In the matter of “The Alexandros T”

This is a remarkable case in more than one respect. The appeal depends upon whether the Court is bound to stay action 2006 Folio 815 (“the 2006 proceedings”) under Article 27 of Regulation 44/2001 of the Council of the European Union (“the Regulation”) and, if not, whether it should do so under Article 28. Before Burton J (“the judge”), the respondents expressly disclaimed any intention to rely upon Article 27 but relied upon Article 28 in support of a submission that the court should stay the 2006 proceedings in favour of proceedings in Greece. The judge refused to grant a stay and gave summary judgment for the appellants against the respondents. The judge granted the respondents permission to appeal to the Court of Appeal on various grounds, including a ground based on Article 27. The Court of Appeal (Longmore, Toulson and Rimer LJJ) held that it was bound to stay the action under Article 27. It also gave some consideration to Article 28 but held that it was not necessary to reach a final conclusion in that regard because of its decision under Article 27. It declined to consider the issues relevant to summary judgment on the ground that, if there was to be a stay, those issues should be determined by the courts in Greece.

I can take the relevant events from the judgment of Longmore LJ in the Court of Appeal. He in turn took them from the judgment of the judge. On 3 May 2006 the vessel Alexandros T sank and became a total loss 300 miles south of Port Elizabeth, with considerable loss of life. Her owners were Starlight Shipping Company (“Starlight”). They made a claim against their insurers, who denied liability on the basis that the vessel was unseaworthy with the privity of the assured, namely Starlight. The insurers also said that Starlight had failed properly to report and repair damage to the vessel in accordance with Class Rules.

Starlight, through their solicitors Messrs Ince & Co, made a number of serious allegations against the insurers which fell into two categories, as summarised by Longmore LJ at para 4: (1) allegations of misconduct by the insurers and their underwriters involving alleged tampering with and bribing of witnesses, in particular the bosun, a Mr Miranda, to give false evidence, coupled with other allegations of spreading false and malicious rumours (described for some reason as “malicious scuttlebutt”) against Starlight in the course of purported investigation of their claims; and (2) deliberate failure by the insurers to pay up under the policy, said to have had consequential financial impact upon Starlight, and to have led to substantial recoverable loss and damage. The insurers also relied upon material non-disclosure.

Those allegations were made before the issue of proceedings and, in particular, in a letter dated 18 July 2006 from Ince & Co to the insurers’ solicitors, who were Hill Dickinson LLP, then Hill Taylor Dickinson, whom I will together call “HD”. On 15 August 2006 Starlight issued the 2006 proceedings in the Commercial Court against various insurers. The first four defendants have been described as the Company Market Insurers (“CMI”) and the fifth to seventh defendants as the Lloyd’s Market Insurers (“LMI”). The policies issued by both the CMI and the LMI contained exclusive jurisdiction clauses. They provided for English law and each party expressly agreed “to submit to the exclusive jurisdiction of the Courts of England and Wales”. Overseas Marine Enterprises Inc (“OME”) were identified in the policies as managers.

In paras 5 to 8 of his judgment Longmore LJ spelled out in some detail the issues between the parties in the 2006 proceedings. It is plain that the points raised by Ince & Co to which I have referred were both pleaded and central to the issues between the parties in those proceedings. Thus, in para 7 Longmore LJ referred to a witness statement in which Mr Crampton of Lax & Co, who were now acting for Starlight, asserted that the allegations made by the insurers in defence of the claim were based on false evidence which they had obtained from the bosun. He also relied upon significant payments said to have been made to the bosun on behalf of the insurers in this connection.

In addition, a witness statement was introduced in support of a proposed amendment of the claim form alleging that Starlight had sustained losses beyond the measure of indemnity in the relevant policy. It was alleged that, but for the failure of the insurers to pay under the policy, Starlight would have purchased a replacement vessel and had lost between US$ 45 million and US$ 47.7 million by way of increased capital cost and chartering losses. However, on 14 December 2007, Tomlinson J refused to allow the amendment on the basis of the decision of the Court of Appeal in Sprung v Royal Insurance [1999] Lloyd’s Rep IR 111, approving the decision in The Italia Express (No. 2) [1992] 2 Lloyd’s Rep 281. As Longmore LJ explained in para 1, as a matter of English law, an insurer commits no breach of contract or duty sounding in damages for failure promptly to pay an insurance claim. (Footnote: ¹) The law deems interest on sums due under a policy to be adequate compensation for late payment; this is so, even if an insurer deliberately withholds sums which he knows to be due under a policy. If parties agree that English law is to apply to a policy of insurance, this principle is part of what they have agreed. English law, moreover, gives no separate contractual remedy to an insured who complains that an insurer has misconducted himself before settling a claim. In either case the remedy of the insured is to sue the insurer and, if no settlement is forthcoming, proceed to judgment. The trial was fixed for 14 January 2008.

On 13 December 2007, which was the day before the hearing before Tomlinson J referred to above, the 2006 proceedings had been settled between Starlight and OME and the LMI for 100% of the claim, but without interest and costs, in full and final satisfaction of the claim. It was a term of the settlement agreement that Starlight would obtain a stay by way of a Tomlin Order, and a Tomlin Order by consent between Starlight and the LMI was accordingly made on 20 December 2007, backdated to 14 December, in these terms:

“Save for the purposes of carrying into effect the terms agreed between the Claimant and the Fifth to Seventh Defendants, all further proceedings between the Claimant and the Fifth to Seventh Defendants shall be stayed with effect from 14 December 2007 or such earlier date as may be agreed between the parties or otherwise ordered hereafter.”

A similar settlement agreement dated 3 January 2008 was made between Starlight and OME and the CMI and a similar Tomlin Order was made on 7 January 2008, but with immediate effect.

In each settlement agreement “the Assured” were defined as being “[OME] and Starlight … as Managers and/or Owners and/or Associated and/or Affiliated Companies for their respective right and interest in the ship Alexandros T”. The CMI settlement agreement then provided:-

“1.

Each Underwriter agrees to pay on or before 18 January 2008 … their due proportions of the sum of US$16m … being 100% of their due proportions of the sum insured being 50% of the US$32m … without interest or costs.

2.

The Assured and Claimant agree to accept the EURO equivalent of each Underwriter’s due proportion of US$16m … in full and final settlement of all and any claims it may have under Policy No 302/CF 000220Z against the Underwriters in relation to the loss of “Alexandros T”, including all claims for interest and costs (including in respect of all costs orders made to date in the proceedings) but without effect to any other insurance policy in which each Underwriter may be involved.

3.

The Assured and Claimant agree to Indemnify each Underwriter against any claim that might be brought against it by any of the Assured’s or the Claimant's associated companies or organisations or by any mortgagee in relation to the loss of “Alexandros T” or under Policy No 302/CF000220Z….

4.

Following the signing of this agreement, and in consideration of the promises herein, the Claimant and the Underwriters will apply to stay the Proceedings as against the Underwriters, the Proceedings to be stayed for all purposes save for the purposes [of] carrying the terms agreed herein into effect, such stay to have effect from the first obtainable date after 27 December 2007

5.

Following the due and proper payment by the Underwriters of the amount specified in paragraph 1 above, the Assured and Claimant and the Underwriters agree to file a consent order dismissing the Proceedings, with no order as to costs.

6.

This agreement is subject to English law and to the exclusive jurisdiction of the High Court in London.”

The LMI settlement agreement provided in similar but not identical terms:-

…

“2.

The underwriters … agree to pay on or before 24 December 2007 … the sum of US$8M … being 100% of their due proportions of the sum insured being 25% of US$32m … without interest or costs …

3.

The Assured and claimant agree to accept the EURO equivalent of US$8M … in full and final settlement of all and any claims it may have under Policy No … against the Underwriters signing below in relation to the loss of “Alexandros T” …

4.

The Assured and Claimant agree to indemnify the underwriters signing below against any claim that might be brought against them by any of the Assured’s or the Claimant's associated companies or organisations or by any mortgagee in relation to the loss of “Alexandros T” or under Policy No …

5.

This agreement is subject to English law and the jurisdiction of the High Court of London.”

After setting out the terms of the settlement agreements, Longmore LJ wryly observed at the end of para 12 of his judgment that one might have expected that to be that, but it was not to be. He described what then happened in paras 13 to 15. More than three years later, in April 2011, nine sets of Greek proceedings, in materially identical form, (“Greece 1”), were issued by Starlight, by OME, by their co-assureds under an associated Fleet Policy and by individual officers of those companies, against the LMI and the CMI, some of their employees or underwriters, and HD and some of their partners and employees (“the HD defendants”). The claims are for compensation for loss of hire and loss of opportunity by Starlight totalling approximately US$ 150 million and for pecuniary compensation due to moral damage amounting to €1 million. The claims also include similarly substantial claims by the other claimants in respect of alleged acts, all done unlawfully and in breach of good faith for the alleged purpose of avoiding the performance by the defendants of their legal obligations.

All the claims rely upon breaches of the Greek Civil and Criminal Code. However the factual allegations, which Longmore LJ noted had been said by the judge to be “entirely familiar”, include the allegation that the appellants were responsible for using false affidavits of witnesses (primarily Mr Miranda) with intention to harm the claimants, described thus by Mr Crampton in a statement summarising the Greek claims:

“The underwriters pursued this criminal effect by intentionally fabricating false evidence with the purpose that the underwriters (who were responsible for the payment of insurance indemnity for the vessel) … avoid paying this insurance indemnity, contrary to their contractual obligations and their legal obligations and in particular contrary to the provisions of the insurance contract, providing for the timely payment of the insurance indemnity.”

They also include the claim that the appellants were asserting and disseminating false information to third parties, although they were aware of their falsity, damaging the claimants’ reputation and credibility

“with the purpose that the underwriters (who were responsible for the payments of the insurance indemnity for the vessel) …avoid paying the insurance indemnity, contrary to their contractual obligation and their legal obligation and in particular contrary to provisions of the insurance contract providing for the timely payment of the insurance indemnity”

Mr Crampton then turned to what he called the “[i]ntentional fabrication of false evidence for defrauding the English court” and “[t]he moral instigation alternatively complicity of the underwriters to perjury and on the defrauding of the court by the underwriters”. He summarised the position in this way in paragraph 20 of his witness statement:-

“The essence of the complaint against the Defendants in the Greek proceedings concerns the allegation that the Defendants obtained false evidence in Greece from the bosun of the Alexandros T, Aljess Miranda … This evidence was then deployed in these proceedings in England and also in the … Greek proceedings.”

There is a substantial section of the Greek pleadings relating to the financial consequences of the failure by the insurers to comply with their obligations under the policy and the way in which they allegedly handled the investigations. In a further set of proceedings, known as “Greece 2”, two additional heads of loss are claimed by Starlight and OME, arising out of substantially the same allegations.

As Longmore LJ put it in para 15, in apparent recognition of the problem raised by the fact that such claims had either not been brought in England or had been ruled out as a matter of English law by Tomlinson J, Mr Crampton, in paragraph 27 of his witness statement, explained that the claims are advanced in two ways in the Greek pleadings: first, that as a result of the underwriters’ intention to avoid payment of the insurance indemnity, eventually resulting in late payment of the policy proceeds, the claimants missed the opportunity to use the policy proceeds to invest in three vessels (not just the one referred to in the 2006 proceedings); and, secondly, that, as a result of the defendants’ actions in acquiring the false evidence of Mr Miranda, his clients were not able to insure the vessels and without insurance they would not have been able to trade them and could not purchase them. He stated that his clients would “amend their pleadings prior to the hearing of the disputes in Greece so as to clarify this head of claim, such that no claim is made in respect of the late payment of the policy proceeds”. The expert evidence from the defendants is that it is not possible to amend the pleadings in the Greek courts, but, treating the proposed draft amendment, which he exhibited, as a clarification, it did not seem to the judge that it in any way cured the defect, if defect there was. It is further said that the insurance of the three potential new vessels was rendered impossible, since all the London insurers refused to quote for the vessels because of the refusal of the defendant underwriters to quote for them and because of the defamatory accusations spread as to the unseaworthiness of the Alexandros T. All these allegations arise out of the alleged manner in which the defendants handled Starlight's claim in respect of the Alexandros T, and, even though the consequences and the consequential losses have expanded, and the claim for moral damages has been included, and although it seems that Starlight now rely on an expanded affidavit of Mr Miranda, the allegations, even though put into the context of Greek law, were said by the judge to be materially identical to those made prior to the settlement agreement. The acts complained of are all said to have constituted delicts under Greek law akin to the torts of defamation and malicious falsehood under English law.

Since the issue of the Greek proceedings, as Longmore LJ explained in para 16 (and the judge at his para 14), the insurers have taken further steps and brought further proceedings in England as follows. By applications issued in the 2006 proceedings on 25 July and 3 August 2011, the CMI and the LMI respectively sought, pursuant to the Tomlin Orders (if necessary after lifting the stay imposed by them) summary relief pursuant to CPR Part 24 by way of declarations and damages against Starlight. The LMI, because permission was given to them to join OME as a third party, also sought summary relief pursuant to Part 24 against OME (which filed an acknowledgment of service and a defence) to enforce the LMI settlement agreement, to which it also was a party.

In addition, fresh proceedings (2011 Folio 702) were commenced by the LMI, without prejudice to their case that sufficient relief could and would be obtained in the 2006 proceedings, against both Starlight and OME, and, after an acknowledgment of service and defence were filed, an application was made under Part 24 for similar relief to the claim in the 2006 proceedings. The LMI also brought fresh proceedings (2011 Folio 1043) against Starlight's co-assured and, again after acknowledgments of service and defence had been filed, sought declaratory relief and damages for breach of the exclusive jurisdiction clause in their insurance policies, by virtue of the issue of the Greek proceedings by those co-assured. Also in fresh proceedings (2011 Folio 894), the CMI brought claims against OME and the same co-assured in respect of similar claims for breach of the exclusive jurisdiction clause in the policy, and in respect of OME by reference to breach of the terms of the settlement agreement. Judgment in default was entered by the CMI against all those defendants on 26 October (amended on 14 November) 2011. Those proceedings are not the subject of this appeal and no issue therefore currently arises with respect to them.

Finally, and by separate application, the HD defendants were joined as defendants in the 2006 proceedings so that, in due course, they too might be able to claim relief by seeking declaratory relief within the original proceedings. Starlight and their associates applied to stay both the 2006 proceedings in their current form and 2011 Folios 702 and 1043.

In summary, the claims made in the various proceedings are these.

The insurers sought to enforce the settlement agreements referred to in the Tomlin Orders and, in a judgment handed down on 19 December 2011, having refused a stay under Article 28, the judge held that they were entitled to summary judgment for (inter alia) a declaration that the matters sought to be raised in Greece were part of the settlement of the claim and that Starlight (and OME) are bound to indemnify the insurers against any costs incurred and any sums that may be adjudged against them in the Greek proceedings.

As stated above, the Court of Appeal held that it was bound to stay the 2006 proceedings and 2011 Folio 702 and 1043 under Article 27, made no final determination of the position under Article 28 and declined to consider the issues of summary judgment. The Court of Appeal also held that it was not too late for the respondents to rely upon Article 27 or Article 28.

In this Court the appellants challenge the correctness of the Court of Appeal’s conclusion under Article 27 and, on the respondents’ cross-appeal, submit that the judge was correct to refuse a stay under Article 28. If the appellants succeed under both articles, the case will have to be remitted to the Court of Appeal to consider the respondents’ appeal from the summary judgment granted by the judge.

The questions for decision under Article 27 are whether, in the events which happened, the Court of Appeal was wrong to hold that it was not too late for the respondents to rely upon Article 27, whether the proceedings in Greece and the proceedings in England involve “the same cause of action”, whether they are “between the same parties” and which court was the court first seised. For reasons which will appear, I will defer consideration of the ‘too late’ point until after consideration of the other issues.

Article 27 must be construed in its context. The immediate context of Articles 27 and 28 is that they form part of Section 9 of Chapter II of the Regulation, which must be read in the light of Recitals 2 and 15 of the preamble. It is apparent from Recital 2 that the Regulation aims, in the interests of the proper functioning of the internal market, to put in place:

“Provisions to unify the rules of conflict of jurisdiction in civil and commercial matters and to simplify the formalities with a view to rapid and simple recognition and enforcement of judgments from Member States bound by this Regulation...”

Recital 15 provides:

“In the interests of the harmonious administration of justice it is necessary to minimise the possibility of concurrent proceedings and to ensure that irreconcilable judgments will not be given in two Member States. There must be a clear and effective mechanism for resolving cases of lis pendens and related actions and for obviating problems flowing from national differences as to the determination of the time when a case is regarded as pending. For the purposes of this Regulation that time should be defined autonomously.”

The mechanism referred to in Recital 15 is provided by Section 9 of Chapter II of the Regulation, which includes Articles 27 and 28:

“Section 9

Lis pendens – related actions

Article 27

1.

Where proceedings involving the same cause of action and between the same parties are brought in the courts of different member states, any court other than the court first seised shall of its own motion stay its proceedings until such time as the jurisdiction of the court first seised is established.

2.

Where the jurisdiction of the court first seised is established, any court other than the court first seised shall decline jurisdiction in favour of that court.

Article 28

1.

Where related actions are pending in the courts of different member states, any court other than the court first seised may stay its proceedings.

2.

Where these actions are pending at first instance, any court other than the court first seised may also, on the application of one of the parties, decline jurisdiction if the court first seised has jurisdiction over the actions in question and its law permits the consolidation thereof.

3.

For the purposes of this Article, actions are deemed to be related where they are so closely connected that it is expedient to hear and determine them together to avoid the risk of irreconcilable judgments resulting from separate proceedings.

Article 29

Where actions come within the exclusive jurisdiction of several courts, any court other than the court first seisedshall decline jurisdiction in favour of that court.

Article 30

For the purposes of this Section, a court shall be deemed to be seised:

1.

at the time when the document instituting the proceedings or an equivalent document is lodged with the court, provided that the plaintiff has not subsequently failed to take the steps he was required to take to have service effected on the defendant, or

2.

if the document has to be served before being lodged with the court, at the time when it is received by the authority responsible for service, provided that the plaintiff has not subsequently failed to take the steps he was required to take to have the document lodged with the court."

The Regulation is the successor to the Brussels Convention on Jurisdiction and the Enforcement of Judgments in Civil and Commercial Matters 1968 (“the Brussels Convention”), in which the equivalent provisions to Articles 27 and 28 were Articles 21 and 22 respectively. The Court of Justice of the European Union (“the CJEU”) has held that the principles developed in its case law with regard to Articles 21 and 22 of the Brussels Convention apply equally to Articles 27 and 28 of the Regulation: see Folien Fischer AG v Ritrama SpA (Case C-133/11) [2013] QB 523 at paras 31 and 32. The CJEU was of course previously the European Court of Justice (“ECJ”). Although some of the decisions to which I refer were made by the ECJ, for simplicity I will refer to all the European decisions as those of the CJEU.

The CJEU has laid down a number of general principles which are of some importance. They include the important principle that a court in a Member State must not grant an anti-suit injunction to restrain the bringing or continuing of proceedings in another Member State, whether to restrain an abuse of process or to restrain proceedings brought or continued in breach of an exclusive jurisdiction clause: see eg Turner v Grovit (Case C-159/02) [2005] 1 AC 101 and West Tankers Inc v Allianz SpA (The Front Comor) (Case C-185/07) [2009] 1 AC 1138.

They also include the following, with specific reference to Articles 27 and 28. First, the purpose of Article 27 is to prevent the courts of two Member States from giving inconsistent judgments and to preclude, so far as possible, the non-recognition of a judgment on the ground that it is irreconcilable with a judgment given by the court of another Member State: Gubisch Maschinenfabrik KG v Palumbo (Case C-144/86) [1987] ECR 4861 at para 8. Second, the objective of Article 28 is to improve co-ordination of the exercise of judicial functions within the European Union and to avoid conflicting and contradictory decisions, thus facilitating the proper administration of justice: see eg The Tatry (Case C-406/92) [1999] QB 515 at paras 32, 52 and 55 and Sarrio SA v Kuwait Investment Authority [1999] 1 AC 32, per Lord Saville at 39F-H.

It is convenient to consider first the position of the CMI claims. The first specific question is whether the 2006 proceedings involve the same cause or causes of action as the Greek proceedings, by which I mean Greece 1 and Greece 2. The principles of EU law which are relevant to the determination of this question are in my opinion clear. They have been considered in a number of cases in the CJEU and are essentially as submitted on behalf of the CMI. They may be summarised in this way.

See also to similar effect Kolden Holdings Ltd v Rodette Commerce Ltd[2008] 1 Lloyd's Rep 434, per Lawrence Collins LJ at para 93 and Research in Motion UK Ltd v Visto Corporation[2008] 2 All ER (Comm) 560, per Mummery LJ at para 36.

It follows that Article 27 is not engaged merely by virtue of the fact that common issues might arise in both sets of proceedings. I would accept the submission on behalf of the CMI that this is an important point of distinction between Articles 27 and 28. Under Article 28 it is actions rather than claims that are compared in order to determine whether they are related.

After discussing Gubisch, The Tatry, Sarrio, The Happy Fellow [1998] 1 Lloyd’s Rep 13 and Haji-Ioannou v Frangos [1999] 2 Lloyd’s Rep 337, Rix J summarised the position clearly and, in my opinion, accurately in Glencore International AG v Shell International Trading and Shipping Co Ltd [1999] 2 Lloyd’s Rep 692 at 697:

“It would appear from these five cases, of which the first two were in the European Court of Justice, and the latter three in the domestic Courts of England, that, broadly speaking, the triple requirement of same parties, same cause and same objet entails that it is only in relatively straightforward situations that art 21 bites, and, it may be said, is intended to bite. After all, art 22 is available, with its more flexible discretionary power to stay, in the case of ‘related proceedings’ which need not involve the triple requirement of art 21. There is no need, therefore, as it seems to me, to strain to fit a case into art 21. The European Court, when speaking in Gubisch (at para 8) of the purpose, in the interests of the proper administration of justice within the European Community, of preventing parallel proceedings in different jurisdictions and of avoiding ‘in so far as it is possible and from the outset’ the possibility of irreconcilable decisions, was addressing arts 21 and 22 together, rather than art 21 by itself.

Thus a prime example of a case within art 21 is of course where party A brings the same claim against party B in two jurisdictions. Such a case raises no problem. More commonly, perhaps, the same dispute is raised in two jurisdictions when party A sues party B to assert liability in one jurisdiction, and party B sues party A in another jurisdiction to deny liability, or vice versa. In such situations, the respective claims of parties A and B naturally differ, but the issue between them is essentially the same. The two claims are essentially mirror images of one another. Gubisch and The [Tatry] are good examples of this occurrence.

On the other hand, Sarrio v KIA is a case where the same claimant was suing the same defendant on different bases giving rise to different issues and different financial consequences, and where liability on one claim did not involve liability (or non-liability) on the other. Haji-Ioannou v Frangos illustrates the situation where even though the cause is the same, and even though there is some overlap in the claims and issues, nevertheless different claims, there the proprietary claim to trace, may raise sufficiently different issues of sufficient importance in the overall litigation for it to be concluded that the objet differs. The authority of The Happy Fellow at first instance may be somewhat shaken by the reservations expressed by Lord Justice Saville on appeal, but it too may be said to illustrate the process of analysing the claims and issues in the respective proceedings to identify whether they are the same. Where, for instance, there is no dispute over a shipowner’s right to limit should he be found liable (a separate question, which need not even be resolved at the time when a limitation action is commenced or a decree given), I do not for myself see why it should be held that the liability action and the limitation action involve the same cause of action for the purposes of art 21.”

How do these principles provide an answer to the question whether the 2006 proceedings involve the same cause or causes of action as the Greek proceedings? It is necessary to consider the claims advanced by the CMI and the LMI separately and, in the case of each cause of action relied upon, to consider whether the same cause of action is being relied upon in the Greek proceedings. In doing so, the defences advanced in each action must be disregarded.

The essential question is whether the claims in England and Greece are mirror images of one another, and thus legally irreconcilable, as in Gubish and The Tatry, in which case Article 27 applies, or whether they are not incompatible, as in Gantner, in which case it does not. Thus in Gantner a claim for damages for repudiation of a contract and a claim for the price of goods delivered before the repudiation could both have succeeded and the fact that a set-off of the damages would make the price less beneficial to the seller did not make them incompatible. And in Maersk Olie & Gas A/S v Firma M de Haan en W De Boer(Case C-39/02) [2004] ECR I-9657 owners of a vessel which damaged a pipeline (owned by Maersk) sought a declaration that they were entitled to limit their liability under the 1957 International Convention relating to the Limitation of Liability of Owners of Sea-going Ships and the Dutch legislation that gave effect to it and that a limitation fund be established. Maersk subsequently commenced proceedings in Denmark claiming compensation for damage to the pipeline. The CJEU held that the causes of action were not the same: see paras 35 to 39.

The CJEU underlined both the principle in Gantner that account should be taken only of the claims and not of the defences advanced and the principle in The Tatry that the cause of action comprised both the facts and the legal rule invoked as the basis of the application. It held on the facts, at para 38, that:

“the unavoidable conclusion is that, even if it be assumed that the facts underlying the two sets of proceedings are identical, the legal rule which forms the basis of each of those applications is different. … The action for damages is based on the law governing non-contractual liability, whereas the application for the establishment of a liability limitation fund is based on the 1957 Convention and the Netherlands legislation which gives effect to it.”

The CJEU thus distinguished Gantner and The Tatry on the basis that in those cases, by contrast, the claim brought in the second set of proceedings mirrored that brought in the first set.

What then is the position on the facts? The CMI advance the claims referred to in para 18 above under three heads, each of which relies upon provisions either of the CMI settlement agreement or the policies. It is convenient to consider the claims under the three heads in this order: indemnity, exclusive jurisdiction and release.

These are based on clause 3 of the settlement agreement set out above. The claims are simple. By clause 3 the “Assured” as defined agreed to indemnify the CMI against “any claim that might be brought against them by any of the Assured’s or the Claimant’s associated companies or organisations or by any mortgagee in relation to the loss of Alexandros T” or under the relevant policy. The CMI say that the Greek proceedings are in respect of such claims and that they are entitled to be indemnified against the consequences of those proceedings. They say that that claim under clause 3 does not give rise to the same claim or cause of action as any claim or cause of action in the Greek proceedings. They say that, on the contrary, it assumes that the Greek proceedings will proceed and that the claimants in Greece may succeed.

I would accept that submission. In my opinion, none of the causes of action relied upon in the Greek proceedings has identity of cause or identity of object with the CMI’s claim for an indemnity. As to cause, the subject matter of the two claims is different. The former are claims in tort (or its Greek equivalent) and the claim for an indemnity is a claim in contract. As to object, that of the Greek proceedings is to establish a liability under Greek law akin to tort, whereas, as for example in the case of a claim on an insurance policy, the object of the CMI’s claim is to establish a right to be indemnified in respect of such a liability. Further, whereas Starlight and its co-assureds and the individual officer claimants in the Greek proceedings are seeking each to recover its or his own loss, the indemnity clause will, if the indemnity claim is otherwise good, entitle the CMI to recover from Starlight not just any sum awarded in Greece to Starlight, but also any sums awarded to any of Starlight’s co-claimants. So the object of the English indemnity claim against Starlight differs from, and is in fact much wider than, the object of Starlight’s claim in the Greek proceedings.

Moreover, the claim for an indemnity in the 2006 proceedings in England does not interfere in any way with the Greek proceedings or vice versa. There is no attempt in Greece to impugn the settlement agreements or the indemnity agreements contained in them. The respondents do not assert, for example, that the indemnities do not apply to some or all of the Greek claims. I would determine this point in favour of the CMI on this simple basis. The CMI’s cause of action for an indemnity under clause 3 of the settlement agreement is not the same cause of action as any of the causes of action relied upon in Greece, which are tortious. The respective causes of action have neither the same object (le même objet) nor the same cause (la même cause).

The same is in my opinion true of the CMI’s claims that the respondents have brought the proceedings in Greece in breach of the exclusive jurisdiction clauses in the settlement agreement and/or in the insurance policies. Clause 6 of the settlement agreement expressly provides that it is subject to English law and the exclusive jurisdiction of the High Court in London. The CMI say that, in bringing the Greek proceedings, the respondents are in breach of clause 6 and that they are entitled to damages as a result. They do not seek an anti-suit injunction to restrain the Greek proceedings. They simply seek a declaration that the claims brought by Starlight and OME in Greece 1 and Greece 2 fall within the scope of the settlement agreement.

Moreover the respondents do not assert in the Greek proceedings that the settlement agreements do not preclude the bringing of their claims in Greece. It may be that the reason they do not advance that argument is that they would be met with the response that a dispute as to the meaning and effect of the settlement agreements is subject to the English jurisdiction clause so that the court in Greece would have to decline jurisdiction. However that may be, they do not in fact advance the argument. It follows that in this respect too the Greek proceedings are not the mirror image of the English proceedings or vice versa and that the cause or causes of action based on an alleged breach of clause 6 of the CMI settlement agreements are not the same cause or causes of action as are advanced by the respondents in Greece. They do not have le même objet et la même cause.

As I see it, the position is the same in the case of the alleged breach of the exclusive jurisdiction clauses in the insurance policies. There is an established line of cases in England to the effect that claims based on an alleged breach of an exclusive jurisdiction clause or an arbitration clause are different causes of action from claims for substantive relief based on a breach of the underlying contract for the purposes of Article 21 of the Brussels Convention and Article 27 of the Regulation: see eg Continental Bank NA v Aeakos Compania Naviera SA [1994] 1 WLR 588 per Steyn LJ (giving the judgment of the Court of Appeal) at 595H-596C; Alfred C Toepfer International GmbH v Molino Boschi Sarl [1996] 1 Lloyd’s Rep 510 per Mance J at 513; Toepfer International GmbH v Société Cargill France [1997] 2 Lloyd’s Rep 98, per Colman J at 106; Sinco per Beatson J at paras 50 and 54; and WMS Gaming Inc v Benedetti Plus Giocolegale Ltd [2011] EWHC 2620 (Comm) per Simon J at para 32.

Those cases support the conclusion that the claims of the CMI in the 2006 proceedings for breach of the exclusive jurisdiction clauses in the insurance policies (or indeed in the settlement agreement) do not involve the same cause or causes of action within the meaning of Article 27 as the respondents’ claims in (or akin to) tort in the Greek proceedings. I understand that this point has been reserved for decision by the Court of Appeal but, as I see it at present, nothing in the relief sought by the CMI offends the principle of mutual trust and confidence which underlies the Regulation: see eg Erich Gasser GmbH v MISAT Srl(Case C-116/02) [2005] QB 1. The CMI do not seek to stop the Greek proceedings or to restrain Starlight and OME from pursuing them. They merely seek declarations as to the true position under the settlement agreements which are both governed by English law and subject to the exclusive jurisdiction of the English courts and under the clauses in the insurance contracts which also provide for the exclusive jurisdiction of the English courts. This has the advantage that the courts with exclusive jurisdiction decide what is the true meaning of the settlement agreements and the jurisdiction clauses.

The same is also, in my opinion, true of the claims based on what are called the release provisions in the CMI settlement agreement. It is said that the provision that the sums agreed to be paid under the CMI settlement agreement are to be paid “in full and final settlement of all and any claims it may have under the policy” precludes the payment of any further sums arising out of the loss of the vessel insured. It is said that, in the light of the agreement, the CMI are entitled to a declaration that the Greek claims fall within the terms of the agreement, that they are entitled to a declaration that the bringing of those claims is a breach of the agreement and that they are entitled to damages for that breach.

The question is whether these claims involve le même objet et la même cause as the claims in the Greek proceedings. In my opinion they do not for the same reasons as in the case of the claims for an indemnity and the claims arising out of the exclusive jurisdiction clauses. The Greek claims are claims in tort and these are contractual claims. The factual bases for the two claims are entirely different. Moreover the object of the two claims is different.

This is to my mind clear in the case of the claims for damages for breach of the release provisions in the settlement agreements and for a declaration that the bringing of the Greek claims is a breach of the settlement agreement. The nature of the claims is almost identical to the nature of the claims for breach of the jurisdiction agreements. In both cases the alleged breach is the bringing of the claims in Greece. Moreover, like the claims for an indemnity, the claim for damages for breach of the settlement agreement assumes that the claims in Greece may succeed. Is the position different in respect of the claim for a declaration that the Greek claims fall within the terms of the release in the settlement agreements? In my opinion the answer is no.

All these claims have the same thing in common. It is that the legal basis for the claims in Greece is different from the legal basis of the claims in England. In Greece the legal basis for the claims is tortious, whereas in England the legal basis of the claims is contractual. It is thus not a case like Gubisch, where, as the CJEU put it at para 15, the same parties were engaged in two legal proceedings in different Contracting States which were based on the same ‘cause of action’, that is to say the same contractual relationship. The ‘cause’ was therefore the same. Equally the ‘objet’ of the actions was the same, namely to determine the effect if any of the contract. As the CJEU put it at para 16, the action to enforce the contract was aimed at giving effect to it, while the action for its rescission or discharge was aimed “precisely at depriving it of any effect”. The question whether the contract was binding lay “at the heart of the two actions”. That is not true here because the object of the English action is to enforce the contract, whereas the object of the Greek proceedings is to establish a different liability in tort.

Lord Mance takes a different view in one respect. So far as the claims for damages for breach of the releases in the settlement agreements, the claims for a declaration and damages for breach of the jurisdictionclauses and the claims for indemnities are concerned, there is no difference between us. However, so far as the claims for adeclaration that the Greek claims fall within the terms of the release in the settlement agreements is concerned, Lord Mance takes a different view. He notes in para 140 the terms in which the claims are pleaded. The formulation in paragraph 18(a) above, which was adopted by the respondents, is in fact derived from the declaration made by the judge. However, to my mind nothing turns on this difference. Moreover, I do not see that it makes any difference that the respondents discharged their obligations under the settlement agreements.

The critical point is that on the facts here the legal basis of the claims in tort in Greece is different from the legal basis of the contractual claims in England. It is true that, if successful, a declaration that the tortious claims have been settled or released will or may afford the appellants a defence to the Greek proceedings but the cases show that defences are irrelevant. Viewed through the perspective of the claims, the two claims are not the mirror image of one another. Even if (contrary to my view) the two sets of proceedings had in this respect le même objetthey did not have la même cause, whereas the cases show that, in order to involve the same cause of action, they must have both le même objet et la même cause.

The position would be different if the CMI were to advance a claim in the English proceedings claiming a declaration that they are not liable to the respondents in Greece. That claim would be the mirror image of the claims being brought by the respondents in Greece and would fall within the principles laid down in Gantner and The Tatry. In fact, after the judge had delivered his judgment, the CMI did, as I understand it, make an application for such a negative declaration in the light of the fact that Starlight and OME had commenced Greece 2. We were told that in the event the application was never determined and that the CMI do not pursue it. It has been confirmed that any such claim has now been abandoned.

For these reasons, subject to a possible reference to the CJEU discussed in paras 58-59 below, I would hold that Article 27 does not apply to any of the causes of action advanced by or against the CMI.

I appreciate that, in reaching these conclusions I have reached a different view from that of the Court of Appeal. Before I express my reasons, I should say that I suspect that the focus of the argument in the Court of Appeal was somewhat different from that in this Court. The reasons are I think twofold. First, in para 40 of his judgment Longmore LJ distinguished Sinco on the basis that the difference between this case and that is that in that case, in contradistinction to this, there was no settlement agreement which could, as he put it, supposedly deny the Greek claimants the right to bring proceedings at all. I do not see that as correct. As explained above, the CMI do not seek to deny the respondents the right to commence proceedings in Greece but merely say that the causes of action in the two sets of proceedings are different.

The second point is perhaps more significant. In para 46 Longmore LJ correctly notes that the CMI’s case is that the bringing of the Greek proceedings is a breach of the jurisdiction clauses in the policies and a breach of the terms of the settlement agreement and, again correctly, states that the primary relief claimed by the CMI in England is a declaration that Starlight will be liable to indemnify the CMI against any costs incurred in the Greek proceedings and any liability in those proceedings. I have already given my reasons for concluding that those are different causes of action from the causes of action in tort relied upon by the respondents in Greece. They are not a mirror image of one another.

As I see it, the Court of Appeal treated the question as a broad one focusing on the overall result in each jurisdiction. This can be seen from paras 47 to 50 of Longmore LJ’s judgment. In paras 46 and 47 he summarised the claims of both the CMI and the LMI. He then said this at paras 48 and 49:

“48.

It is clear that the first 3 paragraphs of the LMI application are in terms an assertion that LMI are not liable in respect of the claims in Greece. CMI's allegation that the Greek claimants are in breach of the settlement agreements is in effect a similar assertion. It may be said that there are other causes of action in the English proceedings which are not exactly mirror images of the allegations in the Greek proceedings but to the extent that they are not, they are essentially the same in the sense that the key assertion in Greece is that there are non-contractual claims and the key assertion in England is that those non-contractual claims have been compromised by the settlement agreements. The claims for damages and indemnity are in any event parasitic on the central contention that, once a settlement had been reached, all matters in issue had been compromised. It is, of course, elementary that Article 27 has regard to causes of action rather than proceedings and that is why it is necessary to concentrate on the allegations relating to the settlement agreement. It is certainly the case that there is a considerable risk of inconsistent judgments if one of the sets of proceedings is not stayed and the rationale behind Article 27 therefore favours a stay if the Greek court was the court first seised.

49.

I therefore conclude that, in so far as the English proceedings assert non-liability by reason of the settlement agreements, there is an identity of issues and the respective causes of action are the same. To the extent that allegations are made in England that the Greek parties are in breach of the settlement agreements or in breach of the exclusive jurisdiction clauses in either the insurance policy or the settlement agreements themselves (and that they should therefore indemnify the insurers for the cost of the Greek proceedings) they are parasitic and dependent on the basic cause of action in England for a declaration of non-liability. They cannot proceed in their own right until the underlying question of the ambit of the settlement agreements as a defence to the Greek actions in tort has been resolved.”

In my opinion that analysis is not consistent with the principles laid down by the CJEU set out above. As already stated, those principles require a comparison of the claims made in each jurisdiction and, in particular, consideration of whether the different claims have le même objet et la même cause without regard to the defences being advanced. As I see it, Article 27 involves a comparison between the causes of action in the different sets of proceedings, not (as in Article 28) the proceedings themselves. In para 48 Longmore LJ recognises that there are causes of action in the English proceedings which are not (as he puts it) exactly mirror images of the allegations in the Greek proceedings but says that, to the extent that they are not, “they are essentially the same in the sense that the key assertion in Greece is that there are non-contractual claims and the key assertion in England is that those non-contractual claims have been compromised by the settlement agreements”. And at the end of para 49 he says that the claims in England cannot proceed in their own right “until the underlying question of the ambit of the settlement agreements as a defence to the Greek action in tort has been resolved”.

I respectfully disagree with that approach. It focuses on the nature of the settlement agreements as a defence to the Greek action in tort, which the authorities in the CJEU show is irrelevant. Given the fact that defences are irrelevant, the analysis cannot involve a broad comparison between what each party ultimately hopes to achieve. The analysis simply involves a comparison between the claims in order to see whether they have the same cause and the same object. In so far as Andrew Smith J treated the question as a broader one in Evialis SA v SIAT [2003] 2 Lloyd’s Rep 377 I respectfully disagree with him, although, as Beatson J observed in Sinco at para 50, Evialis was distinguishable on the facts because the insured had brought a substantive claim in the English proceedings in addition to their claim in the Italian proceedings, which rendered the former a mirror image of the latter. This case can be distinguished on the same basis, at least in the case of the CMI’s claims.

I also note in connection with Sinco that at para 40 Longmore LJ observed that the difference between that case and this was that in that case there was no settlement agreement which could supposedly deny the right of the Greek claimants to bring proceedings at all. For the reasons I have given I would respectfully disagree with that approach. A settlement agreement might be a defence to a claim. It could not deny the right of the Greek claimants to bring proceedings at all.

For these reasons I would hold that Article 27 has no application to the case of the CMI. Moreover, subject to one point discussed at paras 58-59 below, I would not order a reference to the CJEU on this question because the relevant principles are clearly set out in its jurisprudence and are acte clair. In these circumstances, where none of the causes of action in the English proceedings is the same as the causes of action in the Greek proceedings, it is not necessary in the case of the CMI to consider the other issues which might arise, namely the position in relation to other parties and which court was the court first seised for the purposes of Article 27.

Save possibly for two points, the position of the LMI is essentially the same as in the case of the CMI. The first point is that the jurisdiction clause in clause 5 of the LMI settlement agreement differs from that in clause 6 of the CMI settlement agreement in that it does not expressly provide for the exclusive jurisdiction of the High Court in London but merely for the jurisdiction of the High Court in London. However, subject to its detailed provisions, Article 23 of the Regulation provides that, where parties have agreed that a court or the courts of a Member State shall have jurisdiction, that court or those courts shall have jurisdiction and, moreover, that such jurisdiction shall be exclusive unless the parties have agreed otherwise. The question whether the parties had agreed otherwise was discussed by the judge at paras 19 to 23 of his judgment, where he held that the parties had not agreed otherwise and that clause 5 of the LMI was an exclusive jurisdiction clause. No appeal was brought against that part of the judge’s ruling.

The second point is this. I had understood during the argument that the LMI were seeking a negative declaration of the kind which the CMI were not. It now appears that I was mistaken. I understand that the LMI had indicated an intention of doing so if the CMI proceeded with an application for permission to do so but, since they did not, nor did the LMI, who have now expressly stated that, like the CMI, they will not do so. As I see it, in these circumstances the position of the LMI is the same as that of the CMI. The causes of action advanced in England in the 2006 action and in 2011 Folio 702, as summarised on behalf of the LMI, are claims by the LMI against Starlight and OME based on clauses 3, 4 and 5 of the LMI settlement agreement. Those advanced in 2011 Folio 1043 are claims by the LMI against the co-assureds to enforce the English jurisdiction clause in the insurances.

Since, on this basis, the relief sought by the LMI is not a declaration of non-liability, the conclusions and reasoning set out above on the question whether the causes of action are the same apply to it. It follows that I would allow the appeals of both the CMI and the LMI on the Article 27 point.

However these conclusions are subject to the question whether any of the issues discussed above should be referred to the CJEU. Left to myself, I would not refer any of them because the principles of European law are clear and the only question is how they should be applied in the instant case. However, Lord Mance has arrived at a different view from me on the question whether Article 27 applies to the claims by both the CMI and the LMI for a declaration that the Greek claims fall within the terms of the release in the settlement agreements or that under the agreements the tort claims have been settled. In short he is of the view that those claims are essentially for declarations of non-liability. In these circumstances, I have reached the conclusion that the position is the same as I previously considered it to be when I thought that the LMI were seeking a declaration of non-liability. That is that, unless the CMI and the LMI abandon those claims within 14 days, we should refer the question whether the claims for those declarations involve the same cause of action as the claims in Greece within the meaning of Article 27. Lord Neuberger has also given reasons why, absent such abandonment, this question should be referred.

On the other hand, if the CMI and the LMI do abandon those claims, I would allow both their appeals under Article 27 and refuse a mandatory stay of the proceedings under it. If they do not abandon those claims, I would allow the appeals under Article 27 in respect of the other claims but refer the question referred to above to the CJEU and defer a decision on that issue until the CJEU has determined the question.

It is not I think in dispute (and is in any event correct) that a court is only seised of claims by or against new parties from the date that those parties are added to the proceedings. In relation to the 2006 proceedings, the English court was only seised of claims against OME once OME was joined to the proceedings on 20 September 2011 and, for example, to the extent that the LMI in action 2011 Folio 702 are seeking declarations relying on the settlement agreement as a settlement of or defence to Starlight's and OME's claims in the Greek proceedings, the English courts were only seised of that action in 2011. It follows that, in each of those cases the court first seised was the Greek court and not the English court, and that, to the extent thatthe LMI advance claims for a declaration that the Greek claims fall within the terms of the release in the settlement agreement or that under the agreement the tort claims have been settled,unless the English court is the court first seised, they will be entitled to a stay under Article 27. The same is essentially true of the CMI claims.

The question is which court is first seised of what in circumstances where some of the claims brought in England are different from and based on different causes of action from those brought in Greece and one of them in each case, namely the claim for the declaration or declarations referred to above, is based on the same cause of action.

The approach of the parties is starkly different. It is submitted on behalf of the appellants that the answer is to be found in the language of Articles 27 and 30 and is that the court first seised is that in which the proceedings were first brought and that the court remains the court first seised of the proceedings even where those proceedings are subsequently amended by the addition of new claims or otherwise. It is submitted on behalf of the respondents, by contrast, that if a new claim is added by amendment, the court is seised of the proceedings so far as that amendment is concerned when the amendment is made and not at the time of the institution of the original, unamended proceedings. It seems to me that there is considerable force in the appellants’ analysis of the language of the Regulation but the respondents’ case has support both in the English cases and in the textbooks. In the course of this judgment I will consider the issues (interesting as they are) only briefly because I have reached the conclusion that, if the appellants persist in their claims for the declarations referred to in paras 58 and 59 above and this issue is critical for the resolution of the appeal, the proper course is to refer the question to the CJEU.

The case for the appellants can be summarised thus. Article 27 is concerned with “proceedings involving the same cause of action”. So, for the purposes of deciding whether to grant a stay of its proceedings under Article 27, the court must compare the cause or causes of action in each set of proceedings. It is Article 30 that determines when the court is “deemed to be seised” and, by Article 30(1), it provides that (subject to the limited exceptions at the end of Article 30(1) and in Article 30(2)), it is deemed to be seised when the document instituting the proceedings or an equivalent document is lodged with the court. Where the question is which of two courts is first seised, the two dates on which the courts are deemed to be seised are compared and the court deemed to be seised first is the court first seised. The appellants also rely upon the transitional provisions in Article 66, which they say support the proposition that proceedings have only one date upon which they are instituted and is inconsistent with the idea that they can have several such dates as and when new claims are added by amendment.

The appellants say that in this case the answer is that the English court was the court first seised because the Greek court was not seised until some five years later. They say that this is a simple rule which is easy to apply and that there is no warrant in the language of the Regulation for concluding that it was intended that the court should be seised anew each time a new claim is added by amendment, which would be complicated and unnecessary and give rise to endless interlocutory disputes.

The appellants criticise Longmore LJ for asking in para 52 whether it can be said that the English court was first seised of the relevant causes of action now pursued in Greece and for noting that Article 27 only has regard to "causes of action" rather than proceedings. They say that that is inconsistent with Articles 27 and 30 because Article 27(1) uses the word "proceedings" twice and it is used again in Article 30(1). They recognise that for the purpose of deciding whether there is le même objet or la même cause the court must look to the claims made but, for the purpose of deciding which court is deemed to be "first seised" under Article 27, the autonomous test in Article 30 is applied. Finally, they say that Article 30 does not mention "causes of action" and that the Court of Appeal overlooked the word "proceedings" used twice in Article 27, and did not refer to Article 30 at all. Moreover, although the word "proceedings" is not defined in the Regulation, it appears nearly 50 times in the Regulation used as a word of general application. The uses of the word show that issues or causes of action (or claims) may change during the course of the "proceedings".

The appellants further criticise Longmore LJ in the Court of Appeal by reference to paras 53 and paras 64-66. They contrast the reference in para 53 to Article 27 having regard only to causes of action rather than proceedings, with the reference in para 64, with apparent approval, to this quote from the judgment of Saville LJ in The Happy Fellow at pages 17-18:

“… article 21 is concerned with proceedings and article 22 with actions. The questions are whether the proceedings involve the same cause or object or whether the actions are related. It is thus a misreading of the Convention to ask which Court is first seised of issues which are or might be raised within the proceedings or actions. If such were the case, then the articles would achieve precisely the opposite of their intended purpose which is, to achieve the proper administration of justice within the Community ..."

Saville LJ was there considering the position under what is now Article 28. However the appellants say that the word “action” in Article 28 means the same as “proceedings” in Article 27 and that Longmore LJ was correct in paras 64-66 and wrong in para 53.

Although the appellants’ case has to my mind the merit of simplicity and of the avoidance of time consuming and expensive satellite litigation, the respondents say that it is simplistic and contrary to both principle and authority. It is fair to say that there is considerable support in the authorities and the text books for the proposition that the new claims added to the 2006 proceedings, which were founded on the Greek proceedings and thus made second in time, were new claims, that the English court should be regarded as seised of them only when they were added to the 2006 proceedings and that the Greek court was the court first seised within the meaning of Article 27.

In the important case of FKI Engineering Ltd v Stribog Ltd [2011] 1 WLR 3264, which was itself a case on Article 28, the Court of Appeal considered Article 27 and a number of cases decided under it. At para 84 Rix LJ said that the essence of the cases was that, where the “same cause of action” or “the same parties” are introduced only by way of service, or amendment, the relevant proceedings are only “brought” at the time of such service or amendment, not at the time of the institution of the original, unamended proceedings. Neither Mummery LJ nor Wilson LJ expressed a different view.

The respondents also rely upon Sinco per Beatson J at paras 61 to 68 and, in that connection, upon this comment in Briggs on Civil Jurisdiction and Judgments, 5th edition, 2009 at para 2.235, page 327, note 1:

“In [Sinco] the proposition that an English court was first seised of a claim for damages for breach of a jurisdiction clause, which could only have been brought before the English court after the objected-to proceedings were instituted before the foreign court, was rather challenging.”

And in Research in Motion UK Ltd v Visto Corporation [2007] EWHC 900 (Ch), Lewison J said at para 19:

“It is also common ground that the counterclaim is to be treated as an action in its own right for the purposes of the judgment regulation. It seems to me that once RIM's English non- infringement action is out of the way the only relevant proceedings are Visto's counterclaim and the Italian proceedings. Of those two, the Italian court is plainly the first seised. Indeed it cannot be otherwise since the very fact of the Italian claim is part of the foundation of the counterclaim.”

The respondents rely upon Briggs at para 2.235, where, as I read it, their case is supported, although some doubts are expressed as to the desirability of this approach. The respondents also relied upon the 15th edition, 2012 of Dicey, Morris and Collins on The Conflict of Laws at paras 12-060 and 12-069, where they say this:

“12-060. … Each lis between a plaintiff and a defendant has to be considered individually to determine which court was seised of it first in time, and Art. 27 applied accordingly. …

12-069. Where a claim form which has been issued and served is amended by the addition of an additional claim, or by the introduction of a claim or counterclaim against another party, the material question is whether the date of seisin in respect of the additional claim is the date on which the amended claim form is reissued (which may, depending on the circumstances, be only after obtaining the permission of the court), or the date of the original issue. As it is difficult to see how a court can be said to be seised of a claim which has not been made and does not appear in the claim form, it cannot be correct that as long as a claim form has been issued and served, the court already has temporal priority over any issue which may later be added by amendment. It would follow from a conclusion that the court is not seised of the new claim until the amended claim form is reissued that the defendant may be able to pre-empt the amendment by commencing an action of his own in another Member State. The court seised with such pre-emptive proceedings will obviously be regarded as being seised later than the court before which the original action was brought, but institution of the later action may serve to prevent the proposed, and now duplicative, amendment of the original action; and there is no basis in the Regulation for refusing to give effect to a use of the rules which might be characterised as sharp practice.”

Finally, the respondents rely upon Fentiman on International Commercial Litigation, 2010, at para 11.27:

“Principle suggests that an amended claim arising from the same facts as the original claim might be consolidated with the original claim for the purposes of Article 30 but not where the facts arose subsequently. In the latter case it does no violence to the expressions 'actions' or 'proceedings' to differentiate the claims.”

While these expressions of view undoubtedlyprovide strong support for the respondents’ submissions, some of them seem to me to be expressed in a somewhat tentative way and I am not sure that the textbook writers grapple with the points made by the appellants on the language of the Regulation.

However that may be, as indicated earlier, I am of the opinion that this issue is by no means acte clair and, if the appellants maintain their claim or claims in England for a declaration that the Greek claims fall within the terms of the release in the settlement agreements or that under the agreements the tort claims have been settled, I would refer an appropriate question to the CJEU before forming a concluded view with regard to the applicability to that claim or claims. If they abandon them, I would hold that the respondents are not entitled to a stay under Article 27, refuse them a mandatory stay in respect of all the claims and allow the appellants’ appeal.

The question whether those claims which are not within Article 27 should be stayed depends upon whether they should be stayed under Article 28. As stated above, in the exercise of his discretion the judge refused the respondents’ application for a stay under Article 28. The appellants say that he was right to do so. The respondents’ case is that the English court was second seised for the purpose of Article 28 and that a stay should be granted as a matter of discretion.

It is plain from the express terms of Article 28(1) that the discretion in Article 28 is limited to any court other than the court first seised. It follows that, if the English court was first seised, it has no discretion to stay. Article 28 moreover applies to related actions pending in the courts of different member states and, by Article 28(3), actions are deemed to be related where they are so closely connected that it is expedient to hear and determine them together to avoid the risk of irreconcilable judgments resulting from separate proceedings. It is not in dispute in these appeals that the various proceedings are related proceedings for the purposes of Article 28 and I would in any event so hold. The questions remain whether the actions are pending, whether the English court is the court first seised and, if it is not, how the discretion should be exercised.

In Stribog the Court of Appeal considered the correct approach to Article 28. It held that two questions arise, namely (1) whether the two sets of proceedings are related, taking account of any amendments which have been made at the time of the enquiry and (2) which set of proceedings were commenced first? Rix LJ expressed the position clearly at paras 119 and 120. He explained that it is only when there are related and pending actions in separate member states that Article 28 comes into issue. The question whether they are related is, as he put it, the Article 28(3) question. He then said:

“119.

… The question of when seisin occurs and thus which of the courts is the court first seised is the article 30 question. FKI’s submission in effect seeks to roll the two questions together and ask: which of the two courts is the first to be seised of an action which at the time of its seisin was a related action? This is the concept of the ‘first related action’, a concept found in neither article 28 nor article 30. Stribog on the other hand asks: once you have found two related and pending actions and seek to stay one of them, invoking article 28, which of the two courts was the first to achieve seisin of one or other of those actions?

120.

In my judgment, the latter question is the correct one, and is to be preferred to the former ...”

See also per Mummery LJ at paras 40 to 44, where he stressed in particular that the question is whether the court concerned is seised of an action and not of a particular issue in an action. He also stressed that the time at which the comparison between the two actions is made is the time of the hearing of the application for the stay.

Wilson LJ noted at paras 132-134 that Mummery and Rix LJJ asked the relevant questions in a slightly different order: Mummery LJ asked which court was first seised in a pending action before asking whether the actions were related, whereas Rix LJ preferred to ask them in the reverse order. Wilson LJ said that he did not see why the order matters but that Rix LJ seemed to have the terminology of Article 28 on his side. I agree.

The question whether the actions are pending is closely related to the question whether the English court remains first seised. The respondents say that there was no action pending in England when the Greece 1 proceedings were commenced. In the alternative they say that, if the original action is still alive, the English court is not first seised because the claims now brought are entirely new claims, which they say should be equated with new proceedings. I will consider these points in turn.

On the first point, the appellants say, by contrast, that the 2006 proceedings are still on foot, and thus pending, having been stayed but not finally concluded. I would accept the appellants’ submissions. The settlement agreements were in this respect in identical terms. It was a term of them that Starlight would obtain a stay by way of Tomlin Orders. The orders were both in the same terms, which are standard in such cases, and (as quoted in para 7 above) provided that “save for the purposes of carrying into effect the terms agreed … all further proceedings … shall be stayed.” It appears to me that, on the true construction of those orders, the actions remained unstayed for the purposes of carrying into effect the terms agreed and were otherwise stayed.

As I see it, in so far as the actions remained unstayed, it follows that the court remained seised of them, presumably at least until there was no longer any need for the terms agreed to be carried out. It is plain from the language of Article 28(1) that “the court first seised” means the court first seised of the action, which must mean first seised of the proceedings, not of particular claims or causes of action within the proceedings. It seems to me to follow that, in so far as the appellants are seeking to enforce the provisions of the settlement agreements, as they are, the English court remains first seised. I arrive at this conclusion by a construction of Article 28(1) and of the Tomlin Order. The appellants were able to pursue these claims without issuing further proceedings. In this regard I would accept the analysis of the judge at paras 24 to 29. I would adopt the analysis of Sir Andrew Morritt V-C in Bargain Pages Ltd v Midland Independent Newspapers Ltd [2003] EWHC 1887 (Ch) and I would not follow the reasoning of the Court of Appeal in Hollingsworth v Humphrey, (1987) CAT 1244.

What then of the parts of the actions which are stayed under the Tomlin Orders? These would include the claims for breach of the exclusive jurisdiction clauses in the policies of insurance, which do not depend upon the terms of the settlement agreements. The appellants rely upon principles developed by the English courts as a matter of English, not European, law. However, this is in my opinion a permissible approach. Article 30 of the Regulation provides for the circumstances in which a court is deemed to be seised. I recognise of course that the concept of seisin is an autonomous European law device but Article 30 does not make express provision for the circumstances in which it ceases to be seised. In these circumstances, it seems to me to be appropriate for national courts to have regard both to the nature of seisin in European law and to their own procedural rules in deciding whether their courts are no longer seised of a particular set of proceedings.

The appellants rely upon the decision of the Court of Appeal in Rofa Sport Management AG v DHLK International (UK) Ltd [1989] 1 WLR 902, where the Court of Appeal held that a stay of proceedings is not equivalent to a dismissal or discontinuance and therefore that an action in which all further proceedings have been stayed, even if by consent of all parties after a settlement, remains in being. See in particular per Neill LJ at 909H to 910D and 911A-C. He concluded that, for the sake of clarity and certainty, the word ‘stay’ in an order should not be treated as a possible equivalent of a dismissal or discontinuance. Although the action cannot continue without an order of the court, nor can it, he said, be regarded as dead in the same way as an action which has been dismissed or discontinued by order. I agree.

The reasoning in Rofa supports the conclusion that in circumstances in which the 2006 proceedings have been stayed and not dismissed or discontinued the court remains seised of them. It is not and could not be disputed that the court was seised of the proceedings in accordance with Article 30 when the claim form in the 2006 proceedings was issued. It is not suggested that the appellants failed to take any of the steps referred to in Article 30(1) or (2) which would have nullified that effect. The question is whether anything happened subsequently from which it can be inferred that the court was no longer seised. I would answer that question in the negative.

Although Rofa was not a decision on the construction of the Regulation, the correct approach is to consider whether anything occurred which could lead to the conclusion that the approach adopted there should not be applied to the stay incorporated in the Tomlin Orders and, if not, whether there is anything which leads to the conclusion that the court is not still seised of the proceedings. I would answer both those questions in the negative. Although it is true that the CMI settlement agreements contained a provision that, on payment of the settlement sum, the parties would file a consent order dismissing the proceedings, no such consent order was made or filed. The LMI settlement agreement does not contain any such provision. In all these circumstances, I can see no sensible basis upon which it can be said that the English court is no longer seised of the proceedings. There remain significant disputes arising out of the settlement agreements and the insurances.

The second point taken on behalf of the respondents under this head is that, even if the original action is still alive, the claims now brought are new claims, which should be equated with, or treated as, new proceedings. They rely upon this dictum of Rix LJ in Stribog at para 129:

“Seventhly, there is nothing in the ECJ or English jurisprudence to support the judge’s approach in this case. It is possible that the introduction of entirely new causes of action or parties is to be recognised as the bringing of entirely new proceedings, so that the timing of seisin (the article 30 question) has to be looked at from that point of view, as occurs for the purposes of article 27. Even so, it is not clear to me that in this connection article 27 and article 28 work in the same way: for article 27 is worded in terms of the bringing of actions with the same parties and the same cause of action (‘Where proceedings … are brought in the courts’) whereas article 28 is worded in terms of the pendency of related actions (‘Where related actions are pending in the court’) (emphasis added). That emphasises that the article 28 question is asked with relation to pending actions, and not, as the article 27 question is asked, with relation to the bringing of actions. In any event, the judge is in my respectful judgment mistaken to think that any amendment is analogous to the bringing of new causes of action or the addition or substitution of new parties.”

For my part, I would not accept that approach as applied to Article 28. In para 68 above I referred to the statement of Rix LJ at para 84 of Stribog. In para 63 of his judgment in the instant case Longmore LJ quoted para 84, where Rix LJ said that, where proceedings are amended to add new claims, the court is only seised of the relevant proceedings so far as the new claims are concerned at the time of the amendment. Immediately after the quote, Longmore LJ correctly pointed out that those observations were made in relation to Article 27 and not Article 28. He then quoted the second sentence from the above quotation from para 129 of Rix LJ’s judgment.

Longmore LJ then asked whether this tentative expression of view in relation to "the introduction of entirely new causes of action" being tantamount to "the bringing of entirely new proceedings" means, for the purpose of this case, that the Greek courts are to be regarded as first seised of the relevant related action? He said at para 64 that, in his opinion it did not. He gave two reasons. He said that in the first place Rix LJ had already quoted the passage from the judgment of Saville LJ in The Happy Fellow which I set out in para 66 above.

At para 65 Longmore LJ said that, in the second place, Rix LJ provided his tentative response to his tentative view in the remainder of paragraph 129 which he then quoted. That response is to my mind telling. Longmore LJ then expressed his conclusion at para 66. He expressed doubt about Rix LJ’s distinction between entirely new causes of action as opposed to partially new causes of action. However that may be, his conclusion seems to me to be contained in the last two sentences of para 66:

“As Saville LJ said in The Happy Fellow it is a misreading of Article 28 to ask which court is first seised of ‘issues’; it must likewise be wrong in an Article 28 context to ask which court is first seised of ‘causes of action’. That is Article 27 territory because, for the purpose of Article 28, one has to ask which court is first seised of an action, not a cause of action and, still less, an issue.”

On that basis Longmore LJ concluded at para 67 that, if the original English action and the subsequent Greek actions are related, as he concluded they are, it was the English court that was the court first seised. I agree. First, the contrary view seems to me to be inconsistent with the two stage approach to Article 28 adopted in Stribog. As Longmore LJ observed at para 66, in the context of Article 28 it is wrong in principle to ask which court is first seised of a cause of action, because Article 28 is concerned with related actions as a whole. Secondly, I would accept the appellants’ submission that on the facts of this case the claims now brought are not (as Rix LJ put it) entirely new. On the contrary, applying the “broad and common sense approach” favoured by Lord Saville in Sarrio, the claims now brought by the appellants are unquestionably “related” to the original action within the meaning of Article 28.

I would only add in conclusion that it seems to me that it would be very odd indeed if a court which is seised of proceedings and stays those proceedings by way of a Tomlin order on the express terms that it retains jurisdiction to take further steps by way of implementation or policing of the order were prevented from exercising that jurisdiction, either by lifting the stay or otherwise, on the ground that it was no longer seised of the proceedings. It seems to me to be at least arguable that those steps should properly be treated as part of the existing proceedings. They might perhaps be treated as part of the same “procedural unit” as discussed by the CJEU in Purrucker v Valléz-Pérez (No 2)(Case C-296/10) [2011] Fam 312 at para 80. The case was on very different facts but was concerned with two paragraphs in a regulation which were identical to Articles 27 and 30 of the Regulation. In any event to treat the enforcement action as something entirely new seems to me to be wrong.

It is never easy to decide what is an entirely new claim, what is a new claim and what is an expansion of an old claim. These claims are not new or entirely new because they are brought by way of enforcement of the outcome of the original dispute, in the same way as execution on a money judgment. In these circumstances it makes sense to hold that these claims, which largely arise out of the settlement agreements, arise out of the attempts made by the respondents to avoid the effect of those agreements and, in particular, the exclusive jurisdiction agreements. This solution would, as I see it, be consistent with the overall policy of the Regulation to avoid a multiplicity of proceedings. However, I can see that there is scope for argument under this head and, if the issue of first seised were critical to the decision, it might be appropriate to refer an appropriate question to the CJEU. I therefore turn to the issue of discretion on the assumption that the English court is second seised for the purposes of Article 28.

On that assumption, the question arises whether the action or actions should be stayed as a matter of discretion. The judge held that no such stay should be granted. Given that the shape of the case has changed considerably since the matter was before the judge, it appears to me that this Court should consider for itself whether to grant a stay. I have reached the clear conclusion that it should not. I have reached that conclusion essentially for the reasons advanced on behalf of the appellants. They may be summarised in this way.

In Owens Bank Ltd v Bracco(Case C-129/92) [1994] QB 509, at paras 74-79, Advocate General Lenz identified a number of factors which he thought were relevant to the exercise of the discretion. They can I think briefly be summarised in this way. The circumstances of each case are of particular importance but the aim of Article 28 is to avoid parallel proceedings and conflicting decisions. In a case of doubt it would be appropriate to grant a stay. Indeed, he appears to have approved the proposition that there is a strong presumption in favour of a stay. However, he identified three particular factors as being of importance: (1) the extent of the relatedness between the actions and the risk of mutually irreconcilable decisions; (2) the stage reached in each set of proceedings; and (3) the proximity of the courts to the subject matter of the case. In conclusion the Advocate General said at para 79 that it goes without saying that in the exercise of the discretion regard may be had to the question of which court is in the best position to decide a given question.

On the facts here those questions can be considered together. As I see it, the issues are not dissimilar from those considered by Cooke J in Primacom at para 65, where he said this:

“Even if I had found that these two sets of proceedings and the German proceedings were related within the meaning of article 28, 'the strong presumption' which 'lies in favour of the applicant' on an application for a stay would be overridden here by virtue of the terms of the SSFA. Although the ECJ decision in Gasser means that a stay is mandatory where article 27 applies, there is no reason why weight should be given to that decision in the context of article 28, where a discretion is given to the court, the jurisdiction of which has been agreed by the parties as exclusive. It is nothing to the point that an English court could not have issued an anti-suit injunction to prevent the German proceedings (as per C-159/02 Turner v Grovit [[2005] 1 AC 101]). The injustice of giving precedence to proceedings brought in breach of an exclusive jurisdiction clause where the parties have agreed that England is the appropriate forum is self-evident. To breach the clause and to gain the benefit of priority for the German courts by such breach offends justice, where the court has a discretionary decision to make.”

In my opinion, similar considerations apply here. Although the true construction of the settlement agreements and the question whether Starlight and OME are in breach of them is ultimately a matter for the court which finally determines the summary judgment application or for the court at trial, there is a strong argument (to put it no higher) that the Greek proceedings have been brought by Starlight and OME in breach of the settlement agreements, which are subject to the exclusive jurisdiction of the English courts and/or in breach of the exclusive jurisdiction clauses in the insurance contracts.

I would reject the submission that those considerations are impermissible in the light of the decision in Gasser. It was there held that, if the criteria for ordering a mandatory stay under Article 27 are satisfied, then the court second seised must stay its proceedings even if the court second seised has jurisdiction under an exclusive jurisdiction clause falling within Article 23. That conclusion was reached on the basis that, under Article 27, where there are two sets of proceedings which involve the same cause of action and the same parties, the court second seised is obliged to order a stay. The Regulation only permits one set of proceedings to continue. The position is quite different under Article 28, which clearly contemplates that where there are two related sets of proceedings they may proceed in parallel. That conclusion follows from the proposition that the grant of a stay is discretionary and not mandatory. In these circumstances, I can see no reason why, in exercising that discretion under Article 28, the court second seised should not take into account the fact that the parties had previously agreed (or arguably agreed) an exclusive jurisdiction clause in favour of that court. On the contrary, depending upon the circumstances of the particular case, that seems to me to be likely to be a powerful factor in support of refusal of a stay. After all, Recital 14 expressly provides:

“The autonomy of the parties to a contract, other than an insurance, consumer or employment contract, where only limited autonomy to determine the courts having jurisdiction is allowed, must be respected subject to the exclusive grounds of jurisdiction laid down in this Regulation.”

There is a close relationship between the claims in England and the subject matter of the claims in Greece. The natural court to consider the issues raised by the CMI and the LMI is the High Court in England because they raise contractual questions governed by English law and because it is at least arguable that the parties have agreed that they should be decided by the High Court, where the proceedings are more advanced than in Greece. After all, the judge granted summary judgment as long ago as December 2011. The court in Greece will then have the benefit of the decision of the court which, in the Advocate General’s language, is in the best position to decide these issues. Once there is a final judgment of the English courts, it will be recognisable in Greece, as elsewhere in the EU and will assist the Greek court. In this way, the principles of mutual trust upon which the Regulation is founded will be respected and there will be no risk of irreconcilable judgments.

In these circumstances I would uphold the decision of the judge in refusing a stay under Article 28. There is no need for a reference to the CJEU because the question I would have referred does not arise given my conclusion on the exercise of discretion. It was at one time suggested that there is a referable question as to whether Article 28 gives the court second seised a choice between staying the proceedings under Article 28(1) and declining jurisdiction under Article 28(2). However, that suggestion was abandoned before the hearing. I would in any event have rejected it as unarguable. There is no support whatever for it in the language of Article 28 and none of the sources referred to supports the conclusion. The discretion is to stay or not to stay under Article 28(1) and to decline or not to decline jurisdiction under Article 28(2). The Court may thus both refuse to stay and refuse to decline jurisdiction. As the Advocate General explained in Bracco, all depends upon the circumstances.

The remaining question is whether the Court of Appeal was wrong to reject submissions made on behalf of the appellants that it was too late for the respondents to rely upon Article 27. This is another part of the case where the facts seem to me to be startling. The appeal on this point is brought by the LMI and not the CMI but it is I think accepted that, if the appeal succeeds, the CMI will be able to take advantage of it. The most important point raised by this part of the appeal is whether the courts had a discretion to hold that the LMI should not be permitted to rely upon various procedural acts and omissions on the part of the respondents in response to their attempt at a late stage to rely upon Article 27 of the Regulation or whether, once the point was brought to its attention, the Court of Appeal was bound to consider Article 27 (as quoted at para 24 above) because it expressly provides that, where the conditions are satisfied “any court other than the court first seised shall of its own motion stay its proceedings until such time as the jurisdiction of the court first seised is established”.

I have reached the conclusion that the answer is that the appellants were entitled to rely upon the acts or omissions of the respondents and that, having regard to what had happened before Judge Mackie QC and the judge, the Court of Appeal was not bound to take the point of its own motion. Moreover, subject to a possible reference, I would hold that the Court of Appeal should have considered the acts or omissions of the respondents and have held that it was too late for the respondents to rely upon Article 27. The question of the scope of the Court of Appeal’s duty to take the point of its own motion in circumstances of this kind is however an important point on the construction of Article 27 and, if it were necessary for the determination of the appeal, I would refer it to the CJEU. However, if the LMI abandon their claim or claims for a declaration of non-liability a reference will not be necessary for the determination of the appeal. If they do not, my present view is that it will.

The relevant chronology, which I take from the Statement of Facts and Issues, is briefly as follows. I will omit references to the CMI proceedings, in which the applications were heard at the same time as those in the LMI proceedings. By application notice dated 3 August 2011, the LMI applied for wide ranging relief against Starlight to enforce the LMI settlement agreement. By application notice dated 18 August 2011 the LMI sought permission to join OME and to serve OME out of the jurisdiction. As explained earlier, the LMI commenced 2011 Folio 702 against Starlight and OME in order to enforce the LMI settlement agreement. They also commenced 2011 Folio 1043 only against the co-assureds, which was an action founded solely on the exclusive jurisdiction clause in the policies.

On 20 September 2011 the LMI obtained permission from Judge Mackie QC to issue a Part 20 claim against OME in 2006 Folio 815 and, lest it be needed, to serve that Part 20 claim form and the claim forms in 2011 Folios 702 and 1043 out of the jurisdiction and to serve them on Lax & Co in London. The applications were supported by a witness statement by their solicitor, Mr Zavos, in which he referred both to possible stays under Article 27 and Article 28 giving reasons why stays should not be granted. The orders gave notice to each of Starlight, OME and the co-assureds that:

“You may apply within seven days after the date of service of this Order on you to have the Order set aside or varied. This time limit does not apply to an application to dispute the jurisdiction of the Court in respect of which the procedure in CPR Part 11 as modified by CPR Part 58 applies”

No such application was made.

Starlight did not serve evidence within the time provided in the CPR. However, on 4 November 2011 they served evidence which included an express request by Mr Crampton of Lax & Co that the relief sought by the appellants on the merits be denied, alternatively that the matter be referred to a full trial, with provision for disclosure and exchange of witness and expert evidence. On 7 November Starlight, OME and the co-assureds each filed a defence on the merits in the relevant action, having first obtained an extension of time for doing so.

Each of the defences included a paragraph which stated:

“The claims in the Greek Proceedings fall outside the jurisdiction clause in the policy and the jurisdiction clause in the Settlement Agreement. It is respectfully denied therefore that the High Court of Justice of England and Wales has jurisdiction to determine the claims in the Greek Proceedings”

The grounds on which Starlight, OME and the co-assureds opposed the appellants’ claims and applications for summary relief, were in summary that the claims brought in the Greek proceedings (1) did not fall within the scope of the releases contained in the LMI settlement agreement or the CMI settlement agreement; (2) did not fall within the scope of the indemnities contained in the settlement agreements; and (3) did not fall within the scope of the jurisdiction clauses contained in the settlement agreements or in the policies.

Following service of the defences, the LMI applied for summary judgment in all the actions and all the applications were fixed to be heard on 28 and 29 November at the same time as the application for summary relief against Starlight in the 2006 proceedings. In their skeleton argument prepared for those hearings, which were served on 23 November 2011, the LMI included the following:

“71.

There has been no application for a mandatory stay under Article 27 of the Judgments Regulation in respect of the [LMI’s] claims to enforce the jurisdiction clause in the contract of insurance, and to enforce the terms of the [LMI] Settlement Agreement. This is (no doubt) because the claims are different claims from the claims advanced by the Assureds in Greece.”

On 25 November 2011, Starlight, OME, and the co-assureds, through their former counsel, James Drake QC and Emma Hilliard, provided their skeleton argument to the court, which expressly disavowed any application under Article 27, in these terms:

“69.

It is well-established that in order for Article 27 to operate there must, when comparing the two sets of proceedings in issue, be three identities: of parties, of “cause”, and of “objet”: see generally Briggs & Rees, Civil Jurisdiction and Judgments (5th ed 2009) at paras 2-227 to 2-231.

70.

Starlight does not here contend that there is here an identity of ‘cause’ and ‘objet’ between the Greek proceedings and the Insurers’ applications. Although designed to preclude ‘in so far as possible, and from the outset’ a clash of verdicts, the operation of Article 27 (as distinct from Article 28) is highly restricted in its actual operation. Comparison must be made between the claims made in the two actions, regardless of possible defences, to see whether they proceed on essentially the same facts and under the same rule of law.”

In the footnotes they referred to the cases I have discussed earlier, including Gubisch, Gantner and The Tatry.

It is thus plain that before the matter came before the judge the respondents had made a clear and reasoned decision not to rely upon Article 27. Moreover, there is no reason to think that the judge did not consider the points they made and accept them. They relied only on Article 28. They did so pursuant to an application made by application notice dated 24 November 2011. However that application was out of time. So, by further application notices in each action dated 28 November 2011, the respondents applied for permission to make the Article 28 application out of time, and for relief from sanctions pursuant to CPR Part 3.

The sanction referred to was that imposed by CPR Part 11, which provides that a defendant who files an acknowledgment of service and fails to apply to the court within the time allowed under the CPR for an order declaring that it has no jurisdiction or should not exercise any jurisdiction which it may have, “is to be treated as having accepted that the court has jurisdiction to try the claim”: CPR rule 11(5).

As stated in para 19 above, the judge dismissed the stay application under Article 28 and held that the appellants were entitled to summary judgment. He held that (1) each of the claims made by Starlight, OME, and the co-assureds against the appellants in Greece is in breach of the exclusive English jurisdiction agreement in the policies; (2) each of the claims made by Starlight and OME against the appellants in Greece is in breach of the jurisdiction agreements in the settlement agreements which provide for exclusive English jurisdiction; (3) each of the claims made by Starlight and OME against the appellants in Greece is in breach of the terms of the settlement agreements; (4) each of Starlight, OME and the co-assureds is liable in damages to the insurers for breach of contract and under Section 50 of the Senior Courts Act 1981; and (5) each of Starlight and OME is bound to indemnify and hold the insurers harmless against each of the claims in the Greek proceedings pursuant to the indemnities in the settlement agreements.

The judge handed down his judgment on 19 December 2011 and fixed 2 February 2012 for the hearing of consequential applications. In the meantime, on 7 December 2011 Thomas Cooper had replaced Lax & Co as the respondents’ solicitors. On 24 January 2012 draft grounds of appeal were served which included for the first time reliance on Article 27. They were considered in a somewhat amended form by the judge. The judge granted permission to appeal on a number of grounds including the Article 27 point. As to that he said that he would not have given permission on that point alone, as he put it, “not least because the Article 27 case could become the subject of an independent application at first instance at any time hereafter”. He recognised that this would have the effect of turning the Court of Appeal into a first instance court but concluded that it could be argued without the need for further evidence and without a great addition of time.

In the Court of Appeal the appellants relied upon the provisions of CPR Part 11, but the Court of Appeal held that it did not apply because applications under Articles 27 and 28 are not challenges to the jurisdiction. It further held that it was bound to take the Article 27 point of its own motion. The LMI say that the Court of Appeal was wrong on both points.

CPR Part 11 provides, so far as relevant as follows:

(2) A defendant who wishes to make such an application must first file an acknowledgment of service in accordance with Part 10.

A defendant who files an acknowledgment of service does not, by doing so, lose any right that he may have to dispute the court’s jurisdiction.

An application under this rule must -

(b) be supported by evidence.

If the defendant-

he is to be treated as having accepted that the court has jurisdiction to try the claim.

An order containing a declaration that the court has no jurisdiction or will not exercise its jurisdiction may also make further provision including -

(b) setting aside service of the claim form;

discharging any order made before the claim was commenced or before the claim form was served; and

staying the proceedings.

In an action in the Commercial Court such as this CPR 11(4) is varied by CPR 58.7(2) so that the application under CPR 11(1) must be made within 28 days after filing an acknowledgment of service and not 14 days. As I understand it acknowledgments of service were filed in each case.

The position under CPR Part 11 is different from the position under the former Rules of the Supreme Court, under which the equivalent rule, namely RSC Order 12 rule 8(1), did not include an application for a stay. By contrast CPR 11(1)(b) applies to an application for an order that the court should not exercise its jurisdiction. An application for a stay is precisely that. An application for a stay under Article 27 is thus an application within CPR 11(1)(b). The applicant must file an acknowledgment of service and must make an application within 28 days. The respondents did not do that. Nor did they seek an extension of time to so do within the CPR. It is arguable that the effect of CPR 11(5) is that their failure to do so means that they are treated as accepting that the court both has jurisdiction and that it is free to exercise it. The difficulty is that the wording of paragraph (5) may only relate to the existence of the jurisdiction rather than the exercise of it. This point was left open in Texan Management Ltd v Pacific Electric Wire & Cable Company Ltd [2009] UKPC 46 at paras 68 and 69.

However that may be, the LMI rely upon the voluntary submission to the jurisdiction evidenced by the acknowledgment of service and the service of a defence. They also rely upon the clear and unequivocal statement of the respondents’ position in their skeleton argument before the judge. It is plain from the terms of the concession quoted at para 106 above that serious thought had been given to the question both of whether to make the concession and of the basis on which it was to be made. In these circumstances, unless there is some rule of European law to the contrary, it appears to me that the Court of Appeal should have considered whether, in the exercise of their discretion to permit argument on a new point, they should exercise that discretion in favour of the respondents or not. Moreover, it appears to me that, given the clear basis on which the concession was made and, given that the judgment had proceeded on that basis, the Court of Appeal should have held that it had a discretion under CPR rule 11(1) to permit an application under the rule to be made out of time but should have refused to exercise it.

However it is said that on the true construction of Article 27, the court, including on these facts the Court of Appeal, has a duty to consider the application of Article 27 of its own motion whenever the point is taken. This strikes me as extremely improbable. I would accept the submissions of the LMI in this respect.

The CJEU has recognised the importance of national rules of procedure. Thus, for example, in Shevill v Presse Alliance SA (Case C-69/93) [1995] 2 AC 18 the CJEU said:

I would accept the LMI’s submission that Article 27 is part of European law and overrides national law which is incompatible with it. It does not however follow from this proposition that English procedural rules were overridden. A national procedural rule must not impair the effectiveness of Article 27. It must not render the exercise of rights conferred by EU law impossible or excessively difficult: Amministrazione delle Finanze dello Stato v SpA San Giorgio (Case 199/82) [1983] ECR 3595, [1985] 2 CMLR 658. This is the principle of effectiveness, which involves considering whether the rule can operate consistently with Article 27, or whether it is incompatible with it. The procedural rule should not be less favourable than those governing similar domestic actions, which is the principle of equivalence: see eg Interfact Ltd v Liverpool City Council [2011] QB 744, Kapferer v Schlank and Schlick GmbH(Case C-234/04) [2006] ECR I-2585 at paras 19 to 22, Köbler v Austria (Case C-224/01) [2004] QB 848; and Eco Swiss China Time Ltd v Benetton International NV (Case C-126/97) [1999] ECR I-3055.

I refer only to Interfact, where the Court of Appeal refused to exercise its discretion to allow cases to be reopened under CPR 52.17, so as to give a remedy for infringement of a provision of European law. Lord Judge CJ, delivering the judgment of the Court of Appeal, said:

“41.

In general, EU law does not require national courts to disapply their own procedural rules in order to secure the vindication of EU rights. In Kapferer v Schlank & Schick GmbH … the Austrian Supreme Court was seised of an appeal in which the respondent had failed to lodge within the time stipulated a respondent's notice taking a point on jurisdiction under the Brussels Convention. The court referred to the Court of Justice the questions whether it was, nevertheless, bound to take the point of EU law of its own motion and whether EU law required a national court to review and set aside a final judicial decision in circumstances where it later became apparent that the decision of the court was in breach of EU law. The Court of Justice held that a national court is not so bound …

44.

...[Kapferer] establishes as a matter of general principle that EU law does not require a national court to reopen a final judicial decision, even if failure to do so would make it impossible to remedy an infringement of a provision of EU law: see the Kapferer case, at para 21; Amministrazione dell'Economia e delle Finanze and Agenzia delle Entrate v Fallimento Olimpiclub Srl (Case C-2/08) [2009] ECR I-7501, para 23; Asturcom Telecommunicaciones SL v Rodríguez Nogueira (Case C- 40/08) [2010] 1 CMLR 865 para 37...

49.

... The Court of Justice has upheld national time limits and limitation periods on grounds of legal certainty and the need to ensure finality in decision making, even though the effect has been to preclude enforcement of an EU law right: see, for example, Palmisani v Istituto Nazionale della Previdenza Sociale (INPS) (Case C-261/95) [1997] ECR I- 4025; Fantask A/S v Industriministeriet (Ehrvervsministeriet) (Case C-188/95) [1997] ECR 1-6783.”

Finally, I would accept these submissions made by the LMI. Under English law a final judgment on the merits should not be set aside without very solid grounds: Brown v Dean [1910] AC 373 at 374, per Lord Loreburn. Interest republicae ut sit finis litium. This is part of the common tradition of the legal systems of the Member States: Rewe-Zentralfinanz eG and Rewe- Zentral AG v Landwirtschaftskammer für das Saarland(Case C-33/76) [1976] E.C.R. 1989. As quoted above, in Interfact the Court of Appeal rejected the argument that, where an appellate court has a discretion to exercise under national procedural law to allow a final judgment to be challenged on appeal, it must exercise that discretion so as to remedy the infringement of EU law.

In my judgment, there is no sensible basis upon which it can be said that the time limit under CPR 11(4), which can in an appropriate case be extended under CPR 3.1(2)(a), is contrary to EU law. The time limit satisfies the principle of equivalence because it is the same rule that applies in all cases. It fulfils a legitimate aim, namely making sure that points going to whether the proceedings are to be tried on their substantive merits in England are taken promptly and without unnecessary costs. It satisfies the principle of legal certainty because parties need to know where they stand. The absence of a time limit would allow a litigant to take the point years afterwards. Moreover, the time limit does not render the right to apply for a stay under Article 27 (or Article 28) impossible or excessively difficult to exercise. It allows sufficient time for the point to be raised, especially given the express rule permitting an extension of time in appropriate cases.

As to the expression “of its own motion” in Article 27, there are a number of different parts of the Regulation that have a similar provision. On the facts here the potential for a stay under Article 27 was before the courts on at least two occasions. The position was explained to Judge Mackie QC on the without notice application referred to above. There is no reason to think that he did not give consideration to the position. More importantly perhaps the position was explained to the judge in the skeleton arguments to which I have referred. He was given both reasons and authority on the question whether a stay should be granted under Article 27. It seems to me that the judge was entitled to accept those submissions, which were made on the respondents’ behalf by experienced counsel and solicitors.

For these reasons I would hold that the Court of Appeal should have refused to allow the respondents to rely upon Article 27 in the Court of Appeal. That said, I would accept that the meaning and effect of the duty to consider Article 27 of its own motion are matters of some potential importance and I have (somewhat reluctantly) reached the conclusion that they are not acte clair. I would therefore refer an appropriate question to the CJEU if it were necessary in order to resolve the appeal. If the appellants abandon the claims to the declarations referred to in paras 58 and 59 above, such a reference will not be necessary because, for the reasons given above, I would allow the appeals under Article 27 in their entirety.

It seems to me that rather different considerations apply to Article 28 and that the Court of Appeal were entitled to consider Article 28 as part of the appeal from the decision of the judge who had considered it in detail.

For these reasons I would invite the CMI and the LMI to consider whether they wish to pursue their claims for declarations (referred to in paras 58 and 59 above) that the Greek claims fall within the terms of the release in the settlement agreements or that under the agreements the tort claims have been settled. As Lord Neuberger observes, those are the claims described in para 18(a)(1)(i), 18(a)(2)(i) and 18(b)(i) above. They should indicate their position within 14 days of this judgment being handed down. If they persist in their claims, some limited questions should be referred to the CJEU as described above. The decision whether to stay those claims would then await the result of the reference, although I would allow the appeal under Article 27 in respect of the other claims. If they abandon them, I would allow all the appeals of both the CMI and the LMI under Article 27. I would in any event dismiss the respondents’ cross-appeal under Article 28 and I would hold that their application for a stay under Article 28 should be refused as a matter of discretion. The parties should make written submissions on the form of order and costs within 21 days of the handing down of this judgment. Finally, I would like to thank all counsel and solicitors for their assistance in this unusual and in some respects difficult case.

Subject to one point, I entirely agree with Lord Clarke’s reasoning and conclusions. The one point concerns the issue discussed in paras 44-46 and 58-59 of Lord Clarke’s judgment and in Lord Mance’s judgment. That issue is whether (i) LMI’s claim in England for a declaration that the Greek claims “have been settled”, and (ii) CMI’s claim in England for a declaration that the Greek claims “were compromised” (“the English declaration claims”, described in para 18(a)(1)(i), 18(a)(2)(i) and 18(b)(i) of Lord Clarke’s judgment) should be stayed under Article 27.

In my view, if that issue remains live, it should be referred to the CJEU, as I do not regard it as acte clair.

I see the force of Lord Clarke’s view that the English declaration claims do not have “le même objet et la même cause”, if one gives that expression a very narrow effect. I also accept that, particularly in the light of the existence of Article 28, there is good reason to give Article 27 a relatively narrow meaning, as Rix J pointed out in Glencore International AG v Shell International Trading and Shipping Co Ltd [1999] 2 Lloyd’s Rep 692, 697. I also accept that the decisions of the CJEU cited by Lord Clarke at paras 26-28 of his judgment support the contention that Article 27 has a relatively narrow ambit of application.

However, it is also important to appreciate that the fundamental purpose of Article 27, as explained by the CJEU, is to ensure that judgments obtained in one member state are enforceable in other member states, and that the consequence of this is that one should avoid mutually inconsistent judgments. The purpose of Article 27 is to help achieve that end.

It seems to me that, if the Greek court were to give Starlight and OME judgment for a particular sum in respect of its Greek claims, and the English court were to give judgment in favour of LMI and CMI in the form of a declaration that those very claims have been settled or compromised, the two judgments would be incompatible as a matter of principle and logic. It is not possible for a court to award a claimant damages in respect of a claim which has been compromised with the defendant. To put the point another way, to say that a defendant currently owes a claimant damages in respect of a claim which the defendant has settled or compromised with the claimant involves an illogicality. Accordingly, it seems to me that there is a real case for saying that the English declaration claims should be stayed.

The difference between the English declaration claims and CMI’s and LMI’s claims in England for an indemnity and damages for breach of the settlement agreements (“the English indemnity and damages claims”, as described in paras 18(a)(1)(iii), (v) and (vi), 18(a)(2)(iii) and (iv) and 18(b)(ii), (iii) and (iv) of Lord Clarke’s judgment) may appear to be relatively small, but I believe that there is a crucial distinction, as a result of which it is acteclair that the English damages and indemnity claims do not fall foul of Article 27. The crucial difference is that, if those claims were successful, they could not lead to inconsistent judgments in England and Greece.

I accept that, if they were successful, the English indemnity and damages claims could be fairly said to neutralise, at any rate in commercial terms, any benefit to Starlight and OME of a judgment in the Greek claims. However, crucially in my view, success for LMI and CMI in the English indemnity and damages claims would not be logically inconsistent in any way with success for Starlight in the Greek claims. It is not inconsistent (although it is commercially pointless) to say that a defendant is liable to pay a claimant a sum by way of damages, while the claimant is bound to indemnify the defendant in respect of the whole of that sum (or is bound to pay an equivalent sum to the defendant). Indeed, the indemnity is not merely logically consistent with the liability: it is positively meaningless without the liability for damages, and the liability for damages, though rendered nugatory by the indemnity, is not logically inconsistent with the indemnity.

I am in substantial but not complete agreement with the reasoning and conclusions reached in the course of the judgment prepared by Lord Clarke, although, ultimately, as will appear, we agree on the proper disposition of these appeals.

The differences between Lord Clarke and myself relate to the significance and operation of article 27 of the Council Regulation (EC) No 44/2001 (“the Brussels Regulation”) with regard to the respondents’ Greek claims.

I have no difficulty in agreeing with Lord Clarke’s conclusions regarding the English claims made by CMI and LMI for damages for (i) breach of the exclusive jurisdiction clauses in the Settlement Agreements and insurance policies and (ii) indemnity under clauses 3 and 4 of the respective Settlement Agreements. Such claims do not assert that there is no tort liability because of the Settlement Agreements. They assert (i) that the respondents are claiming in the wrong jurisdiction and (ii) that the respondents have agreed to indemnify them in respect of any tort claims (valid or not) by the respondents themselves as well as by others arising from the loss of the vessel.

However, I do not accept the reasoning by which Lord Clarke reaches his conclusions with regard to these claims for damages and the further “release” claims (as Lord Clarke conveniently calls them) which he addresses in paras 40 to 59 of his judgment. This difference becomes important in relation to the first head of the release claims, as I shall show.

One strand of Lord Clarke’s reasoning is that the English claims based on the Settlement Agreements cannot be the mirror image of the Greek tort claims, because they involve contract and tort claims and cannot constitute the same “cause of action”: para 34, third sentence, para 41, second and third sentences and para 43, second and last sentences.

Another strand is that it is relevant or conclusive that the English and Greek claims do not “interfere” with each other, and, in particular, that the Greek claims do not “impugn” the settlement agreements: para 35, first and second sentences and para 37, in its entirety.

Neither of these strands of reasoning is in my opinion sustainable, for reasons which I will explain.

The release claims need a little analysis. There are three heads. The first head is summarised by the respondents themselves and by Lord Clarke (para 18(a)) as involving claims for declarations that the Greek claims fall within the terms of the release. But this head is in fact pleaded by LMI as a claim for a declaration that the Greek claims “have been settled” (application notice, para (1) 1 and 3), while CMI plead that the Greek claims “were compromised” (particulars of additional claim, para 10) and follow this with a claim for a declaration that the Greek claims fall within clause 2 of the CMI Settlement Agreement (particulars of additional claim, para 27(a)). These are clear statements (right or wrong as they may prove to be) that the Greek claims have been settled or compromised within the terms of the Settlement Agreements.

The second and third heads are claims for a declaration that the bringing of the Greek claims was a breach of the release in each of the Settlement Agreements and for damages for such breach. They must stand or fall together. They raise different considerations from the first head.

The English claims that the Greek claims “have been settled” or “were compromised” are in my opinion mirror images of the Greek tort claims. The English pleas mean, and can only mean that the English claimants are not liable for the Greek tort claims. The legal effect of these English statements is (under English eyes and, I am confident, European law) that the Greek claims are no more. If an English court were to give a judgment to that effect, and there was no prior Greek judgment or other reason for non-recognition, the Greek court ought under the Brussels Regulation to accept it.

It cannot make any difference to the application of article 27 that the reason for non-liability is a contractual settlement agreement. The only point of enforcing the contract is to show that there are no valid Greek tort claims. The Greek claims aim to enforce tort liabilities. The first head of the English claims aims to establish that there are no such valid tort liabilities, because they have been settled. The Greek and English claims cannot stand together.

The concepts used in article 27 (such as “cause of action” or the concept of “same object” which one must read into the English text) are autonomous European concepts: Gubisch v Palumbo Case 144/86, [11] and The Tatry Case C-406/92, [47]. In the latter case, the European Court of Justice said that “the ‘cause of action’ comprises the facts and the rule of law relied on as the basis of the action” and that “the ‘object’ of the action” for the purposes of article [27] means “the end the action has in view” [39]-[41]. An analysis of the cases helps to understand what was meant.

Gubisch v Palumbo happened to concern a situation where the mirror image claims were in a general sense contractual. The German claim was for the price of machinery delivered. The later Italian claim by the buyer was, firstly, that there was no liability because he had revoked his offer before it had reached the seller for acceptance – strictly, this was not a contractual claim, but a claim that there was no contract – and, secondly, that, if there was a contract, his consent was vitiated and the contract should be set aside for mistake or on the ground of the seller’s fraud, or, thirdly, that any contract had been discharged on account of the seller’s late delivery.

Both the question referred and the Court of Justice’s summary of the facts embraced all three aspects of the Italian claim: see e.g. judgment [2] and [4]. The subsequent reasoning and the answer given refer to mirror image claims, one seeking enforcement, the other seeking rescission or discharge, of a contract: see [13] and [15] and the Court’s answer. The Court said [17] that “it must be held that the two actions have the same subject-matter, for that concept cannot be restricted so as to mean two claims which are entirely identical”. The absence of express reference at these points to the first Italian claim (that no contract had ever been concluded) cannot mean that the Court was drawing any distinction between that claim and the other two. On the contrary, the inference is that it saw it as posing no different issue.

It could not have made any difference to the Court of Justice’s conclusions if, instead of or in addition to some or all of the pleas actually made in the Italian proceedings, the Italian claimants had alleged that the contract had been rescinded or discharged under some separate subsequent agreement, whether, for example, by novation or by some compromise relating to the parties’ past dealings or outstanding issues. Nor, in a situation in which concurrent contract and tort claims are possible (see e.g. Henderson v Merrett Syndicates Ltd [1994] UKHL 5; [1995] 2 AC 145), could it be crucial to the application of article 27 whether the foreign claim was being pursued in contract or tort, when the later English claim asserted a settlement agreement wide enough to cover both.

Lord Clarke cites at para 28(iii) a useful encapsulation by Cooke J in JP Morgan Europe Ltd v Primacom AG [2005] 2 Lloyd’s Rep 665, [42], of the meaning of the expression “legal rule” or “rule of law” which the Court of Justice used in The Tatry Case C-406/92, [39]. Cooke J suggested that, in investigating “cause”, it was necessary, after looking at the basic facts, to look at “the basic claimed rights and obligations of the parties”. Here, the basic claimed rights and obligations of the parties are, in Greece, that the English claimants are liable in tort, and, in England, under the first head which asserts that the Greek claims have been settled, that there is no or no further liability for the Greek claims.

The way in which article 27 was applied in The Tatry is also of interest. Having said that the cause of action comprises the facts and the rule of law relied on as the basis of the action [39], the Court of Justice went on:

“40.

Consequently, an action for a declaration of non-liability, such as that brought in the main proceedings in this case by the shipowners, and another action, such as that brought subsequently by the cargo owners on the basis of shipping contracts which are separate but in identical terms, concerning the same cargo transported in bulk and damaged in the same circumstances, have the same cause of action.”

Here, the English claim that the Greek claims fall within the release and have been settled or compromised concerns, and seeks to negative, the same tort claims as the Greek actions seek to enforce.

It can make no difference that the Greek claimants have not sought, pre-emptively, to refer to, address or “impugn” in their Greek claims a possible defence (the Settlement Agreements) that might be raised in the Greek proceedings. One would not expect them to do so, any more than the German claimants in Gubisch v Palumbo addressed or would be expected to address every or any of the multiple arguments that the Italian claimants later deployed. The fact that the English claims do not seek directly to “interfere” with the Greek claims is also irrelevant. It would anyway be impermissible to claim in England an injunction restraining the Greek proceedings, but, quite apart from that, article 27 and the principle in Gubisch v Palumbo do not depend upon one set of proceedings seeking directly to prevent another. They derive from the principle that Member States must recognise each other’s judgments, and the aim of avoiding inconsistent judgments.

As to the “same object”, the end which the Greek and English proceedings have in view is the same in each case, to decide the issue of liability for the torts alleged in Greece. That this is what is meant by the “same object” is clear from both Gubisch v Palumbo and The Tatry. The matter is directly addressed in the latter case in paras 42 to 45:

“42.

The question accordingly arises whether two actions have the same object when the first seeks a declaration that the plaintiff is not liable for damage as claimed by the defendants, while the second, commenced subsequently by those defendants, seeks on the contrary to have the plaintiff in the first action held liable for causing loss and ordered to pay damages.

43.

As to liability, the second action has the same object as the first, since the issue of liability is central to both actions. The fact that the plaintiff's pleadings are couched in negative terms in the first action whereas in the second action they are couched in positive terms by the defendant, who has become plaintiff, does not make the object of the dispute different.

44.

As to damages, the pleas in the second action are the natural consequence of those relating to the finding of liability and thus do not alter the principal object of the action. Furthermore, the fact that a party seeks a declaration that he is not liable for loss implies that he disputes any obligation to pay damages.

45.

In those circumstances, the answer to the fifth question is that, on a proper construction of Article 21 of the Convention, an action seeking to have the defendant held liable for causing loss and ordered to pay damages has the same cause of action and the same object as earlier proceedings brought by that defendant seeking a declaration that he is not liable for that loss.”

The reference in [44] to a party’s claim for a declaration of non-liability implying that it disputes any obligation to pay damages is equally applicable to the present English claims that the Greek tort claims fall within the release or have been settled or compromised. The English claims imply that the Greek claims are disputed.

In short, the issue of liability is central to both the Greek and the English proceedings here, as it was to the Dutch and English proceedings in The Tatry. Not merely the same cause of action but also the same object is involved in the present case, as it was in The Tatry. The two sets of proceedings would, if pursued to judgment, lead to judgments which were legally and directly incompatible. It is therefore necessary under article 27 to consider whether it is the Greek or the English courts which fall in this connection to be regarded as first seised.

The second and third heads are more elusive. Claims for a declaration that the bringing of the Greek claims was a breach, and for damages for the breach, of the release in the Settlement Agreements may on one view be seen as little different from the claims made under the first head. But I have come to the conclusion that this would be wrong. The second and third heads postulate, and for present purposes at least we must accept, that the releases contain some positive continuing promise which the respondents by their Greek claims are now breaching. The terms of the releases were in each case (clause 2 in the case of CMI, clause 3 in the case of LMI) that the respondents would accept underwriters’ due proportion of the relevant payment “in full and final settlement of all and any claims it may have under Policy No. …. against the Underwriters in relation to the loss of Alexandros T”. One must make the assumption, for present purposes, that the Greek tort claims fall within this agreement.

The difficulty is that the agreement was performed, in the sense that there was not merely an accord, but an accord and satisfaction. All policy claims were thus not just agreed to be settled, but they actually were settled, and, if and to the extent that that is the nature of the second and third heads of English release claim, they would not in reality differ from the first head.

The question therefore arises, what if any outstanding promise could there be left to perform which the second and third heads claim to enforce? I have come to the conclusion that the acceptance of the sums paid “in full and final settlement” involves, certainly very arguably, a continuing outstanding promise not further to pursue claims of the nature identified in clauses 2 and 3 respectively.

Even after the settlement, the pursuit of such claims could cause CMI and LMI loss. Most obviously, such loss could consist in the costs of defending the Greek claims. If they let the Greek proceedings go undefended, it could, subject to issues arising from the potential recognition of any Greek judgment under the Brussels Regulation, include the amount of any judgment awarded against them in the Greek proceedings. Likewise potentially, though subject to additional questions arising from any potential issue estoppel or application of the rule in Henderson v Henderson (1843) 3 Hare 100, even if they unsuccessfully defended the Greek claims.

Accordingly, the second and third heads of release claims, analysed as I have analysed them, are outside the scope of article 27. As regards the first head, the remaining issue is whether the Greek or the English courts fall for the relevant purpose to be regarded as first seised. In so far as the first head of release claims was added into the pre-existing English proceedings by an amendment made after the Greek proceedings were begun, is it to be viewed discretely as a new claim of which the English court is second seised? Or does it fall to be viewed as part, by amendment, of a single set of English proceedings commenced well before any Greek proceedings?

I agree with Lord Clarke at para 60 that a court is only seised of claims by or against new parties from the date that those parties are added to the proceedings. In relation to the 2006 proceedings, the English court was only seised of claims against OME once OME was joined to the proceedings on 20 September 2011 and, as against OME therefore, the English courts were only seised of the first head of release claims made by CMI and LMI in 2011. Since the first head of release claims is in my opinion the mirror image of the Greek tort claims, article 27 must, on that basis, apply to preclude the pursuit of the first head of release claims as against OME in England.

The respondents submit that article 27 also applies to preclude the pursuit in the English proceedings of the first head of claim against Starlight, which was party to the English proceedings from their outset. The Court of Appeal accepted this submission. CMI and LMI challenge it. Lord Clarke has in his paras 61 to 71 set out and discussed the respective submissions.

To my mind, the sense of the Regulation as well as the case law and the academic guidance all point in one direction. The chronological priority contemplated by the Regulation cannot be gained, or subverted, by the addition by amendment of a new claim in proceedings otherwise second brought (any more than it can be affected by the addition of new claimants or defendants, as Lord Clarke accepts: para 60). To the authorities under the current Regulation to which Lord Clarke refers, I would only add that similar thinking is to be found under the predecessor provisions of Article 21 and 22 of the Brussels Convention in the decisions at both levels in Grupo Torras SA v Shekh Fahad Al-Sabah [1995] 1 Lloyd’s Rep 374, 418-419 (Mance J) and [1996] 1 Lloyd’s Rep 7, 24 (CA).

It follows that the conclusions I would reach, were all the issues to be finally decided now, would be that:

In relation to the conclusion expressed in para 161ii, we are all in agreement in our conclusions. Any differences in reasoning regarding article 27 are irrelevant, and no reference is necessary.

As to para 161i, Lord Clarke would reach the opposite conclusion to that which I have expressed and he considers in the light of my judgment that a reference is called for, if the English appellants persist in their first head of release claims. With the latter view I agree. The differences between Lord Clarke’s and my reasoning are not, I believe, simple differences regarding the application to facts of clear principles of European law. I might by myself have thought that all the relevant principles of European law were clear, but I certainly do not dissent from the proposition that the differences, being material to our respective conclusions, require a reference. If the appellants wish to persist in, rather than abandon, the first head of release claims, there should accordingly be a reference as Lord Clarke suggests.

Ultimately, therefore, although by different reasoning, Lord Clarke and I arrive at the same conclusions regarding the appropriate disposition of these appeals.