Stribog Ltd v FKI Engineering Ltd

The issues in this appeal are (a) whether the English courts can stay these proceedings, and (b) if so, whether they should stay them.

The ground on which a stay is sought from the English courts is that “related actions” within the meaning of Article 28 in Section 9 (“Lis pendens-related actions”) of Council Regulation (EC) No 44/2001 on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters (the Judgments Regulation) are pending in the courts of two Member States: one action in Germany in the Landgericht Lübeck (the German Action) and the other action in the United Kingdom in the English Commercial Court (the English Action).

Under Article 28 there is discretion to stay a related action that is pending:-

Under Article 27, which was relied on at first instance but is not pursued on appeal, a stay of pending proceedings is mandatory in specified circumstances:-

On 21 May 2010 Burton J dismissed an application for a stay of these proceedings. The application was made by the claimant in the German Action, Stribog Limited, which was the first of the parties to institute proceedings against the other. (Stribog was formerly called DeWind Limited and was referred to as DWL in the judgment of Burton J.)

The English Action, in which Stribog is defendant, was instituted second in time. Burton J held that he had no discretion to grant the stay sought by Stribog, as, in his view, the English Court was the court “first seised” within the meaning of Article 28.1.

On Stribog’s appeal against the decision of Burton J there is some common ground:-

The particular area of dispute in this case arises from FKI’s contention that the two actions were not “related actions” at the date when the English Action was instituted. The German Action and the English Action only became related actions when, subsequently, Stribog introduced a new issue in its German Action.

Chronologically the German Court is the court first seised of legal proceedings between these parties. Stribog relies on that fact as giving the English Court jurisdiction to stay this related action.

FKI’s position is that, after the English Action was instituted, Stribog introduced into the German Action, by a process equivalent to amendment of its case, a new issue. It related to an Assignment Agreement governed by German Law and pleaded by FKI in the English Action. FKI say that Burton J was right to conclude that, in those circumstances, the English court was “first seised” within the meaning of Article 28.1. Burton J held that where a first action, which is not related to a second action when the second action is instituted, is subsequently altered or amended, the court of the second action is the court first seised for the purposes of Article 28.1: see paragraph 39 of his judgment [2010] EWHC 1160 (Comm). Stribog contends that there is no such rule or principle and that the judge’s reasoning is contrary to the language and the purpose of Article 28.1.

Authorities on various aspects of the interpretation and operation of Articles 27 to 30 of the Judgments Regulation were cited to Burton J and to this court. None of them is directly on the point of the interpretation of Article 28.1, read in conjunction with Article 30, and the application of those Articles to the particular facts. The cases cited were Grupo Torras SA & Anor v Sheikh Fahad Mohammed Al-Sabah & Ors [1996] 1 Lloyd’s Law Rep 7; The “Happy Fellow” [1998] 1 Lloyd’s Law Rep 13; Underwriting Members of Lloyd’s Syndicate 980 and others v Sinco SA [2009] 1 All ER (Comm) 272; and Nordea Bank ASA & Ano v Unicredit Corporate Banking SpA & Anor [2011] EWHC (Comm) 30.

By his order dated 21 May 2010 Burton J dismissed Stribog’s stay application. The ground of the application was that a related action (the German Action) between the same parties had already been brought by it in the Landgericht Lübeck, which was accordingly first seised. Burton J disagreed, holding that the English Court was first seised.

He refused permission to appeal, as did Sir Richard Buxton on 22 June 2010. Moore-Bick LJ granted permission to appeal at the hearing of a renewed application on 6 October 2010.

In July 2005 the claimant FKI Engineering Limited, a UK company wholly owned by FKI Limited and involved in the wind power generation business, sold its shares in a German company, DeWind Gmbh, to an English company now called DeWind Holdings.

In August 2005 DeWind Gmbh agreed to sell its business assets to Stribog. The sum of 33m euros is still outstanding as the purchase price due under a Business Transfer Agreement (BTA) between those parties dated 1 August 2005. It is expressly governed by English law and contains a clause which, in an eccentric excess of particularity, names Milton Keynes as the exclusive place of jurisdiction.

On 29 August 2008 insolvency proceedings in relation to DeWind Gmbh began in the Lübeck local court. On 9 October 2008 Dr Klaus Pannen was appointed as insolvency administrator. On 31 August 2009 the administrator purported to assign, by an Assignment Agreement to FKI, various claims by DeWind Gmbh against various parties. Claims against Stribog were included. Clause 6 provided that the Assignment Agreement was to be governed and construed in accordance with German law and that the sole venue for disputes should be Hamburg.

The question of the validity of the Assignment Agreement was not raised by Stribog in the German Action instituted in Lübeck on 18 September 2009. Stribog’s claim was for a declaration of non-liability to FKI for any claims or rights, with the exception of any potential purchase rights or claims plus interest under the BTA. The Particulars of Claim in the German Action referred to FKI’s wrongful assertions that they had rights against Stribog. On 9 October Stribog submitted to the German Court a statement making it clear that purchase price claims on the part of FKI arising from the BTA were not included. Reference was made to the Assignment Agreement to FKI by the insolvency administrator of the alleged claims of DeWind Gmbh.

On 21 January 2010, FKI started the English Action in the Commercial Court against Stribog. It was said to be brought by FKI as assignees of DeWind Gmbh pursuant to the Assignment Agreement for Stribog’s breach of the BTA and seeking payment of the balance of the purchase price under the BTA; alternatively, damages for its breach and interest.

At that point Stribog had not raised any issue in the German Action about the Assignment Agreement being void under German Law. (It has yet to serve its defence in the English Action. FKI have served their defence in the German Action). It is said by FKI that Stribog first raised the issue of invalidity in a further voluntary statement sent by its Hamburg lawyers to the court in Lübeck on 12 February 2010. The statement was by way of further information rather than by a formal modification of the proceedings, though equivalent to an amendment in that it raised an issue. It asserted that FKI are not entitled to any rights or claims against Stribog arising from the Assignment Agreement, because it was void under German law. It failed to satisfy the German principle of specification, which requires assigned claims to be specified or specifiable. If Stribog wins on that issue, the English Action will fail and the German Action will succeed.

On 15 February 2010 Stribog issued its application for the English Action to be stayed either under Article 27.1, which is no longer pursued, or under Article 28. 1, to which I now turn.

Section 9 of the Judgments Regulation tackles the problem of pending actions in more than one Member State and the risk of conflicting judgments. Should the actions be allowed to proceed in each Member State? If not, which courts should have jurisdiction to stay?

The scheme of Section 9 has three notable features.

The first is the concept of a court being “deemed to be seised” of proceedings as defined in Article 30. The provisions of Articles 27 and 28 are only engaged if the courts of different Members States are “deemed to be seised” of proceedings. That is determined in accordance with Article 30, which defines the circumstances in which a court shall be “deemed to be seised.” The definition enables the court to decide when it is deemed to be seised of proceedings. Article 30 does not use the expression “first seised.” That is found in the other Articles in the Section stating the grounds on which proceedings must or may be stayed.

The circumstances in which a court is “deemed to be seised” of proceedings enable it to determine the order in which the courts in different Member States are “deemed to be seised” of their proceedings and therefore which of them was “first” seised. As is obvious, if no court is deemed to be seised of an action, or if the court in only one Member State is “deemed to be seised”, the scheme in Section 9 is not engaged. There is no need for a scheme of priority of actions, unless there are proceedings pending in the courts of two Member States competing for first place. (See, for example, the recent case of SK Slavia Praha-Fotbal v Debt Collect London Limited [2010] EWCA Civ 1250 in which, as matters turned out, the courts in only one Member State, namely the United Kingdom, were “deemed to be seised” within the meaning of Article 30 at the material date, the plaintiff in proceedings in Prague having failed, before the English proceedings were instituted, to take a step, namely payment of a court fee, required for service of the Prague proceedings on the English defendant.)

The second feature of Section 9 is the mandatory nature of a stay where proceedings brought in the courts of different Member States involve the same cause of action and are between the same parties. Under Article 27 the court not first seised must stay its proceedings of its own motion and it must decline jurisdiction in favour of the court first seised where its jurisdiction is established. Although Article 27 is no longer a live issue in this case, it is relevant to an understanding of the scheme of Section 9.

The third feature is the discretionary nature of stay provisions where “related actions” are pending in the courts of different Member States. Under Article 28.1 the court not first seised may stay its proceedings. The court first seised cannot stay its proceedings. Article 28 provides that actions are deemed to be related actions.

Thus, on this application for a discretionary stay, there is both a competition and a comparison. The competition is to decide which of the courts “deemed to be seised” was “first seised.” The result of that competition determines which court may stay its related proceedings. Logically the competition for jurisdiction must be decided before the court looks to see whether there are grounds for a stay. A comparison of the German Action and the English Action is then necessary to decide whether there are grounds for a stay on the basis that they are related actions in the defined sense carrying a risk of irreconcilable judgments.

With those general points in mind I turn to the judgment under appeal.

Burton J concluded that the English court was first seised and that it has no discretion to stay the English Action based on the BTA which contains an exclusive English jurisdiction clause and is expressly excluded from the claim by Stribog in the German Action.

As Burton J saw it, the key point was that the validity of the Assignment Agreement, which is governed by German Law, was not raised by Stribog as an issue in the German Action until after the English Court was already deemed to be seised of the English Action in which the Assignment Agreement was pleaded. Its alleged invalidity only became an issue in the German Action when it was introduced into it after the English Action was instituted. Only at that point did the actions in the two Member States become “related actions” as defined in Article 28.

The judge rejected Stribog’s submission that the alleged invalidity of the Assignment Agreement, though not actually raised before 15 February 2010, was embryonic, inherent or intrinsic in the German Action from its inception. The nub of the judge’s reasoning is summarised in the following passage of his judgment :-

Mr Isaacs, appearing for Stribog, submits that the judge was wrong to hold that the amendment of the German Action after the institution of the English Action made the English Court first seised. The judge erred in not applying a two stage process required by Article 28.

The first stage is to identify whether an Article 28 situation exists at all by asking whether related actions are pending in different Member States. If so, Article 28 applies and there is jurisdiction to grant a stay.

The second stage is to identify the court that has jurisdiction to grant a stay. That takes you to Article 30 to determine when the courts in the two Member States were “deemed to be seised” and in what order. The jurisdiction to stay is that of the court that is not first seised. The judge did not apply this step. Instead, he identified the court in which the Assignment Agreement was first pleaded. That was the English Court.

Mr Isaacs submits that Burton J’s approach involved an incorrect application of Article 28. The Article is concerned with being deemed to be seised of actions rather than of issues. As Saville LJ said in The Happy Fellow at pp17-18:-

Stribog’s alternative submission is that, even if the issue approach adopted by the judge is correct, the German court was first seised, because the invalidity of the Assignment Agreement was in fact first raised in the German Action.

On either approach, the English Court has jurisdiction to grant a stay, which should be granted to avoid the risk of irreconcilable judgments in different Member States.

In the absence of authority directly on the novel point raised by this case, it is understandable that other authorities on Section 9 are cited, especially when they relate to the effect of amendments to proceedings. At the same time one should be cautious about placing too much weight on passages in authorities which, while in the same area of law and couched in general terms, are not directly in point. Indeed, though not doubting their correctness, I did not find them helpful in unravelling a different case. I prefer to go first to the overall context and scheme of Section 9, to read closely the language of Articles 28 and 30 and to apply the provisions step-by-step to the facts of this case.

The first step is the deemed seisin of proceedings by courts in two Member States. The court from which a stay is sought must look at the position in its courts and in the courts of the other Member State concerned and decide whether they are deemed to be seised of an action, not, as was observed in The Happy Fellow, seised of a particular issue in an action. The German Court is deemed to be seised of the German Action. The English Court is deemed to be seised of the English Action.

The second step is chronology. The court must determine the date when the courts of each Member State were deemed to be seised of the action instituted in them. The German courts were deemed to be seised of the German Action on 18 September 2009. The English courts were deemed to be seised of the English Action on 21 January 2010.

The third step is competitive. The German courts were deemed to be seised on 18 September 2009, which was before 21 January 2010 when the English courts were seised of the English Action. Purely in terms of chronology the German Court was first seised and the English courts were not therefore first seised.

The fourth step is comparative. The English courts, as the courts not first seised, have to compare the proceedings in the two Member States to see whether they are related actions. If they are not related actions, the question of a stay does not arise. There is no risk of irreconcilable judgments. In making the comparison the court looks at all the circumstances, not just at the circumstances at the date of the institution of the English Action, in order to see how closely related the actions are and whether there is a risk of irreconcilable judgments in separate proceedings. In this case it is agreed that the actions were related at the time of the issue and hearing of the application. There is a risk of irreconcilable judgments in these actions if the English courts decide that the Assignment Agreement is valid and the German Courts decide that it is void (or vice versa). There is therefore a discretion in the English courts to stay the English Action

The fifth and final step is whether the discretion should be exercised to grant a stay. That turns on the circumstances at the date of the hearing and is dealt with below.

Authorities were cited by Mr Mark Templeman QC appearing for FKI on the effect of the amendment of pleadings in the case of pending proceedings covered by Article 27 and its predecessor Article 21 of the Brussels Convention. Contrary to the view of the judge, who relied on them in paragraph 39 of his judgment cited earlier, I do not think that they are directly in point. They can be distinguished.

As Mr Isaacs pointed out, the analogy with Article 27, under which a stay is mandatory, and the cases on it are not sound. Article 27 looks back to the outset of the proceedings, to the position when “proceedings involving the same cause of action and between the same parties are brought [my emphasis] in the courts of different Member States.” A comparison of the documents is made to see whether the subject matter of the claims is the same.

Article 28, under which a stay is discretionary, looks to the current position and to matters that post-date the onset of the proceedings to consider the relationship between actions that are pending to see if they are related and if there is a risk of irreconcilable judgments. Actions are compared to see if they are related rather than claims compared to see if they are the same. See Research in Motion UK Ltd v. Visto Corporation [2008] EWCA Civ 153 [2008] 2 All ER (Comm) 560 at [35] [36] and [40].

I agree with Mr Isaacs that the decision in Grupo Torras is irrelevant. In the context of the addition of new parties, the court rejected the application of the doctrine of relation back of an amendment. It was a decision on Article 21 of the Convention, which was the predecessor of Article 27. Further, the Convention did not have the uniform interpretation of the date on which the court is deemed to be seised now imposed by Article 30.

I also agree that Sinco is irrelevant. An amendment was made to add a new claim in an action seised first in point of time. The amendment was subsequent to the seisin of the second court and, as a matter of domestic law, would normally relate back to the commencement of the first action. Beatson J said obiter at paragraphs 60-68 that the relation back should be disregarded in determining which court was first seised of the proceedings for the purpose of Article 27.

Mr Templeman contends that the English Action should not be stayed, as Stribog had expressly excluded the subject matter of the English Action from the German Action and it remains excluded, as the German Action has not been amended. Further, the claim in the English Action is within the exclusive jurisdiction of the English courts and outwith the jurisdiction of the German courts. The degree of connection between the English Action and the German Action is small and there will be a real risk of delay if the English Action is stayed to await the result of the German Action.

In my judgment, the discretion to stay the English Action should have been exercised by the judge, so that the English Action is stayed until the German courts decide whether the Assignment Agreement is valid or void.

I do not see how the discretion could reasonably be exercised in any other way. That issue is central to both actions. A determination in favour of Stribog would dispose of both actions. The issue is one of German law, which should be tried in Germany where the proceedings are more advanced. Unless there is a stay, costs will be wasted by duplication of proceedings with a risk of conflicting judgments.

I would allow the appeal. The German court was first seised within the meaning of Articles 28 and 30: the German Action was commenced before the English Action. They were related actions within the meaning of Article 28 at the date of the application for stay and at the date of the hearing of the application when the decision had to be made. The fact that they were not related actions when the English Action was commenced does not affect the chronology of deemed seisin within the meaning of Article 30, or the question of which court was first seised within the meaning of Article 28, or the existence or exercise of the discretion.

The English Action may be stayed by the English Court, as it was not the court first seised. It is proper to exercise the discretion to stay it until the issue of the validity of the Assignment Agreement and the assignments thereunder is decided by the German Courts.

I would add that the result does not seem to me to be unjust, inconvenient or surprising. The validity of the Assignment Agreement will be decided by the German Courts in accordance with German Law by which, as the parties agree, it is governed.

The question raised by this appeal is how one can tell which of two related actions is first seised for the purposes of article 28 of the Judgments Regulation. It will be recalled that article 28 provides:

The appellant, Stribog Ltd (formerly DW Limited), submits that the court merely applies to the related actions the mechanical test as to seisin found in article 30, viz –

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 steps he was required to take to have service effected on the defendant, or

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.”

In other words, to gloss these provisions broadly, a court is deemed to be seised when its jurisdiction is invoked: either by the lodging with the court of the document instituting the proceedings (in England, the issue by the court of the claim form), or, in jurisdictions where service may precede the lodging of such a document with the court, by the receipt of that document by the authority responsible for service; and in both cases subject to the proviso that the claimant does not subsequently fail to do what he has to do to complete the double requirement of issue and service. Thus where issue precedes service, service must in due course be effected; and where service precedes issue, the claim form must in due course be lodged with the court.

It is possible that these provisions contain their own complexities, but we are not concerned with any such complexities here. It is common ground that (i) Stribog commenced its action in Germany against FKI Limited and its subsidiary FKI Engineering Limited (“FKI”) on 18 September 2009; (ii) FKI commenced its action in England against Stribog on 21 January 2010; and (iii) each court was seised of the respective action on those dates. It is also common ground that these actions are now (ie as at the time of Stribog’s application in the English action for a stay pursuant to article 28) pending, related actions.

Stribog therefore submits that the only question is which court, for the purposes of the article 30 test of seisin, was seised first: the German court, or the English court? The answer, Stribog submits, is given by combining the test of seisin in article 30 with the doctrine of chronological precedence inherent in article 28’s concept of first seisin (to be derived from its language of “the court first seised”). On this basis, the answer must be that Germany is the court first seised, and that it is only the English court, as the court second seised (“any court other than the court first seised”), that has the power to stay its proceedings.

FKI, however, submits that such an approach is overly mechanical, and also unpredictable, not clear and effective, and encourages forum-shopping, and was not what was intended by the Judgments Regulation. One must not ask: Which court was first seised of one or other of the actions, now recognised as related actions? Rather, what is intended and required is to discover which court was first seised of an action which at that time of seisin may be recognised as a related, or at any rate potentially related, action. FKI therefore in effect submits that what has to be sought and identified is not merely “the court first seised” but the “first related action”. This is because, so it is submitted, article 28 is speaking of the court first seised of a related action.

The issue is engendered by some special facts. When Stribog commenced its action in Germany it claimed against FKI a declaration of non-liability in relation to possible claims which FKI might make against Stribog arising from an assignment to FKI of a variety of claims which DeWind GmbH might be able to assert against Stribog. DeWind GmbH entered insolvency proceedings, and its administrator, Dr Pannen, had, by an assignment agreement dated 31 August 2009, assigned to FKI all such possible claims by DeWind GmbH against Stribog. That was an assignment governed by German law, concerning the assets in the form of claims of a German company in German insolvency proceedings. FKI was interested in such claims, as I understand the matter, to protect its position under an earlier agreement it had made for the sale of shares in DeWind GmbH to a company within the Stribog group. FKI had become concerned that the assets of DeWind GmbH had been stripped out of it by its new owners and passed to or through Stribog. The assigned claims arose under various provisions of German insolvency law and tort law, and by reference to claims for restitution of assets and intellectual property rights. One such claim which DeWind GmbH had against Stribog arose under the agreement itself by which DeWind GmbH had sold its assets to Stribog. This was achieved by a Business Transfer Agreement (the “BTA”) dated 1 August 2005. The sum of €33 million was still outstanding under the BTA when DeWind GmbH became insolvent. The BTA was governed by English law and contained an English jurisdiction clause.

FKI’s German action was precipitated by US proceedings which FKI commenced in California on 2 September 2009, as assignees of DeWind GmbH, against various members of the Stribog group. These proceedings followed a letter before action written by US attorneys in which FKI gave notice to the ultimate holding company of the Stribog group of the potential claims which might be brought arising out of the insolvency proceedings.

The essential matters about the German proceedings as originally pleaded to be noted for present purposes are as follows. (1) It was a general claim for a declaration of non-liability with respect to any claims or rights asserted against Stribog by FKI. (2) It stated that FKI wrongly asserted claims or rights against Stribog deriving from DeWind GmbH by way of assignment. The assignment dated 31 August 2009 was specifically referred to. (3) However, it made an exception of any purchase price claims under the BTA. Its original pleadings six times referred to this exception. They stated: “with the exception of any potential purchase price claims…under the [BTA]”; “Purchase price claims of [FKI] that might exist under the [BTA] are explicitly not the subject matter of this Action”; “As already explained, possible purchase price claims of [FKI] under the [BTA] are no subject matter of the action”; “For any other claims that might result from the [BTA] the parties agreed on “Milton Keynes” as being the exclusive place of jurisdiction. Any such claims are therefore not subject matter of the present action for a declaratory judgment”; “Insofar as the allegation is raised there [in the US letter before action] that [FKI] had claims against [Stribog] which did not result from the [BTA], such allegation is without any factual or legal basis”; and “The claims asserted by [FKI] do not exist. [FKI] have no rights of avoidance nor any other claims, with the exception of potential price claims under the [BTA] that do not need to be decided in these proceedings”. It is surmised that this exception was made either because of the English jurisdiction clause in the BTA and/or because the inclusion of a claim for non-liability for the outstanding purchase price of €33 million would have added very significantly to the fee payable with respect to the action.

FKI observes that there is nothing in these pleadings which expressly throws doubt on the validity of Dr Pannen’s assignment; and submits that on the contrary it is a necessary inference of the pleadings that the assignment is recognised as valid. Stribog responds to the effect that there was no express acceptance of the validity of the assignment, and that in any event at that time Stribog did not as yet have a copy of it.

In the English action FKI claimed as assignee under Dr Pannen’s assignment of DeWind GmbH’s rights under the BTA to payment of an outstanding purchase price of €33 million. By letter dated 15 January 2010, ie in the week before the English action was commenced, Stribog’s solicitors confirmed that they were authorised to accept service but continued:

As of the time of the current proceedings for a stay of the English action, Stribog had not yet filed a defence in the English action, but, in the light of the development mentioned immediately below in the German action, it could be predicted that one at least of its defences would be that the assignment of DeWind GmbH’s claim under the BTA is void and thus ineffective.

That is because on 12 February 2010 Stribog served a voluntary statement which alleged, for the first time, that the assignment agreement made by Dr Pannen was void because it did not satisfy the “requirements of the German principle of specification for assignments”. By that time Stribog had obtained a copy of the assignment agreement.

As recorded by Burton J, it is common ground (on the expert reports of German law filed by the parties) that the February statement did not constitute an actual amendment to Stribog’s pleadings: in the sense that no permission to amend was sought. Stribog submitted that it was simply the spelling out of what had already been instrinsic within the German proceedings. FKI, however, submitted that it was a de facto amendment. I would be prepared to regard the February statement as an amendment, but I am nevertheless satisfied that, on the common accord of the parties’ experts, it was in any event not such an amendment as to raise a new cause of action or constitute the German claim a new claim. At most it introduced, or made explicit, a new issue, namely as to the validity of the assignment. As such that issue affected all the possible claims which FKI might be able to make as assignee of DeWind GmbH. If the assertion of invalidity of the assignment were to succeed, it would follow that FKI could not succeed in its claim as assignee under the BTA, or in any of its other potential claims. However, there was no move in the February statement to extend the German proceedings to cover a declaration of non-liability in respect of a claim under the BTA. That matter remained an exception to the subject matter of the German action.

It was argued before the judge on behalf of Stribog that, even before the February statement, the potential issue of the invalidity of the assignment was embryonic or intrinsic within the German action. However, the judge rejected that submission. He said (at paras 26(vii) and 27 of his judgment):

Stribog issued its application to stay the English action on 15 February 2010. It is common ground that as at that time the English and the German actions were related actions. What divides the parties, however, is when, and the significance of when, they became related actions. FKI submits that as at 21 January 2010, when the English action was commenced, the two actions were not related. They were not related because as at that time the potential claim under the BTA was expressly excluded from the subject matter of the German action. It was only by reason of the February statement, submits FKI, that the two actions became related. Therefore it was the English court that was the court “first seised” of the related actions. Stribog submits that this is the wrong answer to the wrong question. The question is not which court first became seised of the issue which is common to the related actions, but which court first became seised of one of the actions which, when a stay is requested, are pending and related.

What is a related action? The answer is given in article 28 (3):

The judge, Burton J, agreed with FKI. Therefore, the English court was the court first seised and had no power to stay the English action. Hence Stribog lost its application to stay, and appeals to this court. The judge expressed the kernel of his decision in these words (at para 39 of his judgment):

The judge’s principle is expressed in the first sentence cited. He then appears to state that the facts of this case presents an a fortiori example of that principle.

The facts are indeed special to the case, but such twists and turns tend to be rather typical of transnational litigation.

The relevant parts of article 28 have been cited above. However, article 27 also needs to be addressed in order to take account of the argument relied on by FKI and accepted by the judge.

Article 27 provides:

Thus, both articles 27 and 28 invoke the concept of “the court first seised”. However, article 27 requires the same cause of action and the same parties, whereas article 28 only requires the actions to be “related” in the sense defined in article 28(3). Article 27 arises where more than one action “are brought”, and article 28 arises where more than one action “are pending”, in the courts of different member states. Article 27 allows no discretion: any court other than the court first seised must stay its proceedings and ultimately decline jurisdiction. Article 28 on the other hand allows a discretion, but only to any court other than the court first seised, to stay its proceedings. The common principle is that it is the court first seised which takes precedence, and any other court seised subsequently to that court either must or may stay its proceedings. Article 27, because it requires the “same cause of action” requires an analysis of the cause of action brought in each action. Article 28, because it requires the actions to be “related”, requires some general and less precise analysis to determine whether there is a risk of irreconcilable judgments.

In this connection it is also necessary to cite two of the paragraphs from the recitals to the regulation, viz (11) and (15):

Recital (15) is clearly directed to articles 27-30 of the Regulation, even if recital (11) is of more general application. Both parties accept that these paragraphs express purposes of the Regulation to enact “highly predictable” rules and to minimise the possibility of concurrent proceedings. As the European Court of Justice repeated in Francesco Benincasa v. Dentalkit Srl [1997] 1 L Pr 559:

For all that, the parties differ, however, as to where such principles take the issue which divides them in this case.

It is disputed whether there are any authorities placed before us which bear directly on the issue in this case. However, there have been authorities which discuss what is said to be the connected issue of how one identifies the court first seised of two actions involving “the same cause of action and between the same parties” within article 27. It is established that the “same cause of action” involves the double concept of “cause” and “objet”. It is no longer suggested that the German and the English actions come within this concept.

The essence of those cases is 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.

There are also authorities to which we have been referred which discuss the concept of “related actions”. Both parties claim to obtain support from them.

We have been referred to the following authorities, either directly or, by their citation in other judgments to which we have been referred, indirectly.

In The Maciej Rataj [1995] 1 Lloyd’s Rep 302, the European Court of Justice held that article 21 of the Brussels Convention (the predecessor of article 27) operated on a party by party basis. Thus where the same cause of action was involved in separate proceedings brought in different member states but not all the parties in one action were also involved in the other action, the action in the court second seised had to be stayed only to the extent that the same parties were involved. That action could continue for or against other parties not involved in the first action. Although that might lead to the fragmentation of proceedings, nevertheless article 22 (now article 28) –

As I ventured to say in Glencore International AG v. Shell International Trading & Shipping Co Ltd [1999] 2 Lloyd’s Rep 692 at 697:

In Grupo Torras SA v. Sheikh Fahad Mohammed Al-Sabah [1996] 1 Lloyd’s Rep 7 (CA) the English litigation arose out of proceedings brought against a number of individuals who it was said had defrauded the Kuwait Investment Office (KIO) in one or other of its manifestations. The primary claimant, Grupo Torras, a Spanish company, was indirectly controlled by KIO. It had commenced its proceedings in England, for the purpose of the Brussels Convention, by service on the defendants (Dresser UK Ltd v. Falcongate Freight Management Ltd [1992] QB 502 (CA)). At that time, under the Brussels Convention, there was no equivalent to article 30. The concept of seisin had been developed in jurisprudence, did not have an autonomous meaning, but was regarded as essentially depending on service, not issue (see Zelger v. Salinitri (Case 129/83) [1984] ECR 2397 for the ECJ jurisprudence, and Dresser for the equivalent rule in England). The relevant defendants to the English action were served between 15 April and 21 June 1993. There were almost contemporaneous proceedings in Spain between some of the English parties. There was a dispute as to whether the Spanish court or the English court was the court first seised. The proceedings had been commenced in Spain against Grupo Torras in June 1992 but it was not until May or even October 1993 that those proceedings had been formally amended to introduce as claimants in Spain the relevant defendants in England (the so-called amended “Quail action” in the Spanish proceedings). The court of appeal held that the Zelger v. Salinitri test meant that the Spanish court was not seised of the amended case against Grupo Torras until October 2003. Therefore the English court was the first court seised. It was irrelevant that, upon formal amendment, that amendment related back to the time of the original commencement of the Spanish action (at page 24). It was said that that was for the purposes of both articles 21 and 22 (articles 27 and 28): ibid left hand column, although the decision appears to have been that “The English Court was the Court first seised under art. 21” (right hand column). That was a decision on what constituted seisin, but one can readily see that if article 21 is subject to the rule in The Maciej Rataj, then an article 21 situation could not arise until the relevant same party was served for the purposes of a second set of proceedings involving the same cause of action.

At first instance, in Grupo Torras [1995] 1 Lloyd’s Rep 374 at 418 Mance J said this:

Question (a) is for present purposes the article 28(3) question: are the German and English actions related? That question has to be settled at the time of the application (or, as Mance J states, the time of judgment). Question (b) is now the article 30 question: at any rate when applied to article 27 (21), what is being asked is when an action involving the same parties and the same cause of action was seised by either court. But how does this apply to article 28 and related actions? Mance J continued:

In “The Happy Fellow” [1997] 1 Lloyd’s Rep 130, Longmore J was concerned with litigation which arose out of a collision between two vessels in the mouth of the Seine. The first action was commenced in Le Havre by at latest 22 December 1995 by those interested in The Happy Fellow, against the owners of the Darfur. On 14 March 1996 the owners of the Darfur commenced a limitation action in England against the claimants in Le Havre and others. On 3 May 1996 one of the French claimants sought to stay the English proceedings under articles 21 and/or 22 on the basis that the Le Havre court was the court first seised. Longmore J held that article 21 was not involved, but that the two actions were related for the purposes of article 22, and he therefore stayed the English action, since the English court was the court second seised, so far as the single applicant was concerned. Nevertheless, the English claimants had submitted that England was the court first seised, on the basis that the English court was the court first seised of the issue of limitation (the claim under international convention whereby a shipowner of a vessel liable for more than its total value to one or more claimants may require those claimants to share rateably up to the limit of a limitation fund quantified under that convention). Longmore J considered that the French court would regard itself as seised of limitation issues in the Le Havre action as from its commencement: therefore the French court was the court first seised even on the premise of the English claimants’ argument that it was necessary to ask which was the court first seised of a particular issue. However Longmore J continued (at 136, right hand column):

That, it seems to me, is a decision of a judge of the Commercial Court, Longmore J as he then was, on the issue which is before us. Even if it was the English action which had first specifically raised the issue of limitation, nevertheless as at the time when the English court had been requested to consider an application to stay under article 22 there were two related actions before it and the simple fact of the matter was that the French court was the court first seised of one of those actions. The reasoning is very brief and does not bind this court, but the decision is there.

The “Happy Fellow” went on appeal to this court: [1998] 1 Lloyd’s Rep 13 (CA). The principal arguments in this court were as to the separate issue as to whether the French and English actions were related at all (whereas in the present case it is common ground that the German and English actions are related). However, in that context the English claimants relied on the fact that, since the decision before Longmore J, they had admitted liability in the Le Havre proceedings: they therefore submitted that there was now no risk of irreconcilable judgments. This court, in the judgment of Saville LJ, rejected that submission. He said (at 17):

Although that issue arose in somewhat different circumstances, it seems to me that Saville LJ’s reasoning is entirely consistent with Longmore J’s view that the question of first seisin must be judged as at the time when article 22 is invoked. Saville LJ is saying, similarly, that the question of related actions must be judged as at the time when article 22 is invoked (“on the basis of the material put before the Court first concerned with this question”).

As for first seisin itself, Saville LJ did not deal separately with Longmore J’s second and wider reason for regarding the Le Havre court as the court first seised, but simply founded himself on Longmore J’s first reason, by asking himself whether the limitation issue was inherent in the Le Havre action (even if not explicit). He also regarded this as being reasoning concerned with whether the actions were related, as well as with the question of first seisin. Thus he reasoned (at 17/18):

In my judgment, this reasoning is strongly supportive of Stribog and its appeal. Saville LJ reasons in effect that (i) the question of “related actions” raised by article 22 is a question raised about actions not issues; (ii) the question of first seisin is therefore concerned with actions not issues; (iii) an action which does not expressly raise an issue may still be a related action if a common issue constituting the actions related for the purposes of article 27(3) exists inherently within it; and (iv) a related action may be an action of which the court seised of it is the “court first seised” even though it is not that action, but another action, which first explicitly raises a particular issue (where that issue is capable of being inherent in the former action).

Lord Saville of Newdigate, as he had now become, revisited these fields in Sario v. Kuwait Investment Authority [1999] 1 AC 32. The issue there was confined to disagreement as to whether a first Spanish action and a subsequent English action were related for the purposes of article 22. It was common ground that if the actions were related, it was the Spanish court which had first seisin. Therefore the issue in the present case was not there involved. Nevertheless, Lord Saville was dealing with an argument that the test of related actions was a comparatively narrow one. The court of appeal had restricted the concept to “primary” issues limited to facts necessary to establish a cause of action. In emphasising the concept’s width, by reference to The Maciej Rataj, Lord Saville said this (at 41F):

In the present case, the argument which proceeds from a combining of the test of related actions and the test of first seisin so as to ask the compendious question of which court first became seised of an action which was at that time related to another action (ex hypothesi of which another court was not yet seised) to my mind flies in the face of the simple test of related actions sought by Lord Saville. It also flies in the face of his anxiety to avoid satellite litigation. Lord Saville’s approach essentially involves simply looking at the two actions in the round, and not by reference to any refined analysis of causes of action, and asking whether they are related by reference to the concept of the risk of conflicting judgments.

Research in Motion UK Ltd v. Visto Corporation [2008] EWCA Civ 153, [2008] 2 All ER (Comm) 560 concerned the issue of whether actions were related, and in that context an issue of first seisin. The English claimant RIM commenced an action in England for a declaration of non-infringement of the English defendant’s UK patent (“action 1”). It next commenced proceedings in Italy, seeking revocation of the defendant’s corresponding Italian patent (“action 2”). The defendant in England then served a counterclaim which included a claim for infringement of the English patent and a claim that RIM was abusing both English and Italian law in its claims (“action 3”). At that stage the English judge declined jurisdiction in respect of the defendant’s counterclaim of abuse of Italian law or at any rate stayed that aspect of the counterclaim. That appears to be a reference to the powers of a court second seised pursuant to art 28(1) and (2). RIM then discontinued its claim in action 1, leaving only action 2 and what remained in action 3, viz the counterclaim for infringement and for abuse of English law. It was common ground that action 2 was the first in time of those actions. However, the defendant now appealed against the order of the judge. It did so on the basis that at that earlier time the English court seised of action 1 had been the court first seised as between actions 1 and 2 and thus not in the position to decline jurisdiction in respect of or to stay an aspect of the counterclaim pursuant to the Regulation.

In the court of appeal the argument appears to have been confined to article 28 and the question of related actions. This court did not consider that the English and Italian actions were related and therefore dismissed the appeal (at [37]-[39]). In doing so, it stressed that the article 28(3) question of related actions was not entirely mechanical but required assessment. It would seem to follow that the judge’s decision survived simply as a matter of his exercise of a domestic or common law discretion. This court plainly thought that that was a sensible result: see at [33] where it is regarded as absurd that the English court should decide whether there had been abuse of process, involving bad faith or gross negligence, in Italy.

In these circumstances, it is not at all clear what light Research in Motion can throw on the issue in this case. Nevertheless, Mr Barry Isaacs on behalf of Stribog relied on it for the submission that an issue may be inherent in a related action even before it is made explicit in it. For the purposes of this submission the judge cited much of the following passage from the judgment of the court delivered by Mummery LJ:

In my judgment, Research in Motion is of limited use in the present context, because this court’s conclusion there was that the actions were not related and therefore article 28 did not apply. However, three matters are nevertheless of some assistance. The first is the recognition by this court that the test of first seisin is mechanical, even if the test of related actions is not (see para 37). The second is that, because it is appropriate to look at defences as well as claims for the purpose of the related actions test, the nature of an action’s relationship with another action may develop over time. The third is that, despite that potential for development, there is no suggestion that the mechanical and chronological test of first seisin has to be applied in a staged way, to each action as it develops, as though each development was a separate action for the purposes of seisin and thus of first seisin.

The next case is one to which the judge paid particular attention, although it involved article 27 rather than article 28, and, to the extent that it raised an issue of first seisin, proceeded on an obiter basis since it was held that article 27 did not apply. In Underwriting Members of Lloyd’s Syndicate 980 v. Sinco SA [2008] EWHC 1842 (Comm), [2009] 1 All ER (Comm) 272 (“Sinco”), Beatson J was concerned with an English action which was issued first (on 15 January 2007) albeit served second (in June 2007), by which time a Greek action had been both issued and served by the English defendant on the English claimants (in April 2007). In the meantime the still unserved English claim form had been amended on 21 June 2007 in response to the Greek proceedings, in order to introduce a new claim (the “jurisdiction claim”) for damages for breach of the exclusive jurisdiction clause contained in the parties’ contract, and it was that amended claim form which was then served in Greece. That amendment could be made without permission, because the claim form was still unserved. For the rest, in England the claimant underwriters brought claims against Greek insurance brokers in respect of their operation of insurance binder agreements; while in Greece the brokers also made claims arising out of the operation, and for unlawful termination, of the binders. The article 27 issue, however, related only to the jurisdiction claim, and the brokers’ application to stay that claim failed on the ground that the basic requirement of “same cause of action” was missing.

Beatson J then turned (at [55]) to article 30 and the subject of first seisin. It will be borne in mind that the jurisdiction claim was an entirely new cause of action; and that it arose out of circumstances (the Greek action) which were not yet in being at the time of the issue of the unamended claim form: for that went back to January 2007, whereas the Greek action had been brought only in April 2007. Beatson J opined:

Counsel for the English claimant (Mr Phillips) had submitted that the action should not be broken into parts for the purposes of individual claims; that the English court was therefore the court first seised; and that for that reason as well as the absence of the same cause of action, the English court had no power to stay the jurisdiction claim. In support of that submission, Mr Phillips contended that such a rule would discourage forum shopping. After some discussion of the problem, Beatson J concluded as follows:

That, it seems to me, is a somewhat diffident view that on the facts of that case the English court was not the court first seised of the jurisdiction claim. It may be possible to agree, in the case of a new cause of action which could not even have existed at the time of original issue of the English action, and which was the sole claim in issue for the purposes of article 27, that that cause of action was not involved in proceedings which had been “brought” and of which the English court had seisin until 21 June 2007. However, that seems to me to have very little if anything to do with the raising of a new defence to liability (by reference to an issue as to the validity of the assignment) in proceedings for a declaration of non-liability, where ex hypothesi every potentially available defence is open to such a claimant, where the new defence may raise a new issue but does not raise a new cause of action, and where the question is not whether a court can be said to have seisin at all of proceedings “involving the same cause of action” even before that cause of action has been pleaded, but is which court has first seisin of “related actions…pending in the courts of different Member States”.

In Nordea Bank Norge ASA v. Unicredit Corporate Banking SpA [2011] EWHC 30 (Comm) Gloster J expressly adopted Burton J’s approach in this case, which she regarded as having also been supported by Mance J in Grupo Torras, by Beatson J in Sinco, and by this court in Research in Motion (see at her paras 72ff). Thus she put the matter in this way:

In Nordea Bank, there was a dispute between a vessel’s charterer, Maritima Fluviale (“MF”) and her owner, Vasonia Shipping (“Vasonia”) concerning alleged non-payment of hire under their charter. MF had provided Vasonia with a performance guarantee under which a bank (“Unicredit”) guaranteed MF’s hire payments. The charter dispute was referred to arbitration in London. The first court proceedings (the “Initial Proceedings”) were commenced by MF, in Italy, on 2 November 2009, against Unicredit for an interim injunction to prevent Unicredit paying out under the guarantee pending the resolution of the arbitration dispute. Vasonia was not added as a defendant to this action until 18 December 2009. Even so, the Initial Proceedings did not involve any substantive issue or purport to decide anything on its merits. The second set of proceedings was commenced in England, on 17 November 2009, by Vasonia against Unicredit under the guarantee. The third set of proceedings (the “Final Proceedings”) was issued on 12 April 2010, by MF, by a fresh writ in Italy, against Unicredit and Vasonia, for a substantive determination of Vasonia’s rights under the guarantee and MF’s liability to Unicredit with respect to collateral lodged with the bank. There were issues under both articles 27 and 28. As for article 28, the issues, essentially between Unicredit and Vasonia, were whether the English and the Italian proceedings were related, and, if so, which court was first seised. Gloster J decided: (i) the Initial Proceedings in Italy and the English action were not related (at paras 77/80). (ii) The Final Proceedings were separate proceedings and did not “relate back” to the Initial Proceedings (at para 81). (iii) The only related proceedings were the English action and the Italian Final Proceedings, and it followed that the English court was the court first seised (at para 82).

In my judgment, these decisions do not appear to raise the problems raised by the present case. Once Gloster J had decided that the Initial Proceedings, both at the time of their issue and thereafter, were not “related” to the English action (the article 28(3) question), it necessarily followed that the English court was the court first seised of the only related actions, namely the English action and the Italian Final Proceedings (the only remaining article 30 question).

However, in the course of her analysis Gloster J appears to have regarded the issues before her as involving, principally, the question whether the Initial Proceedings and the English action were related at the time of the bringing of the English action (see her para 73 cited above). For my part, however, I would regard only Burton J’s judgment in this case as supporting her approach.

In my judgment the solution to this case depends in part on a proper analysis of its facts, and in part on the maintenance of a proper distinction between the concepts of related and pending actions on the one hand, and of first seisin on the other.

As to the facts of the case, I consider, first, that in theory any defence is implicit in a declaration of non-liability. That is one of the advantages of such a declaration, and perhaps one of the reasons why, traditionally, such declarations have been looked upon with disfavour by the common law. In Kolden Holdings v. Rodette Commerce Ltd [2008] EWCA 1468, [2008] 2 All ER (Comm) 289 at [7], Lawrence Collins LJ adverted to that traditional view:

In the present case, however, it is not clear that Stribog is concerned with any improper forum shopping. It is true that FKI, like Stribog itself, is an English company: however, the subject matter of the litigation is the claims of a German company, DeWind GmbH, which is in insolvency proceedings in Germany. Moreover, a primary concern of Stribog’s application under the Regulation is to canalise the issue of the validity of the German assignment within the German action. There is nothing wrong with that aspiration. Furthermore, the original exception made for the purchase price claim under the BTA reflected Stribog’s recognition of the BTA’s English law and jurisdiction provisions.

Secondly, I do not agree with the judge that the issue of the validity of the assignment was excluded from Stribog’s original particulars of claim. The most that could be said is that Stribog recognised that a valid claim under the BTA by FKI as assignee properly belonged in England. Stribog simply had not said one way or another whether the assignment was susceptible to challenge or not. Therefore, while an allegation that the assignment was invalid may be a new allegation, it is not an allegation which is in any way inconsistent with a general claim of non-liability.

Thirdly, while I am perfectly happy to recognise that the February statement introduced, or for the first time made explicit, a new issue, namely the validity of the assignment, and that such a plea can be regarded as a de facto amendment, in the sense that, without such a plea, there would be no reason to think that the validity of the assignment was in issue in the German action, and that that is so whether or not formal leave for the making of such an amendment would be necessary: nevertheless, that is very far from considering that such an amendment is for any relevant purpose analogous to other forms of amendment, such as the plea of an entirely new cause of action, especially one which does not arise out of the same or substantially the same facts but would constitute a “new claim” for limitation purposes, or such as the substitution or addition of a new party. It is quite understandable that where such new causes of action or new parties are in question, the proceedings have to be regarded as new proceedings. So, just as under the Brussels Convention, when seisin depended on service, not issue, the service of one party did not count as providing the court with seisin in respect of another party who had not been served (see The Maciej Rataj and Grupo Torras v. Al-Sabah at 20/21), so now the issue of proceedings involving one cause of action may not give the court seisin for the purposes of another cause of action in respect of which the court’s jurisdiction had not as yet been invoked. However, the introduction of a new defence is quite another matter. A limitation defence, for instance, may be pleaded or not, where available. If it is not pleaded, it does not enter into the litigation. If it is pleaded, it enters into the litigation. But it is not like a new cause of action or a new party. It does not constitute new proceedings.

Moreover, even where a new party is substituted, eg by reason of assignment, one must be cautious: the new party may still, for the purposes of article 27 and the concept of “the same parties”, not count as a new party; that is the ratio of Kolden Holdings v. Rodette Commerce.

Fourthly, the exclusion of the BTA purchase price claim from the German proceedings is not tantamount to the exclusion from those proceedings of any issue as to the validity of the assignment to FKI. Thus, the German action may well put in issue the validity of the assignment to FKI, while at the same time recognising that, if FKI holds a valid assignment, the purchase price claim under the BTA must be litigated outside the German action and in England. Thus, it is possible to recognise that a clear statement that such and such an issue is excluded from an action prevents such an action from being recognised as related to another action as long as that exclusion holds good. But if so, the question that then arises is how far that exclusion goes. It seems to me that it is perfectly consistent for Stribog to say at one and the same time that (i) the question of the validity of the assignment is within its German action and (ii) the question of FKI’s purchase price claim under the BTA, should FKI have a valid assignment, lies outside the German action and should be litigated in England.

I turn next to the article 28 concept of related actions. It is only when there are related and pending actions in separate member states that article 28 comes into issue (“Where related actions are pending in the courts of different Member States…”). If there is only one action, there cannot be related actions. Once there are at least two actions pending in different member states, then it is possible to ask whether they are related and also which of the courts is first seised. The question whether they are related is the article 28(3) question. 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?

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

First, at the time of seisin of the first of two related actions, there is only one action. Article 28 cannot operate. The question of related actions operates only when there are two related actions pending in different member states.

Secondly, whereas it is reasonably easy to state which is the court first seised of two related actions, the concept of which of two related actions was first seised by one or other court as a related action (the notion of the “first related action”) is a difficult and problematical concept. Any court asked to decide the article 28 question would have to find not merely whether two actions were related (the article 28(3) question) but at what time they first became related. Take this case. The German action, when it is issued, stands by itself. It claims against FKI a general declaration of non-liability in respect of possible claims by DeWind GmbH assigned to FKI, with the exception of a possible claim by FKI as assignee of DeWind GmbH’s purchase price claim under the BTA. Then FKI brings that purchase price claim in England. At that moment the two actions are not related, although they are potentially related because it remains to be seen whether Stribog accepts the validity of the assignment. Then, in the German action, Stribog denies the validity of the assignment of DeWind GmbH’s claims. It denies the validity generally, and not merely so far as concerns the BTA claim. It does not go back on its exclusion of the BTA claim: that remains to be pursued under the BTA in England. Thus the first time there is any explicit relationship between the two actions, it is brought about in the German action. In theory, the question of invalidity of the assignment, raised for the first time in the German action, might elicit a response from FKI which concedes that invalidity. That is a question of German law entirely distinct from the question of English law under the BTA purchase price claim. If the point is conceded, then the actions do not look as though they can be related, for there should be no dispute but that the English action should fail. If, however, FKI denies the invalidity of the assignment, then there becomes for the first time an explicit clash on which there might be conflicting judgments in Germany and England.

So which is the court first seised of a related action? Is it the German court, because it is only by reason of the February statement that one can speak of a related action at all? Or is it the English court because it is the actual claim by FKI in England which first sets up a potential clash which is then only secondarily created by the allegation of invalidity of the assignment made in the German action? Or is it the German court with respect to Stribog’s claim in its original form, because it is the declaration of non-liability which is the original source of all potential dispute?

I find these questions impenetrable, and I do not think that it could have been intended that courts and litigants should become embroiled in them.

Thirdly, I bear in mind Lord Saville’s approach in both The “Happy Fellow” and Sario v. KIO, to the effect that the Convention/Regulation rules for lis pendens were designed to be approached in a broad, commonsense way, free of over-sophisticated analysis or encouragement to satellite, interlocutory, litigation. That approach is plainly supported by the Regulation’s preamble. Rules should in general be highly predictable and the rules relating to lis pendens should provide a clear and effective mechanism for resolving relevant cases. In my judgment, the approach of FKI in this case is the antithesis of these requirements. It turns two relatively straightforward questions – Are the actions related? If so, which court was first seised? – into a morass of impenetrable sophistication.

Fourthly, I do not, with respect, agree with the judge, or FKI’s renewed submission, that it is Stribog’s approach which is inconsistent with predictability or a clear and effective mechanism. The reason given (at para 39 of the judgment below) is that a decision in Stribog’s favour would enable a litigant, as it were, to write its own passport by unilaterally changing the nature of its action. However, it is in my judgment a hopeless counsel to seek to discipline litigants’ attempts to identify the true matters in issue between them by castigating this as some kind of improper forum shopping and on that ground to rearrange the Regulation’s rules. I do not in any event understand why it is said that the nature of the German action has been changed. It is an action for a declaration of non-liability. One might wish in general to deplore the nature of such a claim, but as things stand it is too late to do so. A plea that FKI’s assignment is invalid is wholly consistent with such a claim. There is no retraction of the exclusion of a claim under the BTA from the German action. It is simply that the invalidity of the assignment catches the assignment of claims under the BTA as it catches the assignment of DeWind GmbH’s claims generally.

Fifthly, I bear in mind the doctrine of The Maciej Rataj that article 22 (28) should be given a broad interpretation. It seems to me that FKI’s attempt to classify the relationship of related actions on a step by step basis is not consistent with that doctrine.

Sixthly, I do not consider Stribog to be conducting improper forum shopping. I agree that it would be better if the Regulation’s rules were to be found and interpreted in a way which discourages and does not encourage forum shopping. However, I do not consider that Stribog’s interpretation of the rules does encourage forum shopping, or, if it does, it does so only to the extent dictated by the essential structure of the rules themselves. By that I mean to advert to their essential structure whereby precedence is given to the litigant who puts a court of his own choice first into the saddle (the doctrine of first seisin). Of course, the claimant has to find a court with jurisdiction sanctioned by the Regulation, without which he cannot get to first base. But if he can find such a court with jurisdiction he is entitled to rely on the doctrine of first seisin if it should emerge, which may well depend on the serendipity of litigation, that there is subsequently brought into existence in another member state another action which turns out to be related to his own. But he can always say, that he was first in the field. I recognise that if such a system operated without the claim for a declaration of non-liability, then it would only be a litigant with a positive claim who could launch such pre-emptive litigation. As it is, the negative claim is recognised as legitimate for such purposes.

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 courts…”) (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.

Eighthly, the authority closest to this case appears to me to be The “Happy Fellow”, where, without perhaps expressly deciding the issue which has arisen in this case, both Longmore J and this court proceeded on the basis that the correct approach was to ask which of the two actions was commenced first, once it had been appreciated that the two were related as of the time that the English court was asked to make a decision for the purposes of article 22 (28).

Finally, the doctrine of article 28 merely raises a discretion to stay. That discretion is given in the interest of avoiding irreconcilable judgments in courts of different member states. One can imagine a system in which either court might be given a discretion: however, the rule is to confine the discretion to one court, and in that case one understands why that should be a court other than the court first seised. Nevertheless, such are the differing circumstances in which that discretion may arise, that the way in which that discretion might be exercised cannot and in my judgment should not be relegated to a rule which eradicates the discretion. Thus, I do not with respect agree with Ognall J in Virgin Aviation Services v. CADA Aviation Services [1999] I L Pr 79 that “where actions are related the first duty of the court is to stay the proceedings” (see also Owens Bank Ltd v Bracco [1994] QB 509 at 541/2 (ECJ, Case C-129/92 at paras 75-79 in the Opinion of Advocate General Lenz) and Centro Internationale Bank AG v. Morgan Grenfell Trade Finance Limited [1997] CLC 870 at 891/2). Of course, the discretion is given to avoid irreconcilable judgments, and should be used, where it properly can, for that purpose. However, that should not be taken as a reason for deploring a finding that one rather than another court is the court first seised, of which there is a flavour in the judge’s para 41. So, in this case, the finding that the German action is the court first seised, and that there is therefore a discretion in the English court to stay the English action, is not a reason to stay the claim under the BTA for any longer than is necessary to resolve the issue of the validity of the assignment where it most properly can be resolved, namely in the German court. Otherwise the baby of the BTA claim (which it has always been recognised should properly be adjudicated in an English court under English law) would be thrown out together with the bathwater of the issue of the validity of the assignment.

For these reasons, in addition to those of Mummery LJ, I would allow this appeal. The German court was the court first seised. The English court was the court second seised. The English court therefore has a discretion to stay the English action. It should do so to avoid irreconcilable judgments on the question of the validity of the German assignment, which it is plain should preferably be decided in the German action, and be decided first. However, the stay should last no longer and be no more extensive than is necessary for that purpose. If for any reason the question of the validity of the German assignment is not properly progressed in Lübeck, it should be open to FKI to return to the commercial court in London to argue that it should be permitted to make progress with the underlying claim under the BTA.

I agree that this appeal should be allowed and that the English proceedings should be subjected to the limited stay which my Lords have identified.

The only difference which I can detect in my Lords’ judgments is that, whereas Mummery LJ, at [40] to [43], prefers to ask “which court was first seised of a pending action?” before asking “are the actions related?” Rix LJ, at [119] to [120] and at [125], prefers to ask the questions in the reverse order. I do not see why the order matters; but Rix LJ seems to have the terminology of Article 28 on his side.

Subject thereto, I agree with both judgments. In effect Burton J decided that the question was “which court was first seised of the issue which has rendered the actions related?”. With respect, it was impermissible for him thus to rewrite paragraphs (1) and (3) of Article 28. He was clearly concerned that any other construction of them would reward manoeuvres cynically designed to thwart resolution of an issue properly first raised in the courts of a particular state. But the discretion conferred by the article takes care of that important concern. In the present case the fact that the assignment agreement dated 31 August 2009 is governed by German law outweighs any manoeuvring on the part of Stribog in the discretionary exercise so as to yield the limited stay.