When the European Commission launched the process of revising the Brussels I Regulation in 2000,1 it identified five main objectives, one of which was to strengthen jurisdiction clauses. The efficacy of jurisdiction clauses had been weakened by the decision of the Court of Justice of the European Communities in Gasser2 where the Court held that the lis pendens provisions of the Brussels I regime were to apply even when the jurisdiction of one of the competing adjudicators lied in a jurisdiction clause. As a result, if proceedings were initiated first before a court of a Member State other than the chosen one, any court seized subsequently, including the chosen court, had to stay proceedings until the court seized first would rule on its jurisdiction. While the courts of some Member States might be able to decide relatively quickly on their jurisdiction, some others are notoriously slow. Proceedings initiated first in such jurisdictions could therefore significantly delay the resolution of the case, and hurt the party willing to argue the case before the agreed court.

This strategy, which came to be known as the Italian Torpedo,3 was denounced throughout Europe. It became one of the targets of the European lawmaker, and the Brussels I Recast eventually addressed the issue. Article 31 of the new Regulation4 (hereafter the ‘Brussels I Recast’ or the ‘Brussels I bis Regulation’) introduces a remedy aiming at protecting jurisdiction clauses. The remedy is to create an exception to the application of the lis pendens doctrine when parallel proceedings are pending before a court designated by a jurisdiction clause and the court of another Member State. In such a case, priority is given to the chosen court by obliging any other court, whenever seized, to stay proceedings until the chosen court has ruled on its jurisdiction,5 and to decline jurisdiction if the chosen court retains jurisdiction.6

In this paper, I shall first discuss whether the new remedy will genuinely address the issue raised by Gasser and offer the required protection to choice-of-court agreements (1). I will then focus on certain particular case scenarios where the operation of the new remedy might raise issues: conflicting and complex jurisdiction clauses (2).

1 THE FLAWS OF THE NEW REMEDY

Gasser was widely perceived as one of the worst decisions of the European Court of Justice. Unsurprisingly, therefore, the new remedy introduced by Article 31(2) and (3) of the Brussels I Recast has been well received by the vast majority of legal commentators.7

It may well be, however, that Gasser was not as bad as it has been portrayed, and that the issue that the new remedy attempts to address is more complex than it seems. The essential flaw of the new remedy is that it ignores the fact that, in most cases where a party initiates proceedings in another forum than the one designated by the jurisdiction clause, that party challenges either the existence or the applicability of the clause in the particular dispute.

Indeed, this was exactly the case scenario of Gasser. Despite the heavy criticisms that this ruling has received, it was not an obvious case. This is because whether the parties had actually agreed on any court was unclear. The dispute had arisen between an Austrian seller, Gasser, and an Italian buyer, MISAT. Gasser had sent invoices to MISAT which all included a clause providing for the jurisdiction of Austrian courts. MISAT initiated proceedings in Italy. Gasser later sued in Austria, arguing that the jurisdiction clause had conferred jurisdiction on Austrian courts. MISAT, however, challenged the jurisdiction of Austrian courts, on the ground that the jurisdiction agreement was non-existent. The actual acceptance by the Italian party of the jurisdiction clause appearing on the invoices was unclear,8 and it may well be that MISAT genuinely thought that there had not been an actual agreement on jurisdiction.

Gasser was a case where the existence of the jurisdiction agreement was disputed. But the applicability of the agreement could equally be challenged. For instance, the party relying on the jurisdiction agreement could argue that it was assigned to the defendant as a consequence of its involvement in the performance of the contract, which may or may not be the case depending on the applicable law.

Article 31(2) reads:

Without prejudice to Article 26, where a court of a Member State on which an agreement as referred to in Article 25 confers exclusive jurisdiction is seised, any court of another Member State shall stay the proceedings until such time as the court seised on the basis of the agreement declares that it has no jurisdiction under the agreement.

Article 31(2) mentions three times the term “agreement” as if assessing its existence did not require judicial determination. The truth of the matter, however, is that in most cases the existence of this “agreement” will be disputed by the parties. This raises two issues:

The first is whether it is a good idea to give priority to the chosen court on the mere ground that its jurisdiction is alleged by one party.9 This is doubtful. There is no empirical evidence that when the existence or applicability of a jurisdiction clause is disputed, it is more often the case than not that the clause actually exists or applies.10 While jurisdiction agreements are clearly good for legal certainty and international commerce, this is not enough to justify that parties merely alleging the existence of a clause be preferred over parties disagreeing. In truth, the issue raised by Italian torpedoes never was so much the fact that the non-chosen court would be asked to rule on its jurisdiction and on the enforceability of the jurisdiction clause, but rather that it would take so long for doing it. A better remedy would have been to establish a uniform procedure for disposing efficiently of jurisdictional challenges,11 but this is not the path that the European lawmaker chose to follow.

The second issue is more technical. It is how to determine whether the agreement exists for the purpose of Article 31(2). In practical terms, the issue would arise before a non-chosen court that would have been seized first, and before which the defendant would argue that the parties agreed on the jurisdiction of another court. Would the court seized first have jurisdiction to rule on the existence and applicability of the jurisdiction agreement?12

One could argue that accepting that the first court rules on the existence of the jurisdiction agreement would amount to largely reinstall Gasser, and that the European lawmaker clearly wanted to overrule this precedent. To fully advance the policy of protecting jurisdiction agreements, article 31(2) should thus be understood to refer not to existing agreements, but rather to allegedly existing agreements. Such interpretation, however, would facilitate greatly and indeed potentially encourage strategic behavior.

A better rule, then, might be to set up a specific standard of proof to be applied by the first court to find the likely existence of a jurisdiction agreement, which would trigger the application of article 31(2). The idea will be familiar to English lawyers, as specific standards of proof already apply in the United Kingdom to jurisdictional issues.13 Unsurprisingly, the question of the articulation of national standards of proof with the Brussels regime was referred to the European Court of Justice by the English House of Lords in Shevill.14 The European Court answered that, as the goal of the Brussels regime was not to unify rules of procedure of Member States, the issue was to be settled by application of the national law of the court seized.15 An English court seized first would thus only decline jurisdiction if satisfied that the claimant arguing that an agreement grants jurisdiction to another court has “much the better of the argument” 16 than the other party.

As always with references to national law, the issue with the Shevill ruling on this point is that courts in other Member States will have different standards of proof. Indeed, in a number of European national systems, courts either are not allowed to, or in practice do not, rule separately on jurisdiction. The idea of a special standard of proof to be applied to the establishment of jurisdiction is unknown. In these Member States, courts will rule both on jurisdiction and the merits, rather later than sooner, irrespective of the strength of their jurisdiction.

In jurisdictions where no special standard of proof applies to the establishment of jurisdiction, the court seized first would simply assess whether the alleged choice-of-court agreement exists and applies in the case at hand. A mere allegation of its existence would thus be enough to found the jurisdiction of the court to rule on its jurisdiction.

2 THE SCOPE OF THE NEW REMEDY

A) Conflicting Jurisdiction Clauses

The remedy introduced by Article 31(2) and (3) of the Brussels I Recast is designed to protect jurisdiction clauses in cases where only one jurisdiction clause is at stake. It is not meant to apply in cases where several clauses might apply.

In a number of case scenarios, the claims of the parties could indeed be argued to fall within the scope of several jurisdiction clauses that might provide for the jurisdiction of different courts. One could first think of the well-known case scenario of the battle of forms. The parties agreed on the essentials of the contract, and then sent to each other forms that might contain different jurisdiction clauses, typically in favor of the sender’s court. But one could also think of cases where the parties concluded several contracts, which contained different jurisdiction clauses. For instance, the parties could have concluded a master contract and a number of contracts implementing it. Both the master agreement and the contracts implementing it could contain different jurisdiction clauses. In each of these case scenarios, each of the parties could argue that its claims fall within the scope of a different clause and thus initiate proceedings in a different court. The result would be parallel proceedings pending as a consequence of conflicting jurisdiction clauses.

