IF litigation arising from the same set of facts is pending in the courts of two EU member states, which court has to give way? In Starlight Shipping Co v Allianz Marine & Aviation Versicherungs AG; The Alexandros T [2013] UKSC 70, [2014] 1 All E.R. 590, the somewhat surprising view of the Supreme Court was that the answer will often be “neither”. The proceedings arose from the loss of the Alexandros T off the coast of South Africa. In 2006, her owners (Starlight) sued her insurers in England. Those proceedings settled pursuant to Tomlin orders, and the settlement agreements contained English exclusive jurisdiction clauses. That, one might have thought, would be the end of the matter, but in 2009 Starlight launched a swathe of proceedings in Greece against the insurers, reiterating the same allegations that had been raised and settled in England. In 2011, the insurers applied to the English courts to enforce the terms of the 2006 settlements, and brought new proceedings in England for damages, an indemnity and declarations concerning the breach of that settlement.

At least three sets of proceedings were therefore on foot: how were the English courts to resolve the risk of conflicting judgments? Since the courts were exercising jurisdiction under Council Regulation (EC) No 44/2001 (OJ 2001 L 12 p. 1) (“Brussels I Regulation”), questions of natural or appropriate forum were irrelevant; vexatious relitigation could never lead to the grant of an anti-suit injunction; and an exclusive jurisdiction clause was no bulwark against prior proceedings in another forum. If justice were to be achieved, it had to be done through the blunt lis pendens rules in Arts. 27 and 28 of the Regulation.

At first instance, no argument was offered about Art. 27. Instead, Burton J. considered that Art. 28 was applicable and that a stay should nonetheless be refused: [2011] EWHC 3381 (Comm); [2012] 1 Lloyd's Rep. 162. The Court of Appeal reversed that decision, holding that the English proceedings should be stayed pursuant to Art. 27, as they were “essentially the same” cause of action as the Greek proceedings, which were prior in time: [2012] EWCA Civ 1714; [2013] 1 Lloyd's Rep. 217. Even if the original 2006 proceedings were earlier than the Greek proceedings, the enforcement of the settlement agreement was a new, subsequent, cause of action.

In turn, the Supreme Court reversed the Court of Appeal. Lord Clarke (with whom Lord Sumption and Lord Hughes agreed) held that the English and Greek claims were not mirror images of each other, and were therefore not legally irreconcilable. The Court of Appeal had taken too broad a view of what counted as the “same cause of action” for the purposes of Art. 27. Since the English claim to enforce the settlement agreement (including its indemnity, release, and jurisdiction clauses) was contractual, and the Greek claim was delictual, the two proceedings lacked “le même objet et la même cause”, to use the language of the French text of the Regulation held by the CJEU to be generally applicable in Case C-144/86, Gubisch Maschinenfabrik AG v Palumbo [1987] E.C.R. 4861, and Case C-406/92, The Tatry [1994] E.C.R. I-5439.

Moreover, his Lordship held that the Court of Appeal incorrectly took into account the insurers’ reliance on the English settlement as a defence to the Greek action. After the decision in Case C-111/01, Gantner Electronic GmbH v Basch Exploitatie Maatschappij BV [2003] E.C.R. I-4207, matters raised by way of defence can, so we are told, have no bearing on whether Art. 27 is engaged. Thus, “[g]iven the fact that defences are irrelevant, the analysis cannot simply involve a broad comparison between what each party ultimately hopes to achieve”. Since the causes and objects were different, Art. 27 did not oblige the English court to stay its proceedings. So far as Art. 28 was concerned, while the claims were clearly “related” and the English court was second in time, the existence of an exclusive jurisdiction clause in favour of England was a “powerful factor” in favour of refusing a stay on discretionary grounds. To the extent that the English proceedings fell within Art. 28 and not Art. 27, they could therefore continue.

Lord Neuberger and Lord Mance agreed with the disposition of the appeals, but took a different view about the applicability of Art. 27 to the insurers’ claim for a declaration that the claims in the Greek proceedings had been released by the 2006 settlement agreement. Their Lordships considered that – contrary to the majority – Art. 27 was applicable here. In Lord Neuberger's view, if the Greek court were to uphold Starlight's claim, and the English court were to grant the insurers a declaration that that claim had in fact been settled, the two judgments would be logically incompatible. This, said his Lordship, was conceptually different from the insurers’ claim for an indemnity, which would only be commercially – but not logically – inconsistent with any Greek judgment in favour of Starlight. The former, but not the latter, aspect of the insurers’ claim therefore fell within Art. 27.

Lord Mance likewise considered that Art. 27 was engaged by the claimed declaration of non-liability. The claims before the Greek and English courts had the same “cause” because they were mirror images of each other: the one asserted, and the other denied, the existence of liability. Likewise, the proceedings had the “same object”, namely deciding the question of liability for the wrongs alleged in Greece. Given the autonomous definition of “cause of action” in the Regulation, the reason for the denial of liability (i.e. the existence of the settlement agreement) was irrelevant. The disjuncture drawn by the plurality between the Greek “tortious” and English “contractual” claims was therefore wrong as a matter of principle.

The result reached by the Supreme Court is, with respect, practical and just; yet practicality and justice are not hallmarks of the Brussels I Regulation nor of its interpretation by the CJEU. Three unsatisfactory aspects of the decision are worth noting. First, Lord Mance was surely correct that the majority's narrow approach to the meaning of “cause of action” is wrong. As the decision in the Tatry shows, Art. 27 cannot in fact be avoided simply by artful pleading. There, an action in rem and an action in personam were indeed held to be the same “cause”. The question certainly cannot turn on domestic definitions of “contract” or “tort”. In France, an action for professional negligence must be pleaded as a breach of contract, yet if it were relitigated in England as a tort claim, it cannot be treated as a different “cause of action” for the purpose of Art. 27. Moreover, while Lord Neuberger's distinction between “commercial” and “logical” inconsistency is attractive from the point of view of the English courts, it is not clear that this is so from the point of view of a third state in which the English and Greek judgments might simultaneously be enforced. The inevitable consequence of the Supreme Court's approach is a heightened risk of practically inconsistent judgments – the very thing the Brussels I Regulation purports to avoid.

Second, the fundamental difficulty in the Alexandros T is the lack of priority given to jurisdiction clauses in the Regulation's lis pendens rules. Lord Clarke's reference to the jurisdiction clause as a “powerful factor” weighing against a stay under Art. 28 is entirely legitimate; yet one cannot help but suspect that the existence of the clause was also a strong motivation for the majority's narrow construction of Art. 27. The revised rules on lis pendens in Art. 31 of the recast Brussels I Regulation (Regulation (EU) No 1215/2012 (OJ 2012 L 351 p. 1)) – which oblige any courts other than the chosen court to stay their proceedings – will solve the immediate problem, but they might also have encouraged a less instrumentalist approach to the assessment of what counts as the “same cause of action”.

Finally, the decision casts unnecessary doubt on an important question which will persist under the recast Regulation: namely, determining when a court is “seised” if the pleadings are amended to include a new claim. Lord Clarke considered that the “court first seised” meant the “court first seised of the proceedings, not of particular claims within the proceedings”, and that an amendment might therefore be related back to the time of the original claim. His Lordship considered that the matter was not acte clair, and that a reference to the CJEU would be necessary if the insurers persisted with their amended claims. With respect, the contrary view – that the court is only “seised” of such a claim at the time of the amendment and not earlier – is compelling as a matter of logic and practicality. As Lord Mance pointed out, the text of the Regulation, the existing case law and academic commentators “all point in one direction”. It is therefore especially unfortunate that a topic of such practical importance in cross-border litigation should now be overshadowed with doubt.