French Cour de Cassation Asks CJEU Whether Granarolo Still Stands After Wikingerhof
This post was written by Lea Marion who is a Member of the Paris Bar.
The legal characterization of claims for abrupt termination of established commercial relationships — an action specific to French law (Article L. 442-1, II of the French Commercial Code) — continues to be a fertile ground for uncertainty and doctrinal debate.
Caught between its traditional classification as a tortious claim under French domestic law (Cass. com., 6 February 2007, No 04-13.178) and the temptation of a contractual reading under the aegis of the Court of Justice of the European Union (C-196/15, Granarolo), recent jurisprudence reflects a deeper tension: that of the delicate balance between national legal autonomy and the European objective of harmonization.
Background: A Fragile Consolidation of Dual Qualification
On 12 March 2025, the Paris Court of Appeal, faced with a question regarding the applicable law for an action based on Article L. 442-1, II of the French Commercial Code, reaffirmed that within the EU legal order, such actions fall within the contractual sphere.
In doing so, the Court aligned itself with the CJEU’s Granarolo ruling — interpreting Regulation No 44/2001 — and extended its reasoning to conflicts of laws, following the logic set out in Recital 7 of the Rome I and Rome II Regulations, which call for coherent interpretations alongside the Brussels I bis Regulation.
On the same day, however, the First Civil Chamber of the French Cour de cassation, faithful to its traditional understanding of Article L. 442-1, II, reaffirmed the tortious nature of the action in the context of international litigation outside the scope of EU Regulations.
From these two decisions, a seemingly stable jurisprudential framework appeared to emerge: a dual characterization — contractual under EU law, tortious under French domestic and international private law — reconciling the demands of both legal orders.
Renewed Debate: The Question of Characterization Referred to the CJEU
Where one might have believed the matter settled, the French Cour de cassation, by a decision of 2 April 2025, has reignited the debate by referring a preliminary question to the CJEU.
The dispute involved a French company and a Cypriot company bound by an air transport services agreement governed by the laws of Jersey concluded in 1995 – an agreement falling within the scope of the Rome Convention, as it predates the entry into force of the Rome I Regulation and which, since 1 August 2004, falls under the interpretative jurisdiction of the Court of Justice. The case law developed under the Convention nevertheless remains relevant for the application of Rome I, insofar as the provisions are identical or substantially similar. Following an allegedly abrupt termination without sufficient notice, the Cypriot service provider brought an action before French courts on the basis of Article L. 442-1, II of the French Commercial Code.
In support of its claim, the service provider argued, firstly, that Article L. 442-1, II applied as a mandatory overriding provision, governing any economic relationship involving at least one operator established in France or where performance took place in France; and secondly, that pursuant to Article 4 of the Rome II Regulation, the applicable law should be French law, as it bore the most significant connection to the dispute.
In its judgment, the Cour de cassation observed that the dispute raised a preliminary issue regarding the legal characterization of the action, and seized the opportunity to refer the following question to the CJEU:
Must Article 1(1) of the Rome Convention of 19 June 1980 on the law applicable to contractual obligations, and Regulation (EC) No 864/2007 of 11 July 2007 on the law applicable to non-contractual obligations (Rome II), be interpreted as meaning that a damages action based on the abrupt termination of established commercial relations, assessed under legislative provisions governing restrictive practices in competition, and therefore arising from a statutory obligation to refrain from certain conduct, falls within the scope of non-contractual obligations, independently of any contractual links that may exist between the parties?.
Hopes for a European Alignment on the Tortious Qualification
In its reference, the Cour de cassation subtly highlighted the extra-contractual nature of the action, signaling its preference for recognizing the tortious characterization of abrupt termination under EU law.
In Granarolo, the CJEU had justified the contractual classification, in the absence of a written contract, through a comparative law analysis:
in a significant number of Member States, long-standing commercial relationships established without a written contract may, in principle, be deemed to fall within an implied contractual relationship, the breach of which may give rise to contractual liability.
This approach, however, strained both French law and the CJEU’s own case law. French law, notably, treats the indemnity action — whether or not a contract exists — as based on a statutory provision aimed at ensuring the orderly conduct of economic life, independent of contractual obligations (see Advocate General’s opinion, para 19).
It also overlooked the Brogsitter case, where the CJEU held that it is for the referring court to determine
whether the claims brought seek to obtain compensation for damage the cause of which can be reasonably regarded as a breach of the contractual obligations binding the parties, requiring the interpretation of the contract (C-548/12, para 26).
However, the CJEU’s position evolved significantly in Wikingerhof (C-59/19), a case concerning an alleged violation of German competition law.
The Court there held that
where the applicant relies, in its application, on rules on non-contractual liability, namely the breach of an obligation imposed by law, and it is not necessary to examine the content of the contract between the parties to assess the legality of the defendant’s conduct, the action falls within non-contractual matters within the meaning of Article 7(2) of Regulation No 1215/2012.
It is this Wikingerhof judgment — and French scholarly commentary pointing out that the ruling made no reference to Granarolo, thus raising the prospect of its abandonment — that has led the French Cour de cassation to seek clarification from the CJEU.
The forthcoming ruling from the CJEU could mark a decisive turning point in the litigation of abrupt terminations of established commercial relationships.
The CJEU is faced with an impossible mission. In the absence of a European private law, it is difficult to qualify the issue by means of autonomous notions. The only way to dispense with national qualifications is precisely to circumvent the problem of qualification. The formula, in my view, is to consider that, for the purposes of the rules of international jurisdiction, an issue must be considered contractual when the action necessarily requires, as a precondition, the existence of the contract (not its interpretation), so that it would not be feasible or conceivable without the existence of such a contract
Dear Sixto, since we cannot rely for PIL purposes on a EU taxonomy built for substantive law purposes, as you rightly say, shouldn’t we rather espouse a “pure” PIL analysis here? I proposed this approach in a book published a few years ago (P. Franzina, La giurisdizione in materia contrattuale, Cedam, 2006, especially p. 225 ff.).
My view can be summarised as follows. Autonomous interpretation posits that to determine the meaning of language used in EU legislation (here, the expressions “matters relating to a contract” and “matters of tort” as used in Article 7 point 1 and point 2 of Brussels I bis, respectively) one should look at the EU provisions themselves, and extract all the indicia they appear to offer.
Thus, rather than focusing on the categories themselves (“contract”, “tort”), one might want to look at the connecting factors associated with them. Proximity and predictability suggest that this is a promising approach, because – if the rule is designed to confer jurisdiction on a court that is close to the dispute and easy to foresee – one can assume that a functional relationship exists between the connecting factor, on the one hand, and the class of disputes in relation with which it that factor is meant to operate, on the other.
One can accordingly take the connecting factor as a point of departure, and interpret the scope of the category in such a way that it covers all situations to which the factor can apply “fruitfully”, i.e., in a manner that yields practical results which ensure both proximity and predictability.
Let us look at the connecting factors, then (only the connecting factors: forget for a moment we’re talking of factors that apply to “contracts” and “torts”). The connecting factor in Article 7 point 1 is the performance of an obligation, while that in Article 7 point 2 is a prejudice-causing event.
The first factor is centred on the idea that something was planned and that the planning resulted in some form of expectations (e.g., the creditor’s expectation that he or she will be paid, once the monies become due under the contract): the place of performance is the place where those expectations have, or should have, materialised.
The second factor, instead, revolves around the idea that (unexpectedly, I would add) an event has occurred, which infringed some rights or interests (eg., the place where the traffic accident, which resulted in a broken leg, occurred).
Things being this way, a “purely PIL-based” functional analysis suggests that Article 7 point 1 applies whenever the claimant complains of something that he or she was entitled to expect, but did not materialise, whereas Article 7 point 2 applies to such disputes as relate to events that occurred out of any expectations, where the claimant complains of the prejudice that the event has caused.
If this approach is followed, the characterisation of a claim for the purposes of Article 7 point 1 or point 2 does not depend, strictly speaking, on whether the claim is a about a “contract” in the substantive sense. If the the case of the claimant is that the defendant failed to do what the claimant was expected to receive from the defendant, then that claim (whether the expectation is founded or not, of course) comes with purview of point 1. Conversely, if the claimant seeks some form of relief on account of the fact that the defendant, while owing nothing specifically to the claimant, infringed the latter’s rights, then we’re in the province of point 2.
Arguably, a claim for compensation made on the ground of the abrupt termination of established relationships fits into the description I proposed for point 1, because the claim rests precisely on the fact that the claimant did not get what the “established relationship” with the defendant allegedly entitled him to receive.
It might well be that I’m just nostalgically attached to an approach I devised in my PhD years (!), but I’m still convinced this approach makes some sense. It has the advantage of relying on the (uniform) PIL rules concerned and their own rationales, rather than on (non-uniform) substantive law, and it has, as such, the potential of leading to practical outcomes that are consistent with the goals of EU PIL.
I don’t expect the Court will overrule Granarolo, Brogsitter and the other judgments in this saga (and I don’t certainly expect the Court might be interested in knowing my opinion on this!), but when you see that the distinction between contracts and torts remains a debated issue over the decades, notwithstanding the many rulings rendered on the topic, you are permitted to ask: shouldn’t the Court take this opportunity to reconsider the whole problem from scratch, and maybe come out with something new, rather than trying to fix the difficulties that the existing “official” approach appears to create?
While the main goal of the French supreme court is likely to obtain that Granarolo is overruled to fit with its own analysis, it must be recognised that the mess created by Wikingerhof is an excellent pretext to seek such a ruling.
I am not aware of cases where the CJEU has actually expressly overruled a precedent, but it has distinguished some to the same effect.
The best we can hope for is probably that the CJEU will ‘clarify’ Wikingerhof, and in doing so possibly partly overrule it. Whether this will mean that it will also overrule Granarolo is doubtful, however, as this judgment seemed perfectly sensible to me.
The moment the Court was persuaded to depart from the conclusions it came to, and expressed perfectly well, in Brogsitter and in Holterman, all hope of legal coherence was lost. Where the litigants are party to a contract, and the claim for compensation arises from the the defendant’s alleged disrespecting the obligations (express and implied) of the relationship established by the contract, whether by breach or otherwise, the matter for complaint is one relating to a contract. Wikingerhof was as wrong as wrong can be. If the Court sees it so but cannot bring itself to say so, that would be a matter of shame; if it cannot see that it is so, it will be so much the worse.
I absolutely agree with you, Pietro. The key is that the expectations (not the will) of the contracting parties are not limited to what is expressly or implicitly stated by the parties in the contract, but of the external obligations imposed by law (what in English law are called the implied in law terms) or, in some systems, by good faith or equity. However, dear Adrian, Brogsitter’s problem, in my view, is that he should in turn have come up with an autonomous notion of what it means to ‘interpret the contract’ (maybe impossible mission). For a common law mentality, perhaps a legal obligation, external to the contract, is a matter of contract interpretation (implied in law term), but in many civil law systems basing a claim on a legal obligation is something external to the contract and is not understood as implied in the will of the parties, so there is no need to interpret the contract strictly speaking.
In my opinion, dear Gilles, Granarolo leads to an insoluble vicious circle. If what is intended is an autonomous notion, in order to avoid a ‘national’ qualification of the action, the result is precisely the opposite. The criteria of interpretation that the court uses to infer a tacit obligation are those of some civil law systems, but we do not know why it has chosen those ‘national’ systems. In common law, an implied in fact term can hardly be justified on the basis of good faith or other elements external to the contract itself, beyond the business efficacity test. Do we want to turn the court into a capricious (no reason is given) legislator of European private law? I do not think so, particularly when the problem can be solved with the ordinary tools of a European Private Law that does exist, as Pietro explains.