Court of Appeal applies Brussels I “anti-torpedo” mechanism to “asymmetric” jurisdiction clause

In Etihad Airways PJSC v Flöther [2020] EWCA Civ 1707, the English Court of Appeal reaffirmed that “asymmetric” jurisdiction clauses can engage the “anti-torpedo” mechanisms in the Brussels I Recast Regulation. Although that regulation now has a limited lifespan in the UK, the decision joins a growing body of jurisprudence in favour of that conclusion.

The facts of Etihad Airways

Air Berlin was a major German budget airline. Etihad was an investor in Air Berlin. Air Berlin had a history of financial difficulties and, in 2016, its auditors raised concerns about its ongoing viability.  

Air Berlin sought further financial support from Etihad. In April 2017, Etihad, accordingly, concluded two documents. First, a comfort letter from its CEO to Air Berlin confirming Etihad’s intent to continue to provide financial support. Second, a loan facility under which Etihad agree to advance EUR 350 million.  

The facility agreement was subject to English law, and an “asymmetric” jurisdiction clause in favour of the English courts. Under this, as is usual in such clauses, Air Berlin, as borrower, was bound to bring proceedings only in England whilst Etihad, as lender, retained the freedom to sue in any court which would otherwise take jurisdiction.

The dispute and first instance proceedings1  

Air Berlin went into insolvency and its administrator then commenced German court proceedings against Etihad alleging breaches of the comfort letter and, alternatively, a non-contractual claim connected with the comfort letter.

In response, Etihad commenced proceedings in the English courts arguing that those claims were caught by the jurisdiction clause in the facility agreement. In Etihad’s case, they should therefore have been brought in England and the English court could proceed (even though the German courts were first seised) by virtue of Article 31(2) Brussels I Recast (colloquially known as the “anti-torpedo” mechanism, introduced to reverse the Gasser decision).

At first instance, Jacobs J ruled in favour of Etihad, agreeing that Air Berlin’s claims fell within the jurisdiction agreement in the facility agreement and that Article 31(2) could apply to an “asymmetric” jurisdiction clause. Air Berlin was granted permission to appeal only in relation to the latter issue.

The Court of Appeal on Article 31(2)'s application to “asymmetric” jurisdiction clauses in favour of an EU Member State court.

The Court of Appeal upheld the judge’s decision on this point. Its reasoning was essentially based on three strands

First, Air Berlin relied on a lengthy exposition [47-67] of the evolution of the provisions concerning jurisdiction clauses in the Brussels I Recast to argue that Article 31(2) should be (narrowly) applied to fully exclusive clauses. The Court rejected this, finding that, in fact, party autonomy was at the heart of the provisions and that this supported Etihad [68]

Second, the Court of Appeal found that Etihad’s position was in line with the policy of Article 31(2). Otherwise, the job of abolishing “torpedo actions” would be “half done”. Further, this conclusion was more in line with the wording of Article 31(2). In particular, if a clause was given effect by Article 25, as was conceded in this case, there should be no difficulty in holding that Article 31(2) would apply to the extent it was exclusive [73-76].  

Third, the Court of Appeal rejected an argument that the scope of the Hague Choice of Court Convention (“Hague”), which only extends to exclusive clauses, should influence matters. In short, they were different instruments. Brussels I had a wider scope, and different lis pendens provisions. So consistency of interpretation couldn’t be treated as a “controlling principle” [87-88]. In dealing with this, the Court of Appeal also discussed arguments about whether Hague extended to “asymmetric” jurisdiction agreements. Whilst this was not necessary for its decision, and so it expressly didn’t rule on it, it did point out that there were strong indications (in, for example, the Hartley-Dougachi Report) that this was not the case [85-86]. In the face of those, and the practical point that (for example, in a post-EU/UK transition period context) it may well be the view of an overseas, not the English, court which is important, the tentative (obiter) suggestions in first instance cases that there may at least be arguments that Hague goes beyond “fully” exclusive clauses need to be treated with all due caution.  


An important judgment but, as many will know, the Brussels I Recast has a limited shelf life in the UK due to the impending end of the EU/UK transition period. In future (in the UK and in cases involving the UK) its application will be limited to cases covered by the provisions of the Withdrawal Agreement. It is notable, however, that the Court’s judgment appears to have joined a growing list of national court rulings in favour of this interpretation of Article 31(2) [94].

Beyond the Withdrawal Agreement (and assuming the UK is not permitted to re-accede to the Lugano Convention) the question of parallel proceedings between the UK and EU in cases of “breach” of an English jurisdiction clause will (assuming the UK’s application to re-join the Lugano Convention is rejected) generally be regulated, on the UK side, by the use of an anti-suit injunction. That can potentially help off-set the fact that, if matters came before an EU court, the route to declining jurisdiction will be somewhat less straightforward and, depending on the facts, might involve the application of national law, or Articles 33/34 Brussels I Recast (for more on the end of the EU/UK transition period and the impact on civil and commercial justice in the UK click here).

Stephen Lacey, Counsel PSL, London

If you would like to discuss this case further or have any questions, please get in touch with our Banking Litigation team.

1 For more on the background to Article 31(2), and asymmetric jurisdiction clauses in favour of EU Member State courts and the Brussels I Recast, see our note on the first instance decision here.