Introduction

This judgment gives the reasons for the court’s unanimous decision, announced on 23 April 2024, to dismiss this appeal. The appeal is from an order made by the Court of Appeal on 29 January 2024 requiring the appellant, RusChemAlliance LLC (“RusChem”), to cease court proceedings in Russia against the respondent, UniCredit Bank GmbH (“UniCredit”), in circumstances where the parties have agreed, in a contract governed by English law, that any disputes between them shall be settled by arbitration in Paris. The result of this court’s decision is therefore that the Court of Appeal’s order is undisturbed.

The underlying dispute

RusChem is a Russian company which in July and September 2021 entered into two contracts with German companies (together described as “the Contractor”) for the construction of liquefied natural gas and gas processing plants in Russia. Under these contracts RusChem agreed to pay, in stages, approximately €10 billion, including advance payments of around €2 billion. RusChem made the advance payments to the Contractor.

Performance of the Contractor’s obligations was guaranteed by bonds payable on demand. Seven such bonds have been issued by the respondent, UniCredit, a German bank. Each of the contracts contained in these bonds provides (in clause 11) that the bond is governed by English law and (in clause 12) that all disputes are to be settled by arbitration in Paris under the rules of the International Chamber of Commerce (“ICC”). Here is the full wording of these clauses:

“11.

This Bond and all non-contractual or other obligations arising out of or in connection with it shall be construed under and governed by English law.

12.

In case of dispute arising between the parties about the validity, interpretation or performance of the Bond, the parties shall cooperate with diligence and in good faith, to attempt to find an amicable solution. All disputes arising out of or in connection with the bond (which cannot be resolved amicably) shall be finally settled under the Rules of Arbitration of the International Chamber of Commerce (ICC) by one or more arbitrators appointed, in accordance with the said ICC’s Rules. The place of arbitration shall be Paris and the language to be used in the arbitral proceedings shall be English.”

Following Russia’s invasion of Ukraine in February 2022, the European Union imposed sanctions on Russia and on designated Russian legal entities and individuals. The designated entities did not include RusChem. Even so, in May 2022 the Contractor announced that, because of EU sanctions, it could not continue to perform the construction contracts. As a result, RusChem terminated the contracts and requested the return of the advance payments. The Contractor stated that it could not return the advance payments, again giving EU sanctions as the reason.

In October 2022 and April 2023 RusChem made demands on UniCredit for payment under the bonds. UniCredit refused to pay on the ground that payment was prohibited by EU sanctions, in particular article 11 of Council Regulation (EU) No 833/2014 of 31 July 2014 concerning restrictive measures in view of Russia's actions destabilising the situation in Ukraine. UniCredit has not relied on any other reason for its refusal to pay.

The Russian proceedings

On 5 August 2023 RusChem issued proceedings against UniCredit before the Arbitrazh Court of the St Petersburg and Leningrad Region in Russia, claiming payment of €448 million under the bonds. In its statement of claim in those proceedings RusChem relied on article 248.1 of the Arbitrazh Procedural Code, introduced by the Russian Federation in 2020. The effect of article 248.1 is, among other things, to confer exclusive jurisdiction on Russian Arbitrazh Courts over disputes between Russian and foreign persons arising from foreign sanctions; to treat an agreement providing for arbitration of such a dispute outside the territory of the Russian Federation as inoperable. Article 248.2 enables Russian persons affected by foreign sanctions to apply to a Russian Arbitrazh Court for an anti-suit injunction prohibiting the other party from initiating or continuing proceedings before a foreign court or international arbitration tribunal located outside the territory of the Russian Federation.

UniCredit applied to the Arbitrazh Court to dismiss RusChem’s claim on the ground that the parties have agreed that all disputes arising out of the bonds are to be settled by arbitration in Paris under the rules of the ICC.

On 1 November 2023 the judge in the Russian proceedings, SS Saltykova, announced the decision of the Arbitrazh Court to dismiss that application. Judge Saltykova ruled that, by virtue of article 248.1(2)(1) of the Arbitrazh Procedural Code, the dispute falls within the exclusive competence of the Arbitrazh Courts of the Russian Federation, so that the arbitration agreements cannot be enforced. The judge stayed the proceedings, however, and has since adjourned the matter pending the outcome of the present proceedings. Lord Reed PSC has formally expressed this court’s gratitude to Judge Saltykova for taking this course and enabling these proceedings to be dealt with in an orderly way.

These proceedings

These proceedings were begun in the Commercial Court in London by UniCredit on 22 August 2023. The claim was for injunctive and declaratory remedies for RusChem’s commencement and pursuit of the Russian proceedings in breach of the arbitration agreements in the bonds. UniCredit applied without notice for an interim injunction prohibiting RusChem from continuing the Russian proceedings until further order of the court, which was granted on 24 August 2023.

On 8 September 2023 RusChem issued an application disputing the English court’s jurisdiction to hear UniCredit’s claim.

On 22 September 2023 the hearing of RusChem’s challenge to the English court’s jurisdiction and an expedited trial of UniCredit’s claim took place in the Commercial Court before Sir Nigel Teare sitting as a High Court judge. For reasons given in an ex tempore judgment, the judge held that the English court did not have jurisdiction to hear the claim; but he continued the interim anti-suit injunction until the process of appeal from his order had been exhausted: [2023] EWHC 2365 (Comm).

The Court of Appeal granted UniCredit permission to appeal from the judge’s decision and the appeal was heard on 25 January 2024. At the end of the hearing, the court (Bean, Males and Lewis LJJ) announced its decision to allow the appeal and to grant a final anti-suit injunction, with reasons to follow. On 29 January 2024 the order of the Court of Appeal was made granting final relief including a mandatory injunction requiring RusChem to discontinue the Russian proceedings. The reasons for the Court of Appeal’s decision were given by Males LJ in a judgment handed down on 2 February 2024: [2024] EWCA Civ 64; [2024] 1 Lloyd’s Rep 350.

In the Court of Appeal the issues were: (1) whether the English court has jurisdiction over UniCredit’s claim; and (2) if so, whether the Court of Appeal should grant the final injunction claimed by UniCredit or should remit that question to the Commercial Court. In summary, the Court of Appeal decided that the English court has jurisdiction over the claim because: (a) the arbitration agreements in the bonds are governed by English law; and (b) England and Wales is the proper place in which to bring the claim. The Court of Appeal also decided that the question whether to grant a final injunction should not be remitted to the Commercial Court and granted the injunction.

This appeal

On 12 February 2024 this court gave RusChem permission to appeal from the decision of the Court of Appeal on the jurisdiction issue. RusChem was refused permission to appeal on the question whether, if the English court has jurisdiction over UniCredit’s claim, the Court of Appeal was entitled to grant a final injunction rather than remit the matter to the Commercial Court. Permission to appeal on that issue was refused because it did not raise a point of law of general public importance.

The sole issue in this appeal is therefore whether the English court has jurisdiction over UniCredit’s claim. This depends on whether the Court of Appeal was right to decide (a) that the arbitration agreements in the bonds are governed by English law and (b) that England and Wales is the proper place in which to bring the claim. I will refer to these issues, respectively, as “the governing law issue” and “the proper place issue” and will address them in turn.

The governing law issue

Under rule 6.36 of the Civil Procedure Rules (CPR), the claimant may serve a claim form on a defendant out of the jurisdiction with the permission of the court if any of the grounds (commonly known as “gateways”) set out in para 3.1 of Practice Direction 6B applies. The sole ground, or gateway, on which UniCredit relies is that set out in para 3.1(6)(c) of Practice Direction 6B (the “contract gateway”). The contract gateway applies where a claim is made in respect of a contract which is governed by the law of England and Wales.

When applying for permission to serve proceedings out of the jurisdiction, it is only necessary for the claimant to satisfy the court that there is a “good arguable case” that the claim falls within the relevant gateway. The Court of Appeal, however, thought it right to decide this question on a final basis and this court has approached the question in the same way.

As explained in para 3 above, the bond contracts are expressly governed by English law. Where, however, as here, a contract includes an agreement to arbitrate disputes arising out of the contract, it is possible in principle for that agreement to be governed by a different system of law from the rest of the contract. RusChem argues that that is so here and that the arbitration agreements in clause 12 of the bonds are governed by the law of the place which the parties have chosen for the arbitration, that is to say, the law of France.

UniCredit argues that the arbitration agreements are governed by English law because the choice of English law in clause 11 as the governing law applies to clause 12 (the arbitration clause) as well as all the other clauses of the contract. In the courts below this was the only argument that UniCredit advanced on the governing law issue. On this appeal UniCredit raised a suggestion in its written case that, even if the arbitration agreements in clause 12 are governed by French law, UniCredit’s claim still falls within the contract gateway because English law on any view governs the rest of the bond contracts and it can be said that UniCredit’s claim is made in respect of those contracts. If this argument were thought to have any merit, there is no reason why it could not have been made in the courts below. As it is, UniCredit gave no notice that it might seek to raise this new point until after RusChem had filed its written case for this appeal. At the hearing I did not understand counsel for UniCredit to be asking the court to allow UniCredit to rely on this new argument; but if permission to do so had been sought, I would not have thought it right to give it.

I therefore proceed on the basis that, for the purpose of the governing law issue, the only relevant question is whether the arbitration agreements in clause 12 of the bonds are governed by English law.

The principles which determine what system of law governs an arbitration agreement were considered by this court in depth in Enka Insaat ve Sanayi AS v OOO Insurance Company Chubb [2020] UKSC 38; [2020] 1 WLR 4117 (“Enka”). The central issue on that appeal was which system of law governs an arbitration agreement when the law applicable to the contract containing it differs from the law of the seat of the arbitration. According to the common law rules which the court must apply in deciding this question, the arbitration agreement is governed by whichever system of law the parties have agreed will govern it or, in the absence of such an agreement, the system of law with which the arbitration agreement is most closely connected. Whether the parties have agreed on a choice of law to govern the arbitration agreement is ascertained by construing the arbitration agreement and the contract containing it, as a whole, applying the rules of contractual interpretation of English law as the law of the forum.

It is rare for the law governing an arbitration agreement to be separately specified, either in the arbitration clause itself or elsewhere in the contract. It is common, however, in a contract which has connections with more than one country (or territory with its own legal system) to find a clause specifying the law which is to govern the contract. A typical clause of this kind states: “This Agreement shall be governed by and construed in accordance with the laws of [name of legal system]”. Where the contract also contains an arbitration clause, it is natural to interpret such a governing law clause as applying to the arbitration clause for the simple reason that the arbitration clause is part of the contract which the parties have agreed is to be governed by the specified system of law. Thus, in Enka the Supreme Court unanimously held: (1) that a choice of law to govern the contract should generally be construed as applying to an arbitration agreement set out (or incorporated by reference) in a clause of the contract; and (2) that this is so even where the parties have chosen a place with a different system of the law as the seat of the arbitration: see paras 43, 53–54, 60, 170(iv)–(v), 260, 267, 269–271. Additional reasons given for adopting this approach as a general presumption were that it provides a degree of certainty, achieves consistency, avoids complexities and uncertainties, avoids artificiality and ensures coherence: see para 53.

This court in Enka also considered what law applies if the parties have not agreed on a choice of law to govern the arbitration agreement but have chosen a seat of arbitration. The majority held that as a general rule the law applicable in this situation is the law of the seat, even if this differs from the law applicable to the parties’ substantive contractual obligations: see paras 120, 145, 170(viii). In the present case no reliance is placed on this second aspect of the decision in Enka, so there is no need to consider it further.

In Kabab-Ji SAL v Kout Food Group [2021] UKSC 48; [2021] Bus LR 1717 the claimant brought proceedings in England to enforce an arbitration award made in France. The defendant resisted enforcement on the ground that it was not a party either to the contract under which the underlying claim was brought or to the arbitration agreement contained in that contract and therefore had not agreed to arbitration of the claim. The first and main issue on appeal to the Supreme Court was what system of law governed the arbitration agreement. Although the issue arose under section 103(2)(b) of the Arbitration Act 1996, it was common ground that the general principles identified in Enka were applicable to ascertain whether the parties had chosen the law which was to govern their arbitration agreement and, if so, what law they had chosen.

Like the bonds in this case, the contract in Kabab-Ji contained a clause providing for settlement of disputes under the rules of the ICC in Paris and also contained a typical governing law clause, which stated: “This Agreement shall be governed by and construed in accordance with the laws of England”. The phrase “this Agreement” was further defined as consisting of “the terms of agreement set forth herein below …”. The Supreme Court regarded the effect of these clauses as “absolutely clear”: para 39. Even without the further definition, the phrase “this agreement” was “ordinarily and reasonably understood … to denote all the clauses incorporated in the contractual document, including therefore clause 14 [the arbitration clause]”. There was no good reason to infer that the parties intended to except clause 14 from their choice of English law to govern all the terms of their contract. In particular, the choice of Paris as the seat of the arbitration was not such a reason. The law governing the arbitration agreement was therefore English law.

Some commentators have described the general principle recognised in Enka and applied in Kabab-Ji as being that a choice of governing law for the contract containing an arbitration clause amounts to an “implied choice” of law for the arbitration agreement. That is not how the principle was articulated in those cases. Thus, in our judgment in Enka Lord Hamblen and I expressed the principle simply in terms of what, on the proper interpretation of the contract, the parties would reasonably be understood to have agreed and did not use the phrase “implied choice”: see eg para 170(ii)–(iv) and the passages on which that summary was based.

This was deliberate. We pointed out, at para 35, that whether a choice is described as “express” or “implied” is a matter of degree as language may be more or less explicit. If the contract contains a typical governing law clause of the kind contained, for example, in the contract in Kabab-Ji, an element of inference or implication is involved in ascertaining that the parties have chosen that law to govern the arbitration clause. But this is only because the governing law clause does not refer specifically to the arbitration clause. The same can be said about all the other individual clauses of the contract. None of the individual clauses of the contract is specifically referred to in the governing law clause: the inference is merely that the general includes the particular. If it were necessary or relevant to characterise the choice of law for the arbitration agreement signified by such a governing law clause as “express” or “implied”, I think it would be more apt to call it an “express choice” because it is identified by interpreting the express terms of the contact and is not based on any implied term. But it does not matter which description is preferred. The distinction is of no legal significance. As was said in Enka, at para 35, it is “important to keep in mind that whether a choice is described as express or implied is not a distinction on which any legal consequence turns”. The only question of legal relevance is whether, on the proper interpretation of the contractual documents, the parties have agreed on the law which is to govern the arbitration agreement.

Since Enka and Kabab-Ji were decided, the Law Commission in its Review of the Arbitration Act 1996: Final report and Bill (Law Com No 413), published on 5 September 2023, has suggested that the law as stated in Enka is “complex and unpredictable” (para 12.20) and has recommended that the Arbitration Act 1996 be amended to provide that the arbitration agreement is governed by the law of the seat, unless the parties expressly agree otherwise (para 12.77). Depending on what the word “expressly” is taken to add to the word “agree”, this would not by itself alter the law as stated in Enka. The draft clause proposed by the Law Commission, however, includes a further provision that (para 12.78):

“agreement between the parties that a particular law applies to an agreement of which the arbitration agreement forms a part does not … constitute express agreement that that law also applies to the arbitration agreement.”

A Bill is before Parliament which includes a clause in these terms.

In light of this potential legislative change, RusChem sought to raise as a ground of appeal an alternative case that “the principles in Enka should be revisited more generally, and reformulated such that absent clear indications to the contrary, it should be inferred that the implied choice of governing law for the arbitration agreement is the law of the place of the seat chosen for the arbitration”. The invitation to revisit principles which this court has so recently settled (and even more recently affirmed in Kabab-Ji) was declined. Far from being a reason to revisit those principles, the fact that the matter is the subject of draft legislation currently before Parliament is a positive reason why it is inappropriate to do so. Permission to appeal on this ground was consequently refused because it does not raise a point of law “which the court should consider at this time”.

So the question whether the parties have agreed on a choice of law to govern the arbitration agreements in the bonds is to be determined, as it was in Kabab-Ji, by applying the principles identified in Enka.

Applying those principles, the answer to the question is just as clear here as it was in Kabab-Ji. The governing law clause in the bonds is framed in particularly wide terms and covers not only the bond itself but “all non-contractual or other obligations arising out of or in connection with it”. Even if the obligations created by the arbitration agreement were regarded as separate from the bond contract for this purpose, they are on any view “obligations arising … in connection with” the bond. But those additional words are not critical. Even if they are disregarded, the term “this Bond” in clause 11 is reasonably understood to mean the whole bond including clause 12 (the arbitration clause). There is nothing in the wording of the bonds which excepts clause 12 from the choice of English law as the governing law. As was held in Enka, the choice of a different country for the seat of the arbitration does not justify reading “this Bond” as excluding the arbitration agreement in clause 12. The arbitration agreements are therefore governed by English law.

RusChem does not accept that the principles stated in Enka lead to that conclusion. RusChem argues that on the proper interpretation of the bonds: (a) the choice of English law in clause 11 does not apply to the arbitration agreement in clause 12; and (b) the parties have agreed that the arbitration agreement is to be governed by French law. It is not obvious how a reasonable reader of the bonds could attribute this meaning to them. But counsel for RusChem have advanced an argument which is based on certain statements made in the majority judgment in Enka.

Para 170 of that judgment summarised the conclusions reached on the law applicable to the arbitration agreement. This summary included the following points:

“(iv)

Where the law applicable to the arbitration agreement is not specified, a choice of governing law for the contract will generally apply to an arbitration agreement which forms part of the contract.

(v)

The choice of a different country as the seat of the arbitration is not, without more, sufficient to negate an inference that a choice of law to govern the contract was intended to apply to the arbitration agreement.

(vi)

Additional factors which may, however, negate such an inference and may in some cases imply that the arbitration agreement was intended to be governed by the law of the seat are: (a) any provision of the law of the seat which indicates that, where an arbitration is subject to that law, the arbitration agreement will also be treated as governed by that country’s law; or (b) the existence of a serious risk that, if governed by the same law as the main contract, the arbitration agreement would be ineffective. Either factor may be reinforced by circumstances indicating that the seat was deliberately chosen as a neutral forum for the arbitration.”

We are not on this appeal concerned with the second factor referred to in para 170(vi). That factor reflects the principle that an agreement should be interpreted so that it is valid rather than ineffective: see Enka, paras 95–97. Nor is it suggested that the reason why the seat was chosen is a significant consideration. RusChem’s argument focuses on the factor referred to in para 170(vi)(a) of the judgment. This is said to establish an exception to the general rule that a choice of governing law for the contract as a whole will apply to an arbitration agreement in the contract even when a different country has been chosen for the seat of the arbitration. Counsel for RusChem read para 170(vi)(a) as saying that there is an exception to this general rule where the law of the seat treats the arbitration agreement as governed by that country’s law: in this situation it may be inferred that the arbitration agreement was intended to be governed by the law of the seat. They then submit that French law falls within this exception as it provides that arbitration agreements such as those in the bonds are governed by French law. It may therefore be inferred that the arbitration agreements in the bonds were intended to be governed by French law.

A good deal of argument from both sides was directed to whether French law does or does not fall within this putative exception. The expert evidence of French law adduced in this case and in previous cases such as Kabab-Ji shows that the French courts regard questions about the validity of the arbitration agreement as governed by “substantive rules of international arbitration.” The only exception is where a choice of national law to govern the arbitration agreement is contained within the arbitration agreement itself. The “substantive rules” which the French courts apply are rules which they have developed for international arbitration. These rules are different from the French domestic law rules applicable to arbitration agreements contained in the Civil Code, which has been held not to apply to international arbitration. But it is clear that these “substantive rules of international arbitration” are still part of French law. This has been recognised in earlier decisions of this court: see Dallah Real Estate and Tourism Holding Co v Ministry of Religious Affairs of the Government of Pakistan [2010] UKSC 46; [2011] 1 AC 763, para 15; and Kabab-Ji, para 89. The characterisation of the “substantive rules of international arbitration” as rules of French law is further confirmed by the expert evidence in this case and by the judgment of the Cour de Cassation in the parallel French proceedings in Kabab-Ji. In those proceedings the Cour de Cassation held that the Paris Court of Appeal had been right “to consider the existence and efficacy of the arbitration clause, not in the light of English law, but in the light of the substantive rules of French law in international arbitration matters”: see Kabab-Ji (Société) v Kout Food Group (Société) [2023] ILPr 6, para 12.

Counsel for UniCredit have argued that these “substantive rules of international arbitration” do not amount to rules of French law for the purposes of the “exception” contemplated in Enka. They have submitted that what was contemplated in para 170(vi)(a) of the judgment in Enka was a provision of the law of the seat which stipulates that, where the arbitration has its seat in that country, the arbitration agreement will be treated as governed by that country’s law because it is the law of the seat. The relevant French law, they say, does not satisfy this criterion. This is because French law does not link the application of its “substantive rules of international arbitration” to the choice of France as the seat of the arbitration. Rather, the French courts apply those rules (unless the arbitration agreement itself contains a choice of law to govern it) whenever a question arises about the existence, validity or scope of the arbitration agreement, regardless of whether the arbitration has its seat in France or somewhere else.

The language used in para 170(vi)(a) of the judgment in Enka was permissive rather than prescriptive. All that was said was that a provision of the law of the seat of the kind described “may” (not “must” or “will”) “in some cases imply that the arbitration agreement was intended to be governed by the law of the seat”. No attempt was made to suggest when, if at all, such an inference ought to be drawn. It was unnecessary to address that question on the facts of Enka.

It is in any case a mistake, all too frequently made, to treat sentences and phrases in a judgment as if they had textual authority in the same way as an Act of Parliament. As Sir George Jessel MR said succinctly in Hood v Newby (1882) 21 Ch D 605, 608: “You must always look to what was being discussed by the judges as well as to the words used.” It should also be remembered that, as the Earl of Halsbury LC said in Quinn v Leathem [1901] AC 495, 506, “every judgment must be read as applicable to the particular facts proved, or assumed to be proved” and “a case is only an authority for what it actually decides”.

The correct resolution, therefore, of the issue raised on this appeal does not lie in dissecting the particular verbal formulations used in the judgment in Enka but in examining the underlying reasoning. It is essential to understand, first of all, how the point reflected in the summary statement at para 170(vi)(a) of the judgment arose in the context of the arguments in that case.

At paras 65–94 of our judgment in Enka Lord Hamblen and I discussed what we called the “overlap argument” which had been accepted by the Court of Appeal. The thrust of this argument was: (1) that in choosing a place as the seat of the arbitration the parties can be taken to have chosen the law which will govern the arbitration process (known as the “curial law”); and (2) that the curial law is so closely related to the law governing the arbitration agreement that a choice of seat and curial law should generally be understood to be a choice of law to govern the arbitration agreement.

We accepted the first step in this argument (paras 67–68), but not the second. We pointed out that the curial law which governs the arbitration process is conceptually distinct from the law which governs the validity and scope of the arbitration agreement and said that whether a choice of the curial law carries any implication that the parties intended the same system of law to govern the arbitration agreement—and, if so, the strength of any such implication—must depend on the content of the relevant curial law (para 69).

In Enka the chosen seat of arbitration was London and the relevant curial law was therefore the English law governing arbitration contained in the Arbitration Act 1996. It was argued by Enka, and accepted by the Court of Appeal, that the 1996 Act contains provisions which affect substantive rights under the arbitration agreement that are intertwined with, and cannot readily be separated from, procedural provisions of the Act; and that this justifies an inference that, by choosing English law as the curial law, the parties intended their rights under the arbitration agreement also to be governed by English law. We rejected this argument. We agreed that there is a close relationship between provisions of the Arbitration Act 1996 concerned with the arbitration agreement and provisions of the Act concerned with the arbitration process and that the distinction between them is not always clear or easy to draw. But it cannot be inferred that the parties intended their rights under the arbitration agreement to be governed by English law. One conclusive reason is that the 1996 Act contemplates and specifically provides for a situation in which the arbitration agreement will be governed by a foreign law even though the curial law is English law. That makes it impossible to deduce that, just by choosing an English seat and with it English law as the curial law, the parties intended English law to govern their arbitration agreement.

It was in this context, and in support of the overlap argument, that counsel for Enka cited Carpatsky Petroleum Corpn v PJSC Ukrnafta [2020] EWHC 769 (Comm); [2020] Bus LR 1284. This was a claim to enforce in England and Wales an arbitration award made in Sweden. Enforcement of the award was resisted on the ground (among others) that the arbitration agreement pursuant to which the award was made was invalid. That argument depended on the contention that the arbitration agreement was governed by the law of Ukraine. If, as the claimant contended, the arbitration agreement was governed by Swedish law, it was indisputably valid as the Swedish court had already decided that the arbitration agreement was valid under Swedish law.

The judge (Butcher J) held that, having argued in the arbitration and in court proceedings in Sweden that the arbitration agreement was governed by Swedish law, it was not open to the defendant to change its position on the issue. But in case that was wrong he considered what the applicable law was, applying the common law rules. The contract provided for the “law of substance of Ukraine” to apply “on examination of disputes”. The judge, at para 67, interpreted this provision as meaning that Ukrainian law was to apply to the substantive issues which formed part of a dispute between the parties, but held that it was not a choice of law to govern the arbitration agreement itself.

This was a case, therefore, in which there was no choice of law to govern the whole contract including the arbitration clause. It was in this context that the judge then considered whether a choice of Swedish law to govern the arbitration agreement could be inferred from the choice of Sweden as the seat of the arbitration. He reasoned, first, that, by choosing Sweden as the seat, the parties should be taken to have agreed to the application of the Swedish Arbitration Act, including section 48 which provides that:

“where an arbitration agreement has an international connection, the agreement shall be governed by the law agreed upon by the parties. Where the parties have not reached such an agreement, the arbitration agreement shall be governed by the law of the country in which, by virtue of the agreement, the proceedings have taken place and shall take place.” (See Carpatsky, para 70.)

Expert evidence of Swedish law established that, for the purpose of this provision, only an express choice of law to govern the arbitration agreement is sufficient to displace the application of the law of the place of the arbitration. The judge considered that the parties should be taken to have known and agreed that, by failing to make an express choice of law for the arbitration agreement and by providing for a Swedish seat, the effect would be that the arbitration agreement would be governed by Swedish law. The final step in his reasoning, at para 70, was that:

“The parties can be taken to have intended that if Swedish law was to be the governing law of the arbitration agreement when the matter was looked at in Sweden, it should be the governing law of the arbitration agreement wherever it was looked at.”

Butcher J concluded that there was “an implied choice of Swedish law as the law governing the arbitration agreement”; alternatively, Swedish law as the law of the seat applied because it was the law with which the arbitration agreement had its closest connection (para 71).

In Enka, at para 72, Carpatsky was distinguished on the basis that there is no provision in the Arbitration Act 1996 similar to section 48 of the Swedish Arbitration Act (although we noted that the law in Scotland is different as the Arbitration (Scotland) Act 2010 does contain a similar provision).

It was in these circumstances unnecessary to examine the reasoning in Carpatsky further. But in light of that decision Lord Hamblen and I, in our conclusion on the overlap argument, contemplated the possibility that its reasoning could apply in another case. We said, at para 94:

“While a choice of seat and curial law is capable in some cases (based on the content of the relevant curial law) of supporting an inference that the parties were choosing the law of that place to govern the arbitration agreement, the content of the Arbitration Act 1996 does not support such a general inference where the arbitration has its seat in England and Wales.”

This was reflected in the summary at para 170(vi)(a) of the judgment, quoted at para 33 above.

All that was actually decided in Enka was therefore that a choice of seat for an arbitration in England and Wales does not support an inference that the parties are thereby choosing the law of England and Wales to govern the arbitration agreement. The suggestion that it might be possible, based on the content of the relevant curial law, to draw such an inference from a different choice of seat and curial law was obiter and was not explored. No attempt was made to prescribe in advance of a case in which the question arose, when such an inference could properly be drawn. Nor would it have been appropriate to seek to do so.

Although in Enka it was sufficient to distinguish Carpatsky and unnecessary to subject the reasoning in that case to close scrutiny, it is necessary to do so now because it underpins RusChem’s argument on the governing law issue. Transposed to the present case, the reasoning is as follows:

I will assume in RusChem’s favour that the first step in this reasoning is valid—although it attributes to commercial parties and their legal advisers when they are choosing a place of arbitration a degree of legal foresight which goes beyond what it may in practice be realistic to expect. But the argument breaks down altogether at the second step.

At first sight the idea that the question “what law governs the arbitration agreement?” should be answered in the same way in whichever jurisdiction this question is asked seems attractive. Consistency of approach between the courts of different countries is clearly desirable when questions arise about the validity or scope of an arbitration agreement. In an ideal world the situation that occurred in Kabab-Ji, where the English courts held that the arbitration agreement was governed by English law while the French courts held that it was governed by French law,would not occur. One way of avoiding such inconsistency would be for transnational principles to be developed which all national courts apply. This appears to be the aspiration which underlies the approach of the French courts. Another way would be to treat one jurisdiction as what might be called the “lead jurisdiction” whose identification of the law governing the arbitration agreement the courts of other countries will follow. If this approach were to be adopted, the obvious candidate to be regarded the “lead jurisdiction” is the place where the arbitration has its seat, as that is the legal order in which arbitration proceedings are anchored. It is the courts of the seat which have control over the proceedings and are the courts with primary responsibility for deciding questions about the constitution of the arbitral tribunal, the validity of the arbitration agreement or the validity of an award.

Such an approach would, of course, only achieve consistency if it were to be generally adopted. So far as I am aware, there is no jurisdiction in the world which has adopted such an approach. When the idea is examined further, it is evident that there is good reason for this.

At least as desirable as transnational consistency—and best calculated to promote it—is to have a rule which is clear and simple to apply. A rule which treated the arbitration agreement as governed by whichever law the courts of the seat would regard as the law governing the arbitration agreement would be neither clear nor simple to apply. It would have the consequence that, in every case where the parties have chosen a foreign seat for the arbitration, evidence of that country’s law would have to be obtained in order to know what law governs the arbitration agreement. This would introduce significant complication. Particular complication would arise where the relevant foreign law allows the parties to choose the law which is to govern the arbitration agreement (as most legal systems are likely to do) and the contract containing the arbitration agreement also contains a governing law clause. It would then be necessary to determine how the relevant foreign law would answer the question whether the law of the contract or the law of the seat prevails in this situation. That might be a substantial issue as it was, as regards English law, in Enka.

Such considerations show that a rule which treats the arbitration agreement as governed by whatever law the courts of the seat would treat as the law which governs it would in fact be a very unsatisfactory rule for any legal system to adopt. Partly for this reason, it seems improbable that such a rule might become widely adopted. Transnational consistency is far more likely to be achieved in the long term by coalescence around one or other of two default rules: either the rule endorsed by this court in Enka which treats a choice of governing law for the contract as a whole as applying to an arbitration agreement which forms part of the contract (unless the parties specifically agree otherwise); or a rule of the kind recommended by the Law Commission which treats the arbitration agreement as governed by the law of the seat (unless the parties specifically agree otherwise). At present there is no international consensus in favour of either rule. According to Gary Born,International Commercial Arbitration, 3rd ed (2021), pp 553–558, courts in the Netherlands, Japan, India and Australia have adopted a similar approach to this court in Enka. So does the Restatement (Third) of the US Law of International Commercial and Investor-State Arbitration (2019): see §4-14 comment b. But, as noted above, Sweden has adopted the opposite approach.

In deciding this appeal this court is not engaged in a legislative exercise of deciding what would be an optimum rule. That is for the Law Commission and Parliament. The question for us is what the parties to the arbitration agreements in the bonds must be taken to have intended when choosing Paris as the place of arbitration. What this discussion shows, however, is that there is no valid basis for imputing to the parties an intention that, if the arbitration agreements in the bonds would be treated as governed by the French rules of international arbitration when the matter is looked at in France, those rules should govern the arbitration agreements wherever the matter is looked at. It is not enough to justify imputing an intention to contracting parties that it would be a reasonable intention for them to have had. But that is a prerequisite. For the reasons given, an intention that the arbitration agreement should be governed by whatever law a court of the seat would regard as the law which is to govern it would not be a reasonable intention to attribute to the contracting parties (without express words to that effect). Still less therefore is it an intention which they must be taken to have had.

A further objection is that the putative intention attributed to the contracting parties involves an elaborate process of ratiocination that no one transacting business, or commercial lawyer for that matter, would realistically engage in—or could sensibly be expected to engage in—when agreeing on a place as the seat for the arbitration. The parties always have it in their power to agree what system of the law should govern their arbitration agreement. But where there is no language in their contract which would reasonably be understood as recording such an agreement, the court should not strain artificially to find one by attributing to the parties an unrealistic process of reasoning. Instead, the court should apply the rule of law which operates in the absence of party choice.

Having now been required to scrutinise the reasoning in Carpatsky closely, I conclude that it does not stand up on analysis. For the reasons given, even where the law of the seat contains a provision such as section 48 of the Swedish Arbitration Act, no inference can properly be drawn from a choice of seat that the arbitration agreement was intended to be governed by the law of the seat which is capable of displacing the general principles outlined in para 170(iv) and (v) of the judgment in Enka. What was said in para 170(vi)(a) should therefore in future be disregarded.

Thus, in Carpatsky, para 71, the judge was wrong to infer, based on the reasoning that I have just considered, that there was “an implied choice of Swedish law as the law governing the arbitration agreement.” The proper conclusion was that there was no agreement on a choice of law to govern the arbitration agreement in that case. It was therefore necessary to fall back on the rule which applies where the parties have not agreed on a choice of law to govern the arbitration agreement. In Carpatsky the judge took this fall-back rule to be that the arbitration agreement is governed by the law with which the agreement has its closest connection, which—as confirmed in Enka—is generally the law of the seat (see para 23 above). That was not strictly correct, but the error was not material to his conclusion. The closest connection test applies when a question about the validity or scope of the arbitration agreement arises, as it did in Enka, before an award has been made. Where, as in Carpatsky, a question about the validity of the arbitration agreement arises after an award has been made which the successful party is seeking to enforce in England, the matter is governed by section 103(2)(b) of the Arbitration Act 1996. Under that provision, in the absence of party choice, the validity of the arbitration agreement is governed by “the law of the country where the award was made”. By section 100(2)(b), an award is to be treated as made at the seat of the arbitration. So the judge was right to conclude that the validity of the arbitration agreement was governed by Swedish law as the law of the seat.

If the contracts in this case had been in materially similar terms to the contract in Carpatsky except with Paris rather than Stockholm chosen as the place for the arbitration, the proper conclusion would likewise have been that the arbitration agreements in the bonds are governed by French law as the law of the seat. But this would not be because an intention that French law is to govern the validity of the arbitration agreements can properly be inferred from the choice of a French seat combined with knowledge of what a French court would regard as the applicable law. It would be because, in the absence of agreement on a choice of law to govern the arbitration agreements, the law of the seat would apply as the system of law with which the arbitration agreements are most closely connected.

The bond contracts, however, are not in materially similar terms to the contract in Carpatsky. There is a critical distinction. As mentioned earlier, the judge in Carpatsky found that the unusually worded governing law clause in the contract meant only that Ukrainian law was to apply to the substantive issues which formed part of a dispute between the parties and was not a choice of law to govern the arbitration agreement itself. On the proper interpretation of the contractual documents there was therefore no law chosen by the parties to govern the arbitration agreement. By contrast here the bonds contain a governing law clause which, in accordance with the reasoning in Enka, is properly construed as applying to all the provisions of the bonds including the arbitration clauses. As was also held in Enka, this conclusion is not displaced by the choice of a seat of arbitration and curial law which is different from the law chosen to govern the contracts (see para 22 above). The fact that the courts of the seat would take a different view and regard their own law as the law governing the arbitration agreement is not a good reason to reach a different conclusion.

The short answer to RusChem’s argument that the arbitration agreements in the bonds are governed by French law is therefore the correct answer. It is as clear in this case as it was in Kabab-Ji that, applying the rules of contractual interpretation of English law as the law of the forum, the parties have agreed that the arbitration agreements in the bonds are governed by English law. It follows that the Court of Appeal was correct to hold that UniCredit’s claim falls within the contract gateway for service of proceedings out of the jurisdiction.

The proper place issue

Establishing that the claim falls within a gateway does not by itself suffice to obtain permission to serve the claim form on the defendant out of the jurisdiction. The claim must have a real prospect of success, which is not disputed here. In addition, CPR 6.37(3) provides that “[the] court will not give permission unless satisfied that England and Wales is the proper place in which to bring the claim”.

This question most commonly arises when a claimant wishes to bring a substantive claim for relief in the courts of England and Wales and the defendant asserts that there is another available forum which is more appropriate for the trial of the action. The basic principle to be applied in this situation, as stated in the leading case of Spiliada Maritime Corpn v Cansulex Ltd [1987] 1 AC 460, 476, is that the English court should not exercise jurisdiction if there is “some other available forum, having competent jurisdiction, which is the appropriate forum for the trial of the action, ie in which the case may be tried more suitably for the interests of all the parties and the ends of justice”. This principle (often referred to as “forum non conveniens”) applies not only where the court’s permission is required for service out of the jurisdiction but also where the defendant has been served with the claim form in England and Wales but seeks a stay of the proceedings on the ground that the case could more suitably be tried elsewhere. What differs is the burden of proof. When the claim form has been served in England and Wales, the defendant has the burden of satisfying the court that there is another available forum which is clearly more appropriate than England and Wales for the trial of the action. Conversely, if the court’s permission to serve out of the jurisdiction is required, the claimant has the burden of satisfying the court that England and Wales is clearly the appropriate forum for the trial of the action: see Spiliada at pp 476–481.

In Spiliada the House of Lords was not addressing the situation where the parties have agreed on a forum for the resolution of the dispute. In such cases it is not relevant to evaluate whether a forum other than the English court is more appropriate or suitable for the trial of the action. The basic principle applied is “pacta sunt servanda” (agreements must be kept). As Lord Hobhouse pointed out in Turner v Grovit [2001] UKHL 65; [2002] 1 WLR 107, para 25, where a person has a contractual right to be sued only in a particular forum, that person “does not have to show that the contractual forum is more appropriate than any other; the parties’ contractual agreement does that for him”.

The position where there is an agreed choice of court was authoritatively stated by Lord Bingham in Donohue v Armco Inc [2001] UKHL 64; [2002] CLC 440, para 24:

“If contracting parties agree to give a particular court exclusive jurisdiction to rule on claims between those parties, and a claim falling within the scope of the agreement is made in proceedings in a forum other than that which the parties have agreed, the English court will ordinarily exercise its discretion (whether by granting a stay of proceedings in England, or by restraining the prosecution of proceedings in the non-contractual forum abroad, or by such other procedural order as is appropriate in the circumstances) to secure compliance with the contractual bargain, unless the party suing in the non-contractual forum (the burden being on him) can show strong reasons for suing in that forum.”

Where the contractually agreed forum is a court, reasons which may, depending on the circumstances, be of sufficient strength to justify declining to enforce the contractual bargain include, as well as matters such as delay in seeking relief or submission to the jurisdiction of another court, inconvenience and potential injustice that would otherwise result from allowing parallel claims to be litigated in different jurisdictions. It was on that ground that the House of Lords in Donohue held that an anti-suit injunction should not be granted in that particular case.

Where the contractually agreed forum is arbitration, the policy of securing compliance with the parties’ contractual bargain is further reinforced by the strong international policy of giving effect to agreements to arbitrate disputes. The main pillar on which international arbitration rests is the 1958 Convention on the Recognition and Enforcement of Foreign Arbitral Awards, known as the New York Convention, which now has more than 170 state parties and has been implemented through national legislation in almost all contracting states. Article II(3) of the New York Convention provides:

“The court of a Contracting State, when seized of an action in a matter in respect of which the parties have made an agreement within the meaning of this article, shall, at the request of one of the parties, refer the parties to arbitration, unless it finds that the said agreement is null and void, inoperative or incapable of being performed.”

This mandatory rule is implemented in England and Wales by section 9 of the Arbitration Act 1996, which requires the court to stay proceedings brought in breach of an arbitration agreement in such circumstances.

In a case falling within this provision, the court has no discretion in the matter. It must stay the proceedings. If the proceedings are brought abroad rather than in England, the court is not obliged to grant an injunction to restrain the prosecution of the proceedings. But, as in cases where the parties have agreed to submit the dispute to a specified court, strong reasons are required to displace the prima facie entitlement to enforce the contractual bargain. Furthermore, unlike in cases where the contractually agreed forum is a court, the existence or risk of parallel proceedings is not a factor which in itself carries any weight. Not only is this possibility inherent in the choice of arbitration as a method of dispute resolution (given that arbitration proceedings cannot be consolidated with court proceedings or, in the absence of consent, with another arbitration); but to treat it as relevant would be inconsistent with the mandatory policy embodied in article II(3) of the New York Convention.

The Russian Federation is a party to the New York Convention, bound therefore by article II(3) when an action is brought in a Russian court in respect of a matter covered by an arbitration agreement to refer the parties to arbitration if one of them so requests. Here, however, the Russian court is prevented from doing this by the national legislation described at para 6 of this judgment. Yet that legislation does not bind an English court or affect the validity of the relevant arbitration agreements under English law which, as discussed above, is the law by which those agreements are governed. Under English law the arbitration agreements in the bonds are valid, RusChem’s claim for payment under the bonds falls squarely within the scope of those agreements, and it is a breach of contract for RusChem to pursue its claim for payment under the bonds in the Russian courts. RusChem has not attempted in these proceedings to argue otherwise.

In such circumstances, if the parties had chosen an English seat of arbitration, the English court would not hesitate to enforce the parties’ bargain by issuing an injunction to restrain a party over whom it has personal jurisdiction from commencing or continuing foreign proceedings in breach of the arbitration agreement. That has been clear at least since the decision of the Court of Appeal in Aggeliki Charis Cia Maritima SA v Pagnan SpA (The Angelic Grace) [1995] 1 Lloyd’s Rep 87, where Millett LJ said, at p 96, that “the time has come to lay aside the ritual incantation that this is a jurisdiction which should only be exercised sparingly and with great caution”. He observed that, while such an approach has much to commend it where an injunction is sought on the ground of forum non conveniens, there is no good reason for diffidence in granting an injunction to restrain foreign proceedings brought in breach of an arbitration agreement “on the clear and simple ground that the defendant has promised not to bring them”. As Millett LJ further explained:

“The justification for the grant of the injunction … is that without it the [applicant] will be deprived of its contractual rights in a situation in which damages are manifestly an inadequate remedy. The jurisdiction is, of course, discretionary and is not exercised as a matter of course, but good reason needs to be shown why it should not be exercised in any given case.”

This approach has been endorsed and repeatedly followed in later cases, including by this court in Enka, at paras 180–184. Examples of matters which may be relevant to the exercise of the court’s discretion are (as in cases where the contractually agreed forum is a court) delay in applying for an anti-suit injunction or the fact that the applicant submitted to the jurisdiction of the foreign court: see eg Ecobank Transnational Inc v Tanoh [2015] EWCA Civ 1309; [2016] 1 WLR 2231, paras 132–137; SAS Institute Inc v World Programming Ltd [2020] EWCA Civ 599; [2020] 1 CLC 816, paras 113–114.

In The Angelic Grace, Enka and other cases in this line of authority, the parties had agreed to arbitration in England. The question which arises in this case is whether it makes any—and, if so, what—difference to the test which the court should apply in deciding whether to exercise jurisdiction over a foreign defendant to grant anti-suit relief that the seat of arbitration is not within England and Wales.

Both parties have approached this question on the assumption that, in deciding whether England and Wales is the proper place in which to bring a claim for such relief, the test of forum non conveniens as elaborated in the Spiliada case applies. In my view, that is an erroneous assumption. As I have explained, that test is designed to deal with a different situation: one where (a) the claimant wishes to bring a substantive claim for relief in the English courts, (b) the defendant asserts that there is another available forum which is more appropriate for the trial of the action, and (c) no forum has been contractually agreed. The object of the test is to seek to ensure that, in these circumstances, the case is allocated to whichever forum, among those available, is the most suitable place for the trial.

The situation here is different in two significant respects. First, neither party suggests that the courts of England and Wales are an appropriate forum for the trial of the substantive dispute about whether RusChem is entitled to payment under the bonds. Second, the parties have contractually agreed to refer this dispute to arbitration. The English court is therefore not concerned, as it was in Spiliada, with whether England is the forum conveniens but only with whether to enforce the parties’ agreement.

In the argument in these proceedings it has nevertheless been assumed that the Spiliada test should be applied to determine whether UniCredit’s claim to enforce the contractual choice of forum may be brought in England. I do not consider that the test is apt for this purpose. That is because I do not think it right to accept that there is only one court (at most) which can properly exercise jurisdiction over a party for the purpose of preventing that party from breaking its contract to arbitrate a dispute, so that the English court should automatically decline to grant relief unless satisfied that it is clearly the most suitable tribunal to do so. Rather, the appropriate starting-point is that stated by the Court of Appeal in Enka: that in principle “[i]t is desirable that parties should be held to their contractual bargain by any court before whom they have been or can properly be brought”: see [2020] EWCA Civ 574; [2020] Bus LR 1668, para 57 (Popplewell LJ). It should be noted also that the important statement of principle in Donohue, quoted at para 67 above, is expressed in general terms and is not confined to cases where either the forum agreed by the parties or the forum in which proceedings are brought in breach of that agreement is the English court.

Although not directly analogous, two cases are of potential relevance. In Airbus Industrie GIE v Patel [1999] 1 AC 119 victims of an aircraft crash in India, who included British citizens living in London, sued the manufacturer of the aircraft (Airbus) in Texas. Airbus applied in England for an anti-suit injunction to restrain the British claimants from pursuing the Texas proceedings. The House of Lords held that an injunction should not be granted, for reasons given by Lord Goff of Chieveley. He approached the case on the basis that, “[a]s a general rule, before an anti-suit injunction can properly be granted by an English court to restrain a person from pursuing proceedings in a foreign jurisdiction in cases of the kind under consideration …, comity requires that the English forum should have a sufficient interest in, or connection with, the matter in question to justify the indirect interference with the foreign court which an anti-suit injunction entails” (p 138). In that case, although India was the natural forum for the resolution of the dispute, the Indian courts could not grant an anti-suit injunction because they did not have jurisdiction over the British claimants. They were therefore not an alternative available forum for the grant of anti-suit relief. But the House of Lords held that it would be inconsistent with comity for the English court to intervene as the English court had no interest in, or connection with, the matter: see pp 140–141.

The Airbus case is helpful in showing that, where the English court is asked to grant an anti-suit injunction to restrain proceedings in another forum because a third forum is the appropriate forum for the resolution of the substantive dispute, the test for determining whether the English court should exercise jurisdiction is not whether the English court is the most suitable forum for granting anti-suit relief. It is whether the intervention of the English court is consistent with comity. This was held to require that the English forum has a sufficient interest in, or connection with, the subject matter of the case. As Lord Goff emphasised, however, the House of Lords was not concerned in Airbus with cases where the choice of forum is the subject of a contract between the parties (see p 138F); and the requirement to show a sufficient interest or connection was specifically tied to the problem of comity seen as arising in “cases of the kind under consideration” in the Airbus case.

It is easy to see why in Airbus it was considered inconsistent with comity for the English court to interfere, even indirectly, to prevent proceedings which could more suitably be tried in India from being pursued in a court in Texas. The position is very different where an injunction is sought to restrain the pursuit of foreign proceedings brought in breach of an agreement to refer the matter to arbitration. In the first place, it cannot be an objection that the English court is not itself the appropriate forum for the resolution of the substantive dispute since, as Professor Adrian Briggs has pointed out, injunctions to enforce arbitration agreements are always granted by a court which is neither the natural nor the agreed forum, as no court is supposed to be resolving the dispute between the parties: see Adrian Briggs,Civil Jurisdiction and Judgments, 7th ed (2021), para 28.07.

Secondly, when the obligation to refer the dispute to arbitration is the subject of international agreement among the states concerned, considerations of comity have little, if any, role to play. As Millett LJ said in The Angelic Grace, at p 96:

“The courts in countries … party to … the New York Convention, are accustomed to the concept that they may be under a duty to decline jurisdiction in a particular case because of the existence of an … arbitration clause. I cannot accept the proposition that any court would be offended by the grant of an injunction to restrain a party from invoking a jurisdiction which he had promised not to invoke and which it was its own duty to decline.”

As mentioned already, the Russian Federation is a party to the New York Convention. Accordingly, although the legislation described at para 6 above prevents the Russian court from complying with its duty under article II(3) of the New York Convention to decline jurisdiction in this case, there can be no violation of comity in the English court granting an injunction to restrain RusChem from invoking that jurisdiction, and none is suggested. Nor is there any breach of comity as regards the French courts. France, too, is a party to the New York Convention and recognises and respects the policy of upholding agreements to arbitrate. A French court could have no objection to an English court taking steps to enforce the arbitration agreement in this case and the evidence of French law positively confirms that the French courts would have no objection to the grant of an anti-suit injunction by the English court.

A case more directly relevant than Airbus to the situation here is IPOC International Growth Fund Ltd v OAO CT-Mobile LV Finance Group [2007] CA (Bda) 2 Civ; [2007] Bda LR 43, a decision of the Court of Appeal for Bermuda. The claimant in that case applied to the court in Bermuda, where the defendant company was incorporated, for an injunction to require the defendant to discontinue proceedings it had brought in the courts of Russia in breach of agreements to arbitrate the claims in question in Switzerland and Sweden. The judge’s decision to grant such an injunction was upheld by the Court of Appeal. The main issue in the appeal was whether the Bermudian court was entitled as a matter of law to issue the injunction on the basis that it had personal jurisdiction over the defendant; or whether, as the defendant argued, it was also necessary to show that the Bermudian court had a sufficient interest in the matter. The defendant argued that there was no sufficient interest when Bermuda was not the seat of the arbitration.

That argument was rejected. In a judgment given by Sir Murray Stuart-Smith JA, the Court of Appeal held, at para 45, that the court could grant an injunction provided it had personal jurisdiction over the defendant. That was so because the defendant was domiciled within the court’s territorial jurisdiction, which was not a tenuous link. There was no requirement that the court must have a further interest in the resolution of the dispute itself.

It is unnecessary to express a view in this case on whether it will always amount to a sufficient connection to justify intervention by the English court to restrain breach of an agreement to arbitrate in a foreign seat that the English court has personal jurisdiction over the defendant—either because the defendant has been served with the claim form in England and Wales or because, if service out of the jurisdiction is necessary, jurisdiction is established through one of the gateways. As in IPOC International Growth Fund, this is not a case where jurisdiction is based on a tenuous link. There is a substantial connection with England and Wales in the fact that the contractual rights which UniCredit is asking the court to enforce are rights governed by English law.

A question which should be considered is whether bringing a claim for an injunction in the English court is itself compatible with the arbitration agreements. Where parties have agreed that disputes between them should be referred to arbitration in England, the view has been taken that, by choosing England as the seat of the arbitration, they have impliedly agreed that proceedings to uphold that agreement may be brought in the English courts: see Thomas Raphael,The Anti-Suit Injunction, 2nd ed (2019), para 7.58; Sheffield United Football Club Ltd v West Ham United Football Club plc [2008] EWHC 2855 (Comm); [2009] 1 Lloyd’s Rep 167, paras 39–40; Nomihold Securities Inc v Mobile Telesystems Finance SA (No 2) [2012] EWHC 130 (Comm); [2012] Bus LR 1289, paras 45–47; and Enka, para 174. This reasoning does not apply where the parties have chosen a foreign seat for the arbitration. It cannot then be inferred that the parties have impliedly agreed to any proceedings being brought in the English courts.

It is a further question, however, whether the implied negative obligation not to litigate disputes which the parties have agreed to resolve by arbitration should be construed as extending to court proceedings brought to uphold that very agreement or otherwise support the arbitral process. I think it clear that it should not. To construe an arbitration agreement as prohibiting any such proceedings would defeat its purpose. If the obligation on the courts of every contracting state under article II(3) of the New York Convention to refer the parties to arbitration is to be capable of performance, it is obviously necessary that proceedings brought for this purpose should not themselves be treated as contrary to the arbitration agreement. The same applies to any other proceedings brought to enforce the agreement. To make commercial sense of the agreement and give it such efficacy as the parties must intend it to have, an agreement to refer disputes to arbitration must be interpreted as not impliedly prohibiting a party from applying to a court for relief needed either to hold the other party to its agreement or to support the process of arbitration.

As regards interim measures of protection, this principle of compatibility is reflected in the UNCITRAL Model Law on International Commercial Arbitration, 1985 (amended 2006), which has been adopted in, or has influenced the arbitration law of, more than 120 jurisdictions around the world (including England and Wales). Article 9 of the Model Law states:

“It is not incompatible with an arbitration agreement for a party to request, before or during arbitral proceedings, from a court an interim measure of protection and for a court to grant such measure.”

The interim measures contemplated by the Model Law include orders directing a party to “refrain from taking action that is likely to cause … current or imminent harm or prejudice to the arbitral process itself”: see article 17(2)(b). This wording is clearly wide enough to encompass anti-suit injunctions. Indeed, the travaux préparatoires show that the words “or prejudice to the arbitral process itself” were included specifically to make it clear that court proceedings brought in breach of the arbitration agreement, or otherwise used to obstruct the arbitral process, may be restrained by granting interim measures: see Report of the Working Group on Arbitration and Conciliation on the work of its 43rd session (Vienna, 3–7 October 2005), paras 20–26. In addition, it is apparent from article 1(2) of the Model Law that article 9 and article 17J (which recognises the power of the court to grant interim measures) apply even if the place of arbitration is in the territory of another state. Thus, as stated in the Explanatory Note, para 22:

“Article 9 … is ultimately addressed to the courts of any State, insofar as it establishes the compatibility between interim measures possibly issued by any court and an arbitration agreement, irrespective of the place of arbitration.” (Emphasis added.)

Under the Model Law, therefore, it is not incompatible with an arbitration agreement for a court to grant an anti-suit injunction as an interim measure of protection before or during arbitral proceedings, regardless of where the arbitration has its seat. It may be noted too that the ICC Rules of Arbitration, applicable under clause 12 of the bonds, provide at article 28(2) that before an arbitral tribunal is appointed, “and in appropriate circumstances even thereafter, the parties may apply to any competent judicial authority for interim or conservatory measures”.

In Channel Tunnel Group Ltd v Balfour Beatty Construction Ltd [1993] AC 334, the House of Lords accepted that it would not be incompatible with an agreement to arbitrate disputes in Brussels for an English court to grant an interim injunction to restrain a threatened breach of the underlying contract. Lord Mustill said, at p 365:

“The purpose of interim measures of protection … is not to encroach on the procedural powers of the arbitrators but to reinforce them, and to render more effective the decision at which the arbitrators will ultimately arrive on the substance of the dispute. Provided that this and no more is what such measures aim to do, there is nothing in them contrary to the spirit of international arbitration.”

The Channel Tunnel case was decided before section 2(3) of the Arbitration Act 1996 expressly extended the powers of the English court to grant interim relief in support of arbitration proceedings to arbitrations with a foreign seat. Section 2(3) provides that the powers conferred by sections 43 and 44 of the 1996 Act (exercisable in support of arbitral proceedings) apply:

“even if the seat of the arbitration is outside England and Wales or Northern Ireland or no seat has been designated or determined— … but the court may refuse to exercise any such power if, in the opinion of the court, the fact that the seat of the arbitration is outside England and Wales or Northern Ireland, or that when designated or determined the seat is likely to be outside England and Wales or Northern Ireland, makes it inappropriate to do so.”

The 1997 Supplementary DAC Report on the Arbitration Act 1996, at para 18, explains that section 2(3) is:

“based on a very clear policy: the English court should have effective powers to support an actual or anticipated arbitration that does not fall within section 2(1) [which applies where the seat of the arbitration is in England and Wales or Northern Ireland]. However, such powers should not be used where any other foreign Court is already, or is likely to be, seized of the matter, or where the exercise of such powers would produce a clash with any other more appropriate forum.”

Section 2(3) does not apply in this case because the Supreme Court has held that the source of the court’s power to grant an injunction to restrain foreign court proceedings brought in breach of an arbitration agreement is not section 44(2)(e) of the 1996 Act, which confers power to grant an interim injunction in support of arbitral proceedings, but section 37 of the Senior Courts Act 1981: see AES Ust-Kamenogorsk Hydropower Plant LLP v Ust-Kamenogorsk Hydropower Plant JSC [2013] UKSC 35; [2013] 1 WLR 1889, para 48. Section 37 gives the High Court a general power to grant an injunction (whether interim or final) “in all cases in which it appears to the court to be just and convenient to do so”. The powers conferred by section 44 of the 1996 Act are exercisable only “for the purposes of and in relation to arbitral proceedings” and therefore only when such proceedings are on foot or “proposed”: see section 44(3). The court’s power under section 37 of the 1981 Act is not limited in this way and, as the Supreme Court held, may be exercised for the purpose of enforcing the negative promise not to bring court proceedings contained in the arbitration agreement regardless of whether arbitration proceedings are in existence or anticipated.

In AES Ust-Kamenogorsk it was also held that the claim in that case for an injunction to restrain foreign proceedings brought in breach of an arbitration agreement fell within CPR 62.5, which governs the service of an “arbitration claim form” out of the jurisdiction. Where a claim falls within this rule, there is no requirement corresponding to the requirement in CPR 6.37(3) that permission to serve the claim form out of the jurisdiction will not be given unless the court is satisfied that England and Wales is the proper place in which to bring the claim. CPR 62.5 does not apply here because, although a claim for an order under section 44 of the 1996 Act comes within its scope regardless of where the arbitration has its seat, a claim for “some other remedy … affecting … an arbitration agreement” (which includes an anti-suit injunction) is not covered by the rule if the seat of the arbitration is or will be outside the jurisdiction.

I cannot discern any good reason for allowing service out of the jurisdiction on any defendant (subject to the court’s general discretion) of a claim form seeking an order under section 44 of the 1996 Act for interim relief in relation to an arbitration with a foreign seat, and yet imposing an additional test where the claim is for an injunction to restrain breach of the arbitration agreement by a defendant over whom the English court has personal jurisdiction under the contract gateway in CPR PD 6B, para 3.1(6)(c). This apparent anomaly in the procedural rules need not and should not, however, prevent the court from adopting a principled approach to the exercise of jurisdiction. In my opinion, the proper principle to apply in both cases is that expressed in section 2(3) of the 1996 Act. Service out of the jurisdiction should in principle be permitted unless, in the opinion of the court, the fact that the seat of the arbitration is or is likely to be outside England and Wales makes it inappropriate on the facts of the case to exercise the court’s jurisdiction to grant relief aimed at enforcing the arbitration agreement or supporting the arbitral process. This test should be applied consistently with the principle discussed above: that a strong reason needs to be shown as to why in the particular circumstances the court ought not to exercise its jurisdiction to restrain a breach of the parties’ contractual bargain.

I do not consider that the wording of CPR 6.37(3), which requires the court to be satisfied that England and Wales is “the proper place in which to bring the claim”, precludes this approach. Those words are capable of being read, in a case of this kind, against the background of a presumption which treats the courts of England and Wales as the proper place in which to bring the claim for an anti-suit injunction unless the fact that the arbitration has a foreign seat makes it inappropriate to do so.

RusChem contends that England is not the proper place in which to bring the claim made by UniCredit in this case and that the proper place is France. Before the judge, RusChem argued that, by choosing Paris as the seat of arbitration, the parties have chosen to be subject to the supervisory jurisdiction of the French courts, and it is therefore for the French courts to determine whether there has been a breach of the arbitration agreements in the bonds and, if so, what relief to grant. RusChem still maintains this argument. But the judge decided this issue in RusChem’s favour mainly because he considered that substantial justice can be done in an arbitration in Paris. RusChem relies on this finding and argues in the alternative that the proper place in which to bring the present claim is arbitration under clause 12 of the bonds. Both these contentions must therefore be considered.

It is true that in West Tankers Inc v Ras Riunione Adriatica di Sicurtà SpA (The Front Comor) [2007] UKHL 4; [2007] 1 Lloyd’s Rep 391, para 21, Lord Hoffmann described the power to grant an injunction to restrain foreign court proceedings as a valuable weapon in the hands of the court exercising supervisory jurisdiction over the arbitration. In the same case Lord Mance, at para 31, described anti-suit injunctions issued by the courts of the place of arbitration as “a highly efficient means to give speedy effect to clearly applicable arbitration agreements”. In Enka Lord Hamblen and I quoted these statements and, at para 174, described the grant of such injunctions as “A well established and well recognised feature of the supervisory and supporting jurisdiction of the English courts”. In both these cases, however, the parties had chosen England as the seat of arbitration and in neither case was it necessary to delve into the precise basis of the power to grant anti-suit injunctions or its relationship with either the supervisory or the supporting jurisdiction of the court.

It is generally accepted that the courts of the place where an arbitration has its seat have the sole responsibility for supervising the arbitration and the primary responsibility for supporting the arbitration process. As discussed above, where an arbitration has its seat in England, the English court will intervene, absent a strong reason to the contrary, to restrain a party from bringing proceedings in breach of the arbitration agreement. However, it is in fact clear on analysis that the power to grant such relief is not an aspect of either the supervisory or the supporting jurisdiction of the English court.

The precise extent of the court’s supervisory role is defined by national law. But the basic supervisory functions are to intervene in limited circumstances in arbitration proceedings—for example, by appointing an arbitrator in the absence of agreement or dealing with a challenge to the impartiality of an arbitrator—and to provide a forum for establishing the validity of an award or challenging its validity on grounds of lack of jurisdiction or procedural or substantive error. Under articles V(1)(a) and (e) and VI of the New York Convention it is the law of the seat which gives an award the binding force that enables it to be recognised and enforced internationally.

Had arbitration proceedings been commenced in which an issue had been raised about whether the arbitral tribunal had jurisdiction to decide whether UniCredit is liable to pay the sums claimed by RusChem under the bonds, it might be said that for the English court to decide that issue would encroach on the role of the court with supervisory responsibility. But that is not the situation here. No arbitration proceedings have been commenced or proposed. Nor has RusChem advanced any argument that an arbitral tribunal would lack jurisdiction. In particular, it has not been, and could not reasonably be, suggested that article 248.1 of the Russian Arbitrazh Procedural Code, on which RusChem has sought to rely in the proceedings brought in Russia, has any effect on the validity or enforceability of the arbitration agreements as a matter of English law, which (as discussed above) is the law that governs them. Under English law the agreements are valid and enforceable. The only question is whether the English courts can and should exercise their coercive power to enforce them by restraining RusChem from continuing the Russian proceedings. That is not a supervisory function which ought therefore to be left to the courts of the seat. As was rightly said in IPOC International Growth Fund, at para 35:

“The role of the courts of the seat of arbitration is to supervise the arbitration itself. They are not the only courts that can prevent a party breaking his contract to arbitrate.”

As discussed above, the powers exercisable by courts in support of arbitration proceedings include granting interim measures of protection. Such powers may include granting interim anti-suit injunctions—as contemplated, for example, by articles 17(2)(b) and 17J of the Model Law—and may in principle be exercised by courts other than the courts of the seat of the arbitration. In English law, however, it has been authoritatively established in AES Ust-Kamenogorsk that the source of the court’s power to grant anti-suit injunctions is not its jurisdiction to grant interim measures in support of current or intended arbitration proceedings but its general equitable jurisdiction under section 37 of the 1981 Act. The purpose of issuing such an injunction is to enforce the negative promise contained in the arbitration agreement not to bring court proceedings, which applies and is enforceable regardless of whether or not any arbitration proceedings are on foot or proposed which require support.

The fact, therefore, that, in relation to any arbitration which may in future be brought, the parties have chosen to be subject to the supervisory jurisdiction of the French courts is not itself a reason why an English court cannot or should not uphold the parties’ bargain by restraining a breach of the arbitration agreement.

This is not a case where the French court is already, or is likely to be, seized of the matter, nor where the exercise by the English court of its power to grant an anti-suit injunction would or might produce a clash with any exercise of jurisdiction by the French courts so as to give rise to any issue of comity. There is in fact no possibility that the French courts could be seized of the matter. Not only, as is agreed, do the French courts have no power to grant anti-suit injunctions, but uncontradicted evidence which was before the judge shows that the French courts would not have jurisdiction to determine a claim of any kind brought by UniCredit complaining of a breach by RusChem of the arbitration agreements in the bonds.

This evidence is contained in a report from Mathias Audit, a French law professor and practitioner specialising in international arbitration law. His report was originally prepared for a case on materially the same facts: Commerzbank AG v RusChemAlliance LLC [2023] EWHC 2510 (Comm); [2023] 2 Lloyd’s Rep 587. But it was also admitted in evidence at the trial in these proceedings. Professor Audit explains in his expert’s report that the fact that an arbitration has a French seat does not, of itself, confer jurisdiction on any French court to order interim relief. There is evidently no provision of French law comparable to CPR 62.5. Instead, jurisdiction depends on the ordinary French rules which determine when a French court has jurisdiction over a foreign defendant. On the facts a French court would not have jurisdiction over RusChem as, other than the seat of arbitration being Paris, there is no link between the parties or the subject matter of their dispute and France. In particular, RusChem is not established in France but in the Russian Federation; France is not the place of performance of the bonds; nor is France the place where any relief sought would be implemented. There is no suggestion that the position would be any different if the relief sought were final rather than interim relief.

Even when a foreign court would not otherwise have jurisdiction over the defendant, an undertaking by the defendant to submit to its jurisdiction can make the foreign court an available forum: see Dicey, Morris & Collins on the Conflict of Laws, 16th ed (2022), para 12-031; Sharab v Al-Saud [2009] EWCA Civ 353; [2009] 2 Lloyd’s Rep 160. But RusChem has not offered an undertaking to submit to the jurisdiction of the French court; nor could it do so consistently with its position that the Russian courts have exclusive jurisdiction over its claims under the bonds and that initiating or pursuing proceedings before a foreign court in connection with those claims is contrary to Russian law.

The upshot is that the French courts would not have jurisdiction to entertain a claim by UniCredit to enforce the arbitration agreements in the bonds. The French courts are therefore not even an available forum in which to bring such a claim. In any event, as I have explained, even if the French courts were an available forum, there is no reason which can be said to make it inappropriate for an English court to restrain a breach of the arbitration agreements by granting an injunction. In particular, the fact that any arbitration brought would have its seat in France does not amount to such a reason.

I turn to RusChem’s alternative case that the proper place for UniCredit to bring such a claim is in an arbitration commenced under the arbitration agreements in the bonds. RusChem emphasises that arbitrators have power, not merely to award damages, but to make an award ordering a party to refrain from bringing or to terminate court proceedings brought in breach of the arbitration agreement. Under the ICC Rules an arbitral tribunal also has power, as soon as it is constituted and at the request of a party, to order any interim or conservatory measure that it deems appropriate. Further, the ICC Rules provide for the appointment of an emergency arbitrator where a party needs urgent interim or conservatory measures that cannot await the constitution of an arbitral tribunal. Such measures could, again, in theory include an order directing the other party to refrain from bringing or to terminate court proceedings brought in breach of the arbitration agreement.

The judge accepted RusChem’s contention that UniCredit could in these circumstances obtain substantial justice in arbitration proceedings. But the Court of Appeal rejected this suggestion as “an illusion”: para 77. Its main reasons were, first, that any award or interim order made by an arbitral tribunal or emergency arbitrator granting anti-suit relief would not be enforceable in Russia; and, second, that without the protection of an anti-suit injunction from the English court RusChem would be likely to apply for and obtain from the Russian court an injunction to prevent UniCredit from commencing or pursuing an arbitration: paras 76–77. The Court of Appeal also accepted a submission made by counsel for UniCredit that it is an abuse of process for RusChem to assert that the proper forum for UniCredit’s claim is arbitration while simultaneously seeking to pursue proceedings in Russia on the basis that the agreement to arbitrate is unenforceable: para 78.

On this appeal counsel for RusChem criticised the Court of Appeal’s reasoning, denying that RusChem’s stance is an abuse of process and arguing that the assumptions made by the Court of Appeal about the difficulties of obtaining and enforcing an award were unsupported by evidence and in any case do not come close to showing that substantial justice cannot be obtained through arbitration proceedings. They also pointed out that in two materially identical cases, Commerzbank and Deutsche Bank have both commenced arbitrations against RusChem in Paris (see Deutsche Bank AG v RusChemAlliance LLC [2023] EWCA Civ 1144; [2023] Bus LR 1660). Those arbitrations remain on foot and in neither case has RusChem sought an anti-arbitration injunction. RusChem submits that there is nothing stopping UniCredit from likewise commencing an arbitration in Paris.

I do not think it necessary to reach any conclusion on the likelihood that RusChem would seek an anti-arbitration injunction from the Russian court if the injunction granted in these proceedings were lifted, though I see no reason to suppose that it would not. In both the Commerzbank and Deutsche Bank cases anti-suit injunctions were granted by the English court. So no inference can properly be drawn from what has happened in those cases about what RusChem would do in the absence of such an injunction. But the more fundamental reason why substantial justice could not be obtained through arbitration proceedings is that any award or order made by an arbitrator has no coercive force. It is not backed by the powers available to a court to enforce performance of its orders, which include sanctions for contempt of court. An order made by an arbitrator creates only a contractual obligation. RusChem is already under a contractual obligation not to bring proceedings against UniCredit in the Russian courts. That obligation did not deter it from doing so. There is no reason to think that adding a further contractual obligation not to bring such proceedings would have any greater effect. RusChem’s conduct demonstrates that it would not.

The undisputed evidence of French law adduced by UniCredit in this case shows that French courts would have no power to enforce any order made by an arbitral tribunal directing RusChem not to pursue, or to discontinue, proceedings in Russia. It is also clear that, as the Court of Appeal found, such an order would not be enforceable in Russia. RusChem is wrong to say that there was no evidence to support that finding. Evidence of Russian law not challenged by RusChem shows that article 248.1 of the Arbitrazh Procedural Code (described in para 6 above) renders an arbitral award in proceedings falling within that provision unenforceable in Russia. That any arbitration commenced by UniCredit would be regarded as falling within article 248 is not only clear from the evidence of Russian law but has been conclusively established by the decision of the Russian court holding that, by reason of article 248, the arbitration agreements are unenforceable.

In these circumstances the Court of Appeal was right to reject the contention that UniCredit could obtain substantial justice in arbitration proceedings.

It is unnecessary to decide whether the Court of Appeal was also right to characterise RusChem’s contrary argument as abusive. It is, to put it no higher, unattractive for RusChem, whilst contending in the Russian proceedings that the arbitration agreements are invalid and unenforceable, at the same time to be seeking in these proceedings to benefit from the arbitration agreements by arguing that the proper place for UniCredit to bring a claim for redress is in an arbitration commenced under them. In response to a similar argument advanced by the respondent in AES Ust-Kamenogorsk, Lord Mance observed, at para 41, that a party is entitled to benefit by the existence of an arbitration agreement, but normally only by asserting it, eg by commencing an arbitration or applying for a stay of court proceedings in favour of arbitration. RusChem plainly has no intention of doing either. It is sufficient, however, to draw from RusChem’s conduct the obvious conclusion that for UniCredit to seek relief in arbitration proceedings would be wholly ineffectual to prevent RusChem from breaking its agreement to arbitrate.

Accordingly, neither the French courts nor arbitration proceedings are a forum in which UniCredit could obtain any, or any effective, remedy for RusChem’s breach (and threatened further breach) of the arbitration agreements. The fact that the seat of any arbitration would be in France provides no reason why the English court should refrain from upholding UniCredit’s English law contractual rights by granting an anti-suit injunction. Furthermore, even if—contrary to what I consider the correct approach to be—a test of forum conveniens were to be applied, it would yield the conclusion that England and Wales is the proper place in which to bring this claim.

Conclusion

For these reasons, the appeal has been dismissed. The Court of Appeal was entitled to make the order that it did granting final relief to UniCredit which includes a mandatory injunction requiring RusChem to discontinue its Russian proceedings. That order therefore stands.