it mean that if one party starts proceedings in (say) the Courts of Cambodia, the other must permit itself to be sued there? That seems, to say the least, unlikely to have been intended. I do not need to decide what it does mean. But I agree with Mr Lynch that, since it refers to worldwide jurisdiction, it is about as far away from an exclusive England and Wales jurisdiction clause as one could get.

As to sentence (3), Mr Lynch submitted that it catered for claims that the South African ABSA entities pursued directly against reinsurers pursuant to the Direct Claims clause. In my view this is untenable. “claims brought against the insured”, in the context of this policy, must in my view be a reference to the claims against the South African ABSA entities which are insured and indemnified by Reinsurers. The clause serves as a reminder that it does not matter where those underlying claims had been brought. This was the second of two alternative constructions proposed by Mr MacDonald Eggers. Although it is true to say that it seems odd to find a clause with that effect (ie which is relevant only to what risks are covered) in something which is otherwise a law and jurisdiction clause properly so-called, in my judgment the drafting of these provisions as a whole is not of such high quality that apparent illogicalities and infelicities of drafting which arise from a particular reading can be said to give rise to a strong argument that that particular reading was not in fact intended.

The question which remains to be addressed is whether Mr MacDonald Eggers is right to submit that, assuming that the Excess layers are subject to an exclusive E&W jurisdiction clause, either on its proper construction or as a matter of implication, Sentence (2) is subject to a proviso, that proviso being that “where a claim has impacted or would impact either the Excess Insurances and/or the ARR, the Defendants and the Reinsurers are obliged to submit to and to submit any dispute to the exclusive jurisdiction of the Courts of England and Wales.”

I do not consider that this proviso can somehow be read into the contract as a matter of construction. Nothing in the wording used by the parties in this or in any of the other contracts gives any hint of such a proviso. To construe the provision in this way, as opposed to considering the implication of a term, would seem to me to be an exercise in rewriting it, as opposed to reading it in its commercial context.

As far as an implied term is concerned, the test to be applied is established by Marks & Spencer v BNP Paribas [2016] AC 742. The term must be (1) reasonable and equitable and either (2) necessary to give business efficacy to the contract (so that no

Mr MacDonald Eggers’ essential point was that it is very inconvenient, and commercially undesirable, if proceedings on the primary layer and excess layers have to take place in different jurisdictions. It gives a risk of inconsistent decisions, and it increases expense. I agree with all these observations. But in my view that is not enough to justify the implication of the term for which he contends.

First, the term contended for can itself have rather undesirable results. Suppose a claim is just big enough to engage the first excess layer. How much of the excess must be engaged before the entire claim must go to London? 1 ZAR? 1,000 ZAR? 10,000 ZAR? Wherever the line is struck, it is arbitrary, and if, on the facts, that excess were small in comparison with the first layer, the result feels like the tail (here the excess layer) wagging the dog. So this is not a case where implying the term simply sweeps away all odd or apparently uncommercial outcomes: it would do away with some, but create others in their place.

Second, one can postulate circumstances in which, during negotiations, the South African ABSA entities were insistent that all disputes under all layers could and should be litigated only in South Africa, whereas Reinsurers were insistent that all disputes under all layers should go only to London. If both sides were recalcitrant a compromise might be reached which, necessarily, involved the possibility of separate proceedings on different layers in different jurisdictions. Because Courts are not permitted to consider negotiations as an aid to construction, it is important to remember that sometimes parties end up with what seem to be clumsy or commercially undesirable clauses because the parties resolved a disagreement by a compromise. Compromise solutions are often inelegant. It would be wrong for the Court to resolve such a disagreement in a different way by implying a term, just because the court thought that that provided an objectively better solution.

Third, it is fair to observe that the inconvenience for some of the Reinsurers of having to litigate in two jurisdictions reflects nothing more than the fact that they chose to subscribe to two contracts with very obviously different law and jurisdiction clauses. And there is no equivalent problem for the five Reinsurers, C8 to C12, who only subscribed for excess layers.

Finally, the suggested inclusion in the proviso of claims which would or might impact the ARR seems to me to be unworkable or at least highly undesirable. Any claim might end up being one that would impact the ARR: does that mean that every claim in fact has to go to London? If not, what exactly is the test? But I place little weight on this point because it would be possible to improve the implied term by removing the reference to the ARR, and that would neutralise this point.

Standing back I am unable to say that the implied term is necessary to give business efficacy to the Primary Reinsurance Contract (nor to the suite of contracts as a whole), nor that it is so obvious that it goes without saying.

Accordingly, and even on the assumption that Reinsurers are right to contend that the Excess Layer Reinsurances are subject to an exclusive English jurisdiction clause, I conclude that the proviso pleaded by Reinsurers as an implied term should be rejected. Even if Reinsurers contended for a more limited implied proviso that omitted any mention of the ARR policy, I would still have rejected such an implied term.

The position in relation to construction of the jurisdiction clause in the excess layer policies is more difficult.

Given that the third sentence must surely have the same meaning as it does in the primary layer, the question resolves to whether sentence (2) is an exclusive jurisdiction clause or merely an agreement of the parties to submit to the jurisdiction of the English Courts, so that if the counterparty sues them in England they will not object.

I will set out sentences (1) and (2) again, because one of Mr MacDonald Eggers’ submissions focuses on Sentence (1):

Any dispute concerning the interpretation of the terms, conditions, limitations and/or exclusions contained in this policy is understood and agreed by both the insured and the insurers to be subject to England and Wales.

Each party agrees to submit to the jurisdiction of England and Wales to comply with all requirements necessary to give such court jurisdiction.

The task is to construe that, bearing in mind that there is an unambiguously exclusive jurisdiction clause in the ARR and that there is no exclusive jurisdiction clause in the Primary Reinsurances.

Mr MacDonald Eggers for Reinsurers submits that although the word exclusive is not used the proper construction is that this is an exclusive jurisdiction agreement because

the language is in transitive or positive terms, that sense being conveyed by

the placement of the sentence immediately after the choice of law

the obligation “to comply with all requirements necessary to give such court jurisdiction”;

the English choice of law clause indicates that the English jurisdiction clause was intended to be exclusive;

Article 25(1) of the Recast Brussels Regulation effectively creates a presumption that a jurisdiction clause falling within its scope (as this one does) is an exclusive jurisdiction clause.

Mr Lynch for the Defendants submits that:

the language is not transitive: see the approach of Hobhouse J in Cannon Screen v Handmade Films (QB) 11 July 1989 (unreported) and Langley J in The Athena (No 2) [2007] 1 Lloyd’s Rep 280;

the obligation to “comply with all requirements necessary to give such court jurisdiction” reveals that the parties must do something before the court has jurisdiction, which suggests that the agreement has not conferred exclusivity, merely that the agreement is that if one party is sued by the other in England, that party will not take any objection to its jurisdiction;

the permissive language can be contrasted with mandatory language used elsewhere in the suite of contracts;

a non-exclusive construction fits better with the service of proceedings provision which permits service in South Africa;

a non-exclusive construction fits better with clause 13 which says that Reinsurers shall abide by the final judgment of the Court where such legal proceedings are heard.

The parties are agreed that it is perfectly possible for a jurisdiction clause to take effect as an exclusive jurisdiction clause even if the word exclusive is not used.

I begin with the service of suit clause.

As to service of suit clauses, Mr Lynch submitted that such clauses normally operate on the basis that it is in that same jurisdiction that the insurer consents to be sued, as distinct from consenting to be served with proceedings wherever they may be brought, citing BP plc v National Union Fire Insurance Co [2004] EWHC (1132)

I accept that service of suit clauses may provide an indication, and sometimes a strong indication, that the party who consents to be served in a particular jurisdiction consents to the jurisdiction of the courts there, but it seems to me that the weight to be attached to such a provision must depend on the overall context.

Here, the address for service given in the Reinsurance contracts is that of the (then)

Clause 13 of the General Conditions seems to me to be of little assistance either way. It provides that Reinsurers shall abide by the final judgment of the Court, or Appellate Court in the event of an appeal, where such legal proceedings are heard. Mr Lynch’s argument is that the non-specific reference suggests that the parties anticipate that a court in more than one place might be giving final judgments or appeals. But this provision is incorporated from the Primary Layer Reinsurances (which on any view do not have an express exclusive jurisdiction clause), and so in my judgment it carries negligible weight when construing the express L&J clause in the Schedule to the Excess Layer Reinsurance Contracts.

I turn then to the vexed question of the supposed distinction between transitive or non-transitive jurisdiction clauses. As I understand it this is another way of describing the difference between an agreement “to submit disputes to the Courts of X”, said to be transitive, and to constitute an exclusive jurisdiction clause, and an agreement “to submit to the Courts of X”, said to be intransitive, and to amount merely to a nonexclusive jurisdiction clause. On this issue I agree with and propose to follow the approach taken by Males J (as he then was) in BNP Paribas v Anchorage Capital [2013] EWHC 3073 (Comm), when he was considering the proper construction of the following clause:

He said this:

BNPP's case is that this clause is an exclusive jurisdiction clause, while Anchorage contends that … commencement of the New York proceedings was not a breach of the clause because it merely provided for the non-exclusive jurisdiction of the English court, leaving Anchorage at liberty to sue BNPP in New York or, for that matter, in any court anywhere where it can establish jurisdiction over BNPP.

The question whether the jurisdiction clause is exclusive or non-exclusive is the subject of lively debate in the New York proceedings between the English law experts instructed by the parties for the purpose of BNPP's challenge there to the jurisdiction of the New York court. Lord Collins of Mapesbury, the expert instructed by BNPP, maintains that the clause provides for exclusive jurisdiction, while Professor Adrian Briggs, the expert instructed by Anchorage, maintains the opposite. It is of course hard to think of experts of greater eminence in this field. Their reports discuss such matters as the significance of the word “irrevocably”, whether the verb “submit” is used transitively or intransitively and whether that makes any difference, and whether the “contra proferentem” rule has any role to play.

While there have no doubt been cases in which these matters have been accorded some weight in determining whether a jurisdiction clause should be regarded as exclusive or non-exclusive, the leading textbooks have not been impressed with the distinction between transitive and intransitive clauses. Dicey, Morris & Collins on The Conflict of Laws (15th edition 2012) comments at para 12-105 that the distinction between transitive and intransitive verbs to determine whether a clause is exclusive or non-exclusive “would appear to have practically nothing to recommend it”. Professor Briggs is equally scathing:

I must confess that (even with the benefit of a university education) the distinction in the present case between submitting to the jurisdiction in respect of certain matters (intransitive and therefore, so the argument goes, nonexclusive) and submitting disputes in respect of certain matters to the jurisdiction of the court (transitive, and therefore exclusive) is so elusive that it escapes me altogether.

In the end all of these factors are only signposts which may sometimes assist in determining the intention of the parties, while the terms “exclusive” and “non-exclusive” themselves are merely convenient labels. In agreement with Dicey at para 12-105 (“the true question is whether on its proper construction the clause obliges the parties to resort to the relevant jurisdiction, irrespective of whether the word ‘exclusive’ is used”), I prefer to ask the question whether the commencement and pursuit of the foreign proceedings in question are things which a party has promised not to do.

It is clear that the jurisdiction clause in this case constitutes a promise by Anchorage to submit to the jurisdiction of the English court, but there is no equivalent promise by BNPP. BNPP is therefore free, if it wishes, to sue Anchorage elsewhere. In that sense, at least, the clause does not make England the exclusive venue for litigation between the parties and may be regarded as non-exclusive. But the clause does require Anchorage to submit to English jurisdiction, and thus gives BNPP the right to litigate in England, “in respect of any matter arising out of this Agreement, or our services to or Transactions with you under this Agreement” if BNPP chooses to litigate here. 90. The first question, therefore, is whether Anchorage's New York proceedings are “in respect of” such a matter. In my judgment they clearly are and Anchorage does not contend otherwise. On the contrary it accepts, correctly, that the New York proceedings are in respect of “essentially the same issues” as are raised in these proceedings in England.

Drawing together the various arguments, it seems to me that the right approach to the construction of the particular L&J clause in the Excess Layer Reinsurance Contracts here is as follows:

The fact that the parties have chosen English law can be a factor favouring an interpretation of the clause as an exclusive jurisdiction clause;

The distinction between transitive and intransitive clauses is so arcane as to provide little assistance;

In this case the immediate juxtaposition of the choice of law clause which expressly refers to disputes, and the jurisdiction clause in sentence (2), strongly supports an interpretation that reads sentence (2) as referring to the submission of disputes;

Although there can be commercial advantages to non-exclusive jurisdiction clauses, there are in general greater commercial advantages to exclusive jurisdiction clauses, since they provide certainty and predictability to both parties:

The fact that there is a service of proceedings provision in favour of South Africa is best explained by the convenience which it affords to the South African entities in having a single, local, place for service against all Reinsurers, and in terms of the argument as to construction of the jurisdiction clause this factor weighs only lightly in the balance.

Clause 13 is irrelevant because (like sentence (3)) it is concerned not with proceedings between the parties but with underlying proceedings by third parties against the South African ABSA entities and which might give rise to claims for which indemnity is sought against Reinsurers.

Taking those factors together, I consider that the natural and proper reading of sentence 2 is that each party to the Excess Layer Reinsurance Contracts agreed with the others that it would submit any dispute that might arise to the Courts of England and Wales, and to no other court.

I have reached this conclusion without recourse to Article 25 (1) of the Brussels

Reinsurers seek to enjoin the South African proceedings in their entirety. They contend that the Excess Layer claim is brought in breach of contract and falls to be restrained unless there are strong reasons not to do so. They further contend that the Primary Layer claim should be restrained because that claim is brought in breach of contract and falls to be restrained unless there are strong reasons not to do so, alternatively, because that claim is vexatious oppressive or unconscionable."

The Defendants say that no injunction at all should be granted. They say that the Primary Layer claim is not vexatious oppressive or unconscionable and so cannot properly be restrained, and they say that although (ex hypothesi) their Excess Layer claim is brought in breach of contract, there are strong reasons why, nevertheless, it should be permitted to continue.

There is also the possibility of what was described in the course of the hearing as a “halfway house” injunction, restraining the continuation of the South African claim on the Excess Layer Reinsurances but permitting the Primary Layer claim to continue there. The parties had not dealt fully with this possibility in their submissions and I gave directions permitting the parties to provide further submissions, directed to this possible outcome, after the oral hearing had been completed.

I begin with the Reinsurers’ submission that the South African Proceedings on the Primary Layer Reinsurances are vexatious oppressive or unconscionable.

Calver J said on the ex parte application that, even if there was no exclusive English jurisdiction agreement in the primary reinsurances, the result of allowing the proceedings to continue in South Africa would be to allow concurrent proceedings relating to the same subject matter in two different jurisdictions, and would expose the parties to the risks of inconsistent judgments and likely complication with regard to recognition and enforcement of judgments. He said that he was satisfied that granting the injunction had the potential to secure the continuation of only one set of proceedings taking place which is clearly desirable.

I agree that it would clearly be desirable if there were proceedings in only one jurisdiction, but having had the benefit of argument on this point from the ABSA parties I have concluded that the high hurdle of demonstrating that proceedings limited to the Primary Layer Reinsurances would themselves be vexatious or oppressive, has not been made out by the Reinsurers.

The leading case on antisuit injunctions in the absence of an exclusive jurisdiction clause is Deutsche Bank AG v Highland Crusader [2009] EWCA Civ 725. Toulson LJ summarised the relevant principles at [50], setting out the following (8) propositions.

Under English law the court may restrain a defendant over whom it has personal jurisdiction from instituting or continuing proceedings in a foreign court when it is necessary in the interests of justice to do.

It is too narrow to say that such an injunction may be granted only on grounds of vexation or oppression, but, where a matter is justiciable in an English and a foreign court, the party seeking an anti-suit injunction must generally show that proceeding before the foreign court is or would be vexatious or oppressive.

The courts have refrained from attempting a comprehensive definition of vexation or oppression, but in order to establish that proceeding in a foreign court is or would be vexatious or oppressive on grounds of forum non conveniens, it is generally necessary to show that (a) England is clearly the more appropriate forum (“the natural forum”), and (b) justice requires that the claimant in the foreign court should be restrained from proceeding there.

If the English court considers England to be the natural forum and can see no legitimate personal or juridical advantage in the claimant in the foreign proceedings being allowed to pursue them, it does not automatically follow that an anti-suit injunction should be granted. For that would be to overlook the important restraining influence of considerations of comity.

An anti-suit injunction always requires caution because by definition it involves interference with the process or potential process of a foreign court. An injunction to enforce an exclusive jurisdiction clause governed by English law is not regarded as a breach of comity, because it merely requires a party to honour his contract. In other cases, the principle of comity requires the court to recognise that, in deciding questions of weight to be attached to different factors, different judges operating under different legal systems with different legal policies may legitimately arrive at different answers, without occasioning a breach of customary international law or manifest injustice, and that in such circumstances it is not for an English court to arrogate to itself the decision how a foreign court should determine the matter. The stronger the connection of the foreign court with the parties and the subject matter of the dispute, the stronger the argument against intervention.

The prosecution of parallel proceedings in different jurisdictions is undesirable but not necessarily vexatious or oppressive.

A non-exclusive jurisdiction agreement precludes either party from later arguing that the forum identified is not an appropriate forum on grounds foreseeable at the time of the agreement, for the parties must be taken to have been aware of such matters at the time of the agreement. For that reason an application to stay on forum non conveniens grounds an action brought in England pursuant to an English non-exclusive jurisdiction clause will ordinarily fail unless the factors relied upon were unforeseeable at the time of the agreement. It does not follow that an alternative forum is necessarily inappropriate or inferior. (I will come to the question whether there is a presumption that parallel proceedings in an alternative jurisdiction are vexatious or oppressive).

The decision whether or not to grant an anti-suit injunction involves an exercise of discretion and the principles governing it contain an element of flexibility.

I am not persuaded that England is clearly the more appropriate forum for a claim under the Primary Layer Reinsurances. On the contrary, Reinsurers’ defence turns on the date of notification to, or knowledge of, ABSA employees, all of whom are likely to be in South Africa, and the case arises out of underlying claims brought against ABSA in South Africa. I accept Mr Lynch’s submission that the “centre of gravity” of the issues that will be in dispute is South Africa.

The mere fact that there will, on any view, be proceedings on the Excess Layers in this jurisdiction does not in my view render it necessary in the interests of justice to stop the Primary Layer proceedings in South Africa, and were I to do so I consider I would be interfering inappropriately with a claim which is properly before a foreign Court which has jurisdiction to entertain it.

I therefore decline to continue the ex parte injunction insofar as it seeks to stop the claim in South Africa brought on the Primary Layer Reinsurances.

Here we are in the fundamentally distinct territory of proceedings brought in a foreign Court in breach of an exclusive jurisdiction clause.

In relation to an antisuit injunction based on a jurisdiction clause which the Court has found to be exclusive, and abbreviating slightly what was said by Jacobs J in Catlin Syndicate v Amec Foster Wheeler [2020] EWHC 2530 (Comm) the principles are as follows:

The Court has the power to grant an interim injunction in all cases in which it appears to the court to be just and convenient to do so, and any such order

In very broad terms the touchstone is what the ends of justice require. This determination involves an exercise of discretion by the Court. The particular facts of the case are critical to the exercise of this discretion: Emmott v Michael Wilson & Partners Ltd [2018] 1 Lloyd's Rep 299 at [36] per Sir Terence Etherton MR.

The jurisdiction to grant an anti-suit injunction must be exercised with caution because it is one that indirectly affects a foreign court: Société Nationale Industrielle Aérospatiale v Lee Kui Jak [1987] UKPC 12, [1987] AC 871 , 892E per Lord Goff.

As to the meaning of "caution" in this context, it has been described thus in The "Angelic Grace" [1995] 1 Lloyd's Rep 87 at 92 per Leggatt LJ: " The exercise of caution does not involve that the Court refrains from taking the action sought, but merely that it does not do so except with circumspection."

The Court will ordinarily exercise its discretion to restrain the pursuit of proceedings brought in breach of an exclusive jurisdiction clause unless the Defendant can show strong reasons to refuse the relief: The Angelic Grace [1995] 1 Lloyd's Rep 87 ; The Jay Bola [1997] 2 Lloyd's Rep 279 (CA) at page 286 per Hobhouse LJ.

The Defendant bears the burden of proving that there are strong reasons to refuse the relief: Donohue v Armco Inc [2002] 1 All ER 749 at [24]-[25] per Lord Bingham.

If there were no proceeding in South Africa on the Primary Layer Reinsurances, the answer to this question would be straightforward: quite clearly there would be no strong reasons to refuse to restrain the South African Excess Layer proceedings.

So the question really resolves to this: does the fact that there will continue to be proceedings in South Africa on the Primary Layer constitute a strong reason for refusing to restrain the South African proceedings on the Excess Layers?

Mr Lynch accepts that the ABSA entities must demonstrate “strong reasons” against the grant of an injunction. He submits that four factors combine to demonstrate strong reasons:

the majority of Reinsurers under the excess policies also agreed to “worldwide jurisdiction” under the Primary Layer Reinsurances; ii) the Excess Layer jurisdiction clauses are very badly drafted;

there is no commercially coherent reason as to why a different jurisdictional regime should apply to only part of what is a suite of interlocking contracts;

South Africa is the centre of gravity for evidence, witness and the underlying claims.

In my judgment these arguments do not amount to strong reasons against the grant of an injunction. Point (i) is true and explains why those Reinsurers who subscribed to the Primary Layer Reinsurances will have to face those claims in South Africa. I do not see how that weakens their claim to enforce the different agreement which they made in relation to the Excess Layers. As to point (ii) it is true that in various places the drafting leaves much to be desired; but there is no principle of English law which imposes a sliding scale of rewards for good drafting such that a well drafted clause gives rise to a stronger claim than a poorly drafted clause to the same effect. Point (iii), even if true, seems to me to be irrelevant. In any event I am not persuaded of the premise. As I noted above, the reason for different clauses might be a reflection of intransigent negotiating positions, or, as Mr MacDonald Eggers suggests, it might reflect a view of Reinsurers that although they were happy for smaller claims to be resolved in South Africa they wanted to ensure that claims over and above the Primary Layer be resolved in their home jurisdiction. Point (iv) is a point about natural forum and in my view cannot trump the fact that the parties contracted for a different forum.

So I am not persuaded there are strong reasons not to enforce the parties’ agreement.

But, remembering that an injunction is a discretionary remedy, I consider that I should nevertheless pause before granting the injunction, because it is not immediately attractive to adopt a route which will, unless one side or the other backs down, inevitably result in there being proceedings in two jurisdictions. That is self-evidently not an ideal result.

Having paused, the following observations can be made.

If I grant the injunction, there will be proceedings here on the excess layers (and perhaps on the primary layers too) and in South Africa on the primary layers. But on the other hand if I do not grant the injunction the position will be even worse: there will be proceedings in South Africa on the excess layers as well. In other words, there will be two sets of proceedings in any event, and granting the injunction at least reduces the extent of the overlap.

The fact that there will be two sets of proceedings is a consequence of the fact that the parties entered into contracts with different terms. The parties can reasonably expect the Court to enforce the terms which they agree, but they should not expect the Court to try to fashion for them an objectively better agreement. Granting the injunction enforces the contractual promise made.

The sums at stake fall predominantly within the Excess Layer Reinsurances:

The position of C8 to C12 is noteworthy. They are not parties to the Primary Layer Reinsurance and it would seem particularly unfair if they were to be exposed to an excess layer claim in South Africa simply because some of their fellow excess layer Reinsurers had agreed a different L&J provision in the Primary Layer Reinsurance to which C8 to C12 are not parties.

I conclude that there is no strong reason not to grant the contractual anti-suit injunction, but on the contrary there are strong reasons in favour of doing so, and that the appropriate way to exercise my discretion is to continue to enjoin the South African proceedings on the Excess Layer Reinsurances.

The end result is not ideal, because it leaves most of the parties involved in two sets of proceedings, but that reflects the different wording of the agreements which the parties entered into. The result does give effect to and enforce those agreements.