1.

As a line from Game of Thrones has it, ‘what is dead may never die’. Thus, perhaps with the exception of White Walkers, what is dead also cannot be killed. However, can something that has never been alive (or rather never came into existence as a binding EU-law act) be annulled (or rather declared invalid) by the Court of Justice on a preliminary ruling? Alternatively, can the Court provide (binding) interpretation of a non-binding EU measure?

2.

In 2017, the European Banking Authority (‘the EBA’) issued Guidelines on product oversight and governance arrangements for retail banking products. ( 2 ) Thereafter, the French Autorité de contrôle prudentiel et de résolution (Authority for Prudential Supervision and Resolution) (‘the ACPR’) announced in a notice that it complied with those guidelines, thus making them applicable to all financial institutions under its supervision. The Fédération bancaire française (French Banking Federation; ‘FBF’) has sought the annulment of that notice before the referring court, claiming that the EBA did not have the power to adopt those guidelines.

3.

The present case has several layers. On the one hand, there is the issue of whether, by adopting the contested guidelines, the EBA went beyond the scope of its powers under Regulation No 1093/2010. ( 3 ) As complex as it may be to navigate through the thick web of rather technical secondary legislation, this is in fact the easier question.

4.

The much more complicated questions emerge only later: what consequence should such a finding of the lack of competences with regard to a non-binding (or soft-law) measure entail within the preliminary-rulings procedure? Is the Court able to declare a non-binding measure invalid? In systemic terms, can there be a complete disconnect between Article 263 TFEU proceedings and Article 267 TFEU proceedings with regard to non-binding measures? How can the Court’s judgments in Grimaldi, ( 4 )Foto-Frost ( 5 ) and Belgium v Commission ( 6 ) be reconciled in so far as genuine soft-law instruments are concerned? Can non-binding EU measures be subject to the Court’s review under Article 267 TFEU, as follows from Grimaldi, while their (direct) judicial review under Article 263 TFEU is not possible, as most recently confirmed in Belgium v Commission?

5.

Lastly, but rather importantly, all those issues are raised in the specific context in which national law allows for what appears to be, in contrast to the EU level, a much more open access to a direct judicial review of soft-law measures, including national acts ‘implementing’ non-binding EU-law acts. As such, this gives rise to the question of whether national courts also have a duty to refer, in view of Foto-Frost, questions regarding the validity of non-binding EU measures. Or could such a national court simply annul the national implementing measure on its own, since what is (genuinely) not binding can certainly be freely disregarded?

6.

Article 1 of Regulation No 1093/2010 establishes a European Banking Authority. In the version applicable at the time of the adoption of the contested guidelines, it set out the scope of action of the EBA as follows:

‘2.   The [EBA] shall act within the powers conferred by this Regulation and within the scope of Directive 2002/87/EC, [ ( 7 )] Directive 2009/110/EC, [ ( 8 )] Regulation (EU) No 575/2013 of the European Parliament and of the Council, [ ( 9 )] Directive 2013/36/EU of the European Parliament and of the Council, [ ( 10 )] Directive 2014/49/EU of the European Parliament and of the Council, [ ( 11 )] Regulation (EU) 2015/847 of the European Parliament and the Council, [ ( 12 )] Directive (EU) 2015/2366 of the European Parliament and of the Council [ ( 13 )] and, to the extent that those acts apply to credit and financial institutions and the competent authorities that supervise them, within the relevant parts of Directive 2002/65/EC [ ( 14 )] and Directive (EU) 2015/849 of the European Parliament and of the Council, [ ( 15 )] including all directives, regulations, and decisions based on those acts, and of any further legally binding Union act which confers tasks on the [EBA]. The [EBA] shall also act in accordance with Council Regulation (EU) No 1024/2013. [ ( 16 )]

3.   The [EBA] shall also act in the field of activities of credit institutions, financial conglomerates, investment firms, payment institutions and e-money institutions in relation to issues not directly covered in the acts referred to in paragraph 2, including matters of corporate governance, auditing and financial reporting, provided that such actions by the [EBA] are necessary to ensure the effective and consistent application of those acts.

…

5.   The objective of the [EBA] shall be to protect the public interest by contributing to the short, medium and long-term stability and effectiveness of the financial system, for the Union economy, its citizens and businesses. The [EBA] shall contribute to:

For those purposes, the [EBA] shall contribute to ensuring the consistent, efficient and effective application of the acts referred to in paragraph 2, foster supervisory convergence, provide opinions to the European Parliament, the Council, and the Commission and undertake economic analyses of the markets to promote the achievement of the [EBA’s] objective.’

7.

Pursuant to Article 8 of that regulation, entitled ‘Tasks and powers of the [EBA]’:

‘1.   The [EBA] shall have the following tasks:

…

2.   To achieve the tasks set out in paragraph 1, the [EBA] shall have the powers set out in this Regulation, in particular to:

…

…’

8.

Article 9 of Regulation No 1093/2010, entitled ‘Tasks related to consumer protection and financial activities’, reads as follows:

‘1.   The [EBA] shall take a leading role in promoting transparency, simplicity and fairness in the market for consumer financial products or services across the internal market, including by:

2.   The [EBA] shall monitor new and existing financial activities and may adopt guidelines and recommendations with a view to promoting the safety and soundness of markets and convergence of regulatory practice.

…’

9.

Article 16 of that regulation bears the title ‘Guidelines and recommendations’ and states:

‘1.   The [EBA] shall, with a view to establishing consistent, efficient and effective supervisory practices within the [European System of Financial Supervision (ESFS)], and to ensuring the common, uniform and consistent application of Union law, issue guidelines and recommendations addressed to competent authorities or financial institutions.

2.   The [EBA] shall, where appropriate, conduct open public consultations regarding the guidelines and recommendations and analyse the related potential costs and benefits. Such consultations and analyses shall be proportionate in relation to the scope, nature and impact of the guidelines or recommendations. The [EBA] shall, where appropriate, also request opinions or advice from the Banking Stakeholder Group referred to in Article 37.

3.   The competent authorities and financial institutions shall make every effort to comply with those guidelines and recommendations.

Within 2 months of the issuance of a guideline or recommendation, each competent authority shall confirm whether it complies or intends to comply with that guideline or recommendation. In the event that a competent authority does not comply or does not intend to comply, it shall inform the [EBA], stating its reasons.

The [EBA] shall publish the fact that a competent authority does not comply or does not intend to comply with that guideline or recommendation. The [EBA] may also decide, on a case-by-case basis, to publish the reasons provided by the competent authority for not complying with that guideline or recommendation. The competent authority shall receive advanced notice of such publication.

If required by that guideline or recommendation, financial institutions shall report, in a clear and detailed way, whether they comply with that guideline or recommendation.

…’

10.

According to section 1, point 1 of the EBA’s Guidelines on product oversight and governance arrangements for retail banking products, ‘this document contains guidelines issued pursuant to Article 16 of Regulation (EU) No 1093/2010. In accordance with Article 16(3) of Regulation (EU) No 1093/2010, competent authorities and financial institutions must make every effort to comply with the guidelines’.

11.

Point 2 of the EBA’s Guidelines states:

‘Guidelines set the EBA view of appropriate supervisory practices within the European System of Financial Supervision or of how Union law should be applied in a particular area. Competent authorities as defined in Article 4(2) of Regulation (EU) No 1093/2010 to whom guidelines apply should comply by incorporating them into their practices as appropriate (e.g. by amending their legal framework or their supervisory processes), including where guidelines are directed primarily at institutions.’

12.

Point 3 of the guidelines, under the heading ‘Reporting requirements’, reads as follows:

‘Pursuant to Article 16(3) of Regulation (EU) No 1093/2010, competent authorities must notify the EBA as to whether they comply or intend to comply with these guidelines, or otherwise with reasons for non-compliance, by 23.05.2016. In the absence of any notification by this deadline, competent authorities will be considered by the EBA to be non-compliant. …’

13.

Point 5 opens section 2 of the EBA’s guidelines by defining their subject matter:

‘These Guidelines deal with the establishment of product oversight and governance arrangements for both, manufacturers and distributors as an integral part of the general organisational requirements linked to internal control systems of firms. They refer to internal processes, functions and strategies aimed at designing products, bringing them to the market, and reviewing them over their life cycle. They establish procedures relevant for ensuring the interests, objectives and characteristics of the target market are met. However, these Guidelines do not deal with the suitability of products for individual consumers.’

14.

Point 6 of the guidelines sets out their scope of application:

‘These Guidelines apply to manufacturers and distributors of products offered and sold to consumers and specify product oversight and governance arrangements in relation to:

15.

Point 7 of the guidelines reads as follows:

‘Competent authorities may wish to consider applying these Guidelines to other entities in their jurisdictions that do not fall within the scope of the legislative acts referred to above but for which the competent authorities have supervisory responsibilities. In particular, competent authorities may wish to consider applying these Guidelines to intermediaries other than credit intermediaries under the MCD, such as consumer credit intermediaries.’

16.

Point 8 of the guidelines states:

‘Competent authorities may wish to consider extending the same protections set out in these Guidelines in relation to persons other than consumers such as micro-enterprises and small and medium-sized enterprises (SMEs).’

17.

According to point 11 of the guidelines:

‘These Guidelines are addressed to competent authorities as defined in Article 4(2) of Regulation (EU) No 1093/2010 and to financial institutions as defined in Article 4(1) of Regulation (EU) No 1093/2010 (the “EBA Regulation”).’

18.

Point 16, entitled ‘Date of application’, brings an end to the ‘introductory’ sections 1 to 3. That point states that ‘these Guidelines apply from 3 January 2017’.

19.

The actual guidelines are contained in the two following sections: section 4 on ‘product oversight and governance arrangements for manufacturers’ and section 5 on ‘product oversight and governance arrangements for distributors’. The two sections combined contain 12 guidelines, most of them subdivided into further rules.

20.

The Notice of the Authority for Prudential Supervision and Resolution from 8 September 2017, entitled ‘Implementation of the European Banking Authority’s guidelines on arrangements for governance and oversight of retail banking products (EBA/GL/2015/18)’, reads as follows:

‘The Authority for Prudential Supervision and Resolution (ACPR) has declared that it complies with the European Banking Authority’s guidelines on arrangements for governance and oversight of retail banking products (EBA/GL/2015/18) which are annexed to the present notice.

Those guidelines are applicable to credit institutions, payment institutions and to electronic money institutions under the supervision of the ACPR, which must make every effort to comply with them and, pursuant to paragraph 14 of the guidelines, to ensure that their distributors comply with them, in accordance with Article 16 of Regulation (EU) No 1093/2010 of the European Parliament and of the Council of 24 November 2010 establishing the European Banking Authority.’

21.

On 22 March 2016, relying on Article 16 of Regulation No 1093/2010, the EBA adopted Guidelines on product oversight and governance arrangements for retail banking products. The guidelines are addressed to competent national authorities and to financial institutions.

22.

On 8 September 2017, the ACPR, as the competent French supervisory authority in that matter, published a notice on its website. In that notice, the ACPR declared that it complied with those guidelines. It also stated that the guidelines were applicable to the credit institutions, payment institutions and electronic money institutions under its supervision, which were to make every effort to comply with them and to ensure that their distributors also comply with them.

23.

On 8 November 2017, FBF lodged before the Conseil d’État (Council of State, France), the referring court, an application seeking the annulment of the ACPR notice. FBF claims that the EBA’s guidelines, which were made applicable by the notice, are invalid due to the EBA’s lack of competence to issue such guidelines.

24.

The referring court harbours doubts as to the admissibility and the merits of the plea that the contested guidelines are invalid.

25.

The referring court is of the view that the admissibility of such a plea of invalidity for the purpose of a request for a preliminary ruling depends on whether the contested guidelines may be subject to an action for annulment under Article 263 TFEU and whether a professional federation such as FBF could bring such action.

26.

As to the merits of that plea, the referring court observes that the EBA’s guidelines invoke several EU legislative acts but that none of them, apart from Directive 2014/17, make any express provision concerning the governance of retail banking products, which is the area covered by the guidelines. Furthermore, none of those legislative acts contain any provision empowering the EBA to issue guidelines on the governance of retail banking products. However, it follows from Regulation No 1093/2010 that the EBA shall contribute to ensuring that the taking of credit and other risks are appropriately regulated and supervised, and shall contribute to enhancing customer protection. Such objectives are precisely those that the governance of retail banking products helps to achieve.

27.

It is within this factual and legal context that the Conseil d’État (Council of State) decided to stay the proceedings and to refer the following questions to the Court of Justice for a preliminary ruling:

28.

Written observations have been submitted by FBF, the ACPR, the French and Polish Governments, as well as the European Commission. With the exception of the Polish Government, all those parties also presented oral argument at the hearing which took place on 20 October 2020.

29.

This Opinion is structured as follows. I shall start with introductory remarks on the referring court’s questions and the (non‑)binding nature of the contested guidelines (A). I shall then address the questions raised by the referring court in reverse order, beginning with Question 3 to determine whether the contested guidelines were indeed adopted by the EBA within the scope of the latter’s powers (B). Having concluded that the EBA has in fact exceeded its competence, I shall then turn to Questions 1 and 2 and several other elements relating to the general relationship between Articles 263 and 267 TFEU as far as non-binding EU measures are concerned (C).

30.

From a certain perspective, this case is rather simple. If the questions referred were viewed in isolation from the case at hand, and were answered in the order in which they were put, then the answers would not be that difficult.

31.

Question 1 is inadmissible. The present case was submitted as a request for a preliminary ruling under Article 267 TFEU. Within that procedural context, it is an entirely hypothetical exercise to ask whether the same action may possibly be brought as an action for annulment under Article 263 TFEU. In a similar vein, to enquire as to whether or not a professional federation is able to challenge the contested guidelines in those proceedings in not relevant for settling the case before the referring court.

32.

Question 2 is also straightforward. While that question is admissible, the answer can easily be inferred from the existing case-law. Since the judgment in Grimaldi, the Court has consistently insisted that Article 267 TFEU confers on the Court jurisdiction to give a preliminary ruling on both the validity and the interpretation of all acts of the institutions of the Union ‘without exception’. ( 19 ) Non-binding EU measures can thus clearly form the object of a request for a preliminary ruling on validity. ( 20 )

33.

Furthermore, the sub-question common to both Questions 1 and 2, relating to the standing of professional associations, is not relevant in the context of the preliminary-rulings procedure. It is solely for the national court to decide whether to refer a case to the Court pursuant to Article 267 TFEU. ( 21 ) According to the Court, any party may, in proceedings before the national courts, plead the invalidity of an act of the Union and ask that court, which has no jurisdiction itself to declare the act invalid, to submit that question to the Court by means of a reference for a preliminary ruling. ( 22 )

34.

However, for the rest, EU law does not regulate the issue of establishing who ought to be the party before the national court. If the national law allows for such pleas to be raised, it will again form part of the exclusive competence of the national court to decide whether or not it finds it necessary (or is under an obligation) to refer the matter to the Court. Thus, the issue of whether FBF has standing to raise a plea of invalidity before the national court against an EU measure is a matter for national law.

35.

Thus, by answering the questions in the order that they were raised and by dealing with the first two questions in an abstract manner, detached from the facts of the present case, one may easily and immediately arrive at Question 3 and consider that the issue of the EBA’s power to adopt the contested guidelines is the only real question in the case at hand.

36.

Nevertheless, I do not think that such a shortcut is appropriate. It is in fact while seeking to answer Question 3, and to draw the consequences thereof, that one fully understands the scope of Questions 1 and 2, if the latter ones are not approached abstractly, but instead in the specific context of the present case.

37.

Moreover, it is clear that reading the three questions in such a simplified way would not do justice to the referring court. In fact, it is clear from the order for reference, in particular when read together with the elucidating Opinion of the rapporteur public of the Conseil d’État (Council of State) in that case, ( 23 ) that the referring court is aware of the relevant case-law of this Court on the matter. It emerges from those documents that the referring court actually wonders where exactly that case-law may lead it in a specific case such as the one before it.

38.

It is in this context that the referring court questions, in particular, the relationship between Articles 263 and 267 TFEU as far as soft law is concerned, particularly in circumstances where the national court provides for judicial review of non-binding national measures, while the Court does not allow, within actions for annulment, for review of non-binding EU measures. That issue lies at the heart of the question concerning the parallels (or the absence thereof) between the two types of proceedings, especially as far as the position of professional associations in both types of proceedings is concerned. To that issue then connects another one, also identified by the referring court in its order for reference, relating to the fact that under the preliminary-rulings procedure, both the national and European levels of review become intertwined: are national courts under an obligation, pursuant to the judgment in Foto-Frost, to refer questions to the Court on the validity of a non-binding EU measure when reviewing a national measure which made that EU measure applicable at the national level to individual addressees?

39.

Against that background, to the extent that answering Question 3 first will help to clarify the more general issues brought about by Questions 1 and 2, I prefer to start with that third question before returning to those structural issues. In this way, the specific example of the guidelines adopted by the EBA provides a good illustration of the structural issues relating to the review of soft-law measures before the Court of Justice.

40.

Before turning to Question 3, however, another preliminary issue must be dealt with: are the contested guidelines a genuine non-binding measure that does not produce binding legal effects, as argued by all the parties to the proceedings? What may seem to be a commonplace statement is rather an important starting assumption for the discussion that follows. That issue is in fact hidden within the first part of Question 1 given that established case-law excludes genuine soft-law measures from judicial review under Article 263 TFEU. By contrast, if such an act did produce binding legal effects (and thus were a ‘false’ soft-law instrument), it would be reviewable under Article 263 TFEU (at least assuming that the given applicant is directly and individually concerned by that act).

41.

Indeed, it is established case-law that any provisions adopted by the institutions, irrespective of their form, that are intended to have binding legal effects are regarded as ‘challengeable acts’ for the purposes of Article 263 TFEU. In order to determine whether the contested act produces binding legal effects, it is necessary to examine the substance of that act and to assess those effects on the basis of objective criteria, such as the content of that act, taking into account, where appropriate, the context in which it was adopted and the powers of the institution which adopted the act. ( 24 )

42.

Do the contested guidelines produce binding legal effects under that traditional test?

43.

On the one hand, it is true, first, that the contested guidelines are worded in non-mandatory terms. The actual substantive guidelines that are addressed to manufacturers and distributors of banking products, contained in sections 4 and 5 of the guidelines, use the term ‘should’ as opposed to the language of ‘shall’. Second, there is no obligation on the competent authorities ( 25 ) to comply with them. The second subparagraph of Article 16(3) of Regulation No 1093/2010 requires each competent authority to confirm whether it complies or intends to comply with guidelines adopted by the EBA. In the event that a competent authority does not comply or does not intend to comply, it shall inform the EBA, stating its reasons. ( 26 ) Third, when assessing the context and the powers of the agency concerned, the argument is bound to fall within the same kind of circular reasoning that I have already outlined elsewhere: ( 27 ) (i) because the EBA certainly knew that its guidelines are not binding, (ii) it could therefore have no intention of adopting anything that would be binding, and (iii) hence it is evident that the body adopting the act had no intention whatsoever of adopting a binding legal act.

44.

On this basis, it is without doubt possible to conclude that the contested guidelines do not, in themselves, produce binding legal effects.

45.

On the other hand, however, those guidelines are issued in a context of and accompanied by a number of mechanisms which, read in the light of Regulation No 1093/2010 as their legal basis, make them acts capable of being reasonably perceived as inducing compliance by their addressees.

46.

First, in line with the first subparagraph of Article 16(3) of Regulation No 1093/2010, section 1, point 1, of the contested guidelines provides that competent national authorities and financial institutions ‘must make every effort to comply with the guidelines’. Without carrying out a detailed examination into what exactly ‘to make every effort’ specifically entails, it can be presumed that the contested guidelines are not adopted with the intention simply to be disregarded by their addressees, particularly if making such effort is a duty placed upon them.

47.

Second, who exactly is the addressee? Even if the contested guidelines are formally addressed, pursuant to point 11 thereof, to competent authorities and to financial institutions, ( 28 ) it is clear that the financial institutions are those that will end up having to comply with the obligations and, as such, are the genuine addressees. This logic flows from the content of the guidelines in sections 4 and 5, which exclusively targets manufacturers and distributors of banking products.

48.

In systemic terms, the guidelines closely resemble directives: although formally addressed to the Member State, their provisions are in due course meant to govern the conduct of individuals, with the latter having no choice but to apply them. Competent authorities are not the real addressees of those obligations; their task is simply to opt in or to opt out. However, once that decision is made, the initially non-binding nature becomes very much binding, as the ‘nominal addressee’ (the competent supervisory authority) becomes an effective ‘enforcer’. Thus, there is very little choice, or rather none at all, on the part of the real addressees of the guidelines, namely the financial institutions, on whether to comply with them.

49.

In addition, in the light of the combined reading of points 6, 7 and 8 of the contested guidelines, ( 29 ) it is not even clear whether the formal decision by a competent authority not to comply does in fact exempt financial institutions from the duty to ‘make every effort to comply with the guidelines’. Non-compliance by competent authorities could equally only mean that they are not going to enforce those guidelines, without affecting the very duty of the financial institutions. In other words, those guidelines may have a life of their own vis-à-vis financial institutions irrespective of the position adopted by competent authorities.

50.

Third, and in relation to the previous point, when competent authorities decide to comply, financial institutions effectively become bound by the contested guidelines at the national level as a result of their ‘implementation’ or ‘incorporation’ by the competent national authority. In the present case, by declaring in its notice that it complied with the contested guidelines, the ACPR made the content of the latter de facto obligatory for the financial institutions in France. A point worth emphasising in this regard is that, in contrast to a directive, the content of which must first be transposed into a form of national law, the content of the guidelines became applicable through the ACPR notice to all the ‘credit institutions, payment institutions and to electronic money institutions under the supervision of the ACPR’. ( 30 )

51.

The latter point of effective enforcement at the national level should be explained clearly. Once the competent national authority has opted in, compliance with those guidelines becomes enforceable in that Member State. According to the rapporteur public of the Conseil d’État (Council of State) in the dispute in the main proceedings, non-compliance with the guidelines cannot directly lead to the imposition of a sanction. However, the content of the guidelines represents best practices that ought to be adhered to by the financial institutions. If such institutions do not follow those best practices, it might be perceived as bad practice on their part. On that basis, the ACPR could issue individual warnings, the non-respect of which would in essence expose those financial institutions to disciplinary proceedings. ( 31 )

52.

In short, in accordance with the traditional definition of an act of EU law producing binding legal effects for the purposes of Article 263 TFEU, I agree that the guidelines, especially if viewed at the EU level only, are likely to be considered a non-binding, genuine soft-law instrument. However, when the test is applied to cases such as the present one, that traditional approach does not provide much of an answer, but rather eloquently states the nature of the problem.

53.

The test to determine whether an EU-law act is reviewable ought to focus on whether the act can reasonably be perceived as inducing (or even effectively imposing) compliance on the part of its addressee. The result ought to be one of scale, recognising the continuum of legal effects, but then logically focusing closely on the exact effects the contested act has on the legal situation of its addressees. If the problem lies in hybrid forms of governance, then the remedy may also be a hybrid one, logically tailored to the exact type of effects produced and seen as problematic. However, as recently evidenced by the judgment in Belgium v Commission, the Court remains focused on the act and its author, detached from the real life of the legal act and its addressees, thus going round in circles whereby the nature of the act is determined by the intent of its author and vice versa. Ultimately, only one of two results is conceivable in such a binary world: either there will be full binding legal effects or no such effects whatsoever.

54.

I do not wish to reiterate arguments that have failed to convince the Court before. ( 32 ) This is rather to explain how those choices impact on the present case. The EU and national regulatory and judicial levels are joint vessels, at least within the preliminary-rulings procedure. That circumstance adds another layer of complexity to the present case, where the traditional approach of looking only and exclusively at the EU-law act at the EU level becomes problematic: what may perhaps still be construed as soft law when looking only and exclusively at the EBA and the competent national authorities becomes something very different one level down within the Member States. At that level, ‘soft law’ is ‘no longer so soft’, or may even turn into proper ‘hard law’. It should be pointed out that EU law certainly does not preclude that from happening. Quite to the contrary in fact: the entire system is designed to function in precisely that way.

55.

In summary, with regard to their genuine addressees, in the form of financial institutions at the national level, the guidelines are clearly much less ‘soft’ than would be the case if the focus were on the level of competent national authorities. However, as a starting assumption, with regard to the Court’s standard approach vis-à-vis EU measures not producing binding legal effects, the contested guidelines are unlikely to be considered by the Court as binding and, as a consequence, reviewable under Article 263 TFEU.

56.

According to FBF, the contested guidelines regulate product governance. They are deprived of any legal basis in so far as one cannot consider them to be implementing the legislative acts referred to in Article 1(2) of Regulation No 1093/2010 since the latter ones essentially govern corporate governance. In adopting the contested guidelines, it is more likely that the EBA would have taken inspiration from Directive 2014/65/EU, ( 33 ) which regulates the governance of financial products and where the concept of ‘target market’ and the distinction between manufacturers and distributors are central.

57.

The Commission largely takes the same view as FBF. However, it considers that the contested guidelines do not fall entirely outside the scope of the EBA’s competence to the extent that they relate to Article 7(1) of Directive 2014/17. The latter allows for product governance to be regulated, for target markets to be identified, and for a distinction to be made between manufacturers and distributors. As for the rest, however, the guidelines do fall outside the EBA’s competence and should be, according to the Commission, declared invalid.

58.

The ACPR, the French and Polish Governments and the EBA take the opposite view. According to the ACPR, the French Government and the EBA, the EBA is entitled to adopt guidelines beyond the strict scope of the legislative acts referred to in Article 1(2) of Regulation No 1093/2010 because Article 1(3) thereof extends the EBA’s powers to issues not covered by those acts. In any case, product governance and corporate governance are interrelated so that the contested guidelines cannot be considered to be outside the scope of the EBA’s competence. According to the same intervening parties, as well as the Polish Government, the EBA’s action is also lawful because Regulation No 1093/2010 expressly aims to provide consumer protection. The EBA had the power to adopt the contested guidelines in as much as they serve that objective. Thus, according to the ACPR and the EBA, there should simply be a general assessment of the EBA’s power to adopt guidelines. Such a general assessment would lead to the conclusion that the EBA had the power to issue the contested guidelines.

59.

In the following sections, I set out my disagreement with the views of the ACPR, the French and Polish Governments and the EBA. Although the contested guidelines may, if assessed generally, fit roughly within Regulation No 1093/2010 (1), I shall explain why I do not believe such a low-intensity or lenient assessment to be warranted in the context of non-binding measures (2). In view of the incompatibility of the contested guidelines with Regulation No 1093/2010, the remaining question then becomes what should be the formal result of such a finding (3).

60.

The contested guidelines deal with the establishment of product governance for retail banking products. In particular, those guidelines recommend that manufacturers should identify the relevant target markets and ensure that products are appropriate for those markets. ( 34 ) The guidelines also recommend product testing in order to be able to assess how the product would affect its consumers under a wide range of scenarios. ( 35 ) Moreover, manufacturers should select distributors appropriate for the particular target market, and be in a position to provide those distributors with information regarding the characteristics and risks of the product to the consumers. ( 36 ) For their part, distributors of banking products should disclose to the consumer a description of the main characteristics of the products in question and their risks. ( 37 )

61.

In accordance with section 2, point 6, the contested guidelines apply to manufacturers and distributors of products offered and sold to consumers in relation to four provisions of specific legislative acts, namely Article 74(1) of Directive 2013/36, Article 10(4) of Directive 2007/64, Article 3(1) of Directive 2009/110 and Article 7(1) of Directive 2014/17. Competent authorities may also ‘wish to consider applying these Guidelines to other entities in their jurisdictions that do not fall within the scope of the legislative acts referred to above but for which the competent authorities have supervisory responsibilities’. ( 38 )

62.

When comparing the invoked scope of application with the actual content of the guidelines, it seems rather clear that, as essentially suggested by FBF and the Commission, with regard to their legal basis, the contested guidelines go further than what Regulation No 1093/2010 allows for.

63.

Pursuant to Article 16(1) of Regulation No 1093/2010, ‘the [EBA] shall, with a view to establishing consistent, efficient and effective supervisory practices within the ESFS, and to ensuring the common, uniform and consistent application of Union law, issue guidelines and recommendations addressed to competent authorities or financial institutions’.

64.

Under Article 1(2) of Regulation No 1093/2010, the EBA shall act within the powers conferred by that regulation and within the scope of a number of legislative acts laid down therein and of any further legally binding Union act which confers tasks thereon. By virtue of Article 1(3), the EBA shall also act in the field of activities of the various financial institutions referred to in that provision, ‘in relation to issues [that are] not directly covered in the acts referred to in paragraph 2 … provided that such actions by the [EBA] are necessary to ensure the effective and consistent application of those acts’. ( 39 )

65.

It follows rather clearly from that text that, regardless of the type of measure taken, the EBA shall only act within the (substantive) boundaries of those legislative acts. A number of other provisions of Regulation No 1093/2010 pertaining to the EBA’s tasks and powers also confirm that those acts constitute the EBA’s ultimate horizon. ( 40 ) It is thus in the light of the content and scope of those acts that the contested guidelines should be evaluated.

66.

As regards Directives 2013/36, 2007/64 and 2009/110, which are specifically cited by the contested guidelines in relation to the latter’s scope of application, ( 41 ) it appears that they are all legislative acts to which Article 1(2) of Regulation No 1093/2010 refers. ( 42 ) Thus, the EBA is, in principle, allowed to adopt guidelines in order to give flesh to those instruments.

67.

However, there is a clear mismatch between the subject matter of those acts and that of the guidelines. While the latter have set out specific ‘rules’ that concern product governance, the former all relate to corporate governance by providing, in particular, for internal procedures within financial institutions, clear organisational structures with consistent lines of responsibility, and procedures related to risk management and capital requirements. It is unclear how guidelines on product governance contribute to the effective and consistent application of acts concerned with corporate governance, certainly in the short term. While those latter acts deal with risks associated with dysfunctional corporate governance in the long term (thus a structural issue), the contested guidelines strive to regulate internal processes on product governance that have an influence on short-term outcomes. Therefore, the type (and the degree) of risk that is being regulated by the contested guidelines, on the one hand, and by the legislative acts referred to in Article 1(2) of Regulation No 1093/2010, on the other hand, are different.

68.

The difference between product governance and corporate governance as a regulatory subject matter is not merely academic. An analogous example from another area, such as the car industry, may help to illustrate the difference. Corporate governance requires car manufacturers to apply control mechanisms at each stage of the production, to have a transparent organisational structure, to determine clearly who is responsible for what, to have systems in place in order to resolve potential problems, and so on. All those rules are connected to the smooth functioning of the company. By contrast, product governance requires car manufacturers to identify the relevant target market of a new car, to consider how a new model of car fits with the existing range of cars, and whether the presence of too many car alternatives prevents the consumer from making informed decisions. Such ‘rules’ have little to do with the internal functioning of the company; they seek to regulate business decision-making with regard to the quality of the products that are to be offered to customers. In other words, corporate governance rules relate to the quality of internal processes and mechanisms which are there to ensure the smooth functioning of the company. Product governance rules concern business choices that are there, essentially, for the marketing of cars.

69.

In addition, I agree with FBF that the subject matter of those legislative acts is not the same as the subject matter of Directive 2014/65, which specifically and expressly regulates the governance of financial products marketed by providers of investment services. ( 43 ) Thus, the EBA could not lawfully adopt guidelines on the governance of banking products.

70.

By contrast, as the Commission and, to some extent, the referring court acknowledged, Article 7(1) of Directive 2014/17, which is also specifically mentioned to in section 2, point 6, of the contested guidelines, is admittedly linked to products, in particular credit products. It could thus perhaps serve, at least partially, as a proper legal basis for the contested guidelines.

71.

However, Directive 2014/17 does not deal with product governance in the same sense as the contested guidelines. It instead regulates the conduct of credit providers in individual cases and the methodology on deciding whether to grant a credit to a given customer or not. Furthermore, although a few provisions of Directive 2014/17 empower the EBA to adopt certain rules, ( 44 ) none of those provisions specifically concern product governance rules, nor do they refer to the adoption of guidelines in the matters concerned.

72.

I therefore have some difficulty in giving credence to Directive 2014/17 for the purposes of addressing the issue of the EBA’s competence to adopt the contested guidelines. Indeed, even if Directive 2014/17 was in fact recognised as an actual legal basis in the present case, quod non, I wonder where exactly being successful on one out of the four targets may lead in practical terms. The submissions of the Commission illustrate that issue rather well.

73.

The Commission has submitted that the Court should find the contested guidelines invalid to the extent that they relate to (i) Article 10(4) of Directive 2007/64, (ii) Article 3(1) of Directive 2009/110 and (iii) Article 74(1) of Directive 2013/36. At the same time, the Commission also submitted that the Court should state that the contested guidelines are valid to the extent that they relate to Article 7(1) of Directive 2014/17 (or rather, to reproduce the Commission’s proposal in full, that the examination of this point has not revealed any elements capable of casting doubt as to the validity of the guidelines).

74.

I must admit being somewhat at a loss while visualising what the operative part of such a judgment would mean in practical terms. Would it mean that the guidelines remain ‘valid’, but their scope of application would be judicially limited to credit agreements for consumers relating to residential immovable property only? Would the guidelines then remain applicable only to the financial institutions when offering such specific products? Or would one, on this basis, have to delve even deeper into the detail and assess individual guidelines one by one?

75.

In the light of such considerations, the only sensible option seems to be that the guidelines must either stand or fall as a whole. In my view, the contested guidelines as a whole do not fall within the scope of the legislative acts referred to in Regulation No 1093/2010 or the ones conferring specific tasks upon the EBA. The EBA has therefore exceeded its competences in adopting guidelines the subject matter of which is not covered by those legislative acts.

76.

If one were to apply normal scrutiny in this case, then one would arrive at the interim conclusion above. By ‘normal’ I mean the kind of review that the Court carries out commonly, and with the same level of intensity, when reviewing the validity of measures with binding legal effects. ( 45 )

77.

However, it is appropriate to ask whether measures allegedly not generating any binding legal effects ought to be subject to such normal scrutiny in the first place. In short, if they are not binding, why should anybody, including this Court, care about them? Why is it necessary to police whether or not an EU body remained within its scope of competence if nobody has to pay any attention whatsoever to what it advocates? Those considerations should result in, at least with regard to genuine soft-law measures, a rather lenient review, if any review at all.

78.

It is essentially in line with that logic that the ACPR and the EBA maintained at the hearing that the degree of scrutiny of non-binding EU measures should be lower than in the cases of binding acts, embodying merely a general assessment of non-binding measures. That argument essentially suggests that, even if non-binding measures do not strictly fall within the scope of the EBA’s competences, the fact that they may fall roughly within that scope is sufficient to establish the lawfulness of the EBA’s guidelines. In particular, as far as the distribution of competences is concerned, the authors of non-binding measures would benefit from some leeway rather than strict limits being imposed on them. In practice, the setting aside of those measures would then only occur when they are manifestly outside the boundaries of the powers bestowed upon their author.

79.

Indeed, from a certain point of view, product governance is perhaps not that different from corporate governance if one were to view both of them in an abstract, purposive manner. Both have an impact on the stability of the financial system. As the EBA pointed out during the hearing, the overall aim of the guidelines is to recommend internal processes that would eliminate excessive risk since imprudent risk taking was at the root of the financial crisis that emerged in 2008. An effective product oversight and arrangements ensuring that financial products meet the requirements of target markets are likely to enhance the economic performance of the financial institution and to decrease the risk of bankruptcy.

80.

By the same token, as argued by the ACPR, the French and Polish Governments and the EBA, the EBA’s guidelines could still be seen as falling within the mandate of the EBA because Regulation No 1093/2010 is expressly aimed at consumer protection. It is true that consumer protection is an important aim pursued by Regulation No 1093/2010. That aim notably appears in Article 1(5)(f), ( 46 ) and above all in Article 9, which is dedicated to the ‘tasks related to consumer protection and financial activities’. That encompassing aim could perhaps be used to justify the EBA’s action in a wide range of areas.

81.

I certainly acknowledge those arguments. However, I simply cannot subscribe to them, neither in this specific case, nor in general.

82.

As far as the specific guidelines at issue are concerned, I do not think that a generous purposive reading of the scope of the EBA’s action based on vaguely defined ‘overall aims’ should be embraced in order to vindicate the contested guidelines. Such far-reaching regulatory involvement of the EBA based on broadly defined objectives is not tenable in view of Article 9(1) of Regulation No 1093/2010 itself. Under that provision, the EBA’s role appears rather circumscribed – it is to collect and analyse consumer trends, review and coordinate education initiatives and develop training standards. Although that list is not exhaustive, it is indicative of very different types of measures and action foreseen for the EBA under the heading of ‘consumer protection’.

83.

In addition, excessive risk is naturally dangerous. However, every business decision bears a risk. The potential that such hazards come to fruition is not sufficient justification to allow an EU agency to regulate all of them, particularly through ‘mere’ guidelines. Should product governance and corporate governance be considered so interwoven as to justify the EBA’s action as regards the former on the basis of legislative acts regarding the latter, the EBA would then be entitled to adopt rules, possibly limitlessly, in a wide range of situations that have little to do with corporate governance, strictly speaking. Should the EBA also not have the powers to start issuing guidelines on the selection and promotion of personnel, on the selection of software providers, on the functioning of a helpline or on the processes of purchasing furniture? Or perhaps the EBA should be equally able to regulate the safety of seat belts in the company cars used by the top management of a financial institution. If such persons crucial to the stability of the financial sector were not properly protected in the unfortunate case of a car crash, the stability of the sector would be also threatened in a way.

84.

Moreover, in general, there are numerous, rather onerous arguments as to why, if the judicial review of non-binding legal acts were to be carried out under Article 267 TFEU (which the Court insists that it is), that review ought to be a normal, standard type of review.

85.

First, allowing the institutions and, more broadly, the numerous EU bodies to issue legally non-binding measures with limited judicial control (for instance, without any check on competence) would only encourage a further spread of ‘crypto-legislation’ in the form of soft law in the Union. As I explained in my Opinion in Belgium v Commission, ( 47 ) EU bodies are able, through soft law, to create parallel sets of rules which bypass the legislative process and which might have an impact on institutional balance.

86.

Second, it is not only the (horizontal) institutional balance that is likely to be affected by such practices, but, above all, the overall legitimacy of the rules downstream. The proliferation of EU agencies has brought about legitimacy concerns related to the exercise of the delegated administrative power. It is thus necessary to ensure that that exercise does not remain unchecked, even as far as nominally non-binding EU acts are concerned. ( 48 )

87.

Third, these arguments gain particular traction in circumstances where there is already a ‘distinct escape into soft-law instruments’ present in a number of new policy areas, with the banking union and financial supervision being vivid examples of that phenomenon. Within that context, it is paradoxical to see that, while the Treaty of Lisbon finally provided the Court with full default jurisdiction vis-à-vis all acts adopted by EU bodies and institutions, by abolishing the – in terms of legal protection, incomplete – pillar structure, ( 49 ) the same problem may now be replicated internally, with areas of EU action governing the behaviour of individuals being once again effectively excluded from any review, this time by the Court’s own choice.

88.

Fourth, there is an additional argument relating specifically to EU agencies, as opposed to EU institutions. While all EU bodies are indeed subject to the principle of attributed competence, EU agencies tend to be granted, in addition, a specialised and rather narrow mandate. Aside from the constitutional argument why such structure does not call for merely a lenient review, there is also the more pragmatic one. It is the risk of a ‘crowded soft-law house’. In a situation where several agencies or bodies have overlapping mandates, which results in them regulating similar or neighbouring issues (as is, again, the case for banking and finance), to interpret such mandates at a relatively high level of abstraction and with regard to the attainment of abstract goals would only lead to inducing the coming into existence of overlapping or even conflicting soft-law instruments.

89.

I readily acknowledge the singular nature of such a consideration: how can there be, by definition, a conflict between legally non-binding measures? What is not legally binding cannot be in conflict because it is not able to create any legal obligations. As such, what is there to be in conflict?

90.

Fifth and finally, the problem is that the Court has apparently refused to embrace that logic, as most recently confirmed in Belgium v Commission. In my Opinion in that case, ( 50 ) I indeed suggested the possibility to rely on the form of an EU legal measure that would then determine how to look at the substance. Thus, if something is called a ‘guideline’, it will be deemed to produce no binding legal effects whatsoever and everybody will have the right fully to ignore it, because it is a guideline.

91.

However, the Court insisted on the previous approach, within which it has first to be decided, in each individual case and irrespective of the formal label of an act, whether it is a ‘genuine’ or ‘false’ soft-law measure. ( 51 ) However, the necessary consequence of such an approach is that before it is established whether a given measure is a genuine soft-law measure, it cannot be excluded that it does in fact generate binding legal obligations. It is therefore certainly possible that addressees of EU rules of whatever nature find themselves effectively in a position in which they are not able to tell what is formally binding and, as far as their substantive obligations are concerned, which one of the potentially conflicting guidelines they should follow.

92.

As a closing illustration, coming back to the specific example of the EBA, it would appear that the EU legislature itself started to acknowledge that potential issues may arise and demonstrated a willingness to monitor the EBA’s actions more closely. Regulation No 1093/2010 has indeed been amended in the meantime by Regulation 2019/2175. Following that amendment (and although not temporally applicable in the present case), a new Article 60a of Regulation No 1093/2010, entitled ‘Exceeding of competence of the [EBA]’, provides that ‘any natural or legal person may send reasoned advice to the Commission if that person is of the opinion that the [EBA] has exceeded its competence … when acting under [Article] 16 … and that is of direct and individual concern to that person’. ( 52 )

93.

Again, with regard to genuine soft law, all that is simply staggering. Introduced in Chapter V of Regulation No 1093/2010, entitled ‘Remedies’, is a provision which starts from the assumption that when issuing what ought to be legally non-binding guidelines and recommendations under Article 16 of the regulation, or also non-binding opinions under Article 16a, the EBA could exceed its competence, so that there is a need for a remedy. Whatever the systemic rationale behind such a provision may be, it is fair to assume that it would not have been inserted had the EU legislature not come to the conclusion that there was perhaps a problem. The only certainty that remains is that the mailbox of the Commission is unlikely to be too hard hit if, as a condition for a person being allowed, in essence, to write a letter to the Commission, that person needs to be ‘individually and directly concerned’ by the problematic measure, certainly if inspiration for the interpretation of those concepts were to be taken from the established case-law of this Court relating to Article 263(4) TFEU.

94.

In view of all these arguments, I cannot but conclude that it is indeed essential to make non-binding acts adopted by EU agencies subject to normal judicial review, at least with regard to their competences, so that those agencies do not unlawfully interfere with the competences of other EU bodies or institutions.

95.

A final issue, in the context of the judicial review of legally non-binding measures, is that of the type of outcome in circumstances where a request for a preliminary ruling concerning the validity of such a measure has been made. What might the outcome be in a case in which (i) the contested EU act is not legally binding and therefore not subject to an action for annulment under Article 263 TFEU, and thus cannot be formally annulled within those proceedings, but (ii) where a lack of competence on the part of the issuing authority has been identified by the Court following a request for a preliminary ruling under Article 267 TFEU?

96.

Two outcomes are plausible. First, the legally non-binding EU measure could be declared invalid, as is the normal outcome with other types of (binding) EU-law measures found wanting under Article 267 TFEU. Second, a question of validity of such a measure could effectively be rephrased as an issue of interpretation under Article 267 TFEU, relating either to the interpretation of the non-binding EU act itself or the interpretation of the (binding) legislative basis on which that act was adopted (in the present case, Regulation No 1093/2010).

97.

What could perhaps be discarded right away is the first alternative within the second scenario: if the Court were to find that an authority adopted a non-binding EU-law measure in breach of its competence, I fail to see how it would then fall to the Court effectively to sit down and start re-drafting that measure judicially by way of ‘interpretation’ of what should never have been there in the first place.

98.

If I am not mistaken, with respect to the practice of the Court up to now, the Court had never found a non-binding EU measure to be invalid until very recently. ( 53 ) This is partly due to the fact, perhaps, that national courts have usually raised questions of interpretation in relation to such acts. ( 54 ) However, it is also fair to acknowledge that (so far) the Court had always rephrased questions on validity into questions on interpretation. ( 55 )

99.

Such a ‘transformation’ was perhaps most clearly visible in the Grand Chamber judgment in Kotnik and Others, where the Court specifically assessed several questions on validity as if they were questions on interpretation. ( 56 ) Moreover, on two previous occasions, the Court even expressly excluded the possibility to assess the validity of the measure at issue after finding that it had no binding effects. ( 57 )

100.

Thus, in a way, the case-law of the Court could be seen as providing an answer to the somewhat violent opening of this Opinion: what never lived (as law) cannot be put to rest (by the Court). All that the Court can do is state with authority that that animal was never born, since the physiognomy of the mother excludes the possibility of ever giving birth to that creature.

101.

In a way, that solution at least prevents some of the problems that could arise if the Court declares invalid under Article 267 TFEU an act that could not be annulled under Article 263 TFEU on account of that act lacking binding legal effects. The Court could continue with that approach and thus conclude in the present case that the combined reading of Articles 1, 8 and 16 of Regulation No 1093/2010 precluded the EBA from issuing the contested guidelines due to a lack of power on the part of the EBA to that effect.

102.

Moreover, such an answer might appear more appropriate than an answer on validity within Question 3 since, very formally speaking, the referring court has not raised a question on validity. The referring court has merely asked whether the EBA has exceeded its powers by issuing the contested guidelines.

103.

On the whole, therefore, such an outcome is possible. ( 58 ) For all practical purposes, declaring the animal non-existent is not much different to stating that its mother could never have given birth to it in the first place. ( 59 ) That statement does, nonetheless, come at a cost, generating problems of its own. That is why I would still recommend that the Court expressly provide an answer as to the validity of the contested guidelines.

104.

First, although the referring court has phrased Question 3 in terms of competences, a finding of a lack of powers logically leads to the annulment or a declaration of invalidity. ( 60 ) It is rather difficult to see how there could be a declaration of an excess of powers with the act somewhat continuing to hang suspended in mid-air.

105.

Second and more fundamentally, would it then be for the referring court to act on that answer provided by the Court? Drawing consequences from such a statement of the Court, could the national court itself then declare the guidelines invalid? Or would the declaration of an excess of competence at the EU level entitle it to annul the national ‘implementing’ soft-law instrument?

106.

In any case, in systemic terms, any of the hypothetical options for the national court set out above are unsatisfactory from the point of view of distinct normative sets of rules, EU and national ones, and have the potential to undermine the Foto-Frost line of case-law. In principle, it is exclusively for the Court to deal with the (in)validity of EU-law measures. By contrast, national courts may only annul acts of national origin. Generally, providing answers to the effect that ‘an EU-law measure shall be interpreted as precluding another measure’ is suitable for national measures, not for EU measures (whether they be binding or not). That bypass is indeed necessary vis-à-vis national legislation because the Court has jurisdiction to interpret EU law but not to assess national law. Conversely, such a structure is not appropriate where the contested guidelines are EU measures since, in principle, the Court is empowered not only to review but also to annul the latter or to declare them invalid.

107.

Third, the rephrasing of a question on validity into a question on interpretation amounts to a major revision (or at least qualification), of the Grimaldi line of case-law. Contrary to what is stated therein and to what has often been repeated since, ( 61 ) the Court would actually not have jurisdiction to rule on the ‘validity of all acts of the institutions of the EU without exception’. The Court would thus be deprived of the possibility to rule on the validity of non-binding EU acts under Article 267 TFEU. There would therefore be a large and, with the proliferation of soft-law instruments, ever-growing exception to the jurisdiction of the Court, both under Article 267 and Article 263 TFEU.

108.

Fourth and finally, qualifying Grimaldi would in fact mean that, in a way, some coherence between Articles 263 and 267 TFEU would be re-introduced. However, that coherence would effectively be to the lowest common denominator and to the detriment of any effective legal protection provided under EU law. It would mean that, since no effective access to the judicial review of soft-law measures under Article 263 TFEU is available, it is best that there be none under Article 267 TFEU either.

109.

In the closing section of this Opinion (C), I shall explain why and how that logic would be incorrect in structural terms. However, for the purpose of providing an answer to Question 3, I propose that the Court go with the simplest, clearest, and indeed most honest answer: the contested guidelines should be declared invalid in so far as the EBA has acted outside the powers bestowed upon it by Regulation No 1093/2010.

110.

It is only now that one is able fully to appreciate the broader issues raised by the referring court’s Questions 1 and 2. As such, I shall deal with both those questions in turn below. I shall start with the second parts of Questions 1 and 2, which enquire about FBF’s standing before the national court (1). I shall then turn to Foto-Frost and the issue of whether a national court is obliged to make a request for a preliminary ruling to this Court if what it is contemplating largely resembles ‘disabling’ the national application of a non-binding EU act (2). Finally, I shall ponder on the overall relationship between Articles 263 and 267 TFEU with regard to non-binding EU-law measures, which underpin the first parts of Questions 1 and 2 (3). That approach will eventually bring me to a rather unsatisfying conclusion about the overall state of EU law in this area, wondering how (if at all) the judgments in Grimaldi, Foto-Frost, and Belgium v Commission can in fact be reconciled with regard to soft-law instruments (4).

111.

When taken at face value and read separately, ( 62 ) the second part of Question 1 is inadmissible because the present case is not an action for annulment. The second part of Question 2 should be answered to the effect that it is a matter for national law, not EU law, to decide when and how a party to national proceedings may raise a plea of illegality.

112.

However, when reading Questions 1 and 2 together, and placed in the context of the present case, it appears that the referring court is in fact concerned with the ‘TWD scenario’: a situation where an applicant, who could have lodged an action for annulment against an EU act before the Court, but failed to do so, is precluded from challenging the validity of a national measure implementing that act before a national court.

113.

The TWD line of case-law, ( 63 ) as most recently refined in Georgsmarienhütte and Others, ( 64 ) maintains that where a person seeking to challenge an EU measure undoubtedly has standing under the fourth paragraph of Article 263 TFEU, that person is bound to make use of the remedy provided for in that provision by bringing an action before the General Court. Thus, the possibility for a litigant to plead, before the national court hearing its action, the invalidity of provisions in EU acts presupposes that the party in question had no right of direct action under Article 263 TFEU by which it could challenge those provisions.

114.

In the present case, the referring court seeks to know first whether a professional federation, such as FBF, could have challenged the EBA’s guidelines by way of an action for annulment in order to determine, in the circumstances of the present case, whether it is still able to challenge the legality of a national measure (the ACPR notice) making applicable the EBA’s guidelines.

115.

In the context of a case concerning a genuine soft-law measure, I do not think that such considerations are in any way pertinent. That being said, even if such considerations were pertinent, they are not warranted.

116.

First, since the EBA’s guidelines are not binding according to the Court’s traditional approach, ( 65 ) an action for annulment before the Court would in any event be inadmissible due to the nature of the EU measure at issue, regardless of whether the applicant is directly and individually concerned. The approach of the Court up to now has been that if there is a genuine soft-law measure, then there is in fact no challengeable act at all. In practice, such (objective) assessment even precedes any assessment of the (rather subjective) standing of a given person. Thus, in a way, if there is no challengeable act, the review of such measures cannot be sought by anyone under Article 263 TFEU.

117.

Second, and in any case, the TWD line of case-law would simply not be triggered in the case at hand, and, more broadly, hardly ever in any case concerning soft-law measures.

118.

The reason for that is simple: TWD and its offspring require there to be an ‘undoubted right’ to bring an action for annulment under Article 263 TFEU in order for the bar against a review by the same applicant under Article 267 TFEU to be triggered. ( 66 ) However, it is rather impossible to state that a professional federation which protects the interests of its members, such as FBF, would have such undoubted right to bring an action under Article 263 TFEU against non-binding EU measures such as the contested guidelines. Indeed, with regard to soft-law measures, it must always first be assessed whether or not such a measure is in fact a genuine or false non-binding legal act, which in itself is already a rather complex assessment. Next, the question would then be whether an association or other body defending the interests of its members would be individually and directly concerned so as to be endowed with standing. ( 67 )

119.

All that amounts to a type of complex assessment that is very different from the logic and purpose of the TWD exception. ( 68 ) That exception must be properly limited to manifest admissibility. It cannot be extended to potential admissibility, with the actual assessment of conditions under Article 263 TFEU being effectively pencilled into Article 267 TFEU and then imposed on the national judges to examine them in parallel. Nor can it be turned on its head by making it become manifest inadmissibility.

120.

It follows that, under EU law, a professional federation can certainly challenge, by means of a plea of invalidity, EU guidelines intended for the members whose interests it protects, even when they were not of direct and individual concern to it. When and how it may potentially do so under national law and procedure is another matter entirely and for the referring court to determine.

121.

Before turning to the overall structural issue raised by the referring court about the degree of (dis)connection between Articles 263 and 267 TFEU as regards the review of validity of soft-law measures, there is another piece of the puzzle that must be discussed: Foto-Frost.

122.

The judgment in Foto-Frost and its implications for national review of non-binding measures has been addressed by the referring court in its reference order, even if it has not led to a specific question being put to the Court. The present request for a preliminary ruling is indeed the result of the application by the referring court of both the judgments in Grimaldi and Foto-Frost in the specific context of non-binding measures where national law allows for direct review of soft-law measures while EU law does not. In such a scenario, are national courts under a duty to ask the Court about the validity of the contested guidelines in application of Foto-Frost? Or should it be construed as a simple faculty offered by the Grimaldi line of case-law? ( 69 )

123.

Foto-Frost stated that a national court is under the duty to make a reference to the Court on the validity of an EU-law measure unless it considers that the grounds put forward in support of invalidity are unfounded. ( 70 ) The stated rationale for that obligation, which finds no support in the text of the Treaty, ( 71 ) is twofold: first, uniformity is particularly imperative when the validity of an EU act is in question so that there are no divergences between national courts in this regard. Second, the coherence of the system of judicial protection and the EU ‘complete system of legal remedies’ (in particular the Court’s exclusive jurisdiction to declare EU measures void) require that, where the validity of an EU act is challenged before a national court, the power to declare that act invalid must be reserved to the Court. ( 72 )

124.

In contrast to the Commission, I do not see how that logic and the two reasons provided would also be applicable to genuine soft-law measures.

125.

First, as to the requirement of uniformity, it is true that, on a certain reading, those national competent authorities that decide to comply with the contested guidelines should do so in the same manner throughout the Union. In particular, Article 16(1) of Regulation No 1093/2010 empowers the EBA to adopt such guidelines with a view to ‘ensuring the common, uniform and consistent application of Union law’. Thus, the uniformity sought would essentially not be one within the Union and the internal market as a whole, but uniformity across the individual and rather accidental normative clusters limited to the Member States whose competent authorities decided to opt in for the guidelines.

126.

However, is this the kind of uniformity that Foto-Frost aimed at? Should that uniformity also be categorically required with regard to measures where there is by definition no such uniformity, as is the case with measures that merely may (as opposed to must) be applied? As far as the EBA guidelines and recommendations are concerned in particular, Article 16(3) of Regulation No 1093/2010 makes it clear that competent national authorities may decide not to comply. Moreover, if a national court decides, without the assistance of the Court, to annul the national measure implementing an EU non-binding measure, it will, in practical terms, have the same effect as the decision of the competent authority not to comply.

127.

Second, the argument drawn from the coherence of the system of judicial protection and the EU complete system of legal remedies is possibly even more intriguing. In Foto-Frost, the Court essentially suggested that the obligation to make a reference on validity is justified by the fact that the Court itself provides, with regard to EU-law measures, for an effective judicial review that is comparable to that guaranteed by national courts.

128.

However, at present, at the horizontal level, the access to review of non-binding EU-law measures is simply not possible under Article 263 TFEU because there is no challengeable act in the first place, which means that annulment actions are considered inadmissible. At the vertical level, the lack of equivalence in judicial protection appears to be even sharper when faced with (some) national courts, such as the referring court in the present case, that are in fact ensuring (some) effective judicial protection as far as non-binding measures are concerned. ( 73 )

129.

Thus, if anything, the coherence in respect of effective judicial protection requires, with regard to non-binding EU measures, exactly the opposite result: namely, that Foto-Frost clearly cannot be applicable to such measures. I can hardly see any sense in the argument which posits that, because the EU Courts do not provide any effective protection against genuine soft-law measures, that non-existent protection must be centralised before the Court.

130.

If that were to be the case, then I believe this might be one of the rare instances in which the application of Article 53 of the Charter of Fundamental Rights of the European Union (‘the Charter’) could very well come into play. There would be little doubt to my mind that a national legal system that allows individuals access to an effective judicial review against non-binding acts which affect their legal situation indeed ensures a high(er) level of judicial protection compared to that assured by the Court.

131.

In summary, on its logic and purpose, Foto-Frost is not applicable to non-binding EU measures. That means, in practical terms, that a national court is within its right to annul, if it is itself entitled to do so under national law, the national ‘incorporation’ or ‘implementation’ measure that made an EU soft-law measure applicable within the national territory, without first being obliged to submit a request for a preliminary ruling to the Court on that matter. Likewise, the ‘grounds of challenge parallelism’, set out by the Court in Melki and Abdeli, ( 74 ) which imposes a duty to refer, is not applicable to non-binding EU measures. That naturally does not prevent national courts from submitting a request for a preliminary ruling on the interpretation of a non-binding instrument or its parent law. However, the validity-centralisation logic embedded in Foto-Frost is simply not applicable to non-binding EU measures.

132.

Such a conclusion might appear radical, but only to those who internally believe something other than what they are stating externally. If one genuinely believes that EU soft law is not binding and does not generate any legal effects whatsoever, then why would one be appalled at the idea that national courts may do as they please with national measures implementing such mere guidance? Coherence is key here. Either one believes that such measures do in fact produce some effects (but, in that case, access to the EU Courts would then have to be granted), or one believes that there are no legal effects whatsoever. However, then the question becomes: why would there be a problem if a national court annuls it? At best, that court is engaging in a completely futile exercise, killing something that was always dead.

133.

The only option which poses a problem is the one suggesting that, as far as access to the EU Courts under Article 263 TFEU is concerned, an act has no binding legal effects whatsoever, with the consequence of there being no access to the EU Courts at all. However, when the same issue comes under Article 267 TFEU, the same act would then miraculously be resurrected and fully alive, with the Foto-Frost obligation even being triggered. The very troubling image emerging from such a dissonance is that the nature of an EU-law act would change depending on whether it concerns, on the one hand, access to the Court or, on the other, the obligations of national courts.

134.

The general, structural question that began to partially emerge in the previous parts of this section is the following one: to what extent is one able to make the claim that an EU act which is not legally binding (thus not reviewable under Article 263 TFEU) can still be reviewed and potentially even declared invalid under Article 267 TFEU? Indeed, both previous points, namely the issue of the TWD exception, as well as the logic underlying Foto-Frost, are built on the presumption that there is one EU-law system of judicial protection that ought to operate as a whole.

135.

However, there are also a number of arguments suggesting that some degree of dissociation between the two types of procedures is indeed possible, in particular with regard to non-binding EU-law measures.

136.

First, there is already a clear textual difference. Indeed, the first paragraph of Article 263 TFEU limits the scope of that provision to acts that have legal effects vis-à-vis third parties, to which the case-law of the Court added binding legal effects. Only such acts are challengeable under Article 263 TFEU. ( 75 ) By contrast, there is no such limitation in either the text or the established case-law pertaining to the scope of Article 267 TFEU. Indeed, as confirmed on a number of occasions by the Court in its Grimaldi line of case-law, Article 267 TFEU confers on the Court jurisdiction to give a preliminary ruling on the validity and interpretation of all acts of the institutions of the EU without exception. ( 76 )

137.

