Court of Justice of the European Communities (including Court of First Instance Decisions)

JUDGMENT OF THE GENERAL COURT (Third Chamber, Extended Composition)

25 September 2024 (*)

( Energy – Internal market for electricity – Capacity calculation region – Core region – Adoption by ACER of the methodology for cost sharing of redispatching and countertrading – Obligation to state reasons – Determination of the threshold for legitimate loop flows – Article 16(13) of Regulation (EU) 2019/943 )

In Case T‑482/21,

TenneT TSO GmbH, established in Bayreuth (Germany),

TenneT TSO BV, established in Arnhem (Netherlands),

represented by D. Uwer, J. Meinzenbach, P. Rieger, R. Klein and S. Westphal, lawyers,

applicants,

supported by

Federal Republic of Germany, represented by J. Möller and N. Scheffel, acting as Agents,

intervener,

v

European Union Agency for the Cooperation of Energy Regulators (ACER), represented by P. Martinet, E. Tremmel and Z. Vujasinovic, acting as Agents, and by P. Goffinet, L. Bersou and M. Shehu, lawyers,

defendant,

THE GENERAL COURT (Third Chamber, Extended Composition),

composed of F. Schalin, President, P. Škvařilová-Pelzl, I. Nõmm, G. Steinfatt and D. Kukovec (Rapporteur), Judges,

Registrar: I. Kurme, Administrator,

having regard to the written part of the procedure,

further to the hearings on 12 and 13 June 2023,

gives the following

Judgment

1        By their action under Article 263 TFEU, the applicants, TenneT TSO GmbH and TenneT TSO BV, seek the annulment of the decision of the Board of Appeal of the European Union Agency for the Cooperation of Energy Regulators (ACER) of 28 May 2021 confirming ACER Decision No 30/2020 of 30 November 2020 on the proposal of the electricity transmission system operators (‘TSOs’) of the ‘Core’ capacity calculation region, comprising Belgium, the Czech Republic, Germany, France, Croatia, Luxembourg, Hungary, the Netherlands, Austria, Poland, Romania, Slovenia and Slovakia (‘the Core region’), for the methodology for cost sharing of redispatching and countertrading, and dismissing their appeal in Case A-001-2021 (consolidated) (‘the contested decision’).

 Background to the dispute

2        The applicants are two TSOs that operate an electricity transmission system in part of Germany and in the Netherlands, respectively.

3        In accordance with Article 74(1) of Commission Regulation (EU) 2015/1222 of 24 July 2015 establishing a guideline on capacity allocation and congestion management (OJ 2015 L 197, p. 24), all TSOs in each capacity calculation region are to develop a proposal, no later than 16 months after the decision on the capacity calculation regions is taken, for a common methodology for redispatching and countertrading cost sharing (‘the cost sharing methodology’).

4        On 17 November 2016, ACER adopted, in accordance with Article 15 of Regulation 2015/1222, Decision No 06/2016 regarding the determination of capacity calculation regions. Article 1 of and Annex I to that decision list the territories of the Member States included in the Core region.

5        The proposal for a cost sharing methodology of the TSOs of the Core region should have been submitted within 16 months of that decision, that is to say, by 17 May 2018 at the latest.

6        However, the TSOs of the Core region did not submit a proposal for a cost sharing methodology within the time limit referred to in paragraph 5 above. In accordance with Article 9(4) of Regulation 2015/1222, those TSOs informed the national regulatory authorities (‘NRAs’) and ACER that they needed more time to develop such a proposal. ACER informed the European Commission, which consulted with the TSOs, the NRAs and ACER to assist the TSOs in developing that proposal and submitting it for approval as early as possible.

7        On 27 March 2019, in accordance with Article 9(7)(h) of Regulation 2015/1222, the TSOs of the Core region submitted to all the NRAs of that region a proposal for a cost sharing methodology, accompanied by an explanatory document. Those NRAs had a period of six months within which to take a decision on that proposal, in accordance with Article 9(10) of that regulation.

8        On 26 September 2019, at the request of those NRAs, ACER decided to extend the period for them to approve that proposal by six months, that is to say, until 27 March 2020.

9        On 27 March 2020, the Chair of the Core Energy Regulators’ Regional Forum, on behalf of all NRAs of that region, informed ACER that they were not able to reach a decision on the proposal submitted by the same day, the proposal being considered largely incomplete, to such an extent that the NRAs were not in a position to approve it or to request an amendment to it.

10      On the same day, since the NRAs of the Core region had not been able to reach an agreement on the proposal for a cost sharing methodology submitted by the TSOs, ACER declared itself competent to adopt a decision on that proposal, in accordance with Article 5(3) and Article 6(10) of Regulation (EU) 2019/942 of the European Parliament and of the Council of 5 June 2019 establishing a European Union Agency for the Cooperation of Energy Regulators (OJ 2019 L 158, p. 22), and Article 9(11) of Regulation 2015/1222. Under the latter provision, ACER was required to adopt such a decision within six months, in accordance with Article 6(12)(a) of Regulation 2019/942.

11      Following a long period of cooperation, consultations and discussions between ACER, all the NRAs of the Core region and all TSOs of that region on the proposal for a cost sharing methodology submitted by those TSOs and on the amendments made to that proposal during several meetings and rounds of voting, the Board of Regulators of ACER, which is composed of representatives of the NRAs, issued a favourable opinion on that proposal on 18 November 2020, pursuant to Article 22(5)(a) of Regulation 2019/942.

12      On 30 November 2020, ACER adopted, by Decision No 30/2020, the cost sharing methodology (‘Common methodology for redispatching and countertrading cost sharing for the Core CCR in accordance with Article 74 of Commission Regulation (EU) 2015/1222 of 24 July 2015’), as set out in Annex I to that decision (‘the contested cost sharing methodology’).

13      On 29 January 2021, the applicants submitted an appeal to the ACER Board of Appeal against Decision No 30/2020, in accordance with Article 28 of Regulation 2019/942. Other TSOs and NRAs of the Core region also submitted appeals against that decision. On 18 February 2021, the Board of Appeal consolidated all of those appeals.

14      On 28 May 2021, the Board of Appeal adopted the contested decision, by which it upheld Decision No 30/2020 and dismissed in their entirety the appeals brought against it.

 Forms of order sought

15      The applicants claim that the Court should:

–        annul the contested decision;

–        order ACER to pay the costs.

16      ACER contends that the Court should:

–        dismiss the action;

–        order the applicants to pay the costs.

17      The Federal Republic of Germany, intervening in support of the form of order sought by the applicants, submits that the Court should annul the contested decision.

 Law

18      The applicants put forward three pleas in law in support of their action.

19      The first plea alleges an error of law in the determination of the scope of the contested cost sharing methodology; the second plea alleges that the flow decomposition method is unlawful; and the third plea alleges that the threshold for legitimate loop flows (‘the threshold’) was determined incorrectly.

20      ACER contends that all of the pleas raised by the applicants should be rejected as unfounded.

21      The Court considers it appropriate to deal first with the plea of inadmissibility raised by ACER with regard to the statement in intervention of the Federal Republic of Germany and the annex thereto.

 Admissibility of the statement in intervention of the Federal Republic of Germany and the annex thereto

22      ACER contends that the statement in intervention of the Federal Republic of Germany and the annex thereto should be declared inadmissible.

23      According to ACER, the statement in intervention of the Federal Republic of Germany is limited to a few general assertions. The single general reference to the annex to that statement in intervention, containing the arguments supporting the intervention of the Federal Republic of Germany in another case, is not lawful and cannot make up for the absence of any legal arguments in the statement in intervention in the present case.

24      At the hearing, the Federal Republic of Germany was invited to comment on the question of the admissibility of its statement in intervention.

25      Under Article 21 of the Statute of the Court of Justice of the European Union and Article 76(d) of the Rules of Procedure of the General Court, each application is required to state the subject matter of the proceedings and a summary of the pleas in law on which the application is based. According to consistent case-law, it is necessary, for an action to be admissible, that the basic matters of law and fact relied on be indicated, at least in summary form, coherently and intelligibly in the application itself, so as to enable the defendant to prepare its defence and the Court to rule on the action, if necessary without any other supporting information. Whilst the body of the application may be supported and supplemented on specific points by references to extracts from documents annexed thereto, a general reference to other documents, even those annexed to the application, cannot make up for the absence of the essential arguments in law which, in accordance with the abovementioned provisions, must appear in the application. Furthermore, it is not for the Court to seek and identify in the annexes the pleas and arguments on which it may consider the action to be based, since the annexes have a purely evidential and instrumental function (see judgment of 1 June 2022, Algebris (UK) and Anchorage Capital Group v Commission, T‑570/17, EU:T:2022:314, paragraph 299 and the case-law cited).

26      The case-law cited in paragraph 25 above is applicable by analogy to a statement in intervention (see judgment of 14 March 2013, Fresh Del Monte Produce v Commission, T‑587/08, EU:T:2013:129, paragraph 541 and the case-law cited), in respect of which Article 145(2)(b) of the Rules of Procedure provides that it is to contain the pleas in law and arguments relied on by the intervener.

27      In the present case, first of all, it should be noted that the statement in intervention of the Federal Republic of Germany contains certain observations on the first and third pleas raised by the applicants. It should also be noted that, in that statement, the Federal Republic of Germany supports all of the applicants’ pleas. Next, it must be stated that the Federal Republic of Germany produces, as an annex to its statement in intervention in the present case, the statement in intervention that it submitted in the case of BNetzA v ACER (T‑485/21), that case also seeking the annulment of the contested decision, and it refers generally to that statement. Last, it should be noted that the Federal Republic of Germany sets out, in its statement in intervention in the present case, certain considerations concerning the allegedly discriminatory treatment of larger bidding zones in comparison with smaller bidding zones, concerning the prohibition on taking account of relieving flows when allocating costs between the different types of electricity flows, and concerning the priority given to loop flows over internal flows when determining the causes of network congestions.

28      First, the Federal Republic of Germany’s argument that it has been granted leave to intervene must be rejected.

29      In that regard, it is sufficient to note that, in the present case, it is not a matter of calling into question the status of the Federal Republic of Germany as intervener, but of ensuring that its statement in intervention is admissible.

30      Second, the Federal Republic of Germany’s argument that the lodging of a statement in intervention is not an obligation for the intervener, but an option, is also ineffective.

31      Since the Federal Republic of Germany chose to lodge a statement in intervention, it must be ascertained whether that statement satisfies the formal requirements applicable to it, referred to in paragraphs 25 and 26 above.

32      Third, in accordance with the case-law cited in paragraph 25 above, which is applicable, by analogy, to the statement in intervention, the general reference to the statement in intervention submitted in the case of BNetzA v ACER (T‑485/21), annexed to the statement in intervention in the present case, is not admissible and, therefore, cannot be taken into account in order to make up for any absence of the essential arguments in law, which, in accordance with the case-law cited in paragraphs 25 and 26 above, must appear in the statement in intervention.

33      Fourth, as regards the considerations relating to the allegedly discriminatory treatment of larger bidding zones in comparison with smaller bidding zones, it should be noted that, irrespective of the fact that the applicants in the present case did not raise such a plea, those considerations do not enable ACER to prepare its defence or the Court to examine them.

34      In paragraph 4 of its statement in intervention, the Federal Republic of Germany submits as follows:

‘by the [contested decision], the defendant disadvantages larger bidding zones (such as the German and French) in favour of smaller zones, where fewer loop flows occur for physical reasons alone. Disregarding the physical and geographical circumstances, the defendant provides for a uniform threshold of legitimate loop flows. As is always the case with the principle of equal treatment, it is infringed not only when similar situations are treated differently, but also when different situations are treated in the same way. The differences that exist with regard to the unavoidable, legitimate loop flows between the individual network elements and bidding zones involved are levelled (as far as possible) by equal treatment.’

35      It is true that, according to settled case-law, the general principle of non-discrimination or equal treatment requires that comparable situations must not be treated differently and that different situations must not be treated in the same way unless such treatment is objectively justified (see judgment of 16 December 2008, Arcelor Atlantique et Lorraine and Others, C‑127/07, EU:C:2008:728, paragraph 23 and the case-law cited).

36      However, in the present case, the mere allegation of discriminatory treatment of larger bidding zones in comparison with smaller bidding zones, without the slightest indication of the extent to which those bidding zones are different or of the lack of justification for the allegedly equal treatment applied to them, does not enable the Court to rule on such a plea.

37      In those circumstances, irrespective of whether an intervener may raise pleas different from those relied on by the main party which it supports, it must be concluded that the plea alleging discriminatory treatment of the different bidding zones is inadmissible because it is not set out in sufficient detail to enable the Court to rule on it.

38      Fifth, as regards the considerations concerning, on the one hand, the prohibition on netting of burdening flows and relieving flows and, on the other, the incorrect penalisation of loop flows as compared with internal flows, it should be noted that, irrespective of the fact that the applicants, in the present case, did not raise such a plea, those considerations do not enable ACER to prepare its defence or the Court to examine them.

39      In paragraph 5 of its statement in intervention, the Federal Republic of Germany submits as follows:

‘in addition, the methodology prohibits the taking into account of relieving flows when allocating costs between the different types of electricity flows and prioritises loop flows over internal flows when determining the causes of network congestions. Both are contrary to the regulatory objective of Regulation 2019/943 to increase as far as possible congestion-free bidding zones and cross-zonal trade.’

40      The mere assertion that the prohibition on taking account of relieving flows when allocating costs and the penalisation of loop flows by comparison with internal flows is ‘contrary to the regulatory objective of Regulation 2019/943 to increase as far as possible congestion-free bidding zones and cross-zonal trade’, without putting forward any legal arguments whatsoever, does not enable the Court to rule on such a plea.

41      In those circumstances, irrespective of whether an intervener may raise pleas different from those relied on by the main party which it supports, it must be concluded that the pleas based on the prohibition on taking account of relieving flows when allocating costs between the different types of electricity flows and on the priority given to loop flows over internal flows when determining the causes of network congestions are inadmissible because they are not set out in sufficient detail to enable the Court to rule on them.

42      Sixth, by contrast, as regards the observations of the Federal Republic of Germany on the first and third pleas raised by the applicants, they are admittedly very concise, but may, read in the light of the arguments raised by the applicants, be subject to legal assessment by the Court.

43      In the light of the foregoing, it must be concluded that the statement in intervention of the Federal Republic of Germany is only partially admissible, in so far as, for the reasons set out in paragraphs 36 and 40 above, it does not enable the Court to rule on the line of argument based on, in the first place, discriminatory treatment of the different bidding zones, in the second place, an alleged prohibition on netting of burdening flows and relieving flows and, in the third place, an allegedly incorrect penalisation of loop flows as compared with internal flows.

 The first plea, concerning the scope of the contested cost sharing methodology

44      By the first plea, the applicants, supported to that effect by the Federal Republic of Germany, claim that the Board of Appeal erred in law by confirming the scope of the contested cost sharing methodology.

45      First, the applicants submit that the inclusion, within the scope of the contested cost sharing methodology, of all transmission system elements with a voltage level higher than or equal to 220 kilovolts (kV) is incompatible with Article 16(13) of Regulation (EU) 2019/943 of the European Parliament and of the Council of 5 June 2019 on the internal market for electricity (OJ 2019 L 158, p. 54), and Article 74(2) of Regulation 2015/1222, as is apparent from the wording, scheme and purpose of those provisions.

46      According to the applicants, Article 16(13) of Regulation 2019/943 refers to cost sharing for remedial actions relating to congestion ‘between two bidding zones’. In accordance with the definitions of ‘congestion’ and ‘critical network element’ in points 4 and 69 of Article 2 of Regulation 2019/943, Article 16(13) of Regulation 2019/943 refers only to cost sharing for remedial actions relating to critical network elements, to the exclusion of other network elements, which are not considered to be capable of affecting electricity trading capacity.

47      The applicants claim that the concept of ‘actions of cross-border relevance’ in Article 74(2) of Regulation 2015/1222 also covers only remedial actions relating to critical network elements, to the exclusion of internal remedial actions. That is borne out by recital 10 of that regulation and the general scheme of Article 74 of Regulation 2015/1222, which relates to the capacity allocation process.

48      Second, according to the applicants, there is no legal basis for such an extension of the scope of the contested cost sharing methodology. In particular, that wide scope cannot be justified by considerations relating to the ‘polluter pays’ principle or to the alleged incentives that would result from such a determination of the scope.

49      Third, the applicants claim that the extension of the scope to all transmission network elements cannot be justified on operational security grounds either.

50      The Federal Republic of Germany broadly supports the arguments put forward by the applicants, while relying on arguments based on the wording and legislative history of Article 16(13) of Regulation 2019/943.

51      ACER disputes the arguments made by the applicants and supported by the Federal Republic of Germany.

 Preliminary observations

52      As a preliminary point, in the first place, it should be recalled that, in the contested decision, the Board of Appeal rejected the arguments alleging that the scope of the contested cost sharing methodology was unlawful, in essence, in paragraphs 89 to 379 (‘First Consolidated Plea – Excessive scope of the RDCTCS and unlawful determination of XNEs’), in paragraphs 1078 to 1105 (‘Eighth Consolidated Plea – Polluter Pays Principle’) and in paragraphs 1126 to 1191 (pages 203 to 212) (‘Fourteenth Consolidated Plea – ACER exceeded its competence and infringed the principle of conferral’) of that decision.

53      As is apparent, in essence, from paragraphs 196 to 210 of the contested decision, that decision is based, in part, on an interpretation of Article 16(13) of Regulation 2019/943 to the effect that that provision reflects the ‘polluter pays’ principle. Similarly, that provision requires the origin of physical flows that contribute to congestion between zones to be identified without, however, limiting cost sharing to remedial actions carried out with regard to congestions on interconnectors.

54      Similarly, as is apparent, in essence, from paragraphs 173 to 180 of the contested decision, the Board of Appeal considered that the contested cost sharing methodology was in line with Article 74(2) of Regulation 2015/1222, in so far as the actions on the network elements included in the scope of the contested cost sharing methodology were of cross-border relevance.

55      In the second place, examination of the first plea involves determining the legal basis, purpose and scope of the contested cost sharing methodology, as confirmed by the contested decision.

56      First, as regards the legal basis and purpose of the contested cost sharing methodology, it must be noted that that methodology was adopted in accordance with Article 74 of Regulation 2015/1222 and that paragraph 2 of that provision provides for the adoption of cost-sharing solutions for remedial actions of cross-border relevance. The purpose of that methodology is thus to share the costs incurred in activating costly remedial actions, namely redispatching and countertrading, between the TSOs.

57      In that regard, point 13 of Article 2 of Regulation 2015/1222 defines ‘remedial action’ as ‘any measure applied by a TSO or several TSOs, manually or automatically, in order to maintain operational security’.

58      Point 26 of Article 2 of Regulation 2019/943 defines redispatching as a measure, including curtailment, that is activated by one or more TSOs or distribution system operators by altering the generation, load pattern, or both, in order to change physical flows in the electricity system and relieve a physical congestion or otherwise ensure system security.

59      Point 27 of Article 2 of Regulation 2019/943, for its part, defines countertrading as a cross-zonal exchange initiated by system operators between two bidding zones to relieve physical congestion.

60      Congestion, which is a risk to operational security that requires remedial action, is defined in point 4 of Article 2 of Regulation 2019/943 as ‘a situation in which all requests from market participants to trade between network areas cannot be accommodated because they would significantly affect the physical flows on network elements which cannot accommodate those flows’.

61      Congestion is caused by physical flows. There are several types of physical flows defined by the contested cost sharing methodology. Among these, Article 2(2)(a) of the contested cost sharing methodology defines allocated flows as ‘a physical flow on a network element where the source and sink are located in different bidding zones’. Internal flows are defined, in Article 2(2)(o) of the contested cost sharing methodology, as ‘a physical flow on a network element where the source and sink and the complete network element are located in the same bidding zone’. According to Article 2(2)(p) of the contested cost sharing methodology, a loop flow is ‘a physical flow on a network element where the source and sink are located in the same bidding zone and the network element or even part of the network element is located in a different bidding zone’.

62      Furthermore, it should be noted that Article 16(13) of Regulation 2019/943 provides for the sharing of the costs of remedial actions intended to relieve congestion between two bidding zones on the basis of the contribution to that congestion of flows resulting from transactions internal to a zone. That provision provides, inter alia, that ‘when allocating costs of remedial actions between transmission system operators, regulatory authorities shall analyse to what extent flows resulting from transactions internal to bidding zones contribute to the congestion between two bidding zones observed, and allocate the costs based on the contribution to the congestion to the transmission system operators of the bidding zones creating such flows except for costs induced by flows resulting from transactions internal to bidding zones that are below the level that could be expected without structural congestion in a bidding zone’.

63      In that regard, a bidding zone is defined, in point 65 of Article 2 of Regulation 2019/943, as ‘the largest geographical area within which market participants are able to exchange energy without capacity allocation’. At present, the areas forming part of the Core region correspond, in most cases, to the territories of the Member States.

64      Second, as is apparent from Article 5(1) of the contested cost sharing methodology, in order to share the costs of remedial actions, the network element on which each remedial action is actually carried out must be identified.

65      As emphasised in paragraphs 106 to 110 of the contested decision, the scope of the contested cost sharing methodology extends not only to cross-zonal network elements (interconnectors), but also to all internal network elements with a voltage level higher than or equal to 220 kV.

66      The contested cost sharing methodology provides, in Article 3(4) thereof, that all ‘cross-border relevant network elements’ are ‘eligible for cost sharing’.

67      On the one hand, as regards cross-border relevant network elements, the contested cost sharing methodology defines those elements, in Article 2(2)(j) thereof, as ‘a network element identified as cross-border relevant and on which operational security violations need to be managed in a coordinated way’.

68      In that regard, it is apparent from paragraphs 106 to 110 of the contested decision and it is, moreover, common ground between the parties that the concept of ‘cross-border relevant network elements’ must be understood in the same way as the concept included in the definition set out in Article 5 of the methodology for coordinated redispatching and countertrading for the Core region in accordance with Article 35(1) of Regulation 2015/1222, adopted by ACER Decision No 35/2020 of 4 December 2020 on the methodology for coordinated redispatching and countertrading for the Core region (‘the RDCT methodology’), and in Article 5 of the methodology for regional operational security coordination for the Core region in accordance with Article 76(1) of Commission Regulation (EU) 2017/1485 of 2 August 2017 establishing a guideline on electricity transmission system operation (OJ 2017 L 220, p. 1), adopted by ACER Decision No 33/2020 of 4 December 2020 on the methodology for regional operational security coordination for the Core region (‘the ROSC methodology’).

69      The concept of ‘cross-border relevant network elements’ thus refers, first, to all critical network elements, in accordance with Article 5(1) and Article 7 of the day-ahead and intraday common capacity calculation methodologies, adopted by ACER Decision No 02/2019 of 21 February 2019 on the proposals of the TSOs of the Core region for the day-ahead and intraday common capacity calculation methodologies (‘the CCM’), namely, currently, interconnectors and internal network elements determined by TSOs with a power transfer distribution factor, as defined by point 22 of Article 2 of Commission Regulation (EU) No 543/2013 of 14 June 2013 on submission and publication of data in electricity markets and amending Annex I to Regulation (EC) No 714/2009 of the European Parliament and of the Council (OJ 2013 L 163, p. 1), of 5% or higher, and, second and in principle, all internal network elements with a voltage level higher than or equal to 220 kV.

70      According to the definition in point 69 of Article 2 of Regulation 2019/943, a critical network element is ‘a network element either within a bidding zone or between bidding zones taken into account in the [cross-zonal] capacity calculation process, limiting the amount of power that can be exchanged [between zones]’.

71      Critical network elements are therefore either interconnectors or internal elements with a power transfer distribution factor of 5% or higher. According to the contested cost sharing methodology, cross-border relevant network elements are all critical network elements as well as internal network elements with a voltage level higher than or equal to 220 kV.

72      On the other hand, as regards the concept of ‘cross-border relevant redispatching and countertrading actions’, it must be noted that Article 3(1) of the contested cost sharing methodology states that that concept refers to cross-border relevant redispatching and countertrading actions that are determined in accordance with the RDCT methodology and the ROSC methodology.

73      It follows, as confirmed by the parties, that, according to the contested cost sharing methodology, cross-border relevant redispatching and countertrading actions are, in principle, all redispatching and countertrading actions intended to relieve congestion on cross-border relevant network elements.

 Analysis of scope

–       The extent of the scope of the contested cost sharing methodology

74      According to the applicants, the costs of the remedial actions attributed to internal network elements that are not critical network elements should be excluded from cost sharing.

75      In particular, the applicants maintain that the costs attributed to non-critical network elements should be excluded, since congestion on those elements is ‘internal’ congestion, which does not fall within the definition of congestion ‘between zones’.

76      It must therefore be examined whether the applicants are correct in claiming that the sharing of the costs of remedial actions should be limited to critical network elements, as they are set out in the CCM, since those are elements ‘limiting the amount of power that can be exchanged’, according to point 69 of Article 2 of Regulation 2019/943. Accordingly, only congestion on those elements would be congestion between zones for the purposes of Article 16(13) of Regulation 2019/943.

77      First, it must be examined whether critical network elements are the only elements of cross-border relevance within the meaning of Article 74(2) of Regulation 2015/1222.

78      It should be recalled that, according to Article 74(2) of Regulation 2015/1222, the contested cost sharing methodology is to ‘include cost-sharing solutions for actions of cross-border relevance’. Furthermore, Article 74(4)(b) of that regulation establishes that the contested cost sharing methodology must define ‘which costs incurred from using redispatching or countertrading to guarantee the firmness of cross-zonal capacity are eligible for sharing between all the TSOs of a capacity calculation region’, in the present case the Core region.

79      Thus, in order to establish whether the contested cost sharing methodology, as confirmed by the contested decision, is compatible with Article 74(2) of Regulation 2015/1222, it is necessary to examine to what extent the remedial actions, the costs of which that methodology aims to share, are intended to guarantee the firmness of cross-zonal capacity.

80      However, it should be recalled that Regulation 2015/1222 is an act implementing Regulation 2019/943, as is apparent from Article 18(5) of Regulation (EC) No 714/2009 of the European Parliament and of the Council of 13 July 2009 on conditions for access to the network for cross-border exchanges in electricity and repealing Regulation (EC) No 1228/2003 (OJ 2009 L 211, p. 15), which was subsequently replaced by Regulation 2019/943. Regulation 2019/943 is therefore a law that is higher ranking than and subsequent to Regulation 2015/1222. Except where their meaning is clear and unambiguous and therefore requires no interpretation, the provisions of an implementing regulation must be given, if possible, an interpretation consistent with the provisions of the basic regulation (judgment of 28 February 2017, Canadian Solar Emea and Others v Council, T‑162/14, not published, EU:T:2017:124, paragraph 150). Article 74 of Regulation 2015/1222 must therefore be interpreted consistently with Article 16(13) of Regulation 2019/943.

81      In that regard, it must be observed that Article 16(13) of Regulation 2019/943 does not define the network elements on which remedial actions are intended to relieve congestion between zones, or the network elements on which the costs of remedial actions intended to relieve congestion between zones must be shared.

82      Similarly, point 4 of Article 2 of Regulation 2019/943 does not specify the network elements on which physical congestions, including those linked to cross-zonal trade, occur.

83      Article 16(13) of Regulation 2019/943 refers to the costs of the remedial actions activated by the TSOs in order to ensure cross-zonal trade and, subsequently, requires those costs to be shared between the TSOs according to the extent to which flows resulting from transactions internal to zones contribute to the congestion between two zones observed.

84      Consequently, according to the purpose of Article 16(13) of Regulation 2019/943, it is necessary to determine which congestion is to be relieved in a coordinated way in order to ensure cross-zonal trade, which, in accordance with paragraphs 79 and 80 above, will then make it possible to establish whether the remedial actions covered by the contested cost sharing methodology are intended to guarantee the firmness of cross-zonal capacity, within the meaning of Article 74(4)(b) of Regulation 2015/1222.

85      Second, it must be noted that, in order to carry out such an analysis seeking to determine which remedial actions ensure cross-zonal trade and are thus affected by cost sharing, it is necessary to clarify the role of remedial actions in the cross-zonal capacity calculation process, as provided for by the CCM, and in the regional operational security assessment process, as established by the ROSC methodology, in accordance with Article 76(1)(b) of Regulation 2017/1485 (‘the CROSA process’).

86      In the first place, it should be noted that the cross-zonal capacity calculation process begins two days before the delivery of electricity and is used to determine the amount of energy that can be exchanged between zones, within the operational security limits.

87      The capacity calculation process is carried out only on critical network elements. On the one hand, TSOs may limit, before the market is opened, the amount of energy that can be exchanged by the participants in order to comply with the operational security limits. On the other hand, that process also uses remedial actions, such as redispatching and countertrading. Those remedial actions, which could be activated in order to maximise cross-zonal capacity, in accordance with Article 16(4) of Regulation 2019/943, are taken into account by the TSOs, but not yet activated.

88      In fact, no remedial action is activated when calculating cross-zonal capacity and no cost is incurred.

89      In the second place, it should be noted that the costly remedial actions of redispatching and countertrading, which are the subject of the present dispute, arise only in the CROSA process, which is closely linked to the calculation of cross-zonal capacity.

90      In accordance with Article 76(1)(b)(iii) of Regulation 2017/1485, governing the coordinated preparation of cross-border relevant remedial actions, and as is apparent from Article 3(2)(b) of the ROSC methodology, the objective of the CROSA process is to coordinate, validate and implement cross-border relevant remedial actions. As is apparent from Articles 5 and 9 of that ROSC methodology, cross-border relevant remedial actions are all those that are at least sometimes able to address violations of current limits on cross-border relevant network elements, that is to say, in principle, all critical network elements taken into account for the calculation of cross-zonal capacity, and all other network elements with a voltage level higher than or equal to 220 kV.

91      Similarly, according to recital 12 and Article 5(1) of the ROSC methodology, the CROSA process ensures, in principle, the operational security of all network elements with a voltage level higher than or equal to 220 kV following the market outcomes and cross-zonal capacity allocation. At that time, TSOs have detailed information on the energy that will be injected into or withdrawn from the network.

92      The CROSA process starts with the local operational security assessment carried out by each TSO in its network, in accordance with Article 13 and Article 14(1) of the ROSC methodology, leading to the creation of an individual grid model by each TSO.

93      Subsequently, in accordance with Article 18 of the ROSC methodology, individual grid models are to be provided to the regional coordinators and merged by them with a view to the creation of a common grid model for all hours of the day, which includes ‘a Union-wide data set … describing the … [characteristics] of the power system’, in accordance with point 2 of Article 2 of Regulation 2015/1222.

94      When a flow on a network element exceeds the maximum flow, a remedial action must be prepared and implemented in order to comply with operational security limits.

95      As is apparent from recital 10 of the ROSC methodology, remedial action optimisation is provided for as part of the CROSA process.

96      More specifically, that optimisation within the CROSA process, described in Article 2(1)(p) of the ROSC methodology, involves determining specifically, during each iteration, which congestion on which network element, critical or non-critical, must be managed in a coordinated way.

97      Thus, for each hour in the common grid model, out of all the remedial actions available to the TSOs, it is identified which one is the most effective and economically efficient, in accordance with Article 76(1)(b)(iii) of Regulation 2017/1485, and can resolve all congestion on all interconnectors and on all internal network elements with a voltage level higher than or equal to 220 kV in the common grid model, without creating new congestion, in accordance with Articles 20, 23 and 24 and Article 27(1) of the ROSC methodology.

98      Third, in the light of the foregoing, it must be examined whether, as stated by the applicants, only the costs incurred as a result of the remedial actions on critical network elements (and therefore only interconnectors or internal network elements with a power transfer distribution factor of 5% or higher) should be included in the scope of the contested cost sharing methodology. To that end, as stated in paragraph 84 above, it is necessary to determine which congestion is to be relieved in a coordinated way in order to ensure cross-zonal trade.

99      In that regard, in the first place, the mere fact of including within the scope of the contested cost sharing methodology the costs incurred as a result of congestion on network elements with a voltage level higher than or equal to 220 kV cannot be contrary to Article 16(13) of Regulation 2019/943, since that provision merely entails determining which congestion is to be relieved in a coordinated way in order to ensure cross-zonal trade.

100    In the second place, the TSOs may use a maximum level of 30% of the capacity of each critical network element to relieve congestion on that element, provided that, as provided for in point (b) of the first subparagraph of Article 16(8) of Regulation 2019/943, 70% of that capacity remains available for cross-zonal trade, in accordance with Article 16(1) and (8) of Regulation 2019/943.

101    Nevertheless, the fact that congestion may be relieved by using up to 30% of the capacity of a subgroup of network elements does not mean that only the costs of remedial actions carried out on that subgroup of elements must be shared.

102    In the third place, it should be noted that the firmness of the minimum capacity of 70% of each critical network element is guaranteed in the most efficient way by optimising the remedial actions activated on all network elements with a voltage level higher than or equal to 220 kV. Thus, that optimisation of remedial actions on all those elements achieves the objectives of Regulation 2019/943, inter alia the delivery of market signals for increased efficiency and security of supply, in accordance with Article 1(a) of Regulation 2019/943.

103    That optimisation contributes to reducing the costs of remedial actions, thus making it possible to limit the reduction in cross-zonal capacity, as is apparent from Article 16(4) of Regulation 2019/943.

104    In the fourth place, when remedial actions activated by a TSO on internal elements that are not critical network elements form part of the optimal solution necessary to relieve congestion on critical network elements as well, the costs relating to the former elements must be shared between the TSOs in the same way as those relating to the latter. Consequently, not including network elements with a voltage level higher than or equal to 220 kV within the scope of the cost sharing methodology would lead to an unjustified limitation, in particular in the light of Article 16(13) of Regulation 2019/943, which requires all costs of the remedial actions activated to resolve congestion between zones to be shared, with the exception of the costs relating to flows that contribute to the congestion between two bidding zones observed and that do not exceed the threshold.

105    If cost sharing were limited to critical network elements, it would become uncertain, since the costs associated with congestion management would be shared differently on the basis of the element on which a remedial action would be activated. As ACER correctly observes, that would lead to discrimination between the different network elements and, thus, between the TSOs that own those elements, which is not provided for by the legislation at issue.

106    Therefore, flows resulting from internal transactions that contribute to the congestion between two bidding zones observed, described in Article 16(13) of Regulation 2019/943, which reduce cross-zonal capacity on critical network elements, cannot be treated differently, in cost sharing, where those flows pass through non-critical network elements. The remedial actions activated on those two groups of elements are optimised in order to resolve congestion problems.

107    In the fifth place, remedial actions activated on non-critical network elements can sometimes resolve congestion on critical network elements, as ACER observed in response to the Court’s questions.

108    Consequently, when such remedial actions can help resolve such congestion on critical network elements, their costs must be shared, in accordance with the ‘polluter pays’ principle. By contrast, although they may not contribute, at a particular moment, to resolving congestion on critical network elements, that does not mean that the elements on which the remedial actions are activated lose their link with cross-zonal trade. Those remedial actions were nevertheless chosen as part of the CROSA process, having regard to the other congestions and remedial actions, in order to find the optimum solution at regional level. In addition, that congestion, in so far as it is caused by a cross-border flow, namely a loop flow, is of cross-border relevance.

109    Moreover, as provided for in recital 35 of Regulation 2019/943, in an open, competitive market, TSOs should be compensated for costs incurred as a result of hosting cross-border flows of electricity on their networks by the operators of transmission systems from which cross-border flows originate and the systems where those flows end.

110    Thus, the remedial actions on all network elements included in the CROSA process are potentially relevant for cross-zonal trade, according to Article 16(13) of Regulation 2019/943, irrespective of their power transfer distribution factor, that factor being a concept used in the CCM, which cannot determine the sharing of costs in accordance with Article 16(13) of Regulation 2019/943.

111    In any event, it should be noted that, if the TSOs establish that a non-critical network element with a voltage level higher than or equal to 220 kV can never be useful for ensuring regional security and guaranteeing the firmness of allocated cross-zonal capacity, they may exclude it from the CROSA process, in accordance with Article 5(1)(b) and Article 7(3)(b) of the ROSC methodology, and, consequently, from the sharing of costs.

112    In the sixth place, as regards recital 12 of Regulation 2015/1222, which requires coordination between cross-border and internal remedial actions, it should be noted that all remedial actions in the CROSA process are of cross-border relevance, while all other remedial actions are internal. Thus, the CROSA process ensures coordination with internal remedial actions and also ensures the security of ‘internal’ network elements.

113    In the absence of coordinated management through the CROSA process, there is a risk of an operational security violation, endangering cross-zonal trade. It must be emphasised that the functioning of the network is an essential condition for the security of energy supply, since electricity can reach the citizens of the Union only through the network, in accordance with recital 2 of Regulation 2015/1222.

114    Accordingly, congestion ‘between zones’ is all congestion that is currently, in the Core region, managed in a coordinated way as part of the CROSA process. Consequently, the ‘polluter pays’ principle must apply to the costs relating to such coordinated management.

115    It should be noted that coordination and cost sharing do not depend on whether there is a particular exchange or allocated flow on a network element at a given moment, because all remedial actions activated on all network elements, critical and non-critical, with a voltage level higher than or equal to 220 kV potentially contribute to facilitating cross-zonal trade, whereas their actual contribution to facilitating trade may vary depending, inter alia, on network topology, market conditions and particular generation and consumption schedules. Accordingly, it is not possible from the outset to exclude network elements with a voltage level higher than or equal to 220 kV from the scope of the contested cost sharing methodology.

116    In the seventh place, as is clear from paragraph 167 of the contested decision, without it being disputed by the parties, network elements with a voltage level higher than or equal to 220 kV were considered to be cross-border relevant elements in so far as those elements would not be structurally congested in the absence of energy exchanges. Furthermore, as is apparent from that paragraph, all the TSOs of the Core region considered, in the Explanatory Note to the ROSC methodology, that those elements were the most relevant for the CROSA process.

117    In that regard, it is apparent from page 8 of the Explanatory Note to the ROSC methodology, on which, moreover, the parties had the opportunity to comment at the hearing, that the TSOs considered that the most relevant network elements for the CROSA process were the network elements with a voltage level of 220 kV and 380 kV, ‘as these elements [were] used to facilitate the energy exchanges between bidding zones within the European energy system’.

118    In the Explanatory Note to the ROSC methodology, the TSOs explained that network elements with a voltage level higher than or equal to 220 kV facilitate the energy exchanges between bidding zones in the European energy system. However, the TSOs did not establish which elements should be considered when sharing the costs of remedial actions.

119    Moreover, the voltage level of 220 kV was chosen in the Core region because it is a highly meshed network, whereas Article 16(13) of Regulation 2019/943 does not prevent other solutions from being chosen, as clarified at the hearing.

120    Furthermore, it is common ground between the parties that cost sharing in that context is necessary to achieve an integrated electricity market at European level.

121    A fortiori, the scope of the contested cost sharing methodology is not defined geographically and cannot be limited solely to network elements located on the border between two bidding zones, or interconnectors that connect Member States’ networks. By contrast, as stated in paragraph 84 above, Article 16(13) of Regulation 2019/943 requires the identification of which congestion is to be relieved in a coordinated way in order to ensure cross-zonal trade.

122    In the light of the foregoing, the applicants are incorrect in claiming that the contested decision is contrary to Article 16(13) of Regulation 2019/943 because only critical network elements should be included in cost sharing.

123    In the Core region, all congestion relieved by remedial actions activated in accordance with the ROSC methodology, as established by Decision No 33/2020 and confirmed by the contested decision, corresponds to ‘congestion between … zones’ referred to in Article 16(13) of Regulation 2019/943.

124    It should also be noted that, in the Core region, all remedial actions activated in accordance with the ROSC methodology, as established by Decision No 33/2020, contribute to guaranteeing the firmness of cross-zonal capacity, in accordance with Article 74(4)(b) of Regulation 2015/1222.

125    Accordingly, the contested decision also cannot be deemed to be contrary to Article 74 of Regulation 2015/1222, interpreted in accordance with Article 16(13) of Regulation 2019/943. The Board of Appeal was therefore correct in dismissing the appeal brought against the contested cost sharing methodology.

126    The arguments raised by the applicants against the contested decision cannot invalidate that conclusion.

127    First, it is true that, as the applicants maintain, TSOs are legally obliged to ensure operational security irrespective of how costs are shared. Nevertheless, it must be stated that managing congestions on all elements with a voltage level higher than or equal to 220 kV necessarily involves coordination between all the TSOs and that the costs relating to it are shared between them, in order to compensate them for all of their interventions that are necessary to ensure cross-zonal trade.

128    Second, as regards the applicants’ argument that congestion on a non-critical network element is ‘purely internal’ congestion the costs of which cannot be shared on the basis of Article 16(13) of Regulation 2019/943, it must be pointed out that, in applying the contested cost sharing methodology, the only situation in which costs relating to non-critical network elements are shared in accordance with that provision is that in which congestion on those elements is caused by loop flows, as described in paragraph 61 above, that exceed the threshold referred to in Article 16(13) of Regulation 2019/943.

129    According to recital 6 of the contested cost sharing methodology, loop flows that exceed the threshold are the primary contributor to congestion subject to cost sharing. That follows from the finding that congestion caused by internal flows is borne by the owner of the congested network element, whereas, as is apparent from Article 16(13) of Regulation 2019/943, the costs caused by flows resulting from internal transactions that contribute to the congestion between two bidding zones observed must be shared proportionately between TSOs.

130    A loop flow is a ‘cross-border flow’, within the meaning of point 3 of Article 2 of Regulation 2019/943, namely a ‘physical flow of electricity on a transmission network of a Member State that results from the impact of the activity of producers, customers, or both, outside that Member State on its transmission network’, which is not limited to critical network elements.

131    Accordingly, congestion caused by a cross-border flow, such as a loop flow, cannot be classified as ‘purely internal’ congestion that is excluded from the scope of Article 16(13) of Regulation 2019/943.

132    Moreover, if congestion on a non-critical network element is caused solely by internal flows, the costs of remedial actions to address that congestion will be borne, in any event, by the owner of that network element, in accordance with Article 76(1)(b)(v) of Regulation 2017/1485. Thus, contrary to what the applicants claim, the ‘polluter pays’ principle remains the exception to the rule, since that principle applies, de facto, only to loop flows that exceed the threshold, whereas the ‘owner pays’ principle applies to other flows, such as loop flows that do not exceed the threshold and internal flows.

133    In addition, congestion that is not congestion ‘between zones’ for the purposes of Article 16(13) of Regulation 2019/943 is congestion that appears on the network elements excluded from regional operational security coordination, either because that has been decided by the TSOs or because they are elements that are excluded at the outset from the coordination of remedial actions, namely network elements with a voltage level below 220 kV.

134    Third, it should be noted that the only exception to the rule in Article 16(13) of Regulation 2019/943 according to which the costs associated with remedial actions relating to flows resulting from internal transactions that contribute to the congestion between two bidding zones observed must be shared is the exception relating to those flows that do not exceed the threshold, for which the costs must be borne by the owner of the congested network element.

135    In that regard, it should be noted that, contrary to what the applicants claim, the legislature did not intend to exclude a group of network elements from cost sharing, since it provided, in Article 16(13) of Regulation 2019/943, for the sharing of costs incurred through the coordinated relief of congestion between zones.

136    Similarly, the fact that Article 76(1)(b)(v) of Regulation 2017/1485 provides for the possibility of adopting other cost sharing methodologies related to the various remedial actions referred to in Article 22 of that regulation, which complement where necessary the common methodology developed in accordance with Article 74 of Regulation 2015/1222, is irrelevant to the interpretation of Article 16(13) of Regulation 2019/943, which is, moreover, a higher-ranking law.

137    Fourth, the applicants’ argument that the scope of the contested cost sharing methodology, as confirmed by the contested decision, is contrary to Article 291 TFEU or has no legal basis, in so far as ACER extended the scope provided for by the legislation, must be rejected.

138    In particular, coordinated remedial actions on network elements with a voltage level higher than or equal to 220 kV allow cross-zonal trade, help to prevent the limitation of cross-zonal capacity and guarantee the firmness of that capacity, in accordance with Article 16(13) of Regulation 2019/943.

139    In the same way, and given that the scope of the contested cost sharing methodology is consistent with Article 16(13) of Regulation 2019/943, ACER cannot be criticised for extending the scope of that methodology beyond Article 16(13) of Regulation 2019/943 by relying on the ‘polluter pays’ principle.

140    Fifth, the argument raised at the hearing by the Federal Republic of Germany, based on the legislative history of Article 16(13) of Regulation 2019/943, according to which the legislature intended to exclude from sharing between TSOs the costs arising from remedial actions carried out on elements within zones when it refused to modify, