Tell me more, tell me more: the obligation for national courts to reason their refusals to refer to the CJEU in Sanofi Pasteur.

By Jasper Krommendijk (Radboud University, the Netherlands)

On 13 February 2020, the ECtHR found for the fourth time ever a violation of Article 6(1) ECHR for a failure of the highest national court to give proper reasons for its refusal to refer preliminary questions to the Court of Justice of the EU (CJEU) in Sanofi Pasteur. In 2014 and 2015, the ECtHR already determined a breach for a similar omission of the Italian Court of Cassation in Dhahbi and Schipani, while the Lithuanian Supreme Administrative Court was given a rap over the knuckles in 2019 in Baltic Master. This time it was the French Court of Cassation who took the blame in a claim for damages for bodily harm resulting from vaccination against hepatitis B. The company Sanofi Pasteur was found liable by lower French courts and appealed to the Court of Cassation. It unsuccessfully requested the Court to refer questions about the Product Liability Directive 85/374 and, more specifically, the obligation for the victim to prove the damage, defect and causal relationship between defect and damage. The Court dismissed the appeal whereby it explicitly noted that it did so without a need to refer. In this comment I will argue that this judgment shows once again the unpredictability and inconsistency in the case law of the ECtHR. What is needed is a Grand Chamber judgment as well as guidance from the CJEU to tell us more about the exact requirements for national courts.

 Tell me more, national court: the strict obligation to provide reasons

This judgment fits perfectly in a line of cases in which the ECtHR has employed a strict test. Since the ECtHR’s Ullens de Schooten judgment, which is also extensively relied upon in this case, it has been well established that Article 6(1) ECHR requires the highest national courts to give reasons for decisions not to refer a question for a preliminary ruling under Article 267 TFEU. The ECtHR has consistently held that Article 6(1) ECHR does not give a right to have a case referred to the CJEU (see also para. 69 of this judgment). Article 267 TFEU obliges courts to refer a question about the interpretation or validity of EU law when ‘there is no judicial remedy under national law’. The CJEU has formulated three exceptions for highest courts, which are commonly referred to as the Cilfit–exceptions. Firstly, when ‘a decision on the question is not necessary to enable it to give judgment. Secondly, when the CJEU has ‘already dealt with the point of law in question’ (acte éclairé). Thirdly, when ‘the correct application of [EU] law may be so obvious as to leave no scope for any reasonable doubt’ (acte clair). The ECtHR determined that the duty to provide reasons implies that judges must indicate which one of the three Cilfit-exceptions applies (para. 68).

Ullens de Schooten marked a departure from the ECtHR’s previous case law (for an extensive overview, see here). Earlier, the ECtHR applied the criterion of arbitrariness and held that a refusal of a request for such a referral might infringe the fairness of proceedings if it appeared to be arbitrary. The Strasbourg Court had never found a refusal to be arbitrary.

In Dhahbi, the ECtHR determined that an implicit substantive reasoning is not enough. The Italian Court of Cassation had examined the request for a referral implicitly by looking at the merits of provisions referred to by the applicant. The ECtHR, nonetheless, held that ‘it is therefore not clear from the reasoning of the impugned judgment whether that question was considered not to be relevant or to relate to a provision which was clear or had already been interpreted by the CJEU, or whether it was simply ignored […]. The Court observed in this connection that the reasoning of the Court of Cassation contained no reference to the case-law of the CJEU’. In Schipani, the ECtHR did not consider it sufficient that the same Court of Cassation discussed the case law of the CJEU and thus seemingly showed that there was already an acte clair, without however expressing this in so many words and without mentioning the complainants’ request to refer.

In the present case, the ECtHR reasoned similarly and in a rather straightforward way: there has been a very precise request to refer (para. 72), the Court of Cassation only stated that it dismissed the appeal ‘sans qu’il y ait lieu de poser une question préjudicielle’ (para. 74) and without explicitly (or even implicitly) mentioning a Cilfit-exception (para. 75). The outcome of the present case seems logical, but one wonders how it fits into a line of cases in which the ECtHR has been more pragmatic.

Tell me less, national court: room for pragmatism

After the two Italian cases of Dhahbi and Schipani, the ECtHR has been more reluctant in finding a violation, at least until Baltic Master four years later. Possibly as a result of some criticism from CJEU judges (see below), the ECtHR has accepted various practices which at first sight do not differ substantively from the cases just discussed. In the light of a pragmatic approach in the cases discussed below, one especially wonders why the ECtHR has been so formalistic in the present case. This is because the first civil division of the Court of Cassation did refer (more or less the same) questions about the Product Liability Directive on the same day as the decision complained of. This is also acknowledged in the present judgment, albeit very briefly (paras. 64 and 80). The decision to refer was made in a case with a similar set of facts, the same President (Mme Batut) and the same respondent (Sanofi Pasteur) (for the CJEU judgment, see here). The main difference between these cases was that Sanofi Pasteur was found liable by lower French courts in the present case, but not in the case referred to the CJEU, because of the failure to prove a causal link between a defect in the vaccine and the harm. This shows that the Court of Cassation was not unwilling to refer. Rather, it seems, as the French government also argued before the ECtHR, that a referral in the present case was not necessary for the resolution of the case, possibly also because the Directive did not apply rationae temporis since the vaccine had been put on the market before the date of transposition. The ECtHR rejected this argument. It did so, because it was unclear for the ECtHR whether this was indeed the reason for the Court of Cassation (paras. 76-77). (I must admit that this reason is not put forward in a very clear way by the Court of Cassation, as the ECtHR also noted, and my knowledge of French and the specific legal area is not sufficient enough to grasp all details).

Several Dutch courts, for example, easily got away with a relatively short statement of reasons in Chylinski. The ECtHR declared the application inadmissible without paying any attention to the fact that the various Dutch courts complained of did not deal with any CJEU judgment or Cilfit, thereby ignoring its Dhahbi/ Schipani requirements. What’s more, it even considered that a preliminary ruling by the CJEU for settling the dispute would not have arrived on time.

 In Somorjai, the ECtHR applied an ‘arbitrariness’ standard and did not find a breach of Article 6(1), primarily because the applicant had not requested the Hungarian courts in a ‘self-contained and comprehensive manner’. This echoes some of the older case law of the ECtHR which is, however, not explicitly mentioned. In cases like John and Matheis  the ECHR ruled for example that there was no violation of Art. 6 ECHR because the applicant had insufficiently substantiated the importance of a reference for a preliminary ruling. In the present case, the ECtHR noted that the applicant had formulated precise preliminary questions (para. 72). As I will argue below, this heavy reliance on the parties does not fit easily with the way in which the preliminary ruling procedure has been envisaged in the EU Treaties and by the CJEU. It is not uncommon that the parties draft very elaborate questions but without any merit whatsoever. As such, this should not require a national court to tell more than it would normally do.

In Repcevirág Szövetkezet, the ECtHR clearly and explicitly stated that an implicit justification for not motivating may be sufficient even though the ECtHR did not seem to accept such a more pragmatic line in Dhahbi and Schipani. The ECtHR also refers to this judgment in the present case (para. 71).

In  Baydar, the ECtHR seems to go back to the pre-Ullens de Schooten case law and examines whether the refusal is ‘arbitrary or otherwise manifestly unreasonable’. The ECtHR applied a more procedural test and did not require an explicit reference to one of the three Cilfit-grounds. It looked at the entire procedure before the Dutch Supreme Court to assess whether the decision was arbitrary. What the ECtHR particularly valued was the fact that the case was dealt with by three judges and was preceded by a conclusion of an Advocate General to which the applicant could react in writing. With this procedural test, the ECtHR basically examined whether the Dutch Supreme Court had, in principle, properly looked at the case. It did not look at whether it is clear to the parties how the EU law aspects have been looked at and how they have been assessed. In Harisch, the ECtHR accepted the limited reasoning of the German Federal Court of Justice because of the extensive reasoning of a lower German judge, the Court of Appeal. The ECtHR said in its ruling in plural: ‘examining of the proceedings as a whole, the Court notes that the domestic courts provided the applicant with a detailed explanation…’. This quote shows that the ECHR does not only look at the procedure before one court, as in Baydar, but at the entire national judicial proceedings. The ECtHR has not taken such an approach in the present case. The ECtHR did quote from and discussed the opinion of the Advocate General (paras. 20 and 80), but did not see any reason to conclude differently or to adopt a more holistic perspective in line with Baydar and Harisch. It is not immediately clear why not, especially considering the preliminary reference of similar questions made on the same day in a parallel case.

Tell me less, national court: the possibility of a summary reasoning

In the present case, the ECtHR determined that the Court of Cassation had not declared the case inadmissible and had not concluded that it was not based on serious grounds. Hence, the ECtHR suggested that a summary reasoning could not be used in such an instance (para. 73 jo. 71). The ECtHR accepted a summary reasoning in the Dutch cases of Mothers of Srebrenica and Baydar  even though the reasoning employed therein completely ignored Cilfit or any other case law of the CJEU. The Dutch Supreme Court simply held in Baydar: ‘The grievances cannot lead to cassation. Based on section 81 (1) of the Judiciary (Organisation) Act, this requires no further reasoning as the grievances do not give rise to the need for a determination of legal issues in the interests of legal uniformity or legal development.’

One wonders how much different the reasoning of the French Court of Cassation is in comparison with the Dutch Supreme Court? What is more, the ECtHR seems to suggest that the Court would have an easier life, had it dismissed the case for not being based on serious grounds, without any further explanation. In a way, the ECtHR seems to encourage the highest national courts to tell less given its rather easy acceptance of summary reasoning. Such an unsatisfactory result is also visible when considering the earlier mentioned case of Baltic Master in which the ECtHR found a violation. This time the ECtHR (a chamber of three judges with Pinto de Albuquerque as president) did not accept a shortened summary reasoning because there was no legal basis in Lithuanian law for such summary reasons, such as art. 81 of the Dutch Judiciary (Organisation) Act. The Lithuanian court got a slap on the wrist even though it had stated that the interpretation of EU law was clear and thus finds an acte clair in the sense of Cilfit without explicitly labelling it as such. The Lithuanian court also referred to the complainant’s request for referral. In that sense, the Lithuanian judge provide a more substantive reasoning than the Dutch Supreme Court in Baydar.

In sum, the ECtHR’s acceptance of summary reasoning makes it very easy for the highest national courts to escape their referral obligation. It invites courts to take the easy way in complex cases rather than to justify why a referral is not necessary.

Tell me more, ECtHR: time for a Grand Chamber judgment

Sanofi Pasteur is the first level 1 judgment since the ECtHR arrived at an obligation for national courts to provide reasons for non-referral in Ullens de Schooten in 2014. The (registrar of the) Court does not specify the specific reason for this categorisation. It, nonetheless, seems that this was based on the clarification of the case law since the ECtHR gives an overview of the principles in paras. 68-72 and does not come up with something new and does not change its case law (importance level 1 is given to this judgment, which denotes that it of ‘high importance’ and makes ‘a significant contribution to the development, clarification or modification of its case-law’). The question is whether the present judgment truly clarifies the case law. As is evident from this commentary, the useful overview of principles does not camouflage that the case law of the ECtHR on this very issue goes in all directions and has been highly dependent upon individual judges involved. Some judges, such as Sajó and Pinto de Albuquerque, have in the past criticised the ECtHR and have proposed a more intensive test (which the latter of the two subsequently employed in Baltic Master), while others, such as Wojtyczek argued for a more pragmatic approach whereby the seriousness of the infringement is considered. This inconsistency and unpredictability calls for a clear Grand Chamber judgment putting an end to this casuistic and fluctuating case law so as to make clear how much room for pragmatism there is.  The Grand Chamber panel of five judges unfortunately rejected a request of the applicant in Harisch for referral to the Grand Chamber in September 2019.

 Should the Grand Chamber fundamentally change its approach? Not necessarily. The ECtHR rightfully expects the national court to make clear which Cilfit-ground was the reason not to refer. That does, however, mean that it should have been more strict in Baydar and thus require more from the national court in terms of reasoning, while being more pragmatic in Baltic Master or Schipani. It is important that national courts make clear which specific Cilfit-ground informed their decision not to refer, since there is a difference between an acte clair or éclairé and the exception that an answer to the question is irrelevant because it is not necessary for the solution of the dispute. If the latter situation exists, it may be that the national court has not considered the interpretation of EU law (yet). There might thus be still an outstanding question of EU law. This is different when the reason for non-referral is an acte clair or acte éclairé. It is important for litigants, but also lower court judges, to know how the highest national court qualifies a particular legal question. When the matter is considered clair that could be a clear discouragement for further litigation. These considerations are especially relevant from the perspective of the functioning of the preliminary ruling procedure and reasons of procedural economy. While related to Article 6 ECHR, one wonders whether this is still something for the ECtHR to expand on or more something for its counterpart in Luxembourg.

Tell me more, CJEU: time for the CJEU to step in

In addition to offering more clarity, there is every reason for Luxembourg to pronounce itself on the matter, even more so because it concerns an important EU procedure. There has been criticism on the ECtHR’s “interference” with this EU procedure. This is because the ECtHR has turned the preliminary ruling procedure into a mechanism safeguarding the individual right to a fair trial. By contrast, the CJEU has approached the procedure as an instrument of cooperation between courts. The CJEU determined in Cilfit that art. 267 TFEU ‘does not constitute a means of redress available to the parties to a case’. It has consistently held that it is entirely a court-to-court procedure, which is ‘completely independent of any initiative by the parties’. A duty for courts to provide reasons that is dependent upon (the quality of) requests of the parties to refer is thus especially questionably from the perspective of EU law, as the previous quotes show as well. It is perhaps not surprising that there has also been discontent among CJEU judges in Luxembourg. This can be derived from the fact that the ECtHR case law on this issue was one of the two topics discussed in March 2016 at a joint meeting of ECtHR and CJEU judges.

There have been calls for the CJEU to take the matter into its own hands and make clear whether a similar obligation for national courts to provide reasons on the basis of Article 267 TFEU and Article 47 of the Charter, the right to effective judicial protection. So far, the CJEU has however avoided to become involved in this delicate matter. In Aquino, the CJEU managed to circumvent questions of the Brussels Court of Appeal about this. In AFNE it seemed to arrive at a duty to provide reasons for non-referral, albeit it limited this obligation to the exceptional situation  of a declaration of illegality of national law in the area of environmental law.

There is thus need for a principled pronouncement of Luxembourg. This is obviously a delicate matter that the CJEU wishes to avoid in order not to upset the fragile balance with national courts. Nonetheless, the CJEU has recently ruled for the first time  in an infringement procedure that France had breached Article 267(3) TFEU because the highest court (the French Conseil d’État) had failed to refer to the CJEU. Another reason why more clarity from the perspective of EU law is needed is that legal practitioners are concerned with this issue. In the Netherlands, there have been various migration and criminal lawyers who have been trying to denounce the issue of short and summary reasoning for years. In addition to several complaints to the ECtHR, complaints have been made to the European Commission to provoke infringement proceedings.

In conclusion, the present judgment tells us a bit more about the expectations from the ECtHR as to the reasons that national courts should provide when they decide not to refer to the CJEU. At the same time, when viewed in the light of earlier case law, it reveals the inconsistencies and arbitrariness whereby national courts are sometimes asked to tell more and at other times to tell less. It is thus essential that not only the ECtHR, but also the CJEU, tell us all more about the intricate issue that has kept practitioners and scholars occupied.

Jasper Krommendijk is associate professor of international and European Law at Radboud University, the Netherlands. He is currently working on a Veni-research project funded by the Netherlands Organization for Scientific Research (NWO): ‘It takes two to tango. The preliminary reference dance between the CJEU and national courts’ (2017-2021). In this project he examines why national courts refer and what they do with the answers of the CJEU (see here and here for his most recent publications).