OPINION OF ADVOCATE GENERAL

BOBEK

delivered on 6 October 2021(1)

Joined Cases C‑59/18 and C‑182/18

Italian Republic (C‑59/18)

Comune di Milano (C‑182/18)

v

Council of the European Union

Case C‑743/19

European Parliament

v

Council of the European Union

(Action for annulment – Agencies of the Union – European Medicines Agency and European Labour Authority – Seats – Scope of Article 341 TFEU – Competence to decide on the seats of agencies – Decisions adopted by the representatives of the Member States in the margins of Council meetings to determine the location of the seats of agencies – Legal nature – Author of the act – Court’s jurisdiction under Article 263 TFEU – Absence of binding effects in the EU legal order)

I.      Introduction

1.        In November 2017, following the notification by the United Kingdom of Great Britain and of Northern Ireland of its intention to withdraw from the European Union, the representatives of the governments of the Member States selected, in the margins of a meeting of the Council of the European Union (‘the Council’), the city of Amsterdam (Netherlands) to replace London as the new location for the seat of the European Medicines Agency (EMA). In June 2019, the representatives of the governments of the Member States also decided by common accord that the newly founded European Labour Authority (ELA) would have its seat in Bratislava (Slovakia).

2.        In Joined Cases C‑59/18 and C‑182/18 (‘the EMA cases’), the Italian Republic and the Comune di Milano (Municipality of Milan, Italy) respectively challenge the decision of the representatives of the governments of the Member States to relocate the seat of the EMA to Amsterdam. In Case C‑743/19 (‘the ELA case’), the European Parliament challenges the decision of the representatives of the Member States to locate the seat of the ELA in Bratislava (Slovakia).

3.        Those challenges give rise to a number of important questions. First, can a collective decision of the representatives of the Member States be subject to an action for annulment under Article 263 TFEU? Second, do the decisions by the representatives of the Member States concerning the location of the seat of EU agencies fall within the scope of Article 341 TFEU? Third, what is the legal status, under EU law, of decisions of the representatives of the Member States that are not envisaged by the Treaties? Since those constitutional issues are common to the three cases at hand, I will deal with them jointly in this Opinion.

4.        Subsequent to the decision of the representatives of the governments of the Member States, Regulation (EU) 2018/1718 (2) provided that Amsterdam would be the location of the new seat of the EMA. That regulation was challenged by two actions lodged by the Italian Republic (C‑106/19) and the Commune di Milano (C‑232/19), respectively. I deal with those cases jointly in a parallel Opinion, delivered on the same day as the present one. (3)

II.    Legal framework

5.        Article 341 TFEU provides that:

‘The seat of the institutions of the Union shall be determined by common accord of the governments of the Member States.’

6.        On 12 December 1992, the representatives of the governments of the Member States adopted by common agreement, on the basis of Article 216 of the EEC Treaty, Article 77 of the ECSC Treaty and Article 189 of the Euratom Treaty, the Decision on the location of the seats of the institutions and of certain bodies and departments of the European Communities (‘the Edinburgh Decision’). (4)

7.        Article 1 of the Edinburgh Decision sets the respective seats of the European Parliament, the Council, the Commission, the Court of Justice, the Economic and Social Committee, the Court of Auditors and of the European Investment Bank. Article 2 of that decision provided that ‘the seat of other bodies and departments set up or to be set up will be decided by common agreement between the Representatives of the Governments of the Member States at a forthcoming European Council, taking account of the advantages of the above provisions to the Member States concerned, and giving appropriate priority to Member States who do not at present provide the sites for Community institutions’.

8.        Pursuant to Protocol No 6, annexed to the TEU and TFEU by the Treaty of Amsterdam, on the location of the seats of the institutions and of certain bodies, offices, agencies and departments of the European Union (‘Protocol No 6’):

‘The representatives of the governments of the Member States,

Having regard to Article 341 of the Treaty on the Functioning of the European Union and Article 189 of the Treaty establishing the European Atomic Energy Community,

Recalling and confirming the Decision of 8 April 1965, and without prejudice to the decisions concerning the seat of future institutions, bodies, offices, agencies and departments,

Have agreed upon the following provisions …:

Sole Article

(a)      The European Parliament shall have its seat in Strasbourg …

(b)      The Council shall have its seat in Brussels. …

(c)      The Commission shall have its seat in Brussels. …

(d)      The Court of Justice of the European Union shall have its seat in Luxembourg.

(e)      The Court of Auditors shall have its seat in Luxembourg.

(f)      The Economic and Social Committee shall have its seat in Brussels.

(g)      The Committee of the Regions shall have its seat in Brussels.

(h)      The European Investment Bank shall have its seat in Luxembourg.

(i)      The European Central Bank shall have its seat in Frankfurt.

(j)      The European Police Office (Europol) shall have its seat in The Hague.’

III. The background to the disputes and proceedings before the Court

A.      Cases C‑59/18 and C‑182/18 (the EMA cases)

9.        The European Agency for the Evaluation of Medicinal Products was created by Council Regulation (EEC) No 2309/93 of 22 July 1993 laying down Community procedures for the authorisation and supervision of medicinal products for human and veterinary use and establishing a European Agency for the Evaluation of Medicinal Products. (5)

10.      On 29 October 1993, the Heads of State or Government of the Member States decided by common agreement that it would have its seat in London (United Kingdom). (6)

11.      Regulation No 2309/93 was subsequently repealed and replaced by Regulation (EC) No 726/2004 of the European Parliament and of the Council of 31 March 2004 laying down Community procedures for the authorisation and supervision of medicinal products for human and veterinary use and establishing a European Medicines Agency. (7) By that regulation, the European Agency for the Evaluation of Medicinal Products was renamed the European Medicines Agency. The regulation did not contain any provision as to the location of the latter’s seat.

12.      On 29 March 2017, in accordance with Article 50(2) TEU, the United Kingdom notified the European Council of its intention to withdraw from the European Union.

13.      On 22 June 2017, in the margins of a European Council meeting (Article 50), the Heads of State or Government of the other 27 Member States accepted, on a proposal of the President of the European Council and of the President of the Commission, the procedure for the relocation of the seat of the EMA and of the European Banking Authority (‘the selection rules’). (8)

14.      Those rules provided, in particular, that the decision on the future seat of the two agencies should be taken on the basis of a fair and transparent decision-making process with an organised call for offers to host the seat based on specific objective criteria. Six criteria were laid down in paragraph 3 of the selection rules, namely (i) the assurance that the agency can be set up on site and take up its functions at the date of the United Kingdom’s withdrawal from the European Union; (ii) the accessibility of the location; (iii) the existence of adequate education facilities for the children of agency staff; (iv) appropriate access to the labour market, social security and medical care for both children and spouses; (v) business continuity; and (vi) geographical spread.

15.      It was stated in the selection rules that those criteria were based by analogy on criteria for the decision on the location of the seat of an agency set out in point 6 of the Joint Statement and Common Approach on Decentralised Agencies, (9) having special regard to the fact that the two agencies already existed and that the business continuity of the two agencies was vital and had to be ensured.

16.      The selection rules also provided that the decision would be taken by a voting process the outcome of which the Member States agreed in advance to respect. (10) In particular, it was stated that, in case of a tie between the remaining offers within the third voting round, the decision would be taken by the Presidency drawing lots between the tied offers, and the offer drawn would then be considered to be the selected one.

17.      On 30 September 2017, the Commission published its assessment of the 27 offers submitted by the Member States. (11) On 31 October 2017, the Council published a note supplementing the selection rules on practical questions regarding voting. (12)

18.      On 20 November 2017, the offer of the Italian Republic and that of the Kingdom of the Netherlands received the same number of votes within the third voting round. The Kingdom of the Netherlands’ offer was subsequently selected after drawing lots between the tied offers.

19.      As a result, on that same day, the representatives of the Member States selected, in the margins of the 3579th meeting of the General Affairs Council, the city of Amsterdam as the EMA’s new seat (‘the contested decision in the EMA cases’). That decision was announced in the minutes of the Council meeting (13) and was published by means of a press release. (14) In both the minutes and the press release, it was stated that: ‘the Commission will now prepare legislative proposals reflecting [the] vote for adoption under the ordinary legislative procedure with the involvement of the European Parliament. The Council and the Commission are committed to ensuring that these legislative proposals are processed as quickly as possible in view of the urgency of the matter.’

20.      On 29 November 2017, the Commission adopted a proposal for a draft regulation amending Regulation No 726/2004 as regards the location of the seat of the EMA. In the explanatory memorandum, that draft regulation notably stated that the ‘Member States, in the margins of the General Affairs Council (Art. 50), selected Amsterdam, the Netherlands, as the new seat of the [EMA]’.

21.      On that basis, Regulation No 726/2004 was amended by Regulation (EU) 2018/1718 of the European Parliament and of the Council of 14 November 2018. (15) It now provides that ‘the Agency shall have its seat in Amsterdam, the Netherlands’. (16)

22.      Against that background, the Italian Republic and the Comune di Milano initiated proceedings against the Council under Article 263 TFEU.

23.      In Case C‑59/18, the Italian Republic asks the Court to (i) order the Kingdom of the Netherlands, the EMA and any other institution or body to provide all necessary information establishing the suitability of the city of Amsterdam as the new seat of the EMA, (ii) annul the contested decision in so far as it established Amsterdam as the new seat of the EMA, and, (iii) allocate that seat to the city of Milan.

24.      The Council, supported by the Kingdom of the Netherlands and the Commission, asks the Court to dismiss that application as inadmissible or unfounded and to order the Italian Republic to pay the costs. Should the action of the Italian Republic be upheld, it also asks the Court to maintain the legal effects of the contested decision until a new selection procedure takes place.

25.      In Case C‑182/18, (17) the Comune di Milano, supported by the Italian Republic and the Regione Lombardia (Lombardy Region, Italy), asks the Court to annul the contested decision in so far as it established Amsterdam as the new seat of the EMA, and to order the Council to pay the costs.

26.      The Council, supported by the Kingdom of the Netherlands and the Commission, asks the Court to dismiss that application as inadmissible or unfounded and to order the Comune di Milano to pay the costs. Should the action of the Comune di Milano be upheld, it also asks the Court to maintain the legal effects of the contested decision until a new selection procedure takes place.

27.      On 17 April 2018, the Council raised a plea of inadmissibility in each of those cases in accordance with Article 151(1) of the Rules of Procedure of the Court of Justice. By decision of 18 September 2018, the Court decided to reserve its decision on that objection for the final judgment in the matter.

28.      On 2 July 2018, the Vice-President of the Court dismissed the application of the Comune di Milano to suspend the contested decision. (18)

29.      On 19 December 2019, the President of the Court decided to join the two cases for the purposes of the judgment.

30.      Written observations were submitted by the Italian Government, the Comune di Milano, the Regione Lombardia, the Council, the Netherlands Government, as well as the Commission.

31.      The Italian Government, the Comune di Milano, the Regione Lombardia, the Parliament, the Council, the Czech Government, Ireland, and the Spanish, French, Luxembourg, Netherlands, and Slovak Governments, as well as the Commission, presented oral argument at the hearing which took place on 8 June 2021. That hearing was organised jointly for the present Joined Cases C‑59/18 and C‑182/18, Joined Cases C‑106/19 and C‑232/19 (the EMA 2 cases), as well as for Case C‑743/19, Parliament v Council (Seat of the ELA).

B.      Case C‑743/19 (the ELA case)

32.      On 13 March 2018, the Commission submitted to the Parliament and to the Council a proposal for a draft regulation establishing a European Labour Authority (‘the draft regulation’). (19) The seat of that authority was to be mentioned in Article 4. However, during inter-institutional negotiations that took place from January to February 2019, the Parliament and the Council considered that they did not have the elements necessary to decide on the seat of the ELA and thus decided to postpone that choice. As a result, they removed Article 4 from the draft regulation and decided to state the reasons for that decision in a joint statement of the Parliament, the Council and the Commission that would be annexed to the draft regulation once adopted.

33.      On 13 March 2019, in the margins of a meeting of the Permanent Representatives Committee (Coreper I), the representatives of the governments of the Member States approved by common accord the procedure and the criteria for deciding on the ELA’s seat. (20) The selection rules made it clear that the criteria for the seat of the ELA were based on point 6 of the Common Approach annexed to the 2012 Joint Statement by the European Parliament, the Council and the Commission on Decentralised Agencies. (21) Those criteria are the (i) geographical balance, (ii) the date when the Agency can be set up on site after the entry into force of its founding act, (iii) accessibility of the location, (iv) existence of adequate education facilities for the children of agency staff, and (v) appropriate access to the labour market, social security and medical care for both children and spouses.

34.      The selection procedure rules also provided that offers to host the agency should be made to the Secretary-General of the Council and copied to the Secretary-General of the Commission and that they would be published on the Council’s website. The Commission would prepare a general assessment of all the offers and describe how each offer meets the criteria. The General Secretariat of the Council would then distribute that assessment to the Member States and make it publicly available. A political discussion among the representatives of the Member States would subsequently be held in the margins of a meeting of Coreper I. The voting process would later take place in the margins of an EPSCO Council in Luxembourg. It would consist of successive voting rounds that should not involve the drawing of lots. Voting would take place until one offer received the majority of votes. The decision on the seat of the ELA reflecting the outcome of the voting process would then be confirmed by common agreement of the Member States’ representatives at the same meeting. (22)

35.      On 16 April 2019, the Parliament adopted, through a legislative resolution, (23) the draft regulation at first reading. An annex to the resolution contains the abovementioned (24) joint statement of the Parliament, the Council and the Commission. In that statement, those three institutions ‘note that the process for selecting the location of the seat of the European Labour Authority (ELA) was not concluded at the time of the adoption of its founding Regulation. … The European Parliament and the Council take note of the Commission’s intention to take any appropriate steps in order for the founding Regulation to provide for a provision on the location of the seat of the ELA, and to ensure that the ELA operates autonomously in line with that Regulation’. (25)

36.      On 5 June 2019, on the basis of the evaluation carried out by the Commission of the four offers that were made (namely Sofia (Bulgaria), Nicosia (Cyprus), Riga (Latvia) and Bratislava (Slovakia)), the representatives of the Member States considered those offers in the margins of a Coreper meeting.

37.      On 13 June 2019, the Council approved the Parliament’s position at first reading. The proposed draft regulation was thus adopted pursuant to Article 294(4) TFEU. (26)

38.      On the same day, in the margins of that Council meeting, a majority of the representatives of the governments of the Member States cast their vote in favour of the offer made by Slovakia. They thus adopted by common accord Decision (EU) 2019/1199 whereby the ELA would have its seat in Bratislava (‘the contested decision in the ELA case’). (27) It was envisaged that that decision would be published in the Official Journal of the European Union and that it would enter into force on the date of its publication.

39.      On 20 June 2019, the Parliament and the Council adopted Regulation (EU) 2019/1149 establishing the ELA. (28) That regulation contains no provision regarding the location of the seat of the ELA.

40.      That regulation was published in the Official Journal on 11 July 2019. The contested decision establishing the seat of the ELA in Bratislava was, for its part, published in the Official Journal on 15 July 2019.

41.      Against that background, the Parliament initiated proceedings against the Council under Article 263 TFEU. The Parliament claims that the Court should annul the contested decision in the ELA case and order the Council to pay the costs.

42.      The Council, supported by all of the intervening Member States, claims the Court should dismiss the Parliament’s action as inadmissible or unfounded and order the Parliament to pay the costs. The Council also asks the Court, should the latter decide to uphold the Parliament’s action, to maintain the effects of the contested decision as long as is it necessary to determine a new seat for the ELA.

43.      Written observations were submitted by the Parliament, the Council, the Belgian, Czech and Danish Governments, Ireland, and the Greek, Spanish, French, Luxembourg, Hungarian, Netherlands, Polish, Slovak and Finnish Governments.

44.      The Italian Government, the Comune di Milano, the Regione Lombardia, the Parliament, the Council, the Czech Government, Ireland, and the Spanish, French, Luxembourg, Netherlands and Slovak Governments, as well as the Commission, presented oral argument at the hearing which took place on 8 June 2021. That hearing was organised jointly for the present Case C‑743/19, as well as for Joined Cases C‑59/18 and C‑182/18 (the EMA cases), and Joined Cases C‑106/19 and C‑232/19 (the EMA 2 cases).

IV.    Assessment

45.      This Opinion is structured as follows. First, I will consider the acts that can be challenged under Article 263 TFEU, and discuss the extent to which, if at all possible, the decisions of the representatives of the Member States may be subject to an action for annulment before the EU Courts (A). I will then examine the scope of Article 341 TFEU on the basis of which the contested decisions in the EMA and ELA cases were said to be adopted (B). It is only after having examined both of those points in detail that I will be able to assess the contested decisions in order to determine whether they may be challenged under Article 263 TFEU and to ascertain their legal nature under EU law (C).

A.      What is a challengeable act under Article 263 TFEU?

46.      The European Union is a Union based on the rule of law. It has established a complete system of legal remedies and procedures designed to enable the Court of Justice of the European Union to review the legality of acts of the EU institutions. (29) In that context, it is established case-law that actions for annulment, provided for under Article 263 TFEU, are available in the case of all measures adopted by the institutions, whatever their nature or form, which are intended to have binding legal effects. (30)

47.      Actions for annulment under the first paragraph of Article 263 TFEU are therefore admissible if they satisfy two conditions. First, they must have been adopted by EU institutions, bodies, offices and agencies (‘the authorship condition’). (31) Second, they must be directed at acts that produce binding legal effects (‘the binding act condition’). (32)

1.      The authorship condition

48.      The purpose of actions for annulment is to verify the lawfulness of EU action. Thus, the Court has jurisdiction over national acts only in exceptional cases, where those acts are merely preparatory to the adoption of a final decision of an EU institution within a composite administrative procedure, (33) or where that jurisdiction is unequivocally established in the Treaties. (34)

49.      In principle, the Court has no jurisdiction to examine the validity of measures adopted by national authorities. The Court has applied that logic both to measures taken individually by an authority of a single Member State, (35) and to collective acts of the Member States whereby the latter work together in exercising their individual competences.

50.      Examples within the latter category include, in particular, decisions of the representatives of the Member States, typically made in the margins of a Council meeting. (36) According to the Court, ‘acts adopted by representatives of the Member States acting, not in their capacity as members of the Council, but as representatives of their governments, and thus collectively exercising the powers of the Member States, are not subject to judicial review by the [EU Courts]’. (37)

51.      The latter statement was made in 1993 in the judgment in Parliament v Council, which concerned a decision adopted by the Member States at a session of the Council with a view to granting special aid to Bangladesh and which contained in one item of the minutes of a Council meeting. The same approach was recently confirmed by the Court in the specific context of a decision of the representatives of the governments of the Member States appointing members to the Court of Justice pursuant to Article 253 TFEU. (38) In that case, it was unequivocally stated that the relevant criterion used by the Court of Justice to exclude the jurisdiction of the EU Courts to hear and determine an action brought against acts adopted by the Member States in their capacity as Member States is that relating to their author, irrespective of their binding legal effects. (39)

52.      It follows that, in the absence of any provision contained therein that would bestow jurisdiction upon the Court pursuant to Article 273 TFEU, (40) collective decisions of the Member States are not amenable to judicial review under Article 263 TFEU, as follows from the clear wording of the latter and from the intention of the drafters of the Treaties. (41)

53.      The exclusion of the Court’s jurisdiction over national acts is in keeping with the classical distribution of (judicial) tasks within the European Union. The delineation between the judicial competence of the European Union and the Member States is indeed primarily one of formal paternity. It is for the Court to examine the lawfulness of EU acts. It is for national courts to examine the acts of the Member States, possibly in cooperation with the Court when they have an EU law-based content. (42) That distinction logically stems from the power to annul: the Court, in general, does not have jurisdiction to annul national acts, while national courts cannot annul EU acts. (43)

2.      The judicial re-classification of the author?

54.      To a certain extent, the judgment in Parliament v Council does, however, qualify that distinction with regard to collective acts of the Member States. In that case, although informed by the actors (the Member States) that what had been adopted was a ‘decision of the Member States’, the Court did not take that conclusion for granted. Instead, it proceeded to a substantive evaluation of the decision in question in order to determine whether or not those acts were, in reality, decisions of the Council. To that effect, the Court looked at the ‘content and all the circumstances’ in which the collective act in question was adopted. (44)

55.      In adopting that approach, the Court sought essentially to ensure that the appearance of an (unreviewable) decision of the Member States did not have the effect of concealing a (reviewable) decision of the Council. The logic underpinning that approach is that the content of a decision formally made by the Member States and the circumstances relating to its adoption may reveal that the Council is the true author of that decision. Thus, the Court may be led to determine, ‘ex post facto’, the true author of an act in cases where the identity of the author is disputed.

56.      That possibility, namely that the formal author of an act may subsequently be ‘corrected’ by the Court in the light of the content of that act and the circumstances in which it was adopted, calls for several observations.

57.      First, such content-based override to formal authorship can, logically, only be applied in order to determine whether an act was adopted by the Member States or by the Council/European Council. This is due to the double function or ‘double hat’ that the Member States possess in the context of EU integration, depending on whether they are carrying out intergovernmental functions (in their capacity as Member States) or supranational functions (in their capacity as members of the Council or European Council).

58.      As far as the other EU institutions are concerned, doubts as to the identity of the true author of an act simply cannot arise since their component members have no such dual function. What could perhaps arise, in very rare circumstances, would be factual disputes as to the identity of the author of an act in cases where the origin of a given document was unclear. However, at the normative, competence-based level, once the author is known, an act of the Commission will always be an act of the Commission. Likewise, an act of the Parliament will always be an act of the Parliament. Doubts may arise only with regard to acts of the Council and the European Council and to acts of the Member States.

59.      However, that observation reveals the true nature of the test in Parliament v Council. In the light of its wording, it appears to be a factual test (what is the content and what were the circumstances in which the contested act was adopted?). However, its true nature is rather a normative, competence-based override to possible factual situations (who, on a proper construction of the relevant provisions of EU law, was supposed to take a given decision?). Returning to the hat metaphor, the real question to be asked is, which was the proper hat that the Member States should have worn when making such a decision? Reasoning backwards from that finding, it may be necessary ex post facto to re-adjust the author and, in part, the nature of the act adopted by the Member States.

60.      Second, the content-based override to formal authorship is the result of the transposing the identification of the true author of an act of the approach used by the Court to determine the existence of binding legal effects in EU acts that could prima facie appear deprived thereof in view of their form. (45) In the latter context, the Court does not merely look at the ‘shell’ of the measure in question, but rather looks more closely at its actual content and purpose in order to ensure effective judicial protection against those acts of EU bodies that could in fact produce binding legal effects in spite of their (non-binding) form. This approach, applied to authorship, means that the Court will ensure that the act in question is not artificially disguised in an unreviewable act due to the mere statement that the Member States are the authors of the act.

61.      In 2017, the General Court applied that logic and found that, ‘notwithstanding the regrettably ambiguous terms of the EU-Turkey statement’ (46) on a joint action plan designed to strengthen cooperation in terms of managing migration, it was in their capacity as Heads of State or Government of the Member States that the representatives of those Member States had met with the Turkish Prime Minister in the premises shared by the European Council and the Council. (47)

62.      In order to come to the conclusion that the statement in question was not attributable to the European Council, the General Court successively examined several factual elements, such as the nature of the meetings that had previously been held by the Heads of State or Government of the Member States; the presentation of the statement in comparison with the previous statements; the terms used, the content of the statement; the form in which it was published; the documents relating to the meeting where the statement in question was adopted. The General Court went even as far as to consider the dinner and lunch invitations in order to conclude that two separate events (the meeting of the European Council and an international summit) had been ‘organised in parallel in distinct ways from a legal, formal and organisational perspective, confirming the distinct legal nature of those two events’. (48)

63.      In my view, the orders of the General Court demonstrate why transposing the approach taken with regard to the content-based override, used in determining the binding nature of the act, to the decision on its authorship (or rather to override authorship), is problematic. Examining circumstantial, factual evidence makes sense when attempting to establish the existence of binding legal effects because such an analysis focuses on the intrinsic characteristics of an act of an institution that has the necessary powers to adopt a binding act. That is, however, not the case when determining the rather normative issue of whether the Member States or the Council (or the European Council) are supposed to be the author of the act.

64.      Moreover, in practice, examining factual elements for the purposes of such a determination is bound to result in an inconclusive finding as regards the identity of the true author of the act in question. That is because a collective decision taken by all the Member States in the margins of a Council meeting is bound to look, in terms of the facts, very similar to a decision of the Council (or of the European Council). Both decisions are indeed likely to involve the same actors (namely the Member States), have an EU-related content, be published in the Official Journal, and involve to some degree the Commission and the Council, at least through the latter’s secretariat (and its legal service in the event of litigation). (49)

65.      In addition, what is perhaps most confusing is that, if the (true) authorship test were indeed a factual test (who signed a treaty, or who adopted a decision), and not a normative, competence-dependent one (who should rightly have signed a treaty, who should rightly have adopted a decision), then one would normally expect the evidence to be produced, and in particular requested by the EU Courts, to be that which is directly relevant and conclusive for the factual issue to be settled in the relevant proceedings. If the factual dispute before a court of law is about ‘who signed the relevant document’, then most courts would simply ask the parties to produce that (if need be confidential) document in order to see whose signature is on the document. It is only where no document in fact exists, and where that is confirmed by the relevant witness statements or sworn affidavits, that one might ultimately be obliged to engage in judicial review of dinner menus. (50)

66.      In the present cases, the arguments put forward by the Italian Government and the Comune di Milano in the EMA cases, and by the Parliament in the ELA case, constitute a further illustration of the inconclusiveness of the ‘content and circumstances’ approach to determining ‘under which hat’ the Member States have adopted the contested decisions. Questions such as whether the Council’s infrastructure and facilities have been used, or whether other EU institutions have somehow been involved in the decision-making process, are of little help in determining the function under which the Member States (should) have acted.

67.      This point highlights, yet again, the fact that such factual issues will offer little guidance in cases where the real issue is one of competence. The real question is not who was the author, but who should had been the author. Factual evidence will primarily (if not, only) be relevant in those cases where the identity of the formal author is simply unknown, thus where there is no particular indication in the act itself identifying either the Member States or the Council.

68.      Third, there is also the issue of the potential ‘direction’ of the override. Does it work, or rather should it work, both ways? In the 1993 judgment in Parliament v Council, the introduction of the content-based override to formal authorship was apparently based on the idea that the Court should assert jurisdiction under Article 263 TFEU over those acts that have been formally adopted by the Member States and are thus prima facie unreviewable, but which appear to be acts of the Council in disguise. Thus, what had previously formally been excluded was then included again.

69.      The direction of the General Court’s orders discussed above was, however, exactly the opposite. What was, at least in a press statement, formally referred to as an ‘EU-Turkey’ arrangement, was subsequently re-classified by the General Court as having actually been adopted not by the Council (or the European Council), but by the Member States. Thus, what apparently previously fell under the jurisdiction of the EU Courts became excluded from that very jurisdiction.

70.      Fourth and finally, the foregoing discussion underlines the true nature of the test, alluded to above: if the test developed in Parliament v Council as regards the identity under which the Member States were acting in an individual instance is to make any logical sense, it must, by its nature, be normative, not factual. The real issue to be dealt with there does not concern the author (who will always be the same 27 persons in a room), but the function they were exercising when adopting the act (were those 27 persons, in that moment, acting in the Council or acting as the representatives of their governments?). That latter issue is thus inevitably one of power and competence. Does the act at issue pertain to a competence of the Member States (so that it should have been adopted by the Member States acting as the Member States) or to a competence of the Union (so that it should have been adopted by the Member States acting as the Council)?

71.      More precisely, as far as the contested decisions are concerned, should they be attributed to the Council because deciding on the seat of an agency falls within the competence of the Union, and such a decision could therefore only be made by the Parliament and the Council, and not the Member States? That is the position essentially embraced by the Italian Government and the Comune di Milano in the EMA cases. For its part, the Parliament has not gone as far as stating that the contested decision in the ELA case was attributable to the Council on account of the fact that that decision should have been taken by the Council. However, the Parliament has devoted a significant amount of its reasoning to suggest that the competence issue is a relevant element, maintaining that the contested decision can be subject to an action for annulment under Article 263 TFEU.

3.      If known, the author shall, in principle, remain the author

72.      It is certainly true that the (re)classification, on the basis of EU competence, of a collective act of the Member States into an act of the Council would present a number of advantages. First, it would indeed robustly protect the integrity of the EU legal order by ensuring respect for EU competences and for the institutional balance within the Union. Second, it may enhance judicial protection (by the EU Courts) in as much as individual national courts are arguably not fully competent to examine the lawfulness of collective decisions of the Member States, even though they may potentially be competent to examine the participation of their respective governments within the collective decision-making process.

73.      However, such normative, competence-based override to formal authorship also has significant drawbacks. On balance, those drawbacks heavily outweigh the potential benefits.

74.      First and foremost, in contrast to the (binding or otherwise) nature of an act, where substance should naturally prevail over form, authorship is authorship. Again, leaving aside the rare cases in which the authorship is indeed factually unclear, it is difficult to extend the same approach to instances where authorship is factually clear, but it is being disputed by other actors merely in terms of power and competence. Again, if it were not for the double-hatted Council/Member States, the issue would not even arise. Indeed, it would be rather bizarre for someone to seek to initiate, for example, an action against the Commission by suggesting that the act in question was in fact adopted by the Parliament.

75.      Second, with regard to the Member States specifically, it goes without saying that, beyond or outside the framework of EU law, they are sovereign. The rule is therefore that the Member States remain free to act, unless restricted by a specific provision of EU law.

76.      Third, how would EU competence be ascertained within such a context where, by default, the Member States are free to act? To start with, nowadays, virtually any matter can potentially be considered to fall within (a certain type of) EU competence. As regards the vast areas of shared competences, it is not always an easy task to determine in practice whether a certain matter has already been ‘pre-empted’ by the Union or whether the Member States can still adopt autonomous rules. (51) In particular, would it be necessary, for the purposes of pre-emption, for the Commission to have formally launched the ordinary legislative procedure, expressly including the matter at issue within its scope, or would it be sufficient to find that that matter is generally governed by EU law?

77.      In addition, it is unclear whether a competence-based override would apply only to the examination of decisions of the Member States in the margins of the Council or also to decisions of the Member States outside the meetings of the Council. Should it be conclusive whether the same 27 persons met physically or virtually in Brussels, or anywhere else in fact? It is also difficult to tell whether such override could operate only in the context of decisions of all the Member States or also with regard to the decisions made by just a majority of the Member States, or even just a few Member States. In extremis, could even a decision of the Member States collectively participating in the decision-making of another international organisation, which, as to its substance, overlaps with an EU competence, be subject to review by this Court under Article 263 TFEU?

78.      I am certainly not suggesting that that should be the case. What I merely wish to suggest is that in terms of a feasible test, those situations are difficult to distinguish. A competence-based override to authorship is therefore, at least in my view, fraught with too many uncertainties, which render its use rather uncertain and arbitrary, and, on the whole, rather counterproductive.

79.      For these reasons, I do not subscribe to the overall approach put forward by the Italian Government and the Comune di Milano in the EMA cases, suggesting, in essence, that, because a decision on the location of the seat would rightly be for the Union to take, any pronouncement of the Member States on that matter automatically constitutes a decision of the Council.

80.      Instead, I would suggest limiting the logic of the judgment in Parliament v Council to exceptional scenarios, in which the power to make a given decision, within a clearly defined and, typically, already ongoing procedure, undoubtedly rested with the Council, in its capacity as an EU institution. It is only if, within such a clearly circumscribed procedural context, a decision is unexpectedly made which is, formally speaking, a decision of the Member States, that one could indeed contemplate re-classifying that decision, despite its formal authorship, as a decision of the Council and thus as a decision amenable to review under Article 263 TFEU. However, such exceptional scenarios should then be governed by the much narrower logic underpinning scenarios in which there is clear circumvention of the extant rules within ongoing procedures, and should not be deemed to exist solely on the ground that there is an overlap of competences in the abstract, where no procedures are ongoing.

81.      Apart from such an exceptional scenario, it is my view that a formal decision of the Member States is simply a true decision of the Member States that falls outside the scope of review by this Court under Article 263 TFEU. If the formal author is known, it should not be for the Court to change it. The content, the factual circumstances or the lack of competence should offer guidance only in circumstances where there is true factual uncertainty as to whether an act has been adopted by the Member States as such or by the Member States acting as the Council. Those elements are not to be used in circumstances where it is simply a matter of challenges to competence by other actors.

82.      It follows that, as a matter of principle and certainly in view of how the Treaties currently stand, the Court has no jurisdiction under Article 263 TFEU over decisions taken by the representatives of the Member States.

B.      The scope of Article 341 TFEU

83.      According to the Court, decisions of the Member States adopted in the framework of the Treaties under Article 253 TFEU cannot be subject to an action for annulment. (52) Article 341 TFEU is drafted in a similar manner in as much as, first, it also concerns decisions determined by the common accord of the governments of the Member States and, second, the Member States are empowered by that provision to decide on the seat of the EU institutions. It thus logically follows that the approach which applies to Article 253 TFEU also applies to decisions of the Member States correctly taken within the proper scope of Article 341 TFEU.

84.      However, the crucial question is whether the decision on the seat of EU agencies is to be made by the Member States under Article 341 TFEU. Does that type of decision pertain to the Member States, as argued by the Council, or is it a matter for the EU legislature, that is, the Parliament and the Council, as argued by the Italian Government, the Comune di Milano and the Commission in the EMA cases, and by the Parliament in the ELA case? Should the latter be indeed the case, what then is the legal nature of decisions taken by the Member States establishing the location of the seat of EU agencies?

85.      In their submissions, the parties to the ELA case have addressed at length the issue of the scope of Article 341 TFEU and whether it could be a valid legal basis for adopting the contested decision regarding the ELA’s seat. Although those submissions are put forward under the first ground for annulment raised by the Parliament, their respective positions are also valid with regard to the matter of the jurisdiction of this Court. Indeed, according to the Court, the assessment of the admissibility of an application for annulment is bound up with the assessment to be made of the complaints levelled against the contested act. (53)

86.      According to the Parliament, Article 341 TFEU is not an appropriate legal basis for deciding on the seat of EU agencies, that provision being only applicable to institutions. Had the drafters of the Treaties intended to include agencies within its scope, they would have done so expressly. However, Article 341 TFEU has never been amended to that effect. In the light of its own text, Protocol No 6 cannot have been intended to broaden the scope of that provision. Furthermore, not only is the past institutional practice irrelevant for determining the scope of Article 341 TFEU, but the EU legislature has also, in the past, made decisions on the seat of its agencies. It cannot thus be argued that it is for the Member States alone to make such a decision.

87.      Nevertheless, according to the Parliament, the decision to establish an agency is incumbent upon the EU legislature which, to that effect, may rely on the legal basis for the corresponding policy. It is on the basis of Articles 46 and 48 TFEU that Regulation 2019/1149 establishing the ELA was adopted. There is nothing to justify singling out the decision on the seat of an agency from the decision on its creation and the defining of its missions, organisation and functioning. Unless a specific provision of the Treaties were to provide otherwise, the decision on the seat of an agency should thus be made by the EU legislature. Although political and symbolic considerations come into play when deciding on the geographical location of an agency, that matter is not uncommon to the creation of the agency, the defining of its missions and, accordingly, its organisation and functioning. As such, it is logical and consistent that the decision on the seat should be made by the EU legislature. In any event, the political sensitivity of that matter should not alter the competences conferred by the Treaties on the EU institutions.

88.      Apart from some minor variations, all of the intervening Member States embrace the Council’s view that Article 341 TFEU, interpreted in the light of its aim and context, empowers the Member States to decide on the seat of EU agencies. Although, since the Treaty of Lisbon, several provisions of the Treaties now expressly refer to institutions and to bodies, offices and agencies of the Union, that should not lead to a restrictive interpretation of the scope of those provisions that refer only to ‘institutions’. That interpretation is supported by the context and the historical evolution of that provision. In particular, Protocol No 6 sets out the location of the seat of two committees and one agency. For its part, Article 2 of the Edinburgh Decision further confirms that the seat of future EU bodies is to be decided by the representatives of the governments of the Member States. Furthermore, many precedents over several decades confirm that the competence of the Member States to decide on the seat of the institutions also applies to agencies. A restrictive interpretation of Article 341 TFEU would unduly single out agencies and exclude them from the application of the final provisions of the TFEU.

89.      In addition, according to the Council, the competence pertaining to the decision on the seat of an agency is distinct from the competence pertaining to substantive regulation of a certain field. The decision on the seat is fundamentally different in nature because of its political and symbolic dimension. Accordingly, the decision on the seat of an EU agency is not to be taken using the ordinary legislative procedure. Bringing the choice of the seat within scope of the legislative debate would have paradoxical and detrimental consequences. The geographical location of an agency would then become part of negotiations and would have an impact on the substance of the policy rules themselves. Furthermore, the voting rules within both the Parliament and the Council would not allow for a proper geographical balance to be struck.