Second, while the Court stressed the coherence of the system, it also insisted on there being a ‘complete system of remedies’ offered by EU law. According to the Court, the Treaty has established, by Articles 263 and 277, on the one hand, and Article 267, on the other, a complete system of legal remedies and procedures designed to ensure judicial review of the legality of EU acts, and has entrusted such review to the EU Courts. ( 77 ) Ever since the judgment in Unión de Pequeños Agricultores v Council, ( 78 ) the ‘complete system of remedies approach’ has allowed non-privileged applicants to come to the Court under Article 267 TFEU when they could not under Article 263 TFEU because of the high threshold represented by the direct and individual concern condition.

138.

In other words, in order to be complete, the individual procedures must be complementary. That is the foundation stone on which much of the case-law in this area has been construed over the past 30 years, ever since the door (in terms of direct access) was (re)shut in Unión de Pequeños Agricultores. Within this logic, the conditions and access under Articles 263 and 267 TFEU have in fact been precisely dissociated: the relatively limited access under Article 263 TFEU was supposed to be complemented by very open access under Article 267 TFEU, with the national judges effectively becoming the gatekeepers.

139.

In the present case, it is the French Government, but also the referring court which has extensively quoted the Court’s case-law in that respect, that rely on that logic to establish the possibility of requests for preliminary rulings on the validity of non-binding EU measures when annulment actions are not possible.

140.

Third, that complementarity and the ensuing dissociation would also be necessary in order to safeguard some degree of access to an effective remedy that could be compliant with Article 47 of the Charter. Indeed, it would be wholly incompatible with robust requirements which the Court has recently established under that provision, ( 79 ) or even under Article 19 TEU, ( 80 ) if EU-law measures having an impact on the legal position of an individual could structurally never reach the Court.

141.

However, such a dissociation between Articles 263 and 267 TFEU with regard to the review of non-binding EU measures is not without its intellectual challenges, with some of them being rather structurally onerous.

142.

First, at the conceptual level, it is hard to understand, in view of the nature of non-binding EU measures, how a request for a preliminary ruling on the validity of non-binding EU measures can be allowed while an action for annulment cannot. An act which cannot be challenged by any applicants under Article 263 TFEU suddenly becomes a challengeable act open to everyone under Article 267 TFEU.

143.

Moreover, it is established case-law that requests for preliminary rulings which seek to ascertain the validity of a measure constitute, like actions for annulment, a means for reviewing the legality of EU acts. ( 81 ) While their outcomes, in the case of a finding of incompatibility between the EU measure at issue and higher EU law, are slightly different, formally (namely, annulment under Article 263 TFEU as opposed to a declaration of invalidity under Article 267 TFEU), they share the same erga omnes and generally ex tunc legal effects: the EU act at issue must be considered void and shall no longer apply. ( 82 )

144.

Thus, in all logic, either (i) EU acts are binding and they can thus be annulled or declared invalid under either Article 263 TFEU or Article 267 TFEU; or (ii) they are not binding, hence not reviewable, and a fortiori cannot be declared null or invalid under both Articles 263 and 267 TFEU.

145.

Second, if there is no act to be reviewed in the first place, it is not immediately obvious why the logic underpinning the ‘complete system of remedies’ should apply either. The Court’s case-law in that respect is, in essence, a tool for overcoming the lack of (subjective) individual standing in one procedure (Article 263 TFEU) by pointing out where that condition is not present in another procedure (Article 267 TFEU). The aim was to try to re-establish a sort of balance between privileged and non-privileged applicants; it is not really a tool for creating access to the Court under Article 267 TFEU when that access is closed to everyone, including privileged applicants, under Article 263 TFEU.

146.

Moreover, doing so would turn upside down the distinction – that has fuelled a lot of Article 263 TFEU case-law but is absent from Article 267 TFEU – between privileged and non-privileged applicants with regard to the review of non-binding EU measures. While, in Belgium v Commission, a privileged applicant (a Member State) was refused the possibility to challenge a non-binding EU measure as regards the issue of the EU’s lack of competence in adopting such a measure, ( 83 ) that same plea would, by contrast, then be open to any individual applicant, without having to show any direct and individual concern, through Article 267 TFEU.

147.

All of that would then mean, practically speaking, that in order to seek the review of a non-binding EU measure, privileged applicants would end up in the same, or sometimes an even worse, position as individual litigants. For instance, a Member State would have to challenge indirectly the implementation of (presumably its own) national measure before a national court. In the alternative, since it is unclear whether national procedural rules could even allow for such a type of challenge, Member State authorities would have to keep their fingers crossed to the effect that: (a) an individual applicant challenges the implementation measure, (b) the national court allows the State authorities to intervene in the main proceedings, and (c) that court seises the Court with a request for a preliminary ruling (on interpretation or validity) so that the State authorities may eventually plead their case before the Court. The situation would be even more intriguing with regard to another type of privileged applicant, namely EU institutions.

148.

Third and finally, there is also another, rather pragmatic argument as to why it makes limited sense to argue for a broad access to judicial review of non-binding measures under Article 267 TFEU, while access under Article 263 TFEU remains effectively closed: a sensible allocation of the docket within the Court of Justice of the European Union. That argument is not soft-law specific, but rather general to the judicial policy of the last 30 years of – across the board – restrictive access under Article 263 TFEU while allowing for unfettered access under Article 267 TFEU. ( 84 ) The consequence of that judicial policy has been that issues which would have otherwise been dealt with at the level of the judicial system where at present there is both the relevant expertise and the judicial capacity, namely the General Court, have instead arrived directly before the Court of Justice by way of requests for preliminary rulings.

149.

Grimaldi, Foto-Frost and Belgium v Commission: I find it quite difficult, or rather impossible, to apply all three of these judgments simultaneously in the context of a case such as the present one. Contemplating how one ought to reconcile all three judgments with regard to non-binding EU measures is reminiscent of a chess game in deadlock, in which, regardless of the move one decides to play, the result is an inevitable loss of at least one piece.

150.

To my mind, it is very clear which one of the three should be revised. However, that cannot be done through a preliminary ruling. In my view, the present request for a preliminary ruling simply serves to highlight, as a very useful example, what will happen if the Court remains trapped in decades-long dichotomies about the law being either 100% binding or non-existent. Eventually, someone else might feel obliged to step in and ensure the necessary legal protection that the Court is not willing to provide under Article 263 TFEU.

151.

However, with regard to the specific outcome of the present case, the Court has, at least theoretically, three options.

152.

First, it may duplicate the requirement that there be a legally binding act of EU law (that is to say, a challengeable act) in the first place from Article 263 TFEU into the possibility to submit a request for a preliminary ruling on validity under Article 267 TFEU. However, in such a case, the Court would be disregarding the wording of the first paragraph of Article 267 TFEU, overruling Grimaldi, and clearly making Foto-Frost not applicable to such ‘non-acts’. Moreover, there is something to be said about the requirement of ‘coherence’ of legal remedies, which results in there not being, in the interest of that coherence, a remedy at all. However, formally speaking, treating everybody equally badly is indeed, in a way, coherent.

153.

Second, the Court could rephrase the third question of the referring court as essentially asking about the interpretation of the legal basis of the contested guidelines, and not their validity. ( 85 ) That would amount to a partial overruling of Grimaldi, unless it could perhaps be squeezed under a ‘clarification’ of that judgment: when Grimaldi and all the case-law thereafter continued to insist on validity questions being possible with regard to ‘all EU acts without exception’, it has always been understood that what they actually meant was ‘all legally binding EU-law acts without exception’. ( 86 ) Consequently, Foto-Frost would not be applicable to such acts either.

154.

Third, if the Court were to find that the EBA exceeded its competence, it could naturally declare the contested guidelines invalid. That would mean upholding and applying Grimaldi, but indirectly revisiting Belgium v Commission. Of course, formally, the latter precedent and the previous line of case-law that it set in stone in relation to the nature of a challengeable act under Article 263 TFEU would remain intact. However, for all the reasons outlined in the previous section, relating to the degree of possible dissonance between Articles 263 and 267 TFEU, ( 87 ) I wonder for how long that differentiation could reasonably be maintained.

155.

That said, even in spite of all those issues, as long as there is no effective legal protection against potentially detrimental legal effects of non-binding EU measures under Article 263 TFEU, submitting a request for a preliminary ruling on validity under Article 267 TFEU with regard to those same acts remains the only way in which this Court may ensure that there is at least some resemblance of a complete system of remedies provided for in EU law. That is indeed the structurally dissatisfying, yet only possible, conclusion in terms of effective legal protection.

156.

I propose that the Court answer the questions referred for a preliminary ruling by the Conseil d’État (Council of State, France) as follows: