I am much pleased that Dr Brooke Marshall has accepted to write on CJEU Lastre for the blog. Dr Marshall has written the guiding volume on the issue (highlights of the book’s launch are here). True to form, her analysis below is as complete as it is on point, and a most excellent addition to the analysis already out there.
Geert.
CJEU on substantive validity and on asymmetric clauses: what we now know, and what we (still) don’t.
The background to Case C-537/23 Società Italiana Lastre ECLI:EU:C:2025:120, and commentary on it, has already been expertly provided by François Mailhé, Gilles Cuniberti, and Geert van Calster. It is a privilege for me to contribute to that discussion on this excellent blog.
My post confines itself to the questions that the French Cour de cassation asked and the answers which the CJEU gave (and did not give). My analysis draws heavily from my book on the subject, so pinpoint references to that are given in each section below.
The asymmetric jurisdiction clause in issue said:
‘the court of Brescia [(Italy)] will have jurisdiction over any dispute arising from or related to this contract. [SIL] reserves the right to bring proceedings against the purchaser before another competent court in Italy or elsewhere’.
The preliminary questions of the Cour de cassation, in essence, were:
(1) Is a complaint about a clause’s asymmetric character or imprecision an autonomous (EU law) question or a question of the clause’s substantive validity? Or should substantive validity be interpreted restrictively ‘and regarded as relating purely to the material grounds for invalidity, which are principally fraud, error, deceit, violence and incapacity’?
(2) If it is an autonomous question, is a clause like the one in issue compatible with Art 25 of the Recast?
(3) If asymmetry goes to substantive validity, how does the conflict-of-laws rule in Art 25 of the Recast work? Which court’s law applies and does that law include renvoi?
These were all good questions, each deserving an answer. Regrettably, the Court only answered the first two.
The Court’s decisions on the first question:
I. The phrase ‘null and void as to its substantive validity’ must be interpreted autonomously
The Court begins at [30] by resolving several uncertainties attending the process of characterisation and the substantive validity rule, which are relevant to jurisdiction clauses generally. The first is as to which law applies to the interpretation of the phrase ‘null and void as to its substantive validity’ in Art 25(1). It had hitherto been uncertain whether the meaning of that phrase and the scope of the issues it covers was a question to be resolved by the law of the court seised, by reference to an autonomous interpretation or by reference to the law applicable to substantive validity of the clause itself, as designated by the conflict-of-laws rule in Art 25(1).
Sensibly, and unsurprisingly in my and Gilles Cuniberti’s view, the Court decides that this question is to be resolved by an autonomous interpretation. This ensures that issues of substantive validity, governed by national law, do not overlap with other aspects of Art 25(1) which are governed by an autonomous approach. Several of those other aspects are pointed out by the Court at [35], namely: that ‘the parties … have agreed that a court or the courts of a Member State are to have jurisdiction to settle any disputes which have arisen or which may arise in connection with a particular legal relationship’, which the Court describes as requirements as regards a jurisdiction agreement’s ‘substance’; and that the agreement complies with conditions as to form. In this respect, the Court’s decision coheres with its decision in C- 519/ 19 Ryanair DAC, paras 41, 48– 61. There, the Court observed that it is for the court seised to ascertain first whether a jurisdiction agreement satisfies the requirements of form and consensus under the Recast before turning to the separate issue of substantive validity. So far, so good.
Asymmetric Jurisdiction Clauses (OUP 2023), para 6.35.
II. Few issues can be characterized as issues of substantive validity (and asymmetry is not one of them)
At [36] and [37], the Court resolves a second uncertainty of general application, namely as to which kinds of issues can be characterized as issues of substantive validity, adopting a restrictive approach. It rules that ‘that concept covers the general causes of nullity of a contract, namely, in particular, those which vitiate consent, such as error, deceit, violence or fraud, and incapacity to contract’ and that it must be interpreted as ‘referring solely to [a jurisdiction agreement’s] nullity in respect of general causes of nullity of a contract’ (my emphasis). In other words, the issues that we knew were already covered by the substantive validity rule are covered but nothing else.