90.      I disagree with the Council. In my view, Article 341 TFEU does not apply to agencies.

1.      Text

91.      The wording of Article 341 TFEU refers to the seat of the ‘institutions’ without any further specification. However, since the entry into force of the Treaty of Lisbon, the concept of ‘institutions’ has been defined at the constitutional level in Article 13 TEU, which indeed lists the EU institutions. It clearly follows from that legal definition provided by the Treaties that agencies are not institutions.

92.      Such a clear, legal, or even constitutional definition of the term to be interpreted does not preclude additional examination of the general scheme and the purpose of the rules of which that term forms part. However, very persuasive arguments under the heading of context or purpose would be needed, showing that the literal and natural reading of a clearly defined term would, in a given context, result in blatant injustice or absurdity, to justify a judicial disregard of such an unambiguous constitutional definition.

2.      Context and system

93.      In my view, there are, however, no such persuasive arguments to the contrary. In fact, the literal interpretation of Article 341 TFEU finds further confirmation in the context and system of that provision.

94.      First, the Treaty of Lisbon has amended numerous provisions of the Treaties, thereby distinguishing EU institutions from other ‘bodies, offices or agencies of the Union’. Indeed, several provisions concerning the Court’s jurisdiction have been amended to clearly include that latter set within its scope. (54) The Treaty of Lisbon has also extended the prohibition for the European Central Bank (‘the ECB’) to grant credit facility (55) to ‘institutions, bodies, offices or agencies of the Union’ and the Ombudsman’s control of maladministration. (56) Similarly, the Statute of the Court of Justice of the European Union now also distinguishes between institutions, bodies, offices and agencies of the Union. That is most notably the case in Articles 23 and 40 thereof: while the institutions may intervene in all cases before the Court, bodies, offices and agencies of the Union may do so as long as they establish an interest in the result of a case.

95.      In essence, it is clear that the Treaties have been revised recently enough to acknowledge the autonomous existence of ‘bodies, offices or agencies’ of the Union that are distinct from institutions. Within such a constitutional context, the concept of ‘institutions’ cannot be, or certainly can no longer be, simply interpreted broadly.

96.      Yet, according to the Council, the fact that the general and final provisions of the TFEU – of which Article 341 TFEU is part – have not been amended and refer only to ‘institutions’ should not be interpreted as the intention of the drafters of the Treaties to narrow down their scope. Since those provisions possess a horizontal character and therefore have transversal applicability across the Treaty, those general and final provisions should actually be interpreted in a broad manner.

97.      I disagree. There is no systemic, inherent reason why the general and final provisions should be singled out and interpreted differently to the other provisions of the Treaties. Furthermore, in concrete terms, the argument of the Council relating to Article 341 TFEU is not very persuasive.

98.      Article 342 TFEU, which concerns the language regime of the EU institutions, (57) in fact applies only to the latter. Certainly, regulations setting up agencies often declare applicable to those agencies the provisions of Council Regulation No 1 of 15 April 1958 determining the languages to be used by the institutions. (58) However, there are also regulations that leave it for the agency’s management board to decide on the internal language arrangements. (59) More importantly, the language regime of an agency can be different from that in force in the institutions, as confirmed by the Court. (60)

99.      The only exception that seems, at least at first glance, to support the arguments put forward by the Council is the second paragraph of Article 340 TFEU. That provision, which concerns the non-contractual liability of the Union, has been interpreted by the Court in the sense that the term ‘institution’, within the meaning of that provision, encompasses not only the EU institutions listed in Article 13(1) TEU but also all the EU bodies, offices and agencies that have been established by or under the Treaties and are intended to contribute to the achievement of the European Union’s objectives. (61)

100. However, upon closer inspection, such a reading is indeed warranted by both the letter and the logic of the second paragraph of Article 340 TFEU. As to its text, that provision in fact states that ‘the Union shall … make good any damage caused by its institutions or by its servants in the performance of their duties’. Logically, therefore, if ‘servants’, or in other words the individual agents of the Union, are already clearly included in its text, then, a fortiori, all the agencies, authorities, or other bodies that form the spectrum between an institution of the Union, on the one hand, and an individual servant of the Union, on the other, must also be included therein. As to its logic, it would indeed be peculiar to interpret that specific provision, which offers an important individual remedy vis-à-vis wrongs potentially committed by the Union as a legal entity, in such a strikingly formalistic manner which, in essence, suggests that only institutions and servants may be liable, but that agencies or other bodies cannot be, even though they are clearly already implicitly included as part of the spectrum of representatives of the Union, given the express reference to the two extreme ends of that spectrum.

101. As such, the second paragraph of Article 340 TFEU undoubtedly includes agencies. It is, however, simply a very different provision, both in terms of its text and logic. In view of that clarification, I cannot subscribe to the argument that, because Article 341 TFEU comes after Article 340 TFEU, in a cluster of various general and final provisions, both must therefore automatically be interpreted in the exact same way, irrespective of the objective differences.

102. I therefore come to the conclusion that, the fact that the Treaty of Lisbon did not amend Article 341 TFEU, while most other provisions of a similar nature were changed to the effect of expressly including other ‘bodies, offices or agencies of the Union’ within their respective scopes, means that the constitutional definition of institutions laid down in Article 13 TEU can hardly be set aside.

103. Second, that view finds further support at a more systemic level: the institutions of the Union are constitutionally different from bodies, offices and agencies of the Union. Institutions are established by the Treaties themselves. They carry out broad constitutional functions across areas governed by EU law where the Union has competence. Most importantly, their establishment and functions are provided directly by the Treaties themselves. The legal regime of the institutions is thus self-standing and does not require any secondary legislation. It is thus quite logical that a special constitutional procedure, that which is provided for in Article 341 TFEU, would apply in relation to decisions regarding their seats. Where else would one expect to find either the seat, or at least the procedure for establishing the location of the seat, for an institution established by the Treaty than in the Treaty itself? Equally, in view of the historically high sensitivity of that matter, it is understandable that the decision must be made by the consensus of the Member States.

104. By contrast, agencies are normally not created by the Treaties. They are established by secondary legislation, under the ordinary legislative procedure, the implementation of a given EU policy. Their role is largely administrative and clearly limited to a narrowly defined, specialised subject matter. The relevant substantive provisions of the Treaties regarding policies form the legal basis for the creation of those agencies. (62) Thus, in strictly constitutional terms, there is no inherent, systemic need as to why the procedure for the selection of their seat should suddenly need to be determined by a provision of the Treaties.

105. Third, in that regard, I also fail to see how, in conceptual terms, the decision on the location of the seat of an agency would be an issue distinct from the creation of that agency – a position put forward by the Council and the intervening Member States. The issue of the location of the seat is in my view part of the organisation of the agency, which is governed by the appropriate legal instrument establishing that agency. There is no clear structural argument as to why, first, the decision on the seat would follow a different legal regime than the decision on its creation and, second, why that former decision would be made following the same legal pattern as that for institutions, that is through Article 341 TFEU.

106. The Council submits that the decision on the seat of an agency is a political one. Naturally, I am not denying the fact that such a decision does indeed have a political dimension. However, it is unclear why, by implication, the Parliament, which represents European citizens, would be unable to make such a decision or why the Council itself, as the EU institution gathering the representatives of the Member States together at ministerial level, could not decide on that matter. In general, I find the argument that the Parliament is not well-placed to make political decisions rather counter-intuitive to say the least. (63)

107. In any case, the decision on the location of the seat of an agency can hardly be considered a purely political decision, at least if the criteria stated in both of the present cases are to be given any credence. The respective selection rules have laid down several criteria that are all more technical in nature, namely the date when the agency could be set up on site; the accessibility of the location; the existence of adequate education facilities for the children of agency staff and appropriate access to the labour market; and the social security and medical care for children and spouses. (64) Geographical balance appears to be the only criterion that is predominantly political in nature.

108. It thus follows that decisions on the seat of EU agencies may very well, in view of their mixed nature, be taken by the Parliament and the Council within the ordinary legislative procedure, triggered by a proposal from the Commission.

3.      The past

109. As regards the arguments raised by the Council in relation to the genesis of Article 341 TFEU, the impact of Protocol No 6 and of the Edinburgh Decision, and the past institutional practice, none of those arguments appear compelling enough to reverse the interpretation that follows already from the text, as well as the context and system of Article 341 TFEU.

110. First, regarding the evolution of Article 341 TFEU, the content of that provision has remained the same ever since 1951. (65) The only difference lies in the fact that the term ‘institutions of the Community’ has been replaced with ‘institutions of the Union’. Crucially, the Treaty of Lisbon did not amend that provision expressly to include bodies, offices and agencies of the Union. (66)

111. Second, the Council and, within the ELA case, the Luxembourg Government in particular, have extensively relied on Protocol No 6 and on the Edinburgh Decision, adopted in the margins of a European Council meeting in 1992, to argue in favour of the inclusion of agencies within the scope of Article 341 TFEU.

112. As regards Protocol No 6, it is true that it does not set the seat of institutions only, but also that of two committees and one agency, namely Europol. Thus, it is clear that the seat of those bodies has been enshrined in primary law. However, Protocol No 6 does not in any way state that the seats of all (possible, future) agencies are to be determined by the Member States on the basis of Article 341 TFEU, quite the contrary. In adopting a specific protocol (thus in amending the Treaties under the amendment procedures provided for therein), the Member States clearly considered that their collective decision needed to be enshrined in the Treaties in order to produce legal effects under EU law.

113. As regards the Edinburgh Decision, Article 2 of that decision provides that: ‘the seat of other bodies and departments set up or to be set up will be decided by common agreement between the Representatives of the Governments of the Member States at a forthcoming European Council, taking account of the advantages of the above provisions to the Member States concerned, and giving appropriate priority to Member States who do not at present provide the sites for Community institutions.’ (67) It follows from that statement that the representatives of the governments of the Member States wished to reserve for themselves the decision on the seats of agencies (as ‘other bodies and departments’) to the same extent that they are expressly and clearly empowered by Article 341 TFEU to decide on the seats of the institutions.

114. In the light of Article 2 of the Edinburgh Decision, I therefore agree that it indeed appears that the Heads of State or Government might have contemplated expanding the scope of Article 341 TFEU to agencies. Incidentally, it would, in fact, also appear that by seeking to adopt the same article, its drafters admitted that Article 341 TFEU did not encompass agencies (if it did, then why would it be necessary to draft a special protocol to that effect?).

115. However, I disagree on the legal consequences to be drawn from Article 2 of the Edinburgh Decision. In my view, the nature of that specific provision is that of an international agreement between Member States. Since that decision was taken outside the revision procedures laid down in Article 48 TEU, it cannot be accepted as a valid way to amend Article 341 TFEU. It thus has no binding legal effects within the EU legal order (which does not preclude it from having such effects under international law).

116. To be clear, a decision of the representatives of the Member States, such as the Edinburgh Decision, which was not taken under a provision of EU law, has legal value under EU law only to the extent that its content has been formally taken over by EU law, following the procedures laid down by EU law. (68) In the past, that ‘incorporation’ has usually taken the form of the adoption of a protocol, as notable examples of which are Protocol No 22 on the position of Denmark following that Member State’s initial refusal to ratify the Maastricht Treaty (69) or the Protocol on the concerns of the Irish people on the Treaty of Lisbon. (70)

117. So far as concerns the Edinburgh Decision, only Article 1 thereof was incorporated through Protocol No 6, thereby becoming EU law in 1999 with the entry into force of the Treaty of Amsterdam, to which that new protocol was annexed. For its part, Article 2 was, to my knowledge, never incorporated in any formal EU law instrument, not even an (interpretive) declaration. It is thus neither part of primary law, nor of secondary law. (71)

118. It follows that the only value to be ascribed under EU law to Article 2 of the Edinburgh Decision might at best be political, but certainly not legal. At most, it could be taken into consideration as an instrument for the interpretation of the Treaties. (72)

119. In any event, it should be borne in mind that the authors of the Edinburgh Decision did not make Article 2 thereof part of primary law. Therefore, the interpretive guidance provided by that provision cannot in any case run counter to the letter, context, system and purpose of Article 341 TFEU.

120. Third, the Council has relied largely on the previous institutional practice regarding the decision on the seat of agencies to argue that such decisions have often been made by a decision of the representatives of the Member States pursuant to Article 341 TFEU. In a nutshell, in the Council’s view, the fact that there has been a certain institutional practice over the course of the past decades not only confirms that Article 341 TFEU includes agencies, but also justifies that practice today and, apparently, also makes it permanent for the future.

121. I disagree. On the one hand, I simply cannot accept that position at the merely factual, empirical level. In view of all information that has been brought to the attention of this Court in the context of the present proceedings, there appears to be quite a degree of variation in the ways in which the decisions on the seats of EU agencies or bodies were made.

122. Certainly, for a number of agencies in the past, the decision on the location of their seat was made by the Member States, not by the EU regulation establishing them. That was especially the case in 1993 for nine agencies, including the EMA’s predecessor, (73) in 2004 for nine other agencies and offices, (74) in 2009 for the Agency for the Cooperation of Energy Regulators, (75) and most recently for the ELA and the European Cybersecurity Industrial, Technology and Research Competence Centre. (76) It must be noted that some of those agencies also had their seat subsequently mentioned in the regulation establishing them. (77)

123. However, there has also been a practice in the past whereby the seat of agencies was decided by the EU legislature, namely the Council acting initially on its own and then, with the developments of the EU constitutional framework, by the Parliament and the Council. That was, in particular, the case for the very first EU agencies in the 1970s, namely Cedefop (78) and Eurofound. (79) That was also later the case for Frontex, (80) the European Fisheries Control Agency (81) and, more recently, the three agencies that were created in the wake of the global financial crisis (the European Banking Authority, the European Securities and Markets Authority, and the European Insurance and Occupational Pensions Authority). (82) The same is true for those agencies that fall within the scope of the common foreign and security policy (CFSP), which have their seat decided on by the Council (alone), not by the Member States. (83)

124. In addition to these two scenarios, there have also been other, less frequently-used, ways of deciding on the seat of an EU agency. That was, for instance, the case for CEPOL where, somewhat unusually, the decision was actually made by several, changing actors. (84)

125. It is clear the manner in which decisions on the seats of agencies have been taken has varied over the past decades. It is impossible to identify clearly one institutional practice. In fact, ever since the establishment of the first agencies, the institutional practice has constantly evolved.

126. On the other hand, and in any case, even if it were possible to identify a single coherent practice over the years, quod non, the question is, what ongoing normative relevance does such a finding have for the present case?

127. On this point, I am bound to disagree with the Council. To me, in a Union based on the rule of law, it is the current constitutional framework, and not past practice, that is decisive. There is, certainly in my view, a duty, incumbent on the EU legislature, to keep the legislation relevant and reasonably up to date, responsive to social and societal developments. (85) It is, however, quite a different matter, to suggest essentially that the current constitution is to be (re-)interpreted in order to accommodate and to prolong some past practice, regardless of the changed constitutional environment. Simply put, it is not the current constitution that has to be brought into line with past practice, but rather the opposite: the present and future practice must be brought into line with the current constitutional framework.

128. However, that indeed begs the question: what is the current purpose of Article 341 TFEU? That is the issue to which I will now turn.

4.      The (present) purpose

129. According to the Council, an interpretation of Article 341 TFEU that would exclude agencies from its scope would deprive that provision of any effectiveness (effet utile). The seat of the institutions themselves has already been decided by primary law. One cannot consider that the effectiveness which would be left for Article 341 TFEU would involve the decision to change the seat of the institutions since, in the Council’s view, such decisions should be taken following the revision procedures of Article 48 TEU, not Article 341 TFEU.

130. In my view, if the present case demonstrates anything, it is that arguments relating to effectiveness, and any other purposive reasoning in law in general, may be problematic. This is because such reasoning, in the present case as well as in general, may lead one to anywhere one wishes it to lead, based on the purpose that one decides to embrace.

131. The purpose or added value of Article 341 TFEU today depends simply on the interest, aim or value that one decides to ascribe to that provision. Article 341 TFEU does not have any evident effectiveness on its own, read in isolation. That interest or value must come from elsewhere. Indeed, as the various positions of the parties in the present case show, based on the particular value chosen, the result may then lead in opposite directions.

132. First, if importance were bestowed on the political dimension of decisions on the seats in general, as stressed by the Council, and also if it were to be believed that only the Member States are able to make such a decision, (86) then Article 341 TFEU might indeed be interpreted in a broad manner. In this regard, the purpose to be inserted into the provision would effectively be the protection of the decision-making powers of the Member States (Purpose No 1).

133. Second, the narrower interpretation, advocated by the Parliament in the ELA case, would be necessary for the purposes of enhancing the Parliament’s role (and that of EU institutions), and for the reason that the decision on the seat of an agency forms a necessary part of the overall decision to establish an agency to monitor the implementation of EU policies. In such a case, the purpose would be to protect the role of the Parliament, and possibly that of other EU institutions (Purpose No 2).

134. Third, relating to the preceding point, but expressed in somewhat more systemic terms, is the aim of preserving the (internal) integrity of the EU legal order. Decisions that ought to be made in the interest and in the name of the Union should not be allowed to exit that system and effectively be delegated outside the system, with the outcome reached outside that system being re-imposed within that system. That very purpose warrants a reading of Article 341 TFEU as not including agencies (Purpose No 3).

135. Fourth, Article 341 TFEU seems to require unanimity. If the aim to be followed by its interpretation is the maintaining or even enhancing of the feasibility of the EU decision-making procedures, then unanimity amongst the Member States might be more difficult to achieve than in circumstances where the same representatives were to decide sitting as the Council and following the (lower) threshold requirements in the context of that decision-making. That would again provide a reason for keeping the interpretation of Article 341 TFEU as narrow as possible (Purpose No 4).

136. Fifth and finally, the purpose in interpreting Article 341 TFEU could (also) be to uphold the principles of a Union based on the rule of law where ‘the very existence of effective judicial review designed to ensure compliance with provisions of EU law is inherent in the existence of the rule of law’. (87) However, if that were the case, as suggested, in essence, by the Italian Government at the hearing, then Article 341 TFEU is to be interpreted in a manner that is as restrictive as possible because of the ensuing impossibility of judicial review, at least before this Court, that derives from its applicability (Purpose No 5).

137. In general, but also in particular situations such as the present one, I do not consider it wise for a court, including this Court, to pick and choose one purpose over another. Instead, one should remain within the rather clear text and system of the Treaties. However, if one purpose were indeed to be favoured over the others, then, from the point of view of the EU Courts, that purpose would logically have to be to uphold the integrity and function of the EU system as such.

138. Second, be that as it may, I must admit that I do not find the Council’s argument relating to effectiveness convincing. On the one hand, it is indeed true that the seats of the EU institutions have already been determined. Thus, to some extent, Article 341 TFEU has largely exhausted its potential. However, that outcome was already inherent in its subject matter. On the other hand, the fact that agencies cannot be read as falling within the scope of Article 341 TFEU does not deprive that provision of its use. Article 341 TFEU clearly encompasses institutions, ensuring that that provision continues to be  operational and of ongoing relevance in their regard. It indeed remains in force and possibly applicable to a decision on the seat of a new institution (88) and, above all, to a decision to change the seat of an existing one. (89) Thus, it cannot be claimed that Article 341 TFEU is wholly deprived of any content if it were considered not to encompass bodies, offices and agencies of the Union. In any event, the fact that the scope of Article 341 TFEU at present appears to be quite limited is not a valid reason for artificially expanding that scope against its wording.

139. Third and finally, I remain puzzled by the Council’s overall argument according to which Article 341 TFEU must include agencies because decisions on their seats are politically sensitive and, as such, the unanimity of the Member States required under Article 341 TFEU must be preserved.

140. However, at the same time, it seems that the same imperative was apparently lacking in both cases in the present proceedings. When specifically looking at the contested decisions in the present cases, neither of them was meant to be governed by unanimity. Indeed, both sets of selection rules anticipated that the decision on the seat of the EMA and of the ELA would be governed by simple majority voting.

141. In factual terms and as far as the EMA is concerned, the Italian Government stated at the hearing that the decision on the seat was not taken by common accord of all the Member States. The Italian Government repeatedly stressed that it never agreed to that decision.

142. It is at this juncture that the dissonance between the stated, systemic purpose advocated by the Council and the reality becomes too great to ignore. The Council has repeatedly stated that the Court should interpret Article 341 TFEU broadly, to include agencies, in order to preserve the special political nature of the decision-making necessitating unanimity. However, the Council itself then departed, whenever it could, from the very same procedure, choosing simple majority voting instead.

143. For all those reasons, I consider that Article 341 TFEU does not govern the decision on the seats of agencies and other bodies or offices of the Union.

C.      The legal nature of the contested decisions

144. Turning to the contested decisions, it appears that they are not decisions of the Council in disguise. They are true decisions of the Member States over which the Court has therefore no jurisdiction under Article 263 TFEU (A). However, as decisions of the Member States that have been taken outside the framework of the Treaties, they are deprived of any binding legal effects within the EU legal order (B).

1.      Decisions of the Council in disguise?

145. In the EMA cases, the Italian Government submits that, given its content and the circumstances of its adoption, the contested decision was taken by the Council. First, the specific conditions under which the contested decision was adopted prove that it was a decision of the Council since it was made on the premises of the Council and relied on its institutional structures, namely its Presidency, its General Secretariat, its legal service and Coreper. Furthermore, several EU institutions, in particular the European Council and the Commission, participated in the selection process of the new seat of the EMA. In addition, that decision was actually made in accordance with the majority rule. That approach is typical of the decision-making process of the Council, and clearly different to an intergovernmental decision of the Member States acting together, since those are taken unanimously or by consensus (‘common accord’).

146. Second, the Italian Government is of the view that the decision on the seat of EU agencies undoubtedly falls within the exclusive competence of the Union, as follows, in particular, from the Commission’s proposal amending Regulation No 726/2004. Such competence cannot be exercised by an act of the Member States. To conclude that the Court lacks jurisdiction would amount to exempting decisions of the Council from review by the EU Courts every time the Member States consider matters that pertain to EU procedures and competences as in fact being ‘intergovernmental’. The contested decision should therefore be regarded as a decision of the Council.

147. For its part, the Council maintains that the contested decision is a decision that was taken by the representatives of the Member States. That decision was adopted in order to amend the decision of 29 October 1993 of the representatives of the Member States that had picked the location of the previous seat of the EMA. More generally, decisions on the seats of EU agencies do not fall within the competence of the Union, nor, accordingly, the ordinary legislative procedure.

148. In the ELA case, according to the Parliament, the Council is the author of the decision. The decision was signed by the President of the Council. The decision-making process preceding the adoption of the act was based on the administrative structures of the Council, including Coreper.

149. According to the Council, and as affirmed essentially by all the intervening Member States, it is a decision of the representatives of the Member States that was adopted under Article 341 TFEU. The fact that the decision-making process took place on the Council’s premises with the assistance of the Council’s General Secretariat, which is entrusted with administrative support of intergovernmental action, and that an EU institution participated in the bringing into effect of the contested decision does not alter the conclusion that, by its nature, the contested decision is an act of the Member States. Although the Council also convened on 13 June 2019, there were two distinct meetings that were organised following distinct procedures on the legal, protocol and organisational planes. As regards the signature of the contested decision by the President of the Council, the representatives of the governments of the Member States would have entrusted Romania (as the then Member State presiding over the Council) to ensure the correct functioning of the decision-making process.

150. In my view, neither of the contested decisions can be attributed to the Council.

151. First, as explained above, (90) the content and circumstances surrounding the adoption of an act may provide guidance in order to ascertain who is the formal author of that act when that author is unknown. In the present cases, however, it is clear that the contested decisions are officially decisions of the Member States. The applicants’ argument in both cases is again a normative one: in their view, the decisions should rightly have been made by the Council, not by the Member States.

152. As far as the factual level is concerned, starting with the ELA, the contested decision is a formal decision of the representatives of the Member States that was published as such in the Official Journal. (91) Admittedly, the fact that the contested decision was in fact published in the L series of the Official Journal comes with problems of its own. (92) However, and at the same time, those problems do not relate to authenticity. The Official Journal authoritatively stating ‘a decision of the Member States’ at the top of a document is certainly not a mere convoluted press release somewhere online, where one may indeed have doubts about what exactly that press release refers to.

153. Likewise, I see little conclusive value in the argument that the document was formally signed by the President of the Council. Certainly, again, from a formal point of view, it would arguably make sense for such a decision to be signed by all the representatives of the Member States. However, the signature of the contested decision by the President of the Council appears to be in line with the coordinating role that was entrusted to him by the Heads of State or Government in the selection rules. Moreover, as the Council suggested, in essence, at the hearing, I agree that it makes good practical sense to entrust the role of ‘notary public’ for such types of collective decisions to the same country presiding over the European Union at the time of the decision-making.

154. In the EMA cases, the contested decision was, for its part, announced in the minutes of the Council meeting and was published in a press release. (93) It did not give rise to a formal decision of the representatives of the Member States published in the Official Journal. However, despite the lack of such formalisation of the contested decision itself, it is still clearly stated that the latter is a decision of the Member States. Both the minutes and the press release indicate unequivocally that the decision was taken by the representatives of the Member States.

155. Incidentally, the fact that it cannot be attributed to the Council is further confirmed by the subsequent adoption of Regulation 2018/1718. The existence of that regulation shows that, in accordance with the selection procedure rules, (94) the representatives of the Member States were the true author of the contested decision. It was only subsequently that the Council as such participated, together with the Parliament in the ordinary legislative procedure, in order to insert the mention of the new seat in the EMA’s founding regulation that would reflect the vote of the Member States.

156. Furthermore, be it for the seat of the ELA or that of the EMA, the involvement of other institutions in the decision-making process – in particular the Commission through its assessment of the offers, and the Council through its Presidency, General Secretariat and more generally its own premises – can hardly lead to the conclusion that the contested decisions are in fact those of the Council. Again, in view of the dual nature of the Council and the Member States as its component members, it is hardly possible to extend the logic (‘what happens in the building of the Council is the Council’) that could potentially work with regard to other EU institutions (‘what happens in the buildings of the Commission is the Commission’), to that peculiar animal. (95)

157. Moreover, it is indeed established case-law that the Member States are entitled, in areas which do not fall under the exclusive competence of the Union, to entrust tasks to the institutions, outside the framework of the Union, such as the task of coordinating a collective action undertaken by the Member States, provided that those tasks do not alter the essential character of the powers conferred on those institutions by the Treaties. (96)

158. Second, as explained above, the establishment of agencies falls within the competence of the Union, to be decided normally through the ordinary legislative procedure. The same should hold true for the agencies’ seats as part of the ‘creation package’. (97) Logically then, such decisions should be made by the Parliament and the Council, not the Member States.

159. However, the EU competence on this matter concerns only the binding selection of the seat of an agency as a matter of EU law. The same can hardly apply to what is, as to its nature under EU law, a non-binding political statement to the effect that a new EU agency has been established that also mentions where its seat should ideally be located.

160. Additionally, and perhaps rather more significantly, in view of the past and current fluctuations regarding the issue of competence to decide on the seats of agencies, (98) the present cases are far removed from any logic that could perhaps, at least in my view, justify exceptionally the ‘re-classification’, ex post facto, of the authorship of an act by the Court. (99)

161. The previous institutional practice regarding the seats of agencies bears witness to the lingering disagreement amongst the relevant actors about the applicable law. In 2012, the Parliament, the Council and the Commission were indeed of the view that ‘the political decision on an agency’s seat [is] taken by common agreement between the representatives of the Member States meeting at Head of State or government level or by the Council’. (100) In 2017, the Commission stated, in connection with the EMA’s future seat, that ‘the issue of the location of the seat of the Agency falls within the exclusive competence of the Union’. (101) In 2018, the Parliament and the Council removed the provision on the location of the ELA’s seat from the draft founding regulation, (102) seemingly leaving it for the representatives of the Member States to make the decision. In 2019, in a report requested by the Council after a request from the Parliament to revise the procedure on the location of agencies’ seats, the Commission eventually concluded, after noting the existence of various practices since 2012 for selecting that location, that ‘the principles of the Common Approach provide a good framework for the decision-making process on agency seats and for ensuring that the host Member States respond to the agencies’ specific needs’. (103)

162. Within such a context, in the light of the fact that in the current state of the Union’s constitutional framework, the decision on agencies’ seats is rightly to be taken by EU institutions and not by the Member States, the past practice of the Member States to decide themselves on the agencies’ seats, even after the constitutional framework had changed with the Treaty of Lisbon, may perhaps be considered as somewhat praeter legem, but not clearly contra legem. (104)

163. I find it therefore impossible to (re)attribute, ex post facto, the contested decisions to the Council. They are collective decisions of the representatives of the Member States over which the Court has no jurisdiction under Article 263 TFEU.

2.      Legal effects of the contested decisions under EU law

164. The latter conclusion comes nonetheless with a rather important clarification which derives directly from the scope of Article 341 TFEU as set out above. The contested decisions are decisions of the Member States. However, they cannot be considered to be decisions taken under Article 341 TFEU as the latter provision does not empower the Member States to decide on the seats of agencies. It follows that the contested decisions are decisions of the Member States adopted outside the framework of the Treaties.

165. It is true that, with regard to similar decisions of the Member States adopted under Article 253 TFEU, the Court has held that it is ‘immaterial whether the representatives of the governments of the Member States acted within the framework of the Treaties or other legal sources, such as international law’. (105) I will not embark on a discussion of whether that is indeed entirely immaterial for the purposes of the Court’s jurisdiction under Article 263 TFEU, an issue on which I entertain serious doubts. However, addressing that issue is not relevant for the purposes of the present cases, since the present decisions of the Member States are not provided for by the Treaties.

166. It is, however, clear to me that Member States’ action within the framework of the Treaties, and Member States’ action outside the framework of the Treaties, are to be distinguished for the purposes of determining whether a decision of the Member States produces binding legal effects under EU law. Undoubtedly, if a decision of the Member States is provided for by the Treaties, then it does produce binding legal effects under EU law by virtue of the Treaties. By contrast, as rightly stated in the EMA cases by the Commission, and also by the Parliament, a decision of the Member States in a matter where the Treaties do not provide for their action is deprived, failing its incorporation in one way or another into the EU legal order, of any binding legal effects under EU law.

167. Similarly to what has been explained above with regard to the Edinburgh Decision, decisions of the representatives of the Member States that are not taken under a provision of EU law are not part of EU law. They have no binding legal effects in the EU legal order. In order to produce binding legal effects in the EU legal order, their content must be incorporated into EU law, following the procedures of EU law. (106)

168. The same logic naturally applies to the contested decisions. Since they are neither mandated, nor provided for by EU law, be it by Article 341 TFEU or by any other provision of EU law, they have therefore no binding legal effects thereunder. They remain, so to speak, ‘ante portas’ of the EU legal order. Certainly, that does not mean that such a decision of the Member States will not potentially produce binding legal effects under international law. Thus, while the Member States that have agreed to it might be bound in relation to each other by virtue of international law, neither the Member States that did not agree, nor the EU institutions, are bound by the contested decisions as a matter of EU law.

169. Admittedly, there is no denying that the contested decisions have a certain political significance since they give political impetus for further legal action in the EU legal order. One might in particular expect, rebus sic stantibus, that the position set out in those decisions by the Member States will also be the position of the Council in the future. However, the Council itself is not legally bound by those decisions when it subsequently participates in the ordinary legislature procedure. It may change its mind depending on the developments of the situation. Certain elements of the institutional design of a new agency might be subject to political trade-off reached within the legislative process.

170. The contested decisions can only produce binding legal effects in the EU legal order if they become part of EU law in one way or another. In particular, the (non-binding) content of such decisions may eventually be incorporated, following an EU legislative procedure, in (binding) secondary law acts. However, such an ‘incorporation’ measure might then be subject to an action for annulment if the requirements under Article 263 TFEU are fulfilled.

171. The challenge to Regulation 2018/1718, which has incorporated the content of the contested decision in the EMA cases, brought in the parallel cases by the Commune di Milano and the Italian Republic, (107) is an example of the latter scenario. However, the EU ‘incorporating’ measure cannot then be considered to be a mere confirmative measure in the light of the absence of binding legal effects of the decision of the Member States. For the same reason, the decision of the Member States cannot be considered to be an act preparatory to the EU incorporating measure. Logically then, it is the latter that is a binding and thus potentially reviewable act under EU law.

172. With regard to the other scenario, where the content of the decision of the Member States has not been placed in the founding act or in any other relevant binding act of the Union, there is no possibility of triggering Article 263 TFEU at some (later) stage. However, even in that case, the decision of the Member States remains deprived of binding legal effects under EU law. As a consequence, with regard to the ELA, where mention of the seat was not incorporated in an EU law act, the decision to locate the latter’s seat in Bratislava cannot produce any legal effects in the EU legal order until that location is formally decided in an EU measure. In other words, for the purposes of EU law, the location of the ELA’s seat in that city is, for now at least, simply a matter of fact.

173. As a concluding remark, in my view, the Member States can, in general, hardly be precluded from adopting acts outside the framework of the Treaties, including in EU-related matters, if they wish to do so. They remain, after all, sovereign. However, the consequence of going outside is (perhaps not entirely surprisingly, at least given the basic laws of physics) precisely that one finds oneself outside.

174. However, even if outside, one may of course express non-binding statements about what ideally should take place inside. After all, the EU legal order and the Court’s case-law allow various actors, including EU institutions, to issue various non-binding (soft law) measures in order to exhort and to persuade, distinct from the power to adopt acts having binding force. (108) If that is in fact possible for EU institutions bound by the principle of attributed competence, it must a fortiori be applicable to the sovereign Member States, even if it concerns matters pertaining to EU law.

175. Nonetheless, the fact that a decision of the Member States does not fall under the Court’s jurisdiction under Article 263 TFEU does not mean that that decision falls entirely outside the Court’s purview. Leaving aside the scenarios in which they ‘become EU law’, acts of the Member States which are adopted outside the framework of the Treaties indeed remain subject to EU law, but potentially to different forms of action, for example, infringement proceedings (109) or requests for preliminary rulings on interpretation. (110)

176. In summary, it is certainly not precluded that the Member States make known their collective view on the desirability of a new EU agency, its general function and overall limits to its functions, including the desired location of its seat, prior to the legislative procedure within the Union for that purpose being initiated. Indeed, the Member States, even if later united in the Council, might already indicate their political preferences in this way. So too might the Parliament, for that matter. Nevertheless, the establishment of such an agency, including the binding selection of its seat, is a matter of EU law and its own internal legislative procedures, typically that of the ordinary legislative procedure. In some way, that is hardly surprising. After all, the agency to be established is an agency of the Union, not of the Member States.

V.      Costs

177. Under Article 138(1) of the Rules of Procedure of the Court of Justice, the unsuccessful party is to be ordered to pay the costs.

178. In formal terms, it is true that in the present cases, the Comune di Milano and the Italian Republic in the EMA cases, and the Parliament in the ELA case, were unsuccessful inasmuch as it has been established that the Court lacks jurisdiction to examine the contested decisions.

179. However, in view of the facts that (i) it cannot be said that the Council would have been successful on the merits of its case; (ii) this has been a complex, inherently institutional litigation, which has been as much about past cases as is has about clarifying the rules for the future; and (iii) the Court joined some of the cases and then held a joint hearing for all five cases, following which a retroactive earmarking of the exact costs for each individual case might be a rather complex exercise, I would find it fairer and more equitable exceptionally to apply Article 138(3) of the Rules of Procedure and to order (all of) the parties to bear their own costs.

180. Under Article 140(1) of the Rules of Procedure, the interveners must also bear their own costs.

VI.    Conclusion

181. In Joined Cases C‑59/18, Italy v Council, and C‑182/18, Comune di Milano v Council, I propose that the Court should:

–        Hold that it lacks jurisdiction over the Decision of the representatives of the Member States adopted in the margins of a meeting of the Council of the European Union locating the new seat of the European Medicines Agency in Amsterdam (Netherlands);

–        Order the Comune di Milano (Italy), the Italian Republic, and the Council to bear their own costs;

–        Order all of the interveners to bear their own costs.

182. In Case C‑743/19, I propose that the Court should:

–        Hold that it lacks jurisdiction over Decision (EU) 2019/1199 taken by common accord between the Representatives of the Governments of the Member States of 13 June 2019 on the location of the seat of the European Labour Authority;

–        Order the European Parliament and the Council to bear their own costs;

–        Order the intervening parties to bear their own costs.

1      Original language: English.

2      Regulation of the European Parliament and of the Council of 14 November 2018 amending Regulation (EC) No 726/2004 as regards the location of the seat of the European Medicines Agency (OJ 2018 L 291, p. 3).

3      Opinion in Italian Republic and Commune di Milano v Council and Parliament (‘the EMA 2 cases’) (Joined Cases C‑106/19 and C‑232/19).

4      OJ 1992 C 341, p. 1.

5      OJ 1993 L 214, p. 1.

6      Decision taken by common Agreement between the Representatives of the Governments of the Member States, meeting at Head of State and Government level, on the location of the seats of certain bodies and departments of the European Communities and of Europol (OJ 1993 C 323, p. 1).

7      OJ 2004 L 136, p. 1.

8      Council Document XT 21045/17 – Procedure leading up to a decision on the relocation of the European Medicines Agency and the European Banking Authority in the context of the United Kingdom’s withdrawal from the Union.

9      Council Document 11450/12 – Endorsement of the Joint Statement and Common Approach.

10      See paragraph 2 of the selection rules.

11      Commission Assessment of 27 offers submitted by Member States under the procedure leading up to a decision on the relocation of the European Medicines Agency and the European Banking Authority in the context of the United Kingdom’s withdrawal from the Union.

12      Council Document XT 21092/17 – Procedure leading up to a decision on the relocation of the European Medicines Agency and the European Banking Authority in the context of the United Kingdom’s withdrawal from the Union.

13      Council Document 14559/17 – Outcome of the Council meeting (3579th Council Meeting, General Affairs (Article 50)).

14      Press Release of 20 November 2017, ‘European Medicines Agency to be relocated to Amsterdam, the Netherlands’ – https://www.consilium.europa.eu/en/press/press-releases/2017/11/20/european-medicines-agency-to-be-relocated-to-city-country/.

15      OJ 2018 L 291, p. 3.

16      See Article 71a of Regulation No 726/2004.

17      That case was initially introduced before the General Court by the Comune di Milano (Case T‑46/18). On 8 March 2018, the General Court decided, in accordance with the third paragraph of Article 54 of the Statute of the Court of Justice of the European Union and Article 128 of the Rules of Procedure of the General Court, to decline jurisdiction in favour of the Court of Justice so as to allow the Court to rule on that action in the light of the fact that, in Case T‑46/18, the issue of the validity of the same act was raised as in Case C‑59/18.

18      Order of the Vice-President of the Court of 2 July 2018, Comune di Milano v Council (C‑182/18 R, not published, EU:C:2018:524).

19      COM(2018) 131 final.

20      See the note from the President of the Council, containing the selection rules in an annex (Council Document 7491/19 – Procedure for the selection of the seat of the European Labour Authority).

21      See above, footnote 9 of this Opinion.

22      In footnote 2 of the document containing the selection rules, ‘common agreement’ and ‘common accord’ were defined as ‘terms having the equivalent effect of requiring that all Government Representatives are able to express themselves in favour of the solution retained (no abstention)’.

23      P8_TA(2019)0380 – European Parliament legislative resolution of 16 April 2019 on the proposal for a regulation of the European Parliament and of the Council establishing a European Labour Authority (COM(2018) 0131 – C8-0118/2018 – 2018/0064(COD)).

24      Point 32 of this Opinion.

25      That statement was published in the Official Journal (OJ 2019 L 188, p. 131) after the founding regulation was adopted.

26      See Council Document 10360/19 – Draft minutes, Council of the European Union, Employment, Social Policy, Health and Consumers Affairs, p. 5. See also the Press Release of 13 June 2019 of the Council of the European Union.

27      Decision taken by common accord between the Representatives of the Governments of the Member States of 13 June 2019 on the location of the seat of the European Labour Authority (OJ 2019 L 189, p. 68).

28      Regulation establishing a European Labour Authority, amending Regulations (EC) No 883/2004, (EU) No 492/2011, and (EU) 2016/589 and repealing Decision (EU) 2016/344 (OJ 2019 L 186, p. 21).

29      See, for example, judgments of 23 April 1986, Les Verts v Parliament (294/83, EU:C:1986:166, paragraph 23); of 28 March 2017, Rosneft (C‑72/15, EU:C:2017:236, paragraph 66 and the case-law cited); and of 3 June 2021, Hungary v Parliament (C‑650/18, EU:C:2021:426, paragraph 34).

30      See, for example, judgments of 26 March 2019, Commission v Italy (C‑621/16 P, EU:C:2019:251, paragraph 44); of 9 July 2020, Czech Republic v Commission (C‑575/18 P, EU:C:2020:530, paragraph 46 and the case-law cited); and of 3 June 2021, Hungary v Parliament (C‑650/18, EU:C:2021:426, paragraph 37).

31      See, for example, where the Court found that it had no jurisdiction, judgments of 28 April 1988, LAISA and CPC España v Council (31/86 and 35/86, not published, EU:C:1988:211, paragraphs 17 and 18), regarding provisions of an Act of Accession; of 19 March 1996, Commission v Council (C‑25/94, EU:C:1996:114, paragraphs 24 to 28), regarding a measure adopted by auxiliary bodies of an EU institution such as Coreper; and of 17 September 2014, Liivimaa Lihaveis (C‑562/12, EU:C:2014:2229, paragraph 51), regarding a measure adopted by the representatives of the national authorities of several Member States acting in the framework of a committee provided for in a European Union regulation.

32      See, for example, judgments of 31 March 1971, Commission v Council (22/70, EU:C:1971:32, paragraphs 40 to 42); of 4 September 2014, Commission v Council (C‑114/12, EU:C:2014:2151, paragraph 39); and of 20 February 2018, Belgium v Commission (C‑16/16 P, EU:C:2018:79, paragraph 31).

33      See, for example, judgment of 19 December 2018, Berlusconi and Fininvest (C‑219/17, EU:C:2018:1023, paragraph 42 to 50).

34      See, for example, with the Court’s jurisdiction flowing from the second subparagraph of Article 14.2 of the Statute of the ESCB and of the ECB, judgment of 26 February 2019, Rimšēvičs and ECB v Latvia (C‑202/18 and C‑238/18, EU:C:2019:139, paragraph 55).

35      See, for example, judgment of 3 December 1992, Oleificio Borelli v Commission (C‑97/91, EU:C:1992:491, paragraph 9), and orders of 16 May 2008, Raulin v France (C‑49/08, not published, EU:C:2008:286, paragraph 7), and of 21 February 2013, Gassiat v Ordre des avocats de Paris (C‑467/12, not published, EU:C:2013:104, paragraph 8 and the case-law cited).

36      See, for example, Bebr, G., ‘Acts of representatives of the governments of the Member States’, Sociaal-Economische Wetgeving, vol. 14, 1966, p. 528; Pescatore, P., ‘Remarques sur la nature juridique des “decisions des représentants des États membres réunis au sein du Conseil”’, Sociaal-Economische Wetgeving, vol. 14, 1966, p. 579; Mortelmans, K.J., ‘The Extramural Meetings of the Ministers of the Member States of the Community’, Common Market Law Review, vol. 11, 1974, p. 62.

37      Judgment of 30 June 1993, Parliament v Council and Commission (C‑181/91 and C‑248/91, EU:C:1993:271, paragraph 12) (‘Parliament v Council’). My emphasis.

38      See orders of the Court of 16 June 2021, Sharpston v Council and Conference of the Representatives of Governments of the Member States (C‑684/20 P, not published, EU:C:2021:486, paragraph 39), and Sharpston v Council and Representatives of the Governments of the Member States (C‑685/20 P, EU:C:2021:485, paragraph 39). See also orders of the Vice-President of the Court of 10 September 2020, Council v Sharpston (C‑423/20 P(R), not published, EU:C:2020:700, paragraph 26), and  Council v Sharpston (C‑424/20 P(R), not published, EU:C:2020:705, paragraph 26).

39      See order of the Court of 16 June 2021, Sharpston v Council and Conference of the Representatives of Governments of the Member States (C‑684/20 P, not published, EU:C:2021:486, paragraph 40).

40      With the Court’s type of jurisdiction being laid down in those decisions. See, for example, by analogy as far as agreements between Member States are concerned, recital 16 and Article 37(3) of the Treaty establishing the European Stability Mechanism, and judgment of 27 November 2012, Pringle (C‑370/12, EU:C:2012:756, paragraphs 170 to 176).

41      See, to that latter effect, order of the Court of 16 June 2021, Sharpston v Council and Conference of the Representatives of Governments of the Member States (C‑684/20 P, not published, EU:C:2021:486, paragraphs 41 and 42).

42      See, to that effect, judgments of 22 June 2010, Melki and Abdeli (C‑188/10 and C‑189/10, EU:C:2010:363, paragraphs 45 to 47 and 54 and 55), and of 19 December 2018, Berlusconi and Fininvest (C‑219/17, EU:C:2018:1023, paragraph 46).

43      Judgment of 22 October 1987, Foto-Frost (314/85, EU:C:1987:452, paragraph 15).

44      Parliament v Council (paragraph 14). My emphasis.

45      See, to that effect, Opinion of Advocate General Jacobs in Joined Cases Parliament v Council and Commission (C‑181/91 and C‑248/91, EU:C:1992:520, points 17 to 22).

46      Order of 28 February 2017, NG v European Council (T‑193/16, EU:T:2017:129, paragraph 67).

47      See the orders of the General Court of 28 February 2017, NF v European Council (T‑192/16, EU:T:2017:128); NG v European Council (T‑193/16, EU:T:2017:129); and NM v European Council (T‑257/16, EU:T:2017:130).

48      Orders of the General Court of 28 February 2017, NF v European Council (T‑192/16, EU:T:2017:128, point 62); NG v European Council (T‑193/16, EU:T:2017:129, point 63); and NM v European Council (T‑257/16, EU:T:2017:130, point 61).

49      See, for example, order of the Vice-President of the Court of 10 September 2020, Council v Sharpston (C‑423/20 P(R), not published, EU:C:2020:700), and order of the Vice-President of the Court of 10 September 2020, Council v Sharpston (C‑424/20 P(R), not published, EU:C:2020:705) where the same persons represented both the Council and the representatives of the governments of the Member States.

50      I nonetheless readily acknowledge that all this might simply be attributable to a rather narrow and perhaps outdated vision of judicial function, insisting that the judicial review of acts can be carried out only with regard to those that one has in fact seen. However, clairvoyance might absolve one from such mundane physical limitations. See also in this regard judgment of 10 March 2009, Heinrich (C‑345/06, EU:C:2009:140), in the light of the Opinion of Advocate General Sharpston in Heinrich (C‑345/06, EU:C:2008:212, points 70 to 77).

51      To provide a notable recent example: it was only in the judgment of 5 December 2017, M.A.S. and M.B. (C‑42/17, EU:C:2017:936, paragraphs 43 to 47) that the Court drew that consequence from the fact that ‘the protection of the financial interests of the Union by the enactment of criminal penalties falls within the shared competence of the Union and the Member States’ and that, as in judgment of 8 September 2015, Taricco and Others (C‑105/14, EU:C:2015:555), ‘at the material time for the main proceedings, the limitation rules applicable to criminal proceedings relating to [value added tax] had not been harmonised by the EU legislature’.

52      See orders of the Court of 16 June 2021, Sharpston v Council and Conference of the Representatives of Governments of the Member States (C‑684/20 P, not published, EU:C:2021:486), and Sharpston v Council and Representatives of the Governments of the Member States (C‑685/20 P, not published, EU:C:2021:485).

53      See, for example, Parliament v Council (paragraph 15).

54      See the last sentence of the first paragraph of Article 263 TFEU; the second paragraph of Article 265 TFEU; and Article 267 TFEU.

55      See Article 123 TFEU. See also, still in relation to the ECB, Article 282(3) TFEU.

56      See Article 228 TFEU. See also, as regards administration more generally, Article 298(1) TFEU.

57      Pursuant to Article 342 TFEU, ‘the rules governing the languages of the institutions of the Union shall, without prejudice to the provisions contained in the Statute of the Court of Justice of the European Union, be determined by the Council, acting unanimously by means of regulations’.

58      Council Regulation determining the languages to be used by the European Economic Community (OJ, English Special Edition, Series I 1952-1958, p. 59). See, for instance, Article 35 of Regulation 2019/1149.

59      See, for example, Article 25(2) of Council Regulation (EC) No 168/2007 of 15 February 2007 establishing a European Union Agency for Fundamental Rights (OJ 2007 L 53, p. 1).

60      See, in relation to the Office for Harmonisation in the Internal Market (Trade Marks and Designs (OHIM), the predecessor of the European Union Intellectual Property Office (EUIPO), judgment of 9 September 2003, Kik v OHIM (C‑361/01 P, EU:C:2003:434, paragraph 92 to 94).

61      See, for example, judgment of 16 December 2020, Council and Others v K. Chrysostomides & Co. and Others (C‑597/18 P, C‑598/18 P, C‑603/18 P and C‑604/18 P, EU:C:2020:1028, paragraph 80).

62      See, for example, as regards the European Foundation for the improvement of living and working conditions (‘Eurofound’), Regulation (EU) 2019/127 of the European Parliament and of the Council of 16 January 2019 establishing the European Foundation for the improvement of living and working conditions (Eurofound), and repealing Council Regulation (EEC) No 1365/75 (OJ 2019 L 30, p. 74), mentioned Article 153(2) TFEU (complementary competence of the Union in the field of working conditions), and as regards the European Union Aviation Safety Agency (EASA), Regulation (EU) 2018/1139 of the European Parliament and of the Council of 4 July 2018 on common rules in the field of civil aviation and establishing a European Union Aviation Safety Agency, and amending Regulations (EC) No 2111/2005, (EC) No 1008/2008, (EU) No 996/2010, (EU) No 376/2014 and Directives 2014/30/EU and 2014/53/EU of the European Parliament and of the Council, and repealing Regulations (EC) No 552/2004 and (EC) No 216/2008 of the European Parliament and of the Council and Council Regulation (EEC) No 3922/91 (OJ 2018 L 212, p. 1), adopted on the basis of Article 100(2) TFEU which concerns sea and air transport.

63      I stress again in conceptual terms, as to the nature of the decision to be made and the ability to make it. I certainly agree that in terms of the specific decision and its outcome, a political decision by the Parliament, as well as by any other institution or actor for that matter, might potentially differ from that of the Member States. However, the ability to reach a different decision certainly cannot be equated with the inability to make a political decision.

64      In detail above, points 14 and 33 of this Opinion respectively.

65      The predecessors to Article 341 TFEU are, in a reverse chronological order, Article 289 TEC, Article 216 TEEC and Article 77 of the ECSC Treaty.

66      See above, point 91 of this Opinion.

67      My emphasis.

68      When that type of decision is taken on the basis of a specific EU law provision, it is then a part of EU law. See, for example, to that effect, with regard to the seat of EU institutions under the predecessor of Article 341 TFEU, judgment of 1 October 1997, France v Parliament (C‑345/95, EU:C:1997:450, paragraph 34), where the Court examined the compatibility of the Parliament’s vote in question with the Edinburgh Decision.

69      Protocol No 22 was adopted in the wake of the Decision of the Heads of State or Government, meeting within the European Council in Edinburgh on 12 December 1992, concerning certain problems raised by Denmark on the Treaty on European Union, to which that protocol expressly refers.

70      Following the first rejection of the Treaty of Lisbon by the Irish people, the Heads of State or Government of the 27 Member States of the European Union, meeting within the European Council, adopted in June 2009 a Decision on the concerns of the Irish people on the Treaty of Lisbon. That decision envisaged the adoption of a protocol at the time of the next accession Treaty. A Protocol on the concerns of the Irish people on the Treaty of Lisbon was thus adopted and annexed to the Treaties on the occasion of Croatia’s accession (OJ 2013 L 60, p. 131).

71      Under the Treaty of Lisbon, with the establishment of the European Council as an EU institution, decisions of the Heads of State or Government are likely to be gradually replaced by (reviewable) decisions of the European Council. As a result, situations which could potentially lead to legal vacuums in EU law due to ‘unincorporated’ decisions of the Heads of State or Government should eventually disappear.

72      See, to that effect, judgment of 2 March 2010, Rottmann (C‑135/08, EU:C:2010:104, paragraph 40), which concerned the part of the Edinburgh Decision relating to Denmark and also an interpretive declaration annexed to the Treaties.

73      See footnote 5 of this Opinion.

74      Decision taken by common agreement between the representatives of the Member States, meeting at Head of State or Government level, of 13 December 2003 on the location of the seats of certain offices and agencies of the European Union (OJ 2004 L 29, p. 15), regarding the European Union Agency for Law Enforcement Training (CEPOL), the European Food Safety Authority (EFSA), the European Union Agency for Criminal Justice (‘Eurojust’), the European Maritime Safety Agency (EMSA), EASA, the European Union Agency for Railways (ERA), the European Union Agency for Cybersecurity (ENISA), the European Centre for Disease Prevention and Control (ECDC) and the European Chemicals Agency (ECHA).

75      Decision taken by common agreement between the Representatives of the Governments of Member States of 7 December 2009 on the location of the seat of the Agency for the Cooperation of Energy Regulators (OJ 2009 L 322, p. 39), taken ‘having regard to Article 341 [TFEU]’ and signed by the President of the Council. The establishment of that agency was decided by Regulation (EC) No 713/2009 of the European Parliament and of the Council of 13 July 2009 establishing an Agency for the Cooperation of Energy Regulators (OJ 2009 L 211, p. 1).

76      Decision (EU) 2021/4 taken by common accord between the Representatives of the Governments of the Member States of 9 December 2020 on the location of the seat of the European Cybersecurity Industrial, Technology and Research Competence Centre (OJ 2020 L 4, p. 7).

77      See, for example, Article 94(3) of Regulation 2018/1139; Article 79(1) of European Parliament and Council Regulation (EU) 2018/1727 of the European Parliament and of the Council of 14 November 2018 on the European Union Agency for Criminal Justice Cooperation (Eurojust), and replacing and repealing Council Decision 2002/187/JHA (OJ 2018 L 295, p. 138); and Article 25 of Regulation (EU) 2015/2219 of the European Parliament and of the Council of 25 November 2015 on the European Union Agency for Law Enforcement Training (CEPOL) and replacing and repealing Council Decision 2005/681/JHA (OJ 2015 L 319, p. 1).

78      The seat of the European Centre for the Development of Vocational Training (Cedefop) was decided in the founding act (Article 1 of Regulation (EEC) No 337/75 of the Council of 10 February 1975 establishing a European Centre for the Development of Vocational Training (OJ 1975 L 39, p. 1)), with no mention of Article 341 TFEU. That regulation was repealed by Regulation (EU) 2019/128 of the European Parliament and of the Council of 16 January 2019 establishing a European Centre for the Development of Vocational Training (Cedefop) and repealing Council Regulation No 337/75 (OJ 2019 L 30, p. 90). Article 20(3) thereof states that ‘Cedefop shall have its seat in Thessaloniki [Greece]’.

79      See Regulation (EEC) No 1365/75 of the Council of 26 May 1975 on the creation of a European Foundation for the improvement of living and working conditions (OJ 1975 L 139, p. 1). Under Article 4(2), ‘the seat of the Foundation shall be in Ireland’. Under Regulation 2019/127, Article 21(3) provides that ‘Eurofound shall have its seat in Dublin [Ireland]’.

80      See Council Decision 2005/358/EC of 26 April 2005 designating the seat of the European Agency for the Management of Operational Cooperation at the External Borders of the Member States of the European Union (OJ 2005 L 114, p. 13). That decision was taken ‘having regard to’ Article 15, fifth paragraph, of Council Regulation (EC) No 2007/2004 of 26 October 2004 establishing a European Agency for the Management of Operational Cooperation at the External Borders of the Member States of the European Union (OJ 2004 L 349, p. 1). Article 341 TFEU was not mentioned in that decision.

81      Article 18(4) of Council Regulation (EC) No 768/2005 of 26 April 2005 establishing a Community Fisheries Control Agency and amending Regulation (EEC) No 2847/93 establishing a control system applicable to the common fisheries policy (OJ 2005 L 128, p. 1).

82      See, for example, Article 7 of Regulation (EU) No 1093/2010 of the European Parliament and of the Council of 24 November 2010 establishing a European Supervisory Authority (European Banking Authority), amending Decision No 716/2009/EC and repealing Commission Decision 2009/78/EC (OJ 2010 L 331, p. 12). See also, in the context of the banking union, the seat of the Single Resolution Board (Article 48 of Regulation (EU) No 806/2014 of the European Parliament and of the Council of 15 July 2014 establishing uniform rules and a uniform procedure for the resolution of credit institutions and certain investment firms in the framework of a Single Resolution Mechanism and a Single Resolution Fund and amending Regulation (EU) No 1093/2010 (OJ 2014 L 225, p. 1)).

83      See, for example, Article 1(3) of Council Decision 2014/401/CFSP of 26 June 2014 on the European Union Satellite Centre and repealing Joint Action 2001/555/CFSP on the establishment of a European Union Satellite Centre (OJ 2014 L 188, p. 73). Taken on the specific basis of Article 45(2) TEU, see also Article 1(5) of Council Decision (CFSP) 2015/1835 of 12 October 2015 defining the statute, seat and operational rules of the European Defence Agency (OJ 2015 L 266, p. 55).

84      Initially, the decision on the seat was made by the Council in its Decision of 22 December 2000 establishing a European Police College (CEPOL) (OJ 2000 L 336, p. 1). That seat was then ‘decided’ again by the decision of the representatives of the Member States of 13 December 2003 (see footnote 77 of this Opinion). Subsequently, it was the Council which decided, on 8 October 2013, to modify CEPOL’s seat before European Parliament. Accordingly, Regulation (EU) No 543/2014 of the European Parliament and of the Council of 15 May 2014 amending Council Decision 2005/681/JHA establishing the European Police College (CEPOL) (OJ 2014 L 163, p. 5) was adopted. At present, CEPOL’s seat is mentioned in Article 25 of Regulation 2015/2219.

85      See, for example, my Opinion in Confédération paysanne and Others (C‑528/16, EU:C:2018:20, points 138 to 142).

86      Quod non, as explained above in point 106 of this Opinion, but with that assumption made for the sake of the argument.

87      See, for example, judgments of 19 July 2016, H v Council and Others (C‑455/14 P, EU:C:2016:569, paragraph 41), and of 28 March 2017, Rosneft (C‑72/15, EU:C:2017:236, paragraphs 72 and 73).

88      Incidentally, there still appears to be no formal seat for the European Council despite the transformation of the latter from an intergovernmental practice into an EU institution by the Treaty of Lisbon.

89      Whether Article 341 TFEU would suffice to that effect or whether it should be relied on in association with Article 48 TEU, as argued by the Council, does not need to be determined for the purposes of the present cases.

90      Points 72 to 81 of this Opinion.

91      See points 38 and 40 of this Opinion.

92      Relating rather to the nature of the document and its binding force, not necessarily authorship. However, even with regard to the former, the Court has stated over and again that publication in the Official Journal is not relevant. See, for example, to that effect, judgments of 6 October 1987, Demouche and Others (152/83, EU:C:1987:421, paragraph 19), and of 15 June 2017, Lietuvos Respublikos transporto priemonių draudikų biuras (C‑587/15, EU:C:2017:463, paragraph 38).

93      Point 19 of this Opinion.

94      See paragraph 6 of the selection rules.

95      With the principle of attributed competence having repercussions on the authorship as well: if it indeed (factually) happened within the Commission, and the Commission may only act as that very institution, the author must therefore be the Commission – see above, point 58 of this Opinion.

96      See, for example, judgment of 27 November 2012, Pringle (C‑370/12, EU:C:2012:756, paragraph 158 and the case-law cited).

97      See above points 104 to 108 of this Opinion.

98      See above, points 121 to 125 of this Opinion.

99      As suggested above, points 80 and 81 of this Opinion.

100      Point 6 of the Common Approach annexed to the Joint Statement on decentralised agencies (see footnote 9). My emphasis.

101      Commission proposal for a regulation of the European Parliament and of the Council amending Regulation (EC) No 726/2004 as regards the location of the seat of the European Medicines Agency (COM(2017) 0735 final). My emphasis.

102      Point 32 of this Opinion.

103      Report from the Commission to the European Parliament and the Council on the implementation of the Joint Statement and Common Approach on the location of the seats of decentralised agencies (COM(2019) 187 final).

104      As also demonstrated indirectly by the submissions of virtually all the parties to the present cases: while having very different views on the legal points raised by the present cases, sometimes even diametrically opposed views, all of the (institutional) parties were united in agreeing that it is vital for the Court to clarify the rules for the future.

105      Orders of the Court of 16 June 2021, Sharpston v Council and Conference of the Representatives of Governments of the Member States (C‑684/20 P, not published, EU:C:2021:486, paragraph 44), and Sharpston v Council and Representatives of the Governments of the Member States (C‑685/20 P, EU:C:2021:485, paragraph 50).

106      See points 115 to 117 of this Opinion.

107      See the parallel Opinion in Italian Republic and Commune di Milano v Council and Parliament (‘the EMA 2 cases’) (Joined Cases C‑106/19 and C‑232/19).

108      See, for example, judgment of 20 February 2018, Belgium v Commission (C‑16/16 P, EU:C:2018:79, paragraph 26).

109      See, to that effect, judgment of 31 January 2006, Commission v Spain (C‑503/03, EU:C:2006:74, paragraphs 33 to 35). See also Opinion of Advocate General Kokott in Commission v Council (C‑13/07, EU:C:2009:190, points 37 to 41).

110      See, for example, to that effect, judgments of 27 November 2012, Pringle (C‑370/12, EU:C:2012:756, paragraphs 163 and 164), as regards the Treaty establishing the European Stability Mechanism, notably in the light of EU competences, and of 20 September 2016, Ledra Advertising and Others v Commission and ECB (C‑8/15 P to C‑10/15 P, EU:C:2016:701, paragraph 67), in relation to memoranda of understanding between a Member State and the European Stability Mechanism.