Epilogue: The Ever-Evolving Preliminary Legislative Procedure

Table of Contents: 1. Introduction. – 2. Preliminary ruling instead of direct review: interpretation with(out) (limited) facts. – 3. Anonymity before the law: direct or indirect (factual) identifiers removed. – 4. Access to written submissions: default generalised anonymous transparency. – 5. Connecting the dots: towards requests for additional legislation before the Constitutional Court of the European Union?

Abstract: This Epilogue to the Special Section on the 2024 Reform of the Statute of the Court of Justice of the EU brings together the various individual elements of the Reform. It assesses their cumulative impact on the preliminary ruling procedure and the judicial style of the Court. It engages with three elements in particular: what the likelihood that in the future, judicial review of EU legislation before the Union Courts is bound to happen primarily within the preliminary ruling procedure, means for the depth and the scope of that review, in particular in cases where complex, technical or scientific assessment of facts is necessary; what impact removal of all direct and indirect identifiers concerning personal data of all natural persons involved in a case have on the judicial style and level of abstraction embraced in deciding cases; as well as what the now Parliament-ordained generalised anonymous transparency, requiring by default the online publication of all written submission after the delivery of the decision on preliminary ruling, might mean for the judicial process and its outcomes. The overall conclusion is that the institutional and procedural choices made in 2024 confirm that the preliminary ruling procedure, originally conceived as a tool of bilateral judicial cooperation, has increasingly mutated into an essentially vertical legislative process.

Keywords: Court of Justice of the EU – the preliminary ruling procedure – Statute reform – direct actions – access to written pleadings – legitimacy.

1.   Introduction

Academic discussions following the 2024 Reform of the Statute of the Court of Justice of the European Union[1] have focused mainly on the partial transfer of the jurisdiction to hear requests for preliminary rulings from the Court of Justice (‘the Court’) to the General Court (‘the GC’).[2] Rightly so: it was indeed a major development. Additionally, the transfer – or, more accurately, the sharing – of jurisdiction in the six selected areas between the two Union courts can hardly be viewed as the final destination. It is but a testing balloon that will soon determine the direction the future judicial architecture of the EU shall take.

Accounts of the early experience with the new system, captured in this Special Section of the European Papers, are already revealing.[3] Nevertheless, it is far too early to draw definite conclusions. The fact that the GC delivered an answer in Gotek,[4] the first case ever heard under Article 267 TFEU by that jurisdiction, within nine months is an excellent result. Yet, that reference was among the first ‘selected to shine’.[5] It raised a question that could scarcely be characterised as complex. It was resolved on the basis of the written observations submitted by two interested parties, namely the Member State from which the question originated and the European Commission, without a hearing and without an Opinion of an Advocate General. For these reasons, it was perhaps not the most representative example of the complex, technical cases intended to fall within the GC’s new jurisdiction. 

As regards the central feature of this Statute Reform – the sharing of jurisdiction for answering requests for a preliminary ruling between the Court and the GC – I remain doubtful about the suitability of such a procedural arrangement. Sharing jurisdiction between more instances in a procedure the very purpose and legitimacy of which lie in the swift and uniform resolution of a matter of interpretation, which is itself an integral part of other national proceedings, is not an obvious systemic choice. Other means of alleviating the Court’s workload were conceivable. This is, however, not the place for rehearsing arguments set out elsewhere.[6]

Instead, this Epilogue shall examine other elements and consequences of the reform, but above all their cumulative impact. The overall conclusion is that taking the various components of the 2024 Statute Reform together and placing them in the context of the evolution of the preliminary ruling procedure over the past decade confirms the changing nature of that procedure. What was originally conceived as a tool of bilateral judicial cooperation is increasingly becoming an essentially vertical legislative process, with the latest developments quietly acknowledging, or potentially further accelerating, that trend. 

It is not something new to observe that, notwithstanding the diplomatic proclamations about ‘assistance’ or ‘dialogue’, the preliminary ruling procedure has, over the decades, evolved from a mechanism of ‘horizontal’ judicial assistance for individual national courts into a channel for ‘diagonal’ or even ‘vertical’ precedent-setting for the Union as a whole.[7] What is more striking is just how much the 2024 Statute Reform, in terms of procedural and institutional choices it made, acknowledges that evolution.

Three elements contributing to this transformation are outlined in this Epilogue. First, the 2024 Statute Reform is likely to entrench the preference, already voiced some time ago, for cases to come to Luxembourg on preliminary ruling rather than by direct actions. This choice has repercussions for the scope and depth of the judicial review exercised because, absent substantial procedural amendments to the preliminary ruling procedure, that procedure remains ill-suited for in-depth factual and contextual review of EU legislation. Second, while itself not the consequence of the 2024 Statute Reform, the by now generalised anonymisation of all natural persons in judicial proceedings before the Court – including the removal of all direct and indirect identifiers from the case file – nudges the Court towards abstract normative pronouncements unchecked (or even uncheckable) by the truncated (or altogether absent) facts of a case. Third, and rather intriguingly, the 2024 Parliament-ordained default online publication of all written submissions after the preliminary ruling has been delivered moves the system in quite the opposite direction from the anonymity drive. It generates a comparatively novel model of generalised anonymous transparency in judicial proceedings. Such transparency is typically a legitimising factor for democratic legislative processes, but it has not, however, traditionally been a hallmark of judicial proceedings, certainly not in the Continental legal traditions.

2.   Preliminary ruling instead of direct review: interpretation with(out) (limited) facts

In 2002, in Unión de Pequeños Agricultores,[8] the Court made the choice of not revisiting the ‘Plaumann test’[9] governing who is directly and individually concerned in order to have standing under Article 263(4) TFEU. Subsequently, the Treaty of Lisbon introduced a third category of legal standing into that provision to alleviate some of the concerns regularly voiced about limited access to the Union courts.[10] The essential features of the second scenario – namely, the one requiring direct and individual concern – nonetheless remained the same. Despite repeated invitations over the years from numerous academic commentators and several Advocates General to soften its stance,[11] the Court’s position remained essentially the same.[12] If anything, subsequent changes have tended to narrow further the already limited avenues for direct access to the Union courts by obfuscation,[13] turning the issue of the admissibility of an action into an increasingly arduous obstacle course for individual applicants.

The Court’s case law nonetheless insisted that the EU legal order provides a complete system of remedies.[14] Thus, any limitations on access in direct actions had to be offset by broader, or in fact rather unlimited, access via the preliminary ruling procedure. In the wake of the Court’s judgment in Unión de Pequeños Agricultores, the case law increasingly indicated that the appropriate forum for bringing cases questioning this or that element of validity of EU law was the preliminary ruling procedure rather than a direct action.[15] That choice was reinforced by a series of incentives that lowered, or even removed, barriers to bringing validity challenges when routed through national courts under Article 267 TFEU. For instance, in contrast to the strict two-month time limit applicable for bringing an action under Article 263 TFEU, no such temporal constraint exists under Article 267 TFEU.[16] Nor is there any requirement of direct or individual concern thereunder.[17] The sole filter is the baseline criterion of ‘relevance’ for the referring court, namely that the national court must be able to take the Court’s guidance into account in the continuation of the domestic proceedings.[18] Perhaps most strikingly, there need not even be a challengeable act in order for an EU measure to be referred, and potentially declared invalid, under Article 267 TFEU.[19] The result is a clear preference for one type of procedural avenue over another: strict limits in one type of proceeding, and minimal in the other.[20]

At the same time, it was not uncommon for questions explicitly referred as questions of validity to be reformulated by the Court as questions of interpretation, either of the same EU law measure, or as questions of interpretation of the ‘parent’ legislation under which the contested measure was adopted.[21]

In consequence, transitive legal issues, which could have been brought before the Union courts as actions for annulment under Article 263 TFEU, were instead channelled through Article 267 TFEU questioning validity, or even ended up (re)formulated as issues of interpretation.[22]

The extension of the Commission’s authorisation of glyphosate as an active substance, which arose in the parallel cases of Région de Bruxelles-Capitale[23] and Blaise,[24] offers an illustrative example of that preference. In the former, the Région de Bruxelles-Capitale – a federal entity under Belgian law responsible for environment protection – challenged, in an annulment action before the GC under Article 263(4) TFEU, the Commission Implementing Regulation that prolonged glyphosate’s market authorisation.[25] The ‘parent’ Regulation No 1107/2009,[26] under which the Implementing Regulation for Glyphosate was adopted by the Commission, and in particular the principles and procedure instituted by that Regulation, especially the precautionary principle, featured prominently in the arguments of that applicant region. The subject matter of the action, however, remained the validity of the Implementing Regulation. The GC nonetheless dismissed the action by order on the ground that the applicant lacked direct concern,[27] a conclusion which was subsequently endorsed by the Court.

In parallel, however, the Court was seised by a request for a preliminary ruling from the tribunal correctionnel de Foix (Criminal Court of Foix, France), which raised questions concerning the validity of Regulation No 1107/2009 in the context of market authorisation for glyphosate. The background of the national proceeding was rather unusual: Mr Blaise and several others entered shops in the department of Ariège (France) and damaged containers of glyphosate-based weed killers, as well as glass display cases. Those individuals were subsequently subjected to criminal prosecution for damaging the property of others. Before the national court, the accused pleaded the defence of necessity and relied on the precautionary principle, claiming that their actions were intended to alert retailers and customers to the dangers associated with selling weed killers containing glyphosate, to prevent such sales, and to protect public health. In the context of that case, the first-instance criminal court submitted a series of questions on the validity of Regulation No 1107/2009, in an order for reference only a few pages long. Before making its reference, the national court had not, quite understandably, conducted any scientific assessment or established any facts in relation to glyphosate or the procedure for its approval.

In a truly complete system of remedies, a direct and more focused action – such as that brought by the Région de Bruxelles-Capitale, which was, with all due respect to the first-instance court in Foix, far better developed and extensively reasoned – was not permitted to proceed before a Union court vested with full jurisdiction to hear it. At the same time, a case in which the glyphosate authorisation issue was, to put it mildly, underdeveloped in both qualitative and scientific terms (because a first-instance criminal court in Foix is not, for obvious reasons tied to the purpose and scope of the proceedings before it, likely to carry out in-depth fact finding or assess complex scientific evidence) was allowed to proceed directly to the Grand Chamber of the Court through a procedural avenue, the preliminary ruling procedure, where facts and their evaluation lie exclusively with the national court. 

There is no shortage of other technical, often scientific, areas of law in which similar patterns have emerged: regulation of tobacco products;[28] various elements of environmental protection, particularly climate-change litigation[29] or plant protection and eradication of dangerous species;[30] air-pollution disputes by car traffic tied to the aftermath of the dieselgate scandal;[31]various issues of the banking union;[32] or data protection[33] to name only a few examples. In such cases, it matters little whether the difficulty lies in parallel (national and European) challenges to secondary or tertiary (typically implementing) legislation or in uncertainty about the stage within various composite EU/national procedures at which a binding and therefore challengeable act of direct concern for a given individual arises. The common thread is that, in those and other cases, there has been a distinct trend towards a restrictive interpretation of what constitutes a challengeable act, direct concern, or individual concern, in direct actions, with the invitation for those legal issues to be submitted later on as requests for preliminary rulings. 

Case law concerning challenges to EU soft law has pushed the same trend to its extreme. In Belgium v Commission,[34] the Court held that true soft law instruments cannot be reviewed through an annulment action under Article 263 TFEU. Absent binding legal effects, there is no challengeable act and thus nothing to review. Yet, subsequently, a true soft law instrument was annulled by the Court, without much further discussion, in a preliminary ruling in Balgarska Narodna Banka.[35] In Féderation bancaire française[36] (FBF), the Court clarified that the annulment of a soft law measure is indeed possible under Article 267 TFEU. One of the arguments advanced in FBF in favour of allowing a challenge to validity under Article 267 TFEU of an EU law act which, under Article 263(1) TFEU, is not even considered a challengeable act, was the existence of the Union’s ‘complete system of remedies’ established under the EU Treaties.[37] Logically, if the Court wished to keep a say in the judicial review of EU soft law instruments and thereby, indirectly, in the supervision of acts of EU law institutions and bodies, it had to allow for some type of access to it in challenges to such instruments, even at the cost of a coherent system of remedies.[38]

There are certainly sound arguments in favour of the preference for indirect access to the Union courts via the preliminary ruling procedure in the manner it has been promoted by the Court for the last two decades.[39] It enables national judges to act as filters, weeding out repetitive, weak or even spurious challenges to validity. It also allows a case to mature at the national level before proceeding to a review before the Union courts. The EU legislation being challenged is not simply there, freshly adopted and in the abstract, with its impact and real-life consequences still unknown. Rather, it has already been applied to individual cases, revealing its effects and visible impact as a legal rule. Ideally, different takes and approaches will have also been tested before the national courts, providing a more developed and contextually grounded basis for the eventual assessment at the European level.[40]

By analogy to the constitutional/judicial review carried out by national (constitutional) courts, the difference is one between an abstract (when compatibility of laws is being assessed without those applied first) and concrete (when the same exercise is carried out incidentally, after the contested legislation has already been applied to individual cases) review of constitutionality/legality. It is the difference between a legal rule ‘as stated’ in the act, as opposed to ‘as applied’ in subsequent practice.

There are, however, significant costs associated with this mode of case flow. Four merit particular attention. First and foremost, within the preliminary ruling procedure, it is for the national court to ascertain the facts and for the Court to interpret EU law based on those facts and their assessment submitted by the national court in its order for reference. That traditional division of tasks may, however, become problematic in cases of complex technical or scientific assessments. In such circumstances, a national court is unlikely to undertake the scientific assessment itself, certainly not before submitting a request for a preliminary ruling to the Court. Were it to do so, the need, and even the likelihood, of submitting a request to the Court would diminish considerably. Moreover, a referring court is likely to assume, rather reasonably in my view, that it is for the ultimate interpreter of EU law to clarify the technical or scientific dimensions of the issues it requires for the purposes of delivering its ruling. In the context of highly technical, complex legislation, questions of interpretation and the underlying factual or scientific determinations thereafter are often intertwined.[41]

However, when it comes to preliminary rulings, the Court never assesses facts and never calls expert witnesses in practice. Naturally, the Court will rely on the information that the (interested) parties, in particular the European Commission, or the Council and Parliament if an issue of validity of legislation which was approved by them is raised, provide to it. But an institution defending its own position which it subsequently transcribed into the legislation it adopted hardly constitutes an independent scientific body. It is defending its own legislation.

Ultimately, therefore, under such a division of competence in the preliminary ruling procedure, certain cases may devolve into a game of ‘hot potato’, where the Court makes its interpretation conditional on a number of assertions that are left ‘for the national court to verify’. That is the better scenario. Alternatively, an interpretation of EU law may be adopted without any genuine, technical evaluation or assessment ever taking place at any stage, with judicial review resting solely on tacit, or even explicit, deference to the broad discretion enjoyed by the EU bodies or administration.[42]

Incidentally, the subliminal unease with undertaking factual and scientific assessments in the context of the preliminary ruling procedure may be indirectly reflected in the previously mentioned tendency to reformulate questions of validity as questions of mere interpretation.[43] A direct challenge to validity – in particular, again, in technical or scientific areas – requires much deeper discussion of the facts than providing the interpretation of more abstract principle of the parent legislation. Interpretation, particularly when conducted at a high level of abstraction, may be much more selective in what must be articulated for the purposes of a judgment. 

Second, there is the factor of time. For cases to filter through the national levels, the transposition of an EU measure, if any is required, must first take place. The measure must then be applied to individual cases, and only thereafter may judicial proceedings arise through which a challenge is conceivable.[44] This process takes years, during which time the soundness of an EU measure may be contested, potentially even in multiple Member States, with considerable costs to legal certainty and stability. By contrast, in a functioning system of centralised judicial review, the validity of a measure may be determined at a single instance and at a relatively early stage. Such a determination would render most subsequent attempts to raise the issue moot, enabling national courts to rely on the central judgment when dismissing applications for further preliminary rulings on the same point.[45]

Third, there is the issue of participation and voice. In a direct action before a Union court, the author of the contested act, as well as any EU bodies or agencies that participated in its adoption, can present their views. Conversely, in a decentralised and diffuse world of multiple national challenges, EU institutions or agencies are unlikely to be parties to the national proceedings, or even aware that such challenges are taking place.[46] It is true, however, that this problem will disappear if the silent assumption is dissuading national courts from even starting the discussion on the validity of any EU measure, because if they do, there is often nobody they can discuss that issue with at the national level.[47] Should a national court nonetheless insist, they ought to refer the matter to the Court, where the EU institution or body which authored the act will be heard. Yet this dynamic may, in turn, exacerbate the trend towards references that are not fully thought through, as that the national court will not have had the benefit of real discussion about the legal problem within in the national proceedings leading to the reference.[48]

Fourth and finally, alongside the just outlined systemic considerations, namely, what is optimal for the EU’s system of remedies, lies the issue of individual rights. From both a symbolic and practical perspective, a direct action is in the hands of the individual applicant. By contrast, a request for a preliminary ruling remains solely in the hands of a national court, even where the proceedings are before a court of last instance and despite the recently reaffirmed obligation on such courts to provide reasons when refusing to make a request for a preliminary ruling.[49] Metaphorically, therefore, the heart of the EU judicial function – full-jurisdiction review of EU law acts, as to both law and fact – appears to be a remarkably cold heart, kicking potential review as far down the road as possible. Passionate abstract discussions about the virtues of a complete system of remedies aside, in concreto, for individual applicants, the system may appear neither complementary nor complete. 

The doubling of the number of judges at the GC after 2015 could have provided the Court of Justice of the EU with the additional capacity needed to handle technical cases and cases requiring independent scientific assessment, thus alleviating some of the issues just described. In fact, it was for this reason that, when writing in a judicial capacity, I joined the calls for a moderate and partial opening of the criteria of direct and individual concern, at least for certain kinds of applicants.[50] Enabling the GC to conduct direct review within the jurisdiction it already possessed, while the Court retained competence to hear any appeal on points of law, could, in my view, lead to deeper and timely judicial review, particularly of implementing EU legislation where technical expertise and scientific evaluation are indispensable. 

The key consequence of the 2024 Statute Reform is that the additional judicial resources created at the level of the GC are likely to be gradually redirected towards its handling of requests for preliminary rulings rather than direct actions. The direction charted by that reform implies a willingness on the part of the Court to confer further (technical) areas of jurisdiction on the GC, albeit under the centralised control and allocation of the Court itself. 

That is nonetheless likely to mean that cases which could have been brought as direct actions – and assessed on the basis of fully established facts and with proper technical depth, obtained with the help of the parties, witnesses or expert witnesses – will instead arrive as requests for preliminary rulings. In the latter cases, both the national court and the Court/GC may assume that the other will undertake the necessary factual work. Yet, even if the formal position remains that the Court does not make factual findings in preliminary ruling procedures, the reality is that the facts as presented (or rather not presented) by the national court decisively influence the legal interpretation subsequently reached. 

Judging without all the necessary facts is problematic, prone to result in hasty decisions which soon require reconsideration. Even if the task is framed as a mere ‘interpretation’ rather than ‘application’, that interpretation still operates on the basis of factual and contextual assumptions generated from the set of initial facts. The facts, or their absence, thus decisively shape the type, scope and depth of the interpretation that will be given to a legal rule. Finally, interpreting, or rather articulating, the law without a sufficiently developed factual context of individual cases moves the exercise away from adjudication and closer to an activity belonging to a different discipline altogether: legislating.

3.   Anonymity before the law: direct or indirect (factual) identifiers removed

The Court’s recent push for the anonymisation of all natural persons in preliminary ruling proceedings,[51] together with the instructions to the national courts not to include in their orders for reference any factual information or details that could enable the identification of a natural person concerned by those proceeding,[52] and with anonymity to be maintained not only in all written communications with the Court, but also in oral hearings,[53] has likewise not facilitated the provision of deeper context. Combined with the (already) long-standing instruction to keep the orders for reference as short as possible for the reasons of translation,[54]national courts are nudged to submit already edited facts in their orders for reference. The contextual understanding of the facts and context remaining is made no easier by the dystopian proliferation of ABs, XZs, or YTs in the narrative part of each case involving natural persons.[55]

The facts of a case involving natural persons tend to be personal data relating to those persons[56] which may lead to the direct or indirect identification of those persons. The more such facts/data are provided, the easier indirect identification becomes. Conversely, omitting such information forces deliberations onto a higher plane of abstraction. The core mental exercise of judicial subsumption becomes difficult when facts/data are missing. Any interpretation must then proceed on the basis of silent or acknowledged hypotheses. Above all, in adjudication, details matter: a single crucial fact or piece of background information can considerably change the entire course of a case. The metaphorical mental train is put on completely different tracks. 

The business of judging has traditionally consisted of subsuming highly detailed and concrete sets of facts under a pre-existing abstract and general legislative framework. First, the facts are a means – indeed, in more realist or critical legal traditions, the onlymeans – of controlling the correctness of past judicial assessments and of anticipating future ones. It is no coincidence that the ‘distortion of evidence’ or the ‘incorrect ascertaining of the facts of the case’ is considered a deadly judicial sin, leading to almost automatic annulment of a decision on appeal, or even on second appeal, that is to say, even in proceedings which are not supposed to be concerned with ascertaining or interpreting the facts at all.

Second, and perhaps more importantly, the facts of a case not only define its scope but equally constitute the legitimising foundation for any judicial pronouncement on, and in, a given case. In a modern constitutional system, facts matter also from the perspective of the separation of powers. Certainly, it is impossible to suggest, today, that the tenets of say Article 5 of the French Code civil[57] or § 12 of the Austrian Allgemeines bürgerliches Gesetzbuch[58] remain applicable,[59] if they in fact ever truly were.[60] But still, even before the highest national courts, and likewise before the Union courts, at some level, the facts of the individual case as presented remain the basis for judicial intervention. The authority that does not need an individual case in order to make a pronouncement on a given issue tends to be the legislature, but not the judiciary.

Third, the facts of a past case matter not only for legitimising the judicial decision in that case, but also, especially for the highest, precedent-setting courts, for future cases. In precedent-based systems, the facts of an earlier case do not themselves form part of a future ratio. But they define the reach of the newly established rule. Conversely, the fewer the facts and the less the individual context is known, the broader the normative reach of any new rule judicially created. Indeed, it is no coincidence that in systems attributing binding force to prior judicial decisions, typically the common law systems, the reach of a precedent and the ratioit generates is constrained by the facts of the previous case. Only with, often detailed, knowledge of those facts can discussion about the next case unfold: is the next case identical, or at least sufficiently similar on its facts, such that the precedent must be followed (fully or in part), or is the case to be narrowed, distinguished or otherwise treated differently.[61]

Intriguingly, civil law systems – which have, on the one hand, dogmatically denied any broader rule-making role to judicial decisions – have, on the other hand, functionally allowed the highest judicial courts to formulate pronouncements of breadth and scope unimaginable to a common-law judge. The trick enabling this is the so-called ‘legal sentence’, ‘legal proposition’, ‘Leitsatz’ or ‘massima’.[62] That is a condensed rule of behaviour, typically framed (to a common-law lawyer, strikingly) at a high level of abstraction, akin to a legislative proposition, inferred from a case and formulated either by the reporting judge in a given case or by the editor/research service of a court. In a number of civil legal systems, such legal propositions started operating much like legislation, applied on their own, detached from the original case. Crucially, however, when a legal proposition is detached from the facts and context which gave birth to its creation, its reach grows exponentially. It becomes, in functional terms, legislation, albeit legislation originally crafted by a court. 

None of this is really new. Judicial pronouncements of the Court in the preliminary ruling procedure have traditionally been delivered at a comparatively higher level of abstraction, with the resulting generalised ‘legal proposition’ readily detachable from an individual case, whether through formulations in the operative part of the judgment or through individual statements lifted from the reasoning and circulated freely across cases in cluster citations. 

The point made here is rather that the robust anonymisation now applied to all natural persons, together with the removal of any direct or indirect identifiers relating to such persons, is bound to further nudge judicial pronouncements in preliminary rulings towards even greater abstraction. However, articulating legal rules in an abstract and anonymous manner, largely detached from the concrete cases that ought to ground them, and leaving their later application to other institutions, namely courts, tends to be the role of the legislature. The (traditional) legitimacy of the judicial function lies in the detail of the individual case.

4.   Access to written submissions: default generalised anonymous transparency

The issue of anonymisation seems to pull in the opposite direction from another change introduced by the 2024 Statute Reform: the default online publication of all written submissions received by the Court in preliminary ruling procedures after the closure of the case. The limited academic attention[63] paid to this aspect of the Reform may partly be due to the fact that it did not feature in the Court’s original Request. It was introduced in the legislative discussions only at a later stage by the European Parliament. Speculation as to whether the Parliament, or rather certain political factions within it, had grown impatient with the successive failures of the Court to ‘put its own house in order’ by establishing a comprehensive access to pleadings submitted to it, is best left to others. It is, nonetheless, true that the pre-2024 system[64] was a singular one: a third party interested in access to part of the judicial file, typically the written pleadings of an interested party, could not address the Court itself, which is the natural controller of the judicial file,[65] but was instead sent off to one of the (interested) parties to the case, namely the Commission.[66]

What is clear, however, is that extremes produce counter-extremes. One singular regime has now been replaced by another equally unique regime, which is, in comparative terms, again an outliner: a system of online default generalised anonymous transparency. 

First, the publication of all written submissions in a preliminary ruling procedure is to be carried out, by default, on the Court’s website once the case is closed. Interested parties who submitted written observations in a case may object to the publication of their submissions, but they must do so explicitly and within a time limit. The data available indicate that in the majority of cases, objections to the publication of pleadings have already been made by one or more of the parties.[67] As far as the objections against the publication of pleadings by the Member States are concerned, it is intriguing to observe that some of those objecting most frequently are the very Member States which, in their earlier submissions in access-to-documents litigation, had consistently advocated for greater openness and transparency on the part of the EU institution or bodies.[68]

Second, the regime is generalised in the sense that it appears to apply to all preliminary ruling procedures, irrespective of their subject matter or the identity of the interested party who made the submissions. Yet is there truly the same public interest in accessing the pleadings of, say, an asylum seeker, a VAT fraud suspect, or a consumer defaulting on his or her mortgage payment, as there is in knowing the position of a Member State or an EU institution in a rather abstract dispute over EU competences? Traditionally, the key ingredient in deciding whether to grant access to a judicial file or parts thereof by a court hearing the case tends to be the weighing of the competing interests of transparency and access, on the one hand, and confidentiality, on the other, in each individual case or at least within a given category of cases. Under the new system, however, all this is replaced by a generalised presumption that, regardless of the case or of the party concerned,[69] transparency at EU level invariably outweighs all other concerns.[70]

Third, in contrast to the overall direction suggested by the previous two points, what is to be made accessible are anonymisedsubmissions, with both the direct and indirect identifiers of the case or natural persons involved required to be removed.[71] This is where the choice made by the European Parliament for the Court becomes particularly intriguing, both at the technical and philosophical levels.

On the one hand, at present, the Court has now publicly committed itself to the pledge of judicial anonymity, thus generating the (legitimate) expectation of natural persons that their cases can proceed to the European level without their identities ever being revealed. However, in order to fulfil that promise, must the Court not only leave out all direct and indirect identifiers from its own decisions, but also act as a censor of the parties’ submissions? Irrespective of instructions issued, parties will inevitably leave a couple of (in/direct) identifiers in their submissions. Must those be erased, thereby altering submissions actually received by the Court? Or will they be kept intact, at the risk of the Court incurring liability for failing to safeguard the anonymity of a natural person? There is no doubt that the Court is the controller of any personal data published on its own website.[72]

On the other hand, what is the purpose of disclosing judicial documents apart from or in addition to the final judgments of a court? It is fair to assume that the objective is public control, public scrutiny and the transparency of the judicial process, motivated by an implicit or explicit desire for greater openness and overall transparency. Yet, what meaningful control can be exercised, or what transparency achieved for that matter, if elements of the information, or even significant parts thereof, are withheld? 

At this stage, the argument can be taken one level higher in its abstraction. The transparency pursued by such a system becomes a ‘second order’, or more ‘abstract’ openness: what will be disclosed is the structure of reasoning and the reaction of the Court to the arguments put before it, in the abstract. It is not the disposition of a concrete case that is supposed to be scrutinised. Such oversight can be exercised without knowing all or even any personal information about the individuals concerned, whether they are parties, witnesses, or any other third persons named in the national proceedings. Their generic qualities and characteristics suffice.[73]

Furthermore, one could equally argue that, unlike a national court, which is there to resolve the individual case of an individual (natural) person, and whose exercise of public power therefore requires close scrutiny, the role of a Union court, in particular in the preliminary ruling procedure, is different:[74] it is to set a precedent for the entire European Union. That function warrants a greater degree of openness and knowledge of the submissions of each interested party.[75]

All these arguments are possible, and some may even be correct. The point being advanced here is that the new regime of anonymous transparency in the publication of written submissions creates a unique contradictio in adjecto, for which comparative parallels are not readily available. As far as the practice in Continental Europe is concerned, more emphasis is on fostering the value of open and candid exchanges in court, with parties’ submissions disclosed only once the case has concluded and typically only upon an individual request supported by a legitimate interest in the disclosure. Common law systems are generally more open in that regard, allowing access to parties’ submissions more broadly, with some, particularly in the United States, even making such submissions available in their legal databases. That said, those systems also take the idea of full disclosure seriously, without the censor’s black pen to remove personal information and/or all factual identifiers capable of leading to the direct/indirect identification of a natural person. Even the European Court of Human Rights, arguably at the forefront of this kind of openness, grants access only upon an individual request made in individual cases.[76]

Moreover, such a vision of generalised anonymous transparency is likely to further impact both the process of the preliminary ruling procedure and its outcomes. 

As to the process, there are good reasons why the majority of domestic legal systems proceed from the assumption of generalised judicial confidentiality, both during the judicial process and after its closure. The purpose is to allow for candid, thereby open and useful, exchange in both the written pleadings and in the courtroom at the oral hearing. This is not only in order to protect the private sphere of litigants, who will, by necessity, be forced to disclose a lot about their life if they are to argue their case at the appropriate and convincing level of factual detail. Another reasons for the confidentiality in the courtroom is rather utilitarian, pertaining to the judicial (self)interest of typically higher jurisdictions called to (co)shape the law: by providing a ‘free space’ for the parties, a full and open dialogue ensues, which can then assist the court in finding an equitable and reasonable solution. 

By contrast, if the parties and their legal representatives are under a constant public supervision of everything they write or say,[77] knowing that an ‘Online Miranda Warning’[78] will effectively apply (‘Anything you say in this court might be used against you by the public’), the practical utility of such pleadings or hearings for the court and the judicial process may diminish. That is, again, one reason why most jurisdictions neither grant default, generalised access to the written submissions, nor allow external recording or streaming of the court hearings. From a pragmatic standpoint, what serves a court best is a frank exchange with the parties, where the legal representatives feel able to genuinely engage with the bench, answer all its questions candidly, and thereby assist it in the judicial process. If the legal representative knows that he or she is under constant online supervision, they will stick to the pre-prepared script and refuse to engage further. On a similar note, defending a client and/or assisting the court is a discipline distinct from ‘performing for the cameras’ or using the courtroom exchanges to make a point to quite a different, external audience.

As to the outcome, will the default publication of all written submissions induce greater judicial engagement with the arguments of those parties, resulting in better-reasoned judgments and, ultimately, greater legitimacy for the Union courts? Paradoxically, the opposite might equally occur. First, why should the Court continue reproducing and reacting to the submissions of the parties, however already shortened at present for the reasons of translation, if they will, in any event, be publicly available? Anyone can simply read for themselves what the individual views were, thus saving the Court the time and translation costs spent on reproducing parties’ views in decisions. All that is required is for the Court to authoritatively opt for one of the options on the table available before it. Second, if the Court’s potential (mis)understanding of the parties’ views becomes subject to public scrutiny, the incentives for referring to and engaging closely with those submissions may diminish.[79] That may not only be due to the fear of later public criticism for misrepresentation of the parties’ views, but also simply as a consequence of parties’ own (self)censorships: their submissions may cease to contain anything genuinely useful or interesting in the first place.

In summary, the new system of default generalised anonymous transparency may perhaps lead to enhanced quality in the judicial process and its outcomes, and more broadly improve public understanding of, and confidence in, that process. Yet it might equally lead, in some cases, to a (self)imposed censorship in both written submissions and oral hearings, with the parties stating merely the bare minimum necessary and leaving all the rest to the Court (‘Débrouillez-vous!’). Undoubtedly, the public will see more of the documents presented within the judicial process. Whether that will lead to greater understanding of, and trust in, that process, or (rather) provoke constant scrutiny and renewed challenges – and whether, ultimately, it will end up generating greater legitimacy for the Court and the Union’s judicial process – remains to be seen. There is a certain (cynical, but still true) wisdom in the old quip, apparently incorrectly attributed to Otto von Bismark, that laws are like sausages: it is better for the general public not to see how they are made.

To those fears, justified or not, one might always respond[80] with a value-based override: before the Union courts, nothing shall be kept in the judicial shadows! If parties are unwilling to state their arguments in daylight, they ought to remain silent. That is indeed a clear and principled message. The point made here is simply that ‘putting everything online’ is not without side effects. 

Finally, and in any case, for the issue central to this Epilogue, what is equally clear is that the ex ante elaboration of relatively abstract and anonymous rules, detached from the concrete specifics of any individual case, made through a transparent and open process, in which individual preferences voiced and the input given are to be fully displayed and publicly accountable, is characteristic of the legislative process, and not necessarily of the judicial one.

5.   Connecting the dots: towards requests for additional legislation before the Constitutional Court of the European Union?

First, sharing jurisdiction to give preliminary rulings with the GC signals further verticalisation of the preliminary ruling procedure. Furthermore, in structural terms, the pooling of judicial resources with the GC for that particular type of procedure is likely to eliminate the practical need to nuance or revisit Plaumann in order to achieve a more equitable distribution of the docket between the Union courts. Yet, the continued preference for cases to reach Luxembourg under Article 267 TFEU rather than Article 263 TFEU impacts the nature of the judicial review that will be carried out in preliminary rulings. Abstract interpretative statements with limited facts are, however, not always a suitable replacement for a detail-sensitive judicial review of legislation, particularly in cases concerning technical or scientific assessment, which are far from rare in EU law.

Second, the generalised anonymity before the law instituted by the Court for all natural persons further accelerates the trend towards ‘factless’ judging. Names, personal information, direct or indirect identifiers are all to be removed, or in the act of self-censorship by national courts and the parties to the original dispute, not even submitted to the Court in the first place. This pushes the final judicial product – already quite abstract in some cases – to even greater abstraction. Alas, with limited facts only, it might be difficult to foresee the ultimate impact of a legal rule being interpreted or created.

Third, all of this will now be happening within a process in which the external input is to be accessible and accountable: the initial order for reference submitted by the national court; the written submissions of the interested parties; the oral hearing in Grand Chamber cases, which will be web-streamed; and, finally, the publication of the decision of the national court after the preliminary ruling. Even if anonymous, all shall be transparent in the end. 

Putting all these elements together yields a rather interesting picture: a body called upon, in a transparent deliberative process in which the value preferences of its participants are to be disclosed to the public to enhance the legitimacy and overall accountability of that process, to establish fairly abstract rules that will later be applied by someone else to individual cases, and with only that subsequent body possessing full knowledge of the facts and circumstance of the individual cases. If only we had a name for such a sui generis process. Wait, was it not legislating? 

The proposition that the Court has long acted as a de facto legislator is certainly not a new one. Nevertheless, what is fascinating about the 2024 Statute Reform is, if all its elements are taken together and then combined with the evolution of the preliminary ruling procedure over the last few years, how much the essentially legislative function of the Court is at present being openly acknowledged through the institutional and procedural choices adopted in 2024. But even there, a sceptic might conclude that nothing truly new is happening, save for the form finally being brought into line with an already existing substance. 

Perhaps all that is what the discussed ‘constitutionalisation’ of the Court is supposed to mean. After all, are the elements just outlined not characteristic of a constitutional court deciding on abstract review of constitutionality, institutionally akin to a (negative or positive) legislator? Is it not what the future is supposed to hold for the Union judiciary, with the Court turning into a Constitutional Court of the Union, and the General Court becoming the Union’s Council of State?[81]

Constitutionalisation is no doubt a catchy slogan. But as a notion, it is a jellyfish. In terms of designing a judicial system and attributing jurisdiction within it, nobody has ever been able to define where the borderline between ‘constitutionality’ and ‘mere legality’ lies. With legal issues and cases moving freely between the two sets,[82] what over half a century of national experience with constitutional adjudication in Europe has taught us is that the notion of ‘constitutionality’ is unable to demarcate any comprehensive division of jurisdiction between a constitutional court and ordinary courts.[83]

This brings up the issue of very different institutional setups in constitutional review and constitutional justice across Europe: what is a constitutional court, and what does it do, in particular in contrast to supreme courts? If there is at least one element common to the world of mostly specialised and concentrated constitutional review in Continental Europe, then it is that the task of safeguarding uniform interpretation of the law is the job of national supreme courts within their respective areas of jurisdiction. By contrast, the role of constitutional courts tends to be the protection of constitutionality and above all of human rights, often exactly at the expense of the uniformity of national established case law, because effective protection of constitutional rights in the individual case requires it.

In this regard, intriguingly, the newly-worded Article 62 of the Statute of the Court foresees that the decision of the GC rendered on preliminary ruling can be subject to review by the Court where there is a ‘serious risk to unity or consistency of Union law’. That is, however, at least in national judicial systems, the textbook example for the admissibility of extraordinary remedy to a supreme court, not conditions for access to a constitutional court. The latter ones tend to be construed around the plausible claim of violation of constitutional rights. Or is there now supposed to be two-layered unity and consistency of Union law: the ‘normal’ or ‘merely legal’ unity of EU law, that causes the national court to make a reference to the GC, in order to safeguard uniform interpretation of EU law across the Union, but then (or precisely because of that in fact?) comes the second layer of constitutional ‘super-unity’ of EU law, which makes the case to advance to the Court? 

It is correct that such musings are only of limited relevance to the current repartition of competence in the six areas shared on preliminary rulings, where it will be for the Court within the system of ‘guichet unique’ to certify, in each individual case, what is a case of ‘constitutional significance’, by simply not delegating it to the GC in the first place. In comparative terms, again intriguing is that what amounts to a constitutional issue is normally decided ‘positively’: a constitutional court picks up a case or an issue after it was dealt with by (an) ordinary court(s) and it thereby becomes a ‘constitutional’ one. By contrast, in the current EU procedural set-up for areas of shared competence for preliminary rulings, what is an EU constitutional issue shall be defined ‘negatively’, by the ‘constitutional exclusion’ of the GC. Apart from the rather non-flattering implicit message towards the GC inherent in such division of competence,[84] similar instances of ‘jealous constitutionalism’ are not easily to be found in constitutional court designs.[85]

However, if, at some rudimentary level, the job description of a constitutional court were to be defined as examining the compatibility of lower laws with the higher or basic law or constitutional foundations of that system, would anybody be seriously claiming that the Court have not in fact been doing exactly that job within the European Union and previously the Communities for decades already? 

Certainly, the argument advanced is not that the Court is getting a new competence, but that its judicial burden shall be alleviated so that it can become more of a constitutional court. But doing more of what, exactly? More abstract, general, value-based pronouncements and reasoning, not to be concerned by ‘mere’ or ‘ordinary’ EU law any more? Becoming the guardian of fundamental rights in Europe, entering in a (renewed) human rights dialogue with national constitutional courts and the European Court of Human Rights? Is all that what the EU legal system really needs (today)? 

The system of EU law is composed of several constitutional documents (at present essentially the two Treaties and the Charter of Fundamental Rights) and then masses and masses of secondary, often detailed and technical, legislation. Within such a system, the legitimacy for judicial pronouncements has traditionally been based on achieving and maintaining a reasonable degree of uniform interpretation of that entire body of law. While, naturally, the protection of fundamental rights acts as a limitation to any and all of those uniformity endeavours today, it has never been the primary aim for cases to come to Luxembourg.[86]

By contrast, human rights protection has been the central job description of national constitutional courts and the European Court of Human Rights. That is, essentially, the foundation for the frequently misunderstood statement that the Court is not a human rights court.[87] That never meant that the Court would not uphold fundamental rights. It just meant that in the structure and logic of the cases coming to Luxembourg, as with the national highest ordinary jurisdictions, fundamental rights have been acting as limits, not as the primary purpose and legitimising element for adjudication.

What work should be entrusted to a ‘constitutional court’ within a system which derives its purpose and legitimacy from securing uniform interpretation of masses and masses of essentially detailed technical secondary legislation? Should the Court gradually delegate (all of) that task to its lower court, the GC, while seeking to wedge itself into the already ‘crowded house’[88] of fundamental rights protection in Europe, where values, emotions, and contestations run higher and higher? Or should the Court be left to focus just on inter-institutional disputes and policing EU competence in abstract terms, making clear the Union is one of values and constitutional principles, but eventually detaching itself from the Member States’ courts and cases? 

Such institutional choices are possible, even if, for the reasons just explained, the functional needs of the EU legal system and the ensuing systemic legitimacy for EU Courts lie elsewhere. The question remains whether such institutional choices are equally wise in the Union of today. As to their tone and direction, they are reminiscent of the optimistic, constitutionalising language of the Union of some twenty years ago, but much less so of the present Union covered in self-doubt and facing multiple institutional challenges, including those levied against the Court itself.[89]

Within such an environment, the learned judicial wisdom has been not to underestimate the elegance and strength of seemingly technical solutions to narrowly conceived cases. Their structure is inductive, their logic functional, and the solutions they embrace, even if potentially hiding quite some value choices, are more difficult to contest. The latter statements are bound to be immediately attacked by the (post)critical streams of legal thought, insisting that all of this is just window-dressing hiding the real conflict of values and political convictions, which must be revealed and acknowledged. That is perhaps true and certainly good material for an academic article. But it is quite a different matter for a court to willingly chart a similar path, clamping its legitimising cord to factual and detailed cases that have been providing it with quite some degree of (overall institutional) legitimacy, and instead boldly embarking on the tumultuous sea of abstract values, human rights, general principles, and, in this sense indeed, constitutional adjudication.