Recital 22 of the Brussels I Recast expressly excludes the situation of conflicting exclusive choice-of-court agreements from the scope of the new remedy. As a consequence, the general lis pendens rule applies. Recital 22 reads:

22. This exception should not cover situations where the parties have entered into conflicting exclusive choice-of-court agreements or where a court designated in an exclusive choice-of-court agreement has been seised first. In such cases, the general lis pendens rule of this Regulation should apply.

One important reason why the general lis pendens rule should apply in such situations is that the newly introduced remedy would simply not work in presence of more than one chosen court. This is because the remedy consists in giving priority to the chosen court. Such remedy only works if it has one beneficiary. Otherwise, several courts might claim that they should benefit from it, which would amount to saying that none of them would. The regime of conflicting clauses, however, still raises a number of issues.

1) Is there a Need for Defining the Concept of Conflicting Clauses?

Recital 22 explains that the general lis pendens rule applies in cases where the parties entered into conflicting jurisdiction clauses. The existence of conflicting clauses is thus one of the cases where the general rule applies.

The existence of a jurisdiction clause is both a legal and a factual issue. Applicable legal rules define what the legal requirements for such clause to be considered as binding on the parties are, and the facts necessary for such requirements to be met, must be found by the competent adjudicator. The general lis pendens rule will apply and allow the court seized first to rule on its jurisdiction if two conflicting jurisdiction clauses are found to exist.

At first sight, there would thus appear to be a need for defining the concept of conflicting clauses and, possibly, for setting a special standard of proof of the existence of conflicting clauses. The main purpose of the recast with respect to jurisdiction clauses was to avoid situations where a party could delay the proceedings by merely alleging that the court that he had seized had jurisdiction on the matter. The fear was not so much that the court seized would eventually ignore the jurisdiction clause and retain jurisdiction, but rather that it would do so late, after delaying considerably the resolution of the dispute.

In practical terms, however, a defendant who would want to behave strategically and rely on Recital 22 for that purpose would logically have seized his preferred court first. His goal would be to benefit from Recital 22 for avoiding the court chosen in an otherwise enforceable jurisdiction clause. If the other party had already initiated proceedings before the chosen court, the defendant would not gain anything from seizing another court and alleging that a conflicting jurisdiction clause exists. He would still have to go and litigate before the court seized first that he would have been attempting to avoid, as a consequence of the lis pendens rule.

The issue would thus be raised before a court seized first. The plaintiff would argue that this court need not decline jurisdiction in favor of the court seized second, despite the existence of a jurisdiction clause granting it jurisdiction, because of the existence of another clause granting jurisdiction to the court seized first.

The logic of the lis pendens rule is to fully trust, and thus to rely on the court seized first to establish its jurisdiction. As a consequence, it seems unavoidable that the first court should conduct a full analysis of the jurisdictional arguments of the parties.

In Gasser, the Court of Justice held:

48. Moreover, the court second seised is never in a better position than the court first seised to determine whether the latter has jurisdiction. That jurisdiction is determined directly by the rules of the Brussels Convention, which are common to both courts and may be interpreted and applied with the same authority by each of them (see, to that effect, Overseas Union Insurance, paragraph 23).17

Although the intent of the European lawmaker was to overrule Gasser, it is submitted that this statement reveals the general philosophy underlying the Brussels regime and the principle of mutual trust in the application of European rules of jurisdiction. There is no doubt, therefore, that the court seized first should have the power to fully address the argument of the plaintiff that a competing clause exists and grants jurisdiction to the first court. If the court seized second was then compelled, in the meantime, to stay proceedings and wait for the decision of the first court, there would still be room for strategic behavior in courts where getting a decision on jurisdiction takes a long time, such as Italy.

