I. INTRODUCTION
Where commercial parties sue each other concurrently in multiple courts, in light of or in spite of a jurisdiction clause, a threshold question arises. Which court should assess the clause? For cases internal to the European Union (EU), the Brussels I Recast Regulation (Recast) provides a deceptively simple answer.Footnote 1 Article 31(2) of the Recast grants priority to the EU Member State court on which the clause ‘confers exclusive jurisdiction’.
Doubt has existed for some time about whether Article 31(2) of the Recast applies to asymmetric jurisdiction clauses that confer exclusive jurisdiction only in respect of one party's proceedings.Footnote 2 English and German appellate courts recently decided it is an acte clair that it does. The English Court of Appeal and Bundesgerichtshof, in the same case involving Etihad and Air Berlin, declined to refer this question to the Court of Justice of the European Union (CJEU).Footnote 3 Those national courts also expressed opinions, in obiter, about the relevance of their conclusions to the Hague Choice of Courts Convention.Footnote 4 They opined that the convention, which unlike the Recast defines ‘exclusive choice of court agreements’,Footnote 5 does not apply to asymmetric ones.Footnote 6 For proceedings initiated since 1 January 2021, the Hague Choice of Courts Convention is the only multilateral instrument governing jurisdiction in force between the United Kingdom (UK) and the EU. The EU had, at the time of writing, rejected the UK's request to accede to the Lugano Convention, which governs jurisdictional questions arising among EU Member States, and Switzerland, Norway and Iceland. Footnote 7
As is well known, the Recast regulates jurisdictional issues in civil and commercial matters between EU Member State courts. Despite the UK's withdrawal from the EU, the Recast also continues to regulate jurisdictional issues as between EU Member State and English courts in several circumstances. One is in respect of appeals from decisions in proceedings originally initiated in an EU Member State or English court before the end of the transition period on 31 December 2020.Footnote 8 The other is in respect of proceedings, initiated after that date, if proceedings in the same matter, initiated prior to that date, are already pending in an English or EU Member State court.Footnote 9
Though other types of asymmetric jurisdiction clause exist, this article focuses on the most common type of asymmetric jurisdiction clause, which it calls a ‘Rothschild clause’, after the widely discussed French Cour de cassation Rothschild case.Footnote 10 An example of a standard variant of a Rothschild clause, like that in issue in the Etihad litigation, is as follows:
39.1 Jurisdiction
(a) The courts of (Frankfurt am Main), Germany have exclusive jurisdiction to settle any dispute … in connection with this Agreement …
(b) The Parties agree that the courts of (Frankfurt am Main), Germany are the most appropriate and convenient courts … and accordingly no Party will argue to the contrary.
(c) This Clause 39.1 is for the benefit of the [lenders] only. As a result, no [lender] shall be prevented from taking proceedings … in any other courts with jurisdiction. To the extent allowed by law, the [lenders] may take concurrent proceedings in any number of jurisdictions.Footnote 11
This clause, in subsection (a), designates the jurisdiction of Frankfurt, the ‘“anchor” court’.Footnote 12 The anchor court has jurisdiction over proceedings brought by the borrower (sometimes called ‘the non-option holder’) and the lenders (sometimes called ‘the option holder(s)’).Footnote 13 In subsection (c), it gives the lenders an option to seise another court or other courts with jurisdiction. There is also a waiver in subsection (b), by both the borrower and lenders, of their rights to challenge the jurisdiction of the anchor court. Not all standard variants of this type of clause contain such a waiver, and some contain a waiver that applies only to the borrower.Footnote 14
German and English courts have generally considered this variant of clause under Article 25 of the RecastFootnote 15 to mean that the anchor court has exclusive jurisdiction over proceedings commenced by the borrower and non-exclusive jurisdiction over proceedings commenced by the lenders.Footnote 16 As further explained below, Article 31(2) of the Recast engages when an EU Member State court, on which an agreement that is governed by Article 25 confers exclusive jurisdiction, is seised in addition to another Member State court.Footnote 17 It might, therefore, be thought that an anchor court should only assess a Rothschild clause when seised by the borrower over whom it has exclusive jurisdiction, and not when seised by the lenders over whom it has non-exclusive jurisdiction. This might be thought to be the case given that the anchor court has only non-exclusive jurisdiction over the lender. This is not, however, how German and English courts have applied Article 31(2). Instead, they have applied Article 31(2) in the scenario where the borrower has first brought proceedings contrary to the clause in a court which would, but for the clause, have general or special jurisdiction under the Recast,Footnote 18 and the lenders have subsequently seised the anchor court, in an effort to have the borrower bring its proceedings there.
The recent litigation involving Etihad Airways and Air Berlin is illustrative. Those parties had concluded a loan agreement containing a Rothschild jurisdiction clause, designating English courts as the anchor. After Air Berlin ceased operations, its insolvency administrator, Flöther, brought proceedings against Etihad in Berlin, alleging that Etihad owed approximately €2 billion. Six months later, Etihad sought declaratory relief in England. It argued that the claims in the Berlin proceedings were subject to the exclusive jurisdiction of the English court (then an EU Member State court), pursuant to the Rothschild clause in the loan agreement. The German courts and English courts, at first instance and on appeal, held that Flöther should bring his claims in England: the Berlin court was right to stay its proceedings under Article 31(2) of the Recast, once the English court had been seised.Footnote 19
For Rothschild clauses, ‘ubiquitous’ in lending and capital markets,Footnote 20 this conclusion appears to make commercial sense. If Article 31(2) does not apply when a borrower/issuer has first commenced proceedings in a non-designated court, the lender/dealer may be incentivised to commence proceedings in the anchor court precipitously to prevent an action by the borrower/issuer in a non-designated court.Footnote 21 That may, in turn, trigger a chain of events of default and cross-default.Footnote 22 Article 31(2) does not mitigate the need for a lender/dealer to commence proceedings in the anchor court, for the provision will not operate unless that court is seised; it does, however, mitigate the need (if there is one) for a lender/dealer to commence proceedings pre-emptively.Footnote 23
This article explores various lines of analysis which support the conclusion that Article 31(2) applies to the variant of Rothschild clause set out above. It considers the implications of that conclusion, beyond the scenario treated in the case law to date. It argues that while this conclusion is sound, the reasoning in the case law leading to it is reductive. Deciding that Article 31(2) applies to Rothschild clauses by reference to this isolated scenario, and disaggregating them into separate agreements for each party as the English Court of Appeal did, fails to account for the anomaly to which application of Article 31(2) to Rothschild clauses gives rise. Importantly, it ignores how Article 31(2) is to operate in the entirely conceivable scenario where a lender has, pursuant to its option, brought proceedings in a court with jurisdiction under the Recast, first, and the borrower subsequently brings proceedings in the anchor court. For instance, assume in the litigation described above, Etihad, pursuant to its option, had first sued Air Berlin in Berlin and Air Berlin subsequently brought proceedings in England under the anchor part of the clause. Article 31(2) in that scenario either defeats the lender's option entirely or it results in parallel proceedings. A characterisation of a Rothschild clause which treats the clause as one Article 25 agreement is needed for Article 31(2) to apply coherently to it in all likely scenarios.
The analysis proceeds as follows. It examines, first, the function, purpose and scope of Article 31(2) of the Recast, and how it relates to Article 25 and the EU law principle of res judicata (Part II). Part III examines four paths to the conclusion that Article 31(2) applies to Rothschild clauses, drawing on case law and literature that have considered the question. The anomaly created by the application of Article 31(2) to Rothschild clauses, regardless of which path one follows, is investigated in Part IV. A solution, based on an autonomous characterisation of this variant of clause under Article 25, is proposed in Part V. Since this article's objective is to examine the in limine issue of which court is competent to decide on the effects and enforceabilityFootnote 24 of a Rothschild clause under the Recast, it does not concern itself with the question of whether Rothschild clauses are enforceable under that instrument.Footnote 25
II. RECAST, ARTICLES 25 AND 31(2), AND RES JUDICATA
The Recast's core provision on jurisdiction agreements is Article 25. It provides that an agreement between the parties confers jurisdiction on the court or courts of an EU Member State to decide ‘any disputes … in connection with a particular legal relationship’. The concept of a jurisdiction ‘agreement’ is autonomous under EU law,Footnote 26 as is the requirement that it be the subject of consensus in fact between the parties.Footnote 27 Though the requirement of ‘a particular legal relationship’ is also autonomous, it is intimately linked with the issue of scope (ie whether the agreement applies to the dispute in question),Footnote 28 however conceptually distinct the two issues may be.Footnote 29 The issue of scope is a question of interpretation for national courts, to be undertaken, at least according to the majority view, by reference to national law.Footnote 30 Article 25 further provides for a presumption that the ‘jurisdiction’ of the designated court shall be exclusive unless the parties have agreed otherwise.’Footnote 31
Though the Recast does not define exclusive jurisdiction, it does appear to be a concept with a ‘consistent, autonomous meaning’.Footnote 32 An agreement that confers exclusive jurisdiction will, by virtue of Article 25 of the Recast, ‘exclude both jurisdiction as determined by the general principle of the defendant's courts laid down in Article 4 thereof and the special jurisdictions provided for in Articles 7 to 9 thereof’.Footnote 33 This means that the exclusively designated court will have no discretion to decline that jurisdiction,Footnote 34 and Member State courts which would otherwise have general or special jurisdiction have, by virtue of the agreement, no jurisdiction to exercise.Footnote 35
The concept of ‘non-exclusive jurisdiction’ is more obscure; those words do not appear anywhere in the Recast or in the jurisprudence of the CJEU. By reference to the concept of exclusivity just considered, however, non-exclusive (or not exclusive) jurisdiction must mean jurisdiction which does not affect the competence of other fora with general or special jurisdiction.Footnote 36 If that is right, a designated court with non-exclusive jurisdiction, first seised, must exercise that jurisdiction.Footnote 37 A designated court with non-exclusive jurisdiction, second seised, must decline jurisdiction in favour of a court first seised that has general or special jurisdiction, provided the proceedings involve the same parties, the same cause of action and have ‘the same end in view’Footnote 38.Footnote 39
Article 31(2) of the Recast provides for the court on which an Article 25 jurisdiction agreement confers exclusive jurisdiction, if it is seised,Footnote 40 to decide on the effects and enforceability of the agreement designating it.Footnote 41 The Recast allows the exclusively chosen court to do so, even if another Member State court has already been seised.Footnote 42 Once the exclusively chosen court has established jurisdiction in accordance with the agreement, any other Member State court must decline jurisdiction.Footnote 43 Recital 22 of the Recast describes the provision as ‘an exception to the general lis pendens rule’.Footnote 44 This implies that both sets of proceedings must involve the same parties and ‘cause of action’, and have ‘the same end in view’.Footnote 45 The exception, therefore, appears not to apply to related proceedings, where at least one of those aspects in the two sets of proceedings is not the same but where the proceedings are ‘so closely connected that it is expedient to hear and determine them together’.Footnote 46
By introducing Article 31(2), the Recast differs from the Brussels ConventionFootnote 47 and Regulation 44/2001Footnote 48 which preceded it. Those instruments subjected all jurisdiction agreements to the lis pendens rule. According to the CJEU's ruling in Gasser, they required the court first seised to examine its own jurisdiction, considering the jurisdiction agreement designating another Member State's courts, and the designated court to stay its proceedings until the first seised court had concluded that examination.Footnote 49 The court first seised would eventually be required to decline jurisdiction, if it decided that the agreement conferred exclusive jurisdiction and was enforceable. Although the effect of an enforceable agreement conferring exclusive jurisdiction was that the court first seised had no jurisdiction,Footnote 50 some Member State courts took a long time to reach that decision.Footnote 51 This ‘torpedo’ problem, to which Article 31(2) of the Recast responds, is well-known and need not be revisited here.Footnote 52 Suffice it to say that article was introduced to dissuade a party from initiating proceedings in a court which—on the assumption that there was an enforceable exclusive jurisdiction clause designating the courts of another Member State—had no jurisdiction. The strategy behind such a manoeuvre was to delay the resolution of the dispute.
The Recast is said to change the identity of the court that should decide on the effects and enforceability of an agreement, where multiple Member State courts are seised.Footnote 53 There is continuing uncertainty, however, about the extent to which the court first seised must be satisfied that an applicable clause conferring exclusive jurisdiction exists for the obligation to stay its proceedings to engage.Footnote 54
On one view, Article 31(2) means that the court first seised makes no decisions that have binding effect on the clause's effects and enforceability, including whether it confers exclusive jurisdiction.Footnote 55 The court first seised need only be satisfied that there apparently, probably, prima facie or manifestly is a clause that confers exclusive jurisdiction on another Member State court, in order to stay its proceedings when that other court is subsequently seised.Footnote 56 This approach has the advantage of one court—and the court which appears to have been chosen—deciding with binding effect on all elements of the agreement, including the question of whether it is exclusive.
An opposing view, advanced by Professor Dickinson, considers that the court first seised should decide with binding effect on all aspects of the agreement that are governed by autonomous requirements, of which he considers the question of whether a clause confers exclusive jurisdiction to be one.Footnote 57 Provided those are satisfied, the court first seised should stay proceedings to allow the chosen court to decide on those requirements that the Recast subjects to national law. This view derives support from Recital 22, which does not expressly refer to exclusivity as an issue for the chosen court to decide. Since no Member State court is better placed than another to decide a matter which is subject to autonomous EU law,Footnote 58 it is also more efficient for the court first seised to resolve the exclusivity question and be done with it.
Which view is preferable? Is it enough that the court first seised considers there to be a clause that apparently/probably/prima facie confers exclusive jurisdiction for the obligation to stay its proceedings to engage?Footnote 59 (Clearly, a prima facie case that there is an agreement conferring non-exclusive jurisdiction will not be enough.Footnote 60) Or should the court first seised make a decision, with binding effect, that the clause confers exclusive jurisdiction, such that the chosen court must recognise it? The latter view is enticing because exclusivity is the hook on which Article 31(2) of the Recast turns, but the former is likely to be preferred by the CJEU. That is because it is more consistent with what Article 31(2) requires the court first seised to do, namely, to stay its proceedings rather than to decline jurisdiction. A decision of the court first seised to characterise the jurisdiction of the chosen court as exclusive and stay its proceedings does not bind the chosen court:Footnote 61 interlocutory decisions are only recognisable judgments if they regulate or determine the parties’ legal relationship.Footnote 62 That decision does regulate the parties’ legal relationship, but only provisionally.Footnote 63
This characterisation is not decisive of the jurisdiction of the court first seised (it has only stayed its proceedings; it has not (yet) declined jurisdiction).Footnote 64 Nor is it determinative of the designated court's jurisdiction: to the extent that the designated court's jurisdiction depends on an enforceable agreement conferring exclusive jurisdiction, the purpose of Article 31(2) in requiring the court first seised to stay its proceedings is to allow the designated court to decide, in all respects, whether there is one.Footnote 65 This makes sense. If the designated court were ultimately to decide that the clause confers on it non-exclusive jurisdiction, it would be odd if the court first seised were then bound to decline its jurisdiction on account of its own initial ruling that another Member State court had exclusive jurisdiction, in circumstances where that other court does not agree.
Granted, the logic of this approach generates an obvious inefficiency: arguments as to whether the clause confers exclusive jurisdiction will, in effect, be ventilated on at least two occasions. The facts of the litigation in Perella Weinberg Partners UK LLP v Codere SA are illustrative.Footnote 66 The scope for argument is, however, reduced if interpreting or characterising a jurisdiction clause is to be done wholly autonomously. That is to say, there will be less room for argument if the question of whether a given clause confers exclusive jurisdiction—a concept which, as has been seen, has an autonomous meaning—is answered without reference to national law.
Needless to say, if the court first seised has (in error)Footnote 67 declined jurisdiction on account of what it considers to be a clause conferring exclusive jurisdiction on another Member State court, that decision will nonetheless bind the designated court.Footnote 68
III. INTERPRETING AND CHARACTERISING ASYMMETRIC CLAUSES AS EXCLUSIVE
Whether a jurisdiction agreement confers exclusive or non-exclusive jurisdiction under Article 25 is said to be a question of ‘interpretation’ or ‘construction’ going to its effects;Footnote 69 whether a jurisdiction agreement on which Article 25 confers exclusive jurisdiction falls within Article 31(2) is said to be a question of ‘characterisation’.Footnote 70 But where construction stops and characterisation begins is murky, in light of the autonomous meaning of exclusivity, articulated above, and the rebuttable presumption of exclusivity contained in Article 25. Determining whether the presumption that a clause confers exclusive jurisdiction has been rebutted is either a question for national courts applying national law; or it is a question for national courts but is to be determined autonomously by examining the wording of the clause neutrally for clear indications that the parties intended to rebut the presumption.Footnote 71 If it is the latter, then the question is properly one of characterisation rather than construction.Footnote 72 Which of those it is, is uncertain.
In practical terms, this is a problem which affects asymmetric rather than symmetric clauses. By reference to the concepts of exclusive and non-exclusive jurisdiction canvassed above,Footnote 73 and the Recast's rebuttable presumption of exclusivity, characterising symmetric agreements for the purposes of Article 31(2) is simple. Article 25 presumes the jurisdiction of a Member State court designated in an agreement to be ‘exclusive unless the parties have agreed otherwise’.Footnote 74 Accordingly, an agreement that a court presumes to confer exclusive jurisdiction under Article 25 will be characterised as falling within Article 31(2); an agreement that a court characterises as having rebutted the presumption (or as conferring non-exclusive jurisdiction) will not.
The difficulty that Rothschild clauses present is that they are generally considered, for the purposes of Article 25 of the Recast, as conferring both exclusive and non-exclusive jurisdiction, so the rebuttable presumption of exclusivity arguably does little to aid their construction (or, properly put, their characterisation).Footnote 75 As already noted, German and English courts have generally determined a Rothschild clause to mean that the jurisdiction of the anchor court is exclusive under Article 25 when the borrower brings proceedings, and non-exclusive when the lender does. This means that, under the Rothschild clause set out in Part I above, only the courts of Frankfurt will have jurisdiction when the borrower sues, whereas the courts of Frankfurt as well as other courts competent under the Recast will have jurisdiction when the lender sues.Footnote 76 The question, then, is how a clause that confers both exclusive and non-exclusive jurisdiction under Article 25 should be characterised as exclusive for the purposes of Article 31(2). Four possibilities are examined.
A. Considered as a Whole?
One path to Article 31(2) is that, considered as a whole, Rothschild clauses confer exclusive jurisdiction. That was Cranston J's approach in Commerzbank Aktiengesellschaft v Liquimar Tankers Management Inc, a case concerning a Rothschild clause in a loan agreement and related guarantee. The lender warned the borrower that it would bring English proceedings against the borrower for non-payment, unless it received a settlement offer by 16 June 2015. The borrower commenced proceedings against the bank in Greece on that very date, contrary to the clause. The following year, the lender brought proceedings in England. The English High Court held that Article 31(2) applied to Rothschild clauses, characterised autonomously. Cranston J reasoned that:
Considered as a whole, they are agreements conferring exclusive jurisdiction on the courts of an EU member state … That this applies in respect of a claim by the [borrower] alone does not detract from this effect.Footnote 77
This logic solves the problem which arose in the factual scenario of the Commerzbank case, because it requires the court first seised by the borrower to stay proceedings in favour of the anchor court.
The problem with it is that it leads to the anomaly, foreshadowed above, where a lender has, pursuant to its option, brought proceedings in a court with jurisdiction under the Recast first, and the borrower brings proceedings in the anchor court second. In that scenario, it is unclear how a court first seised by the lender pursuant to its option (and thus consistently with the agreement) would not need to stay its proceedings under Article 31(2) to allow for the anchor court second seised by the borrower to rule on it, given that ‘considered as a whole’, the clause is an agreement conferring exclusive jurisdiction under Article 25.Footnote 78
It might be thought that this problem falls away if Cranston J's logic is inverted:
Considered as a whole, they are agreements conferring non-exclusive jurisdiction on the courts of an EU member state. That this applies in respect of a claim by the lender alone does not detract from this effect.
Because, according to this logic, Rothschild clauses as a whole are non-exclusive, Article 31(2) would not engage when either party brings proceedings pursuant to a Rothschild clause, meaning that the lis pendens rule would apply.Footnote 79
This logic means that in the factual scenario which arose in the Commerzbank case, the English anchor court, second seised by the lender under the Rothschild clause, would have needed to stay its proceedings until the Greek court first seised by the borrower, contrary to the agreement, has decided that it had no jurisdiction. That result would weaken the policy aims of Article 31(2), set out in Recital 22.Footnote 80
Characterising Rothschild clauses as non-exclusive creates a further difficulty.Footnote 81 The Greek court first seised must have characterised the agreement as non-exclusive under Article 25, in order to conclude that it need not stay its proceedings under Article 31(2). Having done so, it is unclear how the Greek court could then decline its general or special jurisdiction on the basis that another Member State court has exclusive jurisdiction under Article 25.
This difficulty may explain Professor Fentiman's remark that ‘[t]here is nothing incoherent about concluding that such agreements are non-exclusive for the purposes of Article 31(2)’ [of the Recast], but intractable problems arise in the context of Article 25’.Footnote 82 One way out of those problems is for the Member State court first seised to characterise the agreement as a whole as non-exclusive for the purposes of Article 31(2) but exclusive for the purposes of Article 25. Evidently, one court characterising an agreement as non-exclusive for the purposes of one provision of the Recast and, in the next breath, exclusive for the purposes of another has little to be said for it. Nor does it appear to be an approach that the Recast facilitates.
B. Because of the Borrower's Promise
Professor Merrett argues that the ‘the point is not so much that “considered as a whole” they are agreements conferring exclusive jurisdiction, as the judge put it in Commerzbank. Rather, each obligation can be considered on its own’.Footnote 83 Accordingly, another way of arriving at the conclusion that Article 31(2) will apply is because the borrower has, from the perspective of English law, ‘promised not to sue anywhere other than the chosen jurisdiction’.Footnote 84 From the perspective of German law, one could say that Article 31(2) will apply because the borrower has promised to sue only in the chosen court.Footnote 85 A German or English perspective is, however, only relevant to the extent that the Recast says it is.Footnote 86 As noted above, the CJEU has consistently ruled that an agreement on jurisdiction is an independent concept under the Recast and 2007 Lugano Convention.Footnote 87 It is not a reference to national concepts of jurisdiction by agreement.Footnote 88
The CJEU has not explained whether an agreement conferring jurisdiction is a promissory contract creating rights and obligations among the parties, or rather whether it is a dispositional agreement (a Verfügung) which directly affects the parties’ rights to bring and have proceedings brought before courts with competence under the Recast, but which does not produce obligations between the parties inter se.Footnote 89 As an autonomous EU concept, it is plausible that the juridical nature of an Article 25 jurisdiction agreement is neither. It follows that although the English High Court and Court of Appeal both approved of Merrett's approach in Etihad,Footnote 90 as a matter of EU law, it may not be right to conceptualise a jurisdiction agreement in terms of promise or obligation, as a common lawyer would.Footnote 91
C. Because the Clause Comprises Two Agreements or Two Unilateral Acceptances, and One is Exclusive
A further path to Article 31(2)'s application to a Rothschild clause is to characterise the clause as comprising two jurisdiction agreements or two unilateral acceptances of the anchor court's jurisdiction, one of which is exclusive, with respect to a single dispute within a particular relationship. Disaggregating a Rothschild clause into ‘separate’ jurisdiction agreements was the English Court of Appeal's approach in Etihad Airways PJSC v Flöther. Henderson LJ reasoned:
I can see no difficulty in regarding a composite jurisdiction agreement such as that contained in clause 33 of the Facility Agreement as comprising (a) an exclusive agreement in relation to claims brought by Air Berlin, and (b) a separate non-exclusive agreement in relation to claims brought by Etihad.Footnote 92
Professor Briggs has made a similar suggestion, albeit in a response to a different problem.Footnote 93 He suggests that:
it may be prudent for contracts which were intended to contain such asymmetric provisions to be recast so that they contain two jurisdiction agreements, separately numbered and clearly distinct from each other … Clause X … provide[s] that all proceedings in which A sues B shall be brought only before a particular court; and Clause Y that proceedings brought by B against A may be brought before any of a number of named courts.Footnote 94
Elsewhere, he has argued that one should conceive of jurisdiction agreements generally under the Recast as unilateral waivers,Footnote 95 by which parties unilaterally renounce their privileges of being sued in courts with general or special jurisdictionFootnote 96 and unilaterally accept the jurisdiction of courts that would not otherwise have it.Footnote 97 Persuaded by that analysis, the English Court of Appeal observed in Joint Stock Co Aeroflot – Russian Airlines v Berezovsky, in relation to Article 23 of Regulation 44/2001 that ‘the agreement of a party is not bilateral … but unilateral’.Footnote 98 Subsequent English authority, albeit at first instance, has cast some doubt on whether this approach can be reconciled with the Recast's autonomous requirement that a jurisdiction agreement be the subject of consensus in fact between the parties.Footnote 99
Applying Briggs’ conception to a Rothschild clause means that the lender (B in Briggs’ example, quoted above) waives its privileges of being sued in courts with general or special jurisdiction and accepts the jurisdiction of the anchor court. The borrower (A in his example) merely accepts the jurisdiction of the anchor court.
Framing a Rothschild clause in these terms misrepresents the relative jurisdictional advantages and disadvantages for each party under the clause. It suggests that it is advantageous to leave open the possibility of having to defend proceedings in one of any number of courts with special jurisdiction in addition to one's domicile. It suggests that it is disadvantageous to know in advance that proceedings will only have to be defended in one nominated court, that court often being the place of the lender's domicile, branch or registered office.
It is true that the CJEU has used the language of waiver in connection with jurisdiction agreements. In several early cases, the Court framed the question of whether a party (which was the defendant to the proceedings) had agreed to a jurisdiction agreement as one of waiving the advantage of the rules of jurisdiction under the Brussels Convention.Footnote 100 But there is nothing in the Court's decisions to support the conclusion that the Court was concerned with a waiver of those rules by a party only in its position as defendant. Indeed, in subsequent cases, the Court asked essentially the same question in determining whether a party (which was the claimant to the proceedings) had agreed to a jurisdiction agreement. It asked whether that party had agreed to derogate from the default rules, giving it the advantage of initiating proceedings against the counterparty in a court with jurisdiction under Regulation 44/2001.Footnote 101 It would be more cogent to describe an agreement as a waiver of a party's jurisdictional privilege to initiate proceedings in a court with general or special jurisdiction,Footnote 102 rather than a waiver of its jurisdictional privilege to defend proceedings in one of those jurisdictions.
In all events, so far as Article 31(2) of the Recast is concerned, it is unhelpful to construe a Rothschild clause as an exclusive jurisdiction agreement or acceptance for one party, and a distinct non-exclusive jurisdiction agreement or waiver for the other. For the reasons elaborated later, segregation of one clause into two separate agreements prevents the only apparent solution to the problems which Article 31(2), applied to Rothschild clauses, generates.Footnote 103 That solution depends on the interrelationship between the various aspects of the parties’ agreement.
D. Because it has Exclusive Effect Whenever the Borrower Brings Proceedings
A final path to the conclusion that Article 31(2) applies to Rothschild clauses is to interpret the words of that provision by reference to the exclusive effects that Rothschild clauses have under Article 25 when the borrower brings proceedings.Footnote 104 Article 31(2) provides that ‘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 [its] proceedings’.Footnote 105 As Walker J observed in Perella Weinberg v Codere, the inquiry for which Article 31(2) calls is whether ‘the party seeking to bring proceedings in a court of “another member state” has agreed that the dispute in question is to be subject to the exclusive jurisdiction of a court or the courts of another member state’.Footnote 106 The simple answer to that question, in the case of a Rothschild clause, is that the borrower has so agreed, because the clause has exclusive effect under Article 25 whenever the borrower brings proceedings.Footnote 107
If the borrower brings proceedings first in a Member State court, which is not the anchor court, Article 31(2) will apply when the anchor court is second seised, because the parties have agreed that the designated court should have exclusive jurisdiction over the borrower's proceedings. Once the court first seised has established, to the requisite standard, that the borrower has agreed that disputes brought by it should be subject to the exclusive jurisdiction of another Member State court,Footnote 108 the only remaining question is whether that other court ‘is seised … on the basis of the agreement’.Footnote 109 Again, the only answer to that question is ‘yes’.Footnote 110 The passive ‘is’ in Article 31(2) means that the designated court can be seised by either party on the basis of the agreement. It does not matter that it is seised by the party over whom the court has non-exclusive jurisdiction under that agreement,Footnote 111 as long as both sets of proceedings involve the same cause of action or have the same end in view.Footnote 112
E. Evaluation of Existing Approaches
Of the approaches so far considered, the most compelling is that in Section D: to characterise a Rothschild clause as one agreement with different effects for proceedings initiated by each party. It might be thought that the characterisation in Section D is so similar to the ‘separate agreement’ characterisation, criticised in Section C, and ultimately adopted by the English Court of Appeal in Etihad, that drawing any distinction between them is splitting hairs. The argument made here is that this is a distinction without a difference if only one party commences proceedings, such that only Article 25 of the Recast is engaged. But as soon as both parties initiate proceedings before Member State courts, the distinction becomes material.
Disaggregating the clause into two agreements, as the Court of Appeal did, prevents a solution to the anomaly that the application of Article 31(2) to Rothschild clauses creates.
IV. PROBLEMS WITH APPLYING ARTICLE 31(2) TO ASYMMETRIC CLAUSES
Regardless of which path one takes, Article 31(2) applied to Rothschild clauses creates an anomaly.Footnote 113 This anomalous scenario was foreshadowed in the literature when the Recast first took effectFootnote 114 and was raised by David Joseph QC of counsel in Etihad Airways PJSC v Flöther.Footnote 115 It occurs when the lender first commences proceedings in an EU Member State court with general or special jurisdiction, pursuant to the optional limb of the parties’ agreement. And the borrower subsequently seises the Member State court designated in the anchor limb of the parties’ agreement.
The anomaly is that application of Article 31(2) either defeats the lender's option entirely or it results in parallel proceedings. Either, (A) the Member State court first seised by the lender must stay its proceedings until the anchor court ‘declares that it has no jurisdiction under the agreement’,Footnote 116 which the anchor court cannot do, since it has exclusive jurisdiction under Article 25 of the Recast. Once the anchor court establishes its jurisdiction, the court first seised must decline jurisdiction,Footnote 117 thus defeating the lender's option. Alternatively, (B) the Member State court first seised need not stay its proceedings, on account of the clause conferring only non-exclusive jurisdiction in respect of the lender's proceedings of which it is seised. But the anchor court second seised cannot stay its proceedings, because the clause confers exclusive jurisdiction in respect of the borrower's proceedings of which it is seised, such that there are parallel proceedings.
A. Article 31(2) Defeats the Option
The optional court first seised by the lender may take the view that Article 31(2) requires it to stay its proceedings because a Member State court ‘on which an agreement … confers exclusive jurisdiction is seised’. The parties have agreed that the anchor court should have exclusive jurisdiction when the borrower brings proceedings and the borrower has seised the anchor court. Because the borrower must bring proceedings before the anchor court, the anchor court ‘seised on the basis of the agreement’ cannot declare ‘that it has no jurisdiction under the agreement’.Footnote 118 In other words, the anchor court will have no discretion to decline jurisdiction; such is the effect of exclusive jurisdiction.Footnote 119
The optional court first seised must decline jurisdiction in favour of that court under Article 31(3), once the anchor court ‘has established jurisdiction in accordance with the agreement’. If that is right, the lender's option is worthless if and when the borrower brings proceedings before the anchor court. That will be so regardless of whether the borrower's claims are for positive or negative declaratory relief.Footnote 120
In Section D of the previous Part, it was established by reference to the Etihad litigation that if the borrower has brought proceedings outside the anchor court under a Rothschild clause, and the lender seises the anchor court, then Article 31(2) will apply. That is because the parties have agreed that the anchor court should have exclusive jurisdiction when the borrower sues. Because of the lender having seised the anchor court, that court now ‘is seised’. The fact that the anchor court has been seised by the lender over whom the anchor court has non-exclusive jurisdiction is, according to Etihad, immaterial.
By reverse analogy of reasoning, it might be thought that in the scenario where the lender brings proceedings in a court pursuant to its option and the borrower later brings proceedings in the anchor court that Article 31(2) will not apply, because the parties have agreed that the anchor court should have non-exclusive jurisdiction when the lender sues.
The problem with that line of thinking is it ignores the original reasoning on which the analogy is based. If the original reasoning in Etihad is right, where the lender has brought its claims in the optional court, and the borrower seises the anchor court, Article 31(2) will apply because the parties have agreed that the anchor court should have exclusive jurisdiction when the borrower sues. Because of the borrower having seised the anchor court, that court now ‘is seised’ and the optional court first seised by the lender must stay its proceedings. The fact that the anchor court has been seised by the party over whom the court has exclusive jurisdiction must be immaterial.
B. Article 31(2) Results in Parallel Proceedings
Alternatively, the optional court first seised by the lender may take the view that Article 31(2) does not require it to stay its proceedings when the borrower subsequently sues in the anchor court. The optional court may consider that because the jurisdiction of the anchor court is, for proceedings brought by the lender, only non-exclusive, then Article 31(2) will not apply.
There is no obligation for the optional court first seised to stay where there is an agreement conferring non-exclusive jurisdiction on another Member State court,Footnote 121 which is the effect of the agreement for proceedings brought by the lender.Footnote 122 But equally, there is no possibility for the anchor court second seised to stay its proceedings or decline jurisdiction where there is an agreement conferring exclusive jurisdiction on it, which is what the agreement does for proceedings brought by the borrower. That is a problem.
It might be thought that the likelihood in practice of this being a problem is reduced by the possibility that the proceedings in the optional court may proceed quicker and that any interlocutory decision that court makes as to jurisdiction will bind the anchor court. For a court, seised of jurisdiction pursuant to the option, not to stay its proceedings under Article 31(2), it must first decide that the jurisdiction of the anchor court is non-exclusive so far as proceedings brought by the lender are concerned. That interlocutory decision by the court first seised is arguably a judgment under the Recast.Footnote 123 It is an interlocutory decision that is decisive, rather than merely provisional, of that court's own jurisdiction. Its decision and the reasons leading to it would have res judicata effect for the anchor court.Footnote 124
Even if that is so, that decision of the optional court will only have res judicata effect in the anchor court in respect of claims brought by the lender. The decision of a court seised under the option can only be an interlocutory decision that it has jurisdiction over claims brought by the lender and over set-off, if that were raised by the borrower before the optional court as a defence.Footnote 125
If the borrower chooses not to counterclaim in the proceedings before the court in the Member State where the lender has sued, it is hard to see how the borrower could be prevented from bringing its claims in the anchor court. And it is hard to see how the anchor court would rule that it has no jurisdiction to adjudicate those claims, which is the only way it could deseise of those proceedings.Footnote 126 If, instead, the borrower wanted to counterclaim in the optional court, it is probable, though not certain, that the lender would submit to those claims there.Footnote 127 But if the lender does not submit to the counterclaims, the borrower would be obliged to bring them in the anchor court, assuming the lender is domiciled in a Member State. That is because Article 8(3) of the Recast, which allows a person domiciled in a Member State to be sued ‘on a counterclaim … in the court in which the original claim is pending’, cannot be relied upon by the borrower. That ground of jurisdiction is not available, because proceedings brought by the borrower are subject to the exclusive jurisdiction of the anchor court and exclusive jurisdiction excludes special jurisdiction, including Article 8.Footnote 128 The result either way would be concurrent proceedings that risk culminating in irreconcilable judgments, contrary to one of the Recast's principal objectives.Footnote 129
If the anchor court were to render judgment first in time in respect of the borrower's claims, there would be no basis under the Recast on which the optional court could refuse to recognise that judgment: taking jurisdiction contrary to a jurisdiction agreement is not a basis on which a judgment can be refused.Footnote 130 Misapplication of the rules under the Recast will not provide a basis for refusing to recognise a judgment either.Footnote 131 Accordingly, if the court with optional jurisdiction continues to hear the lender's proceedings and its judgment is rendered first,Footnote 132 the anchor court must recognise it.
V. A SOLUTION
Put shortly, while application of Article 31(2) to Rothschild clauses may solve one problem, it risks creating others. That is not to say that Article 31(2) should not be applied to Rothschild clauses. Indeed, a lender losing the benefit of its option is arguably less pernicious than the torpedo problem Article 31(2) was introduced to solve.Footnote 133 But it does suggest that a solution which preserves the lender's option and mitigates the risk of parallel proceedings is needed, while also dissuading torpedo claims by the borrower.
One solution that suggests itself lies in a different approach to the interpretation or characterisation of Rothschild clauses. It seems that the only way the anchor court second seised could declare that ‘it has no jurisdiction under the agreement’Footnote 134 is if it considers the clause to mean that once the lender exercises its option and brings proceedings first in a court with general or special jurisdiction under the Recast, the effect of that choice is that the anchor court no longer has jurisdiction over proceedings brought by either party. For this solution to be effective, it would require Member State courts to adopt a uniform construction or an autonomous approach to the characterisation of Rothschild clauses of the type under discussion here. This is not a radical proposition: the courts of most Member States, with the notable exception of France, generally already characterise asymmetric clauses of this type that are subject to the Recast or 2007 Lugano Convention in the same fashion under national law, as has been seen.Footnote 135
That autonomous approach would require a Rothschild clause to be characterised as an agreement between the parties conferring exclusive jurisdiction, which one party may unilaterally renounce in its entirety. This approach is similar, in one respect, to the approach French courts take under their national rules of private international law.Footnote 136
Framed differently, but to a similar effect, it would require a Rothschild clause to be characterised as an agreement conferring exclusive jurisdiction that is dependent on a condition subsequent, over which the lender has sole power: the parties have agreed on the jurisdiction of the anchor court to the exclusion of courts with general or special jurisdiction, unless or until the lender exercises its power to sue elsewhere.Footnote 137
However it is framed, exclusivity under this construction is contingent or conditional. It is dependent on the condition subsequent not being fulfilled or the lender not unilaterally renouncing the jurisdiction clause.
Once the lender has renounced the clause conferring exclusive jurisdiction and brought proceedings in a court with general or special jurisdiction, the borrower may bring any counterclaims it has before that court.Footnote 138 Any other Member State court, seised by the borrower, would be required to decline jurisdiction in accordance with the general lis pendens rule.Footnote 139
The characterisation of Rothschild clauses proposed here would lead to a result consistent with that which Article 17(4) of the Brussels Convention produced. Unlike the Recast, which contains a rebuttable presumption of exclusivity, the Convention had deemed the jurisdiction of a court designated in any agreement to be exclusive, but it also made provision for an exclusive agreement to be concluded for the benefit of one party.
The effect of Article 17(4) of the Convention was that the benefitting party had a contingent right to bring proceedings in a court with general or special jurisdiction under the instrument. The right was contingent because if the non-benefitting party first seised the exclusively designated court, the party for whose benefit the agreement was drafted lost its right to bring proceedings in a court with general or special jurisdiction under the instrument.Footnote 140 Equally, ‘if the party for whose benefit the agreement was drafted seised a court with general or special jurisdiction first in time, the exclusive jurisdiction of the nominated court seised second in time by the non-benefiting party [appears to have] had no effect’.Footnote 141 That is because the benefitting party had, in seising another court, renounced the entire exclusive jurisdiction agreement (that is to say, for proceedings brought by either party).Footnote 142 Such was the result of the agreement being for its benefit.
To be clear, it was the Convention, not the jurisdiction clause that produced that result. What is being proposed here is a characterisation of the jurisdiction clause that leads to an equivalent result to the Brussels Convention; not an interpretation of the Recast that leads to an equivalent result to the Brussels Convention.Footnote 143 The latter is not possible, because the Recast does not contain a provision equivalent to Article 17(4).Footnote 144
Characterising Rothschild clauses, as currently drafted, in the manner proposed here is not a perfect solution, especially if both parties have expressly agreed to waive their right to object to the jurisdiction of the anchor court.Footnote 145 A submission by the lender that the jurisdiction of the anchor court has, on the proper characterisation of the jurisdiction clause, been lost looks very challenging indeed when the lender has expressly agreed not to contest that jurisdiction.
A preferable solution lies with drafters:Footnote 146 a clause should provide that the anchor court has no jurisdiction over proceedings brought by the borrower, if the lender has already seised the optional court. It should also provide that any express waiver of the parties’ right to object to the jurisdiction of the anchor court only applies if the option has not been exercised. Further, the clause should make clear that in these circumstances, where the lender has brought proceedings before an optional court, the borrower may bring its counterclaims there. But for the many Rothschild agreements still in circulation, characterisation by Member State courts seems to be the only sensible solution to the problem that applying Article 31(2) of the Recast to Rothschild clauses creates when the lender sues in the optional court first, and the borrower sues in the anchor court, second.
These solutions are unavailable if the approach to characterisation taken by the English Court of Appeal in Etihad Airways PJSC v Flöther finds favour with Member State courts. As explained above,Footnote 147 the Court held that Article 31(2) applied on the basis of an ‘exclusive agreement’ in relation to the borrower's claims, which was separate from the ‘non-exclusive agreement’ that would apply to the lender's claims. The court's separation of a Rothschild clause into two agreements under Article 25 necessarily precludes a solution to Article 31(2) that is based on the interaction between the various components of the parties’ agreement.
VI. CONCLUSION: LONG LIVE LUGANO, OR ALL EYES ON THE HAGUE?
It is unsurprising that appellate courts in England and Germany have concluded that Article 31(2) of the Recast applies to asymmetric jurisdiction clauses, as a means of curbing torpedo claims by borrowers outside the anchor court. But one should not arrive at that conclusion by conceptually separating the clause into two agreements, as the English Court of Appeal did, or encouraging parties to draft their Rothschild clause as two agreements.
Disaggregating the clause, whether conceptually or physically, prevents a solution to the anomaly Article 31(2) creates when a lender sues in the optional court first and the borrower sues in the anchor court second. Application of Article 31(2) means either parallel proceedings between the anchor court and optional court, or that the optional court must stay its proceedings thus defeating the lender's option. The solution proposed by this article, which rests on a uniform characterisation of the clause viewed as a whole, both preserves the lender's option and mitigates the risk of parallel proceedings.
The conclusion that Article 31(2) applies to Rothschild clauses is good news for financiers on the Continent whose clauses designate courts in Luxembourg or Frankfurt as the anchor. But it is little comfort for those financiers whose clauses designate English courts in the anchor. Those agreements now fall squarely outside the Recast.
Moreover, although no court of a Contracting State to the Hague Choice of Courts Convention appears to have conclusively decided the question, at least in the opinion of appellate courts in England and Germany, asymmetric jurisdiction clauses fall outside the jurisdictional provisions of that convention too.Footnote 148
Unless or until the EU changes its mind and allows the UK to accede to the Lugano Convention as it wishes, the result is that the effects in the EU of all Rothschild clauses designating English anchor courts are now governed by the national private international law rules of each Member State. The English High Court, in its former role as an EU Member State court, decided that the CJEU's decision in Gasser continued to apply in respect of a jurisdiction agreement governed by the Lugano Convention. As such, the solution that Article 31(2) of the Recast introduced should not be adopted. Jurisdiction agreements governed by the Lugano Convention, it held, continued to be subject to the lis pendens rule, given that the Convention has not been amended to reflect the Recast's exception,Footnote 149 and instead mirrors the provisions of Regulation 44/2001 which preceded it. As the court first seised, the English court considered that it should examine its own jurisdiction in light of the jurisdiction clause conferring jurisdiction on a Swiss court and should not stay its proceedings until the Swiss court had assessed the clause. The Zurich Handelsgericht, perhaps unsurprisingly, took the opposite view in the same case.Footnote 150 The Swiss view is consistent with some academic commentary that argues that the courts of European Free Trade Association Member States are free to depart from the ruling in Gasser. The basis of that argument is that the requisite equivalence between the 2007 Lugano Convention and Recast, on which the obligation to have due regard to CJEU jurisprudence is founded, is lacking.Footnote 151 If the EU does eventually allow the UK to accede to the 2007 Lugano Convention, it will be interesting to see whether English courts maintain their position on Gasser, once the boot is on the other foot.