That nothing else is covered is a surprising result, given that there had been some sign that the CJEU might favour a broad interpretation in C- 519/ 19 Ryanair DAC, para 50. One of the reasons the Court gives for its approach is to reflect the solution taken in the 2005 Hague COCA Convention (at [40]). It does not say anything more, but the Court seems to be insinuating that since there is some support for the view that the null and void rule under Art 5(1) of the COCA Convention should be interpreted restrictively (see the COCA Convention’s Explanatory Report, para 126), so too should the substantive validity rule under the Recast. I, therefore, share Gilles Cuniberti’s view that the Court may seek to align its interpretation of the null and void rule in the COCA Convention and the substantive validity rule in the Recast.
Asymmetric Jurisdiction Clauses (OUP 2023), para 6.37, cf paras 6.38-6.48, 6.52–6.55
III. A complaint about asymmetry in a jurisdiction clause is an autonomous question (and only an autonomous question)
In light of the Court’s restrictive approach to the characterization of issues relevant to substantive validity, the Court consequentially rules at [51] that ‘the validity of an agreement conferring jurisdiction in view of its alleged asymmetry must be examined having regard not to criteria relating to the causes of an agreement being ‘null and void as to its substantive validity’, within the meaning of the first sentence of Art 25(1) of the Brussels Ia Regulation, but to autonomous criteria which are derived from that Art 25 as interpreted by the Court.’
Seemingly, to justify that autonomous approach, the Court relies on the fact that the former Art 17(4) of the Brussels Convention expressly provided for clauses to the benefit of one party and that asymmetric jurisdiction agreements which favour consumers, insureds etc, and employees are contemplated and permitted by Recast: [48]–[50], [52]. For cogent criticism on the latter point, see Geert van Calster’s post.
Asymmetric Jurisdiction Clauses (OUP 2023), paras 1.27–1.36 cf paras 6.69–6.82
IV. Whether a jurisdiction clause is sufficiently precise is a question to be resolved by applying an autonomous standard
The Court reiterates the requirement that the court or courts on which the agreement confers jurisdiction must be sufficiently precisely identified, and that this applies to Art 25(1) of the Recast, as it did to the Recast’s predecessors: [45]. Any complaint about an asymmetric jurisdiction clause being insufficiently precise must, therefore be resolved autonomously: [47].
Frustratingly, the Court merely repeats the crux of that previous case law without elaboration at [42]:
‘It is sufficient that the clause state the objective factors on the basis of which the parties have agreed to choose a court or the courts to which they wish to submit disputes … Those factors, which must be sufficiently precise to enable the court seised to ascertain whether it has jurisdiction, may, where appropriate, be determined by the particular circumstances of the case being examined by the court.’ (C-387/98 Coreck, para 15; C-222/15 Hőszig, para 43).
Asymmetric Jurisdiction Clauses (OUP 2023), paras 6.05–6.07
The Court’s decisions on the second question:
V. Assessed autonomously, an asymmetric jurisdiction clause is valid under EU law and sufficiently precise where the optional limb can be interpreted as contemplating proceedings only before an EU Member State or EFTA Lugano State court
The Court rules that an asymmetric jurisdiction is in principle valid/admissible under EU law, relying especially on the need to protect party autonomy (recital 19) (see [56], [58], [64], [65]). No surprises there, until one does a double take at [58]: what the Court rules is valid is an asymmetric jurisdiction clause the optional limb of which is limited to courts with jurisdiction under the Recast and the 2007 Lugano Convention. More on that in a moment (see VI).
So far as the requirement of precision of content is concerned, the Court rules at [59] that an asymmetric agreement clause which designates a particular EU Member State or EFTA Lugano State court, on the one hand, and courts which have jurisdiction under the Recast and 2007 Lugano Convention, on the other, ‘satisfies the requirement of precision resulting from Art 25(1) of that regulation and the objectives of foreseeability, transparency and legal certainty, set out in recitals 15 and 16 of that regulation.’ Subject to the reservations I have as to what the Court means by the word ‘designate’ (see VIII below), that seems to me to be a cogent conclusion: ‘any court with jurisdiction’ is an objective factor that is sufficiently precise, assessed from the perspective of a court seised: any court EU Member State or EFTA Lugano State court which would otherwise regard itself as competent has jurisdiction.
Asymmetric Jurisdiction Clauses (OUP 2023), paras 6.09–6.10
VI. Assessed autonomously, an asymmetric jurisdiction clause is not sufficiently precise and contrary to EU law where the optional limb is to be interpreted as contemplating proceedings before a Third State court
The Court rules at [60] that, if, in referring to ‘another competent court … elsewhere’, the agreement ‘must be interpreted’ as meaning that it also designates the courts of one or more Third States, it would be contrary to the Recast. Put another way, the asymmetric clause in issue ‘is incapable of satisfying the requirement of precision, … unless it may be interpreted as designating the court of Brescia and the courts of Member States or States that are parties to the Lugano II Convention as having jurisdiction to resolve disputes between the parties’: [62] (my emphasis). In practical terms, if the words ‘another competent court … elsewhere’ could encompass an English court, for example, because one of the parties has some connection to England, the clause would not be valid under the Recast.
What I find curious about this aspect of the Court’s decision is the lack of clarity in the reasoning it used to get there. It reasoned, at [60], that where the optional limb can be interpreted as contemplating proceedings before a Third State court:
‘that agreement conferring jurisdiction would not be consistent with the objectives of foreseeability, transparency and legal certainty referred to in recitals 15 and 16 of that regulation, to the extent that EU law does not, in itself, make it possible to designate the courts which have jurisdiction, as that designation is, where appropriate, the result of the application of the rules of private international law of third countries.’
While it is true that whether a Third State court is competent is not the result of EU law, it is difficult to see why that matters if the requirement of precision applies to the clause as a whole and is to be assessed only from the perspective of a court seised. Indeed, it is difficult to see why the criterion ‘another competent court … elsewhere’ is not an objective factor or why it is insufficiently precise, assessed only from the perspective of a court seised: any court which would otherwise regard itself as competent has jurisdiction, whether that court is a Third State court or not. Although the optional limb does not, in my view, confer jurisdiction on those courts (only the clause’s first sentence designating the courts of Brescia does that), the optional limb provides a criterion allowing any court that is seised to determine whether it is one of the courts before which the parties had contemplated that the option holder could bring proceedings.
Instead, the Court could have more persuasively justified its decision by elaborating on its previous case law in C-387/98 Coreck and C-222/15 Hőszig to clarify that the requirement of precision applies as much to the parties as it does to the court. Put another way, the Court could have clarified that the requirement of precision is not to be examined purely from the perspective of a court seised; rather, as a result of the ‘objectives of foreseeability, transparency and legal certainty, set out in recitals 15 and 16’ of the Recast, it also requires that the parties be able to foresee from a jurisdiction clause where they may be called upon to defend proceedings. Reference to C-566/22 Inkreal paras 28–29 would have done the job. There, the court said that an agreement conferring jurisdiction coheres with the aim of legal certainty because it ‘helps the applicant to ascertain the court before which he or she may bring proceedings and the defendant to foresee the court before which he or she may be sued, and enables the national court seised to be able readily to decide whether it has jurisdiction’. (The Court did refer to Inkreal but, unfortunately, not to make this point.)
I had foreshadowed the Court’s decision at [60] and [62] as a possible problem for asymmetric clauses of this kind, albeit on the argument that an aspect of the requirement of precision under EU law ought to be that the agreement should allow a reasonably well-informed defendant to foresee the courts before which they may be sued. I had suggested that since it was uncertain what the CJEU may say, it might be thought prudent for drafters to specify (expressly) that ‘any competent court’ is restricted to those within the EU or to an EFTA Lugano State, if that was their intention. In light of the CJEU’s decision, that suggestion may well warrant some serious consideration.
The Court’s decision may also give cause to reconsider non-exclusive jurisdiction clauses which designate one EU Member State or EFTA Lugano State court, to the extent that these impliedly permit both parties to bring proceedings before any other competent court, including Third State courts.
Asymmetric Jurisdiction Clauses (OUP 2023), paras 2.31–2.33, 5.50, 6.10, 6.14–6.24
Unanswered questions
VII. So is the clause in Case C-537/23 Società Italiana Lastre sufficiently precise and compatible with the Recast?
This is not a question for the CJEU to answer and, on my reading, it is not one which the Court did. For a different view, see Gilles Cuniberti’s post. Note that the Court’s careful use of conditional language in [60] and [62] ‘if’ … ‘unless’… One may only speculate on how a French court will now interpret the clause against the backdrop of the CJEU’s rulings.
VIII. Which court(s) does an asymmetric clause ‘designate’? And how does the substantive validity rule work for them?
The Court, is to my eyes, vague or perhaps ambivalent as to which court or courts an asymmetric clause typically ‘designates’ (ie confers jurisdiction on/ prorogates the jurisdiction of). Does it designate (1) only the ‘anchor court’ (to use Richard Fentiman’s words), here Brescia, and merely preserve the special or general jurisdiction that other courts have under the Recast or 2007 Lugano Convention for one of the parties? Or (2) does it ‘designate’/confer jurisdiction on/prorogate the jurisdiction of all those courts? The Court’s reasons at [55] suggest the latter and [59] appears to contemplate both possibilities.
This opacity is regrettable for multiple reasons. One reason is that the conflict-of-laws rule in Art 25(1) for determining a clause’s substantive validity presupposes that the clause confers jurisdiction on/designates only one court. That conflict-of-laws rule necessitates the following interpretation for an asymmetric clause of this kind: only the anchor court is ‘prorogated’ and the jurisdiction under the default rules of the courts the subject of the option is preserved but not prorogated. On that argument: if the anchor court is seised by either party, it should apply the law selected by its conflict-of-laws rules to the substantive validity of the clause (assuming renvoi is included – see IX below). If, instead, another court is seised, it should apply the law selected by the anchor court’s conflict-of-laws rules to the clause’s substantive validity. Whether that other court is seised by the option holder, according to the agreement, or by the non-option holder, contrary to it, is immaterial for the purposes of the rule.
Put another way, if that argument is wrong and an asymmetric clause of this kind ‘designates’ more than just the anchor court, the Court will (again) be called upon to explain how the substantive validity rule is supposed to work for a clause that ’designates’ multiple courts. It is a shame that it did not grasp the nettle and do so in this case.
Asymmetric Jurisdiction Clauses (OUP 2023) paras 5.07–5.09, 5.33, 6.58–6.59
IX. Is renvoi included in the substantive validity rule?
Also unfortunate is that the Court did not clarify whether the conflict-of-laws rule for substantive validity in Art 25(1) includes renvoi as recital 20 suggests. At [33], as Geert van Calster has highlighted, the Court quotes recital 20 in its entirety but for the crucial words ‘including the conflict-of-laws rules of that Member State’. Alignment of the conflicts rule in Art 25(1) with the COCA Convention, which the Court appears to favour (see II above) would militate in favour of including renvoi. What one should make of the Court’s misleading truncation of recital 20 is, therefore, anyone’s guess.
Asymmetric Jurisdiction Clauses (OUP 2023), paras 6.50–6.51
Luxembourg should expect more questions to come…
Brooke Marshall.