Would this be an acceptable outcome? Under the conflicting clause exception, could a party argue, for instance, that an oral jurisdiction agreement was concluded in favor of the court first seized without offering any serious evidence, and thus delay proceedings in the other chosen court until the time the court first seized would rule that no evidence of the existence of this oral clause was brought, and that the formal requirements of Article 25 were not met in any case? One remedy against such delaying tactics would be, again, to impose a certain standard of proof for the allegation of the existence of a jurisdiction clause to produce effect for the purpose of the lis pendens rule. In jurisdictions where no special standard of proof applies to the establishment of jurisdiction, the court seized first will simply assess whether the alleged choice-of-court agreement granting it jurisdiction exists and applies in the case at hand. A mere allegation of its existence will be enough to found the jurisdiction of the court to rule on its jurisdiction. In effect, there will be no need to define the concept of conflicting clauses allowing the court to make such assessment. In slow jurisdictions, and in jurisdictions where a separate ruling on jurisdiction cannot be obtained, the result will be that the mere allegation of the existence of a conflicting clause will suffice to delay the resolution of the dispute in the court that was designated by a clear choice-of-court agreement.

One last issue is whether the court seized second would necessarily have to stay proceedings for the time needed for the court seized first to rule on its jurisdiction. The lis pendens rule directs it to do so.18 But it is unclear whether it applies in a case where the existence of conflicting clauses is disputed. The court seized second could have a different view on the existence of a competing clause granting jurisdiction to the court seized first. It could consider that the alleged clause does not exist and that there is only one jurisdiction clause, which makes the court seized second the only chosen court. From the perspective of the court seized second, therefore, the case would fall within the scope of Article 31(2), and the general lis pendens rule would not apply. That court might therefore consider that it has no reason to wait for the ruling of the court seized first, as the purpose of the introduction of a new remedy was precisely to protect jurisdiction clauses in such a situation. The court seized second could thus consider that it may as well address the argument of the existence of a conflicting clause, whether under a lower standard of proof or not. The result would be that the introduction of the first proceedings would not delay proceedings before the chosen court, but that a risk of conflicting decisions would arise. The Brussels I bis Regulation does not expressly address the issue, but the European Court of Justice has repeatedly expressed its understanding that one of the most important objectives of the Brussels I regime is to avoid conflicting decisions.19 It seems, therefore, that the European Court of Justice would probably summon the principle of mutual trust to resolve the issue and impose on the court seized second the appreciation of the first court on the existence of competing clauses.

Reports of the death of the Italian torpedo might have been greatly exaggerated.20

2) Conflicting Clauses in Separate Contracts

While Recital 22 refers to conflicting exclusive choice-of-court agreements without any qualification, it must be kept in mind that the context is the operation of the lis pendens rule. Lis pendens exists “where proceedings involving the same cause of action and between the same parties are brought in the courts of different Member States”.21 The parties to the proceedings could each rely on different jurisdiction clauses, but that would be a useful move only if the other requirements for lis pendens were met. To start with, the parties should be the same in each proceeding. Most importantly, the proceedings should involve the same cause of action.

This last requirement could raise an issue if the parties were relying on jurisdiction clauses that were included in different contracts and narrowly drafted. In such a case, the parties would have to rely on different contracts not only from a jurisdictional perspective, but also from a substantive perspective. The result could be that, while the claims would arguably be related, they would not have the same cause of action, as they would each be grounded in a different contract, and indeed potentially in a different law.

Imagine that the parties concluded a master agreement and a number of contracts implementing it. The master agreement and implementation contracts contained different jurisdiction clauses which each provided that any dispute arising out of the present agreement/contract shall be litigated in the court of X/Y. The substantive claims of the parties could probably be analyzed each as a breach of the substantive clauses of each contract. However, if the jurisdiction clauses were drafted narrowly, they would only encompass claims grounded in the contract that they would refer to. Each of the parties would thus have to rely on a separate contract both from a jurisdictional and a substantive perspective. The claims would thus be grounded in separate contracts, and thus arguably have different causes of action.

In contrast, jurisdiction clauses drafted broadly to encompass all dispute arising out of, or in connection with, the relevant contract, would not raise the same issue, as they would allow parties relying on a such a clause to make a claim based on the breach of another contract, as long as it would be connected with the contract including the preferred jurisdiction clause. All encompassing jurisdiction clauses are certainly enforceable under the Brussels I bis Regulation, since Article 27 allows the parties to provide for the jurisdiction of courts to settle dispute in connection with “a particular relationship”, and not with a particular contract.22

The Court of Justice has not ruled on the specific issue of whether claims made pursuant to related contracts have different causes of action. However, in Gubisch, it stated in passing: “the same cause of action, that is to say the same contractual relationship”.23 In The Ship Tatry,24 the court accepted that an action brought by one group of cargo owners against a shipowner under separate but identical contracts and another action brought by owners of another part of the cargo against the same shipowner under contracts which were separate from but identical to those between the first group and the shipowner were related and thus governed by the related action rule of the 1968 Brussels Convention. It seems, therefore, that two claims made under two different contracts, whether related or not, would not be considered as having the same cause of action for the purpose of lis pendens under the Brussels regime.25

3) Chosen Courts and Non-Chosen Courts


The last issue is whether the remedy introduced by Article 31(2) of the Brussels I bis Regulation should only benefit courts designated by a jurisdiction clause, or whether it might benefit all courts. Recital 22 explains that the general lis pendens rule should apply in presence of conflicting clauses. As already underscored, there is no doubt that the main objective is to clarify that parallel proceedings initiated in courts which were all chosen cannot possibly be regulated by giving them all priority to rule on their jurisdiction. But what about courts other than the courts designated by each of the conflicting jurisdiction clauses? A party could decide to argue that none of the jurisdiction clauses applies, and that the court of a third member state has jurisdiction under Article 7.1 of the Brussels I Recast. Should this action fall within the scope of the remedy introduced by Article 31(2), or within the scope of the general lis pendens rule?

It could be argued that the rationale for reintroducing the general rule for cases of conflicting clauses was first and foremost that the new remedy would not work: priority could not be given to all competing adjudicators. If so, there would be no reason to apply the general rule to non-chosen courts. They could not benefit from the remedy, and it would be perfectly conceivable to direct them to stay proceedings until all chosen courts have declined jurisdiction.

B) Complex Clauses

Another instance where the operation of the remedy introduced by Article 31(2) of the Brussels I Recast will raise issues is complex clauses. Simple jurisdiction clauses provide for the exclusive jurisdiction of one single court, irrespective of which party initiates the proceedings. But more complex clauses are also conceivable. In particular, asymmetric clauses are common in certain industries such as the banking industry. Asymmetric clauses impose on one of the parties the obligation to sue in one single court, but give a range of options to the other party, who is thus entitled to initiate proceedings in different jurisdictions.

Complex jurisdiction clauses such as asymmetric clauses raise issues in the context of the lis pendens rule, because they grant jurisdiction to several courts. As already noted, the protection afforded by the newly introduced remedy is to grant priority to “the” chosen court over all other courts. Such a remedy can function only if there is one chosen court that can be given priority. If the clause grants jurisdiction to several courts, there are more than one chosen court. Granting priority to all chosen courts amounts to granting priority to none.

There is another reason, however, why it is unclear whether complex clauses are within the scope of the new remedy. This is because Article 31(2) limits its scope to jurisdiction clauses conferring exclusive jurisdiction. The regulation does not define the concept of an exclusive jurisdiction clause, but a first interpretation could be that an exclusive jurisdiction clause must be a clause providing for the jurisdiction of a single court, and that any clause granting jurisdiction to more than one court cannot be regarded as exclusive.26 The consequence of such interpretation would be that Article 31(2) would not apply to jurisdiction clauses granting jurisdiction to more than one court. The general lis pendens rule would thus apply.

A second interpretation, however, could be that exclusivity should be defined from the perspective of each party taken in isolation from the other. A clause could grant exclusive jurisdiction to two different courts if each party was obliged to sue in one single court, though not the same as the court where the other party would be under the obligation to sue. In this view, clauses granting jurisdiction to the “court of the defendant” could be exclusive despite the fact that the court would be different depending on who would initiate the proceedings.27 By contrast, asymmetric clauses such as those found in the banking industry, which give a range of options to the bank, would not be exclusive,28 and would thus fall outside of the scope of Article 31(2).

A third interpretation could be that the meaning of exclusivity is merely that the choice-of-court agreement excludes the jurisdiction of some courts. In that view, the number of courts which are granted jurisdiction is irrelevant; what matters is only that their jurisdiction excludes the jurisdiction of another court, that is the court(s) which would have had jurisdiction pursuant to the default jurisdiction rules of the forum, for instance the default rules of the Brussels I bis Regulation.29 Asymmetric clauses could thus be characterized as exclusive, including for the bank.30 The application of Article 31(2) would not be excluded on the ground that the clause was not granting exclusive jurisdiction.

For the purpose of this article, these distinctions are not completely academic. It is true that, if a jurisdiction clause falls outside of the scope of Article 31(2), the general lis pendens rule will apply. The court seized first will thus have jurisdiction. If a clause falls within the scope of Article 31(2), however, Recital 22 will lead to the same result. This is because Recital 22 not only excludes the application of the priority rule in case of conflicting clauses, but also “where a court designated in an exclusive choice-of-court agreement has been seised first”. The purpose of this last exception might have been to underscore that there was no need to protect a chosen court from the dangers of the lis pendens rule when the chosen court was seized first. But the rule has also the consequence of implicitly providing for the application of the general rule to cases where an exclusive jurisdiction clause, or several exclusive jurisdiction clauses, grant jurisdiction to several courts. The court first seized will have jurisdiction.

However, if the court seized first is a non-designated court, the applicability of Article 31(2) will become essential. If the new remedy does not apply, the general lis pendens rule will apply. By contrast, if the new remedy applies, all non-designated courts will have to stay proceedings until any designated court that was seized of the dispute declined jurisdiction.

1
Report from the Commission to the European Parliament, the Council and the European Economic and Social Committee on the application of Council Regulation (EC) No 44/2001 on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters, COM(2009) 174 final, p. 6.

2
Case C - 116/02 Erich Gasser GmbH v. Misat Srl [2004] I.L.Pr. 7.

3
M Franzosi, ‘Worldwide Patent Litigation and the Italian Torpedo’ (1997) 19 European Intellectual Property Review 382, 384.

4
Regulation (EU) No 1215/2012 of the European Parliament and of the Council of 12 December 2012 on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters (recast).

5
Brussels I bis Regulation, Art. 31(2).

6
Brussels I bis Regulation, Art. 31(3).

7
See, e.g., Sancho Villa, Diana, ‘Jurisdiction over Jurisdiction and Choice of Court Agreements: Views on the Hague Convention of 2005 and Implications for the European Regime’, Yearbook of Private International Law, vol. XII, 2010, p. 399, 417; Gaudemet-Tallon, Hélène & Kessedjian, Catherine, ‘La refonte du Règlement Bruxelles I’ Revue Trimestrielle de Droit Européen 2013, p. 435, 447.

8
Nuyts, Arnaud, ‘The Enforcement of Jurisdiction Agreements Further to Gasser and the Community Principle of Abuse of Right’ in Pascal de Vareilles-Sommières (ed), Forum Shopping in the European Judicial Area (Oxford, Hart Publishing 2007) 55, 62.

9
Briggs, Adrian, ‘What Should Be Done about Jurisdiction Agreements?’ Yearbook of Private International Law, vol. XII, 2010, p. 311, 320.

10
Id., at 321.

11
Id., at 335.

12
See, underscoring that such interpretation would be a logical consequence of the language of article 31(2), Dickinson, Andrew, ‘Surveying the Proposed Brussels I bis Regulation: Solid Foundations but Renovation Needed’ Yearbook of Private International Law, vol. XII, 2010, p. 247, 297; Arnaud Nuyts, ‘La refonte du Règlement Bruxelles I’ Revue Critique Droit International Privé 2013, p. 1, 52.

13
Briggs, Adrian, Agreements on Jurisdiction and Choice of Law (Oxford, Oxford University Press, 2008) 283 (hereinafter ‘Briggs, Agreements’).

14
Case C - 68/93 Fiona Shevill v. Presse Alliance [1995] ECR 415.

15
At para. 39.

16
Briggs, Agreements, 284; Collins, Lawrence (ed), Dicey, Morris and Collins on the Conflict of Laws, 15th ed (London, Sweet & Maxwell, 2012) para. 12-120.

17
Case C - 116/02 Gasser.

18
Brussels I bis Regulation, Art. 29(1).

19
See, eg, Case C-406/92 The Tatry v. The Maciej Ratja [1994] ECR I-5439, para 32.

20
Ratkovic, Tena & Zgrabljic Rotar, Dora ‘Choice of Court Agreements under the Brussels I Regulation (Recast)’ Journal of Private International Law vol. 9, 2013, p. 245, 267.

21
Brussels I bis Regulation, Article29(1).

22
Briggs, Agreements, 285.

23
Case C-144/86 Gubisch Maschinenfabrik v. Palumbo [1987] ECR 4861, at para. 15.

24
Case C-406/92 The Tatry.

25
See Simons, Thomas & Carpaneto, Laura, in Simons, Thomas, Hausmann, Reiner & Queirolo, Ilaria (ed.), Regolamento “Bruxelles I” Commento alla Regolamento (CE) 44/2001 e alla Convenzione di Lugano (Munich, IPR Verlag, 2012) Article 27 para 38.

26
In the words of Adrian Briggs and Peter Rees( Civil Jurisdiction and Judgments, 5th ed (London, Informa Law, 2009) para 2.122): “to give exclusive jurisdiction to two courts is inelegant, even self-contradictory (…) a choice of concurrent exclusive jurisdictions may be taken as a case in which the parties have agreed that the jurisdiction should not be exclusive …”

27
Such clause was validated by the European Court of Justice under Article 17 of the 1968 Brussels Convention, which only applied to exclusive jurisdiction clauses: see Case C-23/78, Meeth v. Glacetal [1978] ECR 2133.

28
Or would only be exclusive for one party. See Hartley, Trevor, Choice-of-court Agreements under the European and International Instruments (Oxford, Oxford University Press, 2013) para. 7.34: “Such an agreement is in effect exclusive as regards one party and non-exclusive as regards the other”.

29
See Magnus, Ulrich, in Magnus, Ulrich & Mankowski, Peter( ed.), Brussels I Regulation, 2nd revised ed (Munich, Sellier, 2011) Article 23 para. 145-146; Geimer, Reinhold & Schütze, Rolf, Europäisches Zivilverfarhensrecht 3rd ed. (Munich, Beck, 2010) Article 23 para. 166.

30
The only exception could be asymmetric clauses leaving all default options open to the bank. This would be the case, for instance, of a clause providing that the bank could sue in any court of competent jurisdiction, which would mean all courts having jurisdiction under default jurisdiction rules. Such clause would not exclude the jurisdiction of any court, and thus could be characterized as exclusive even in this third view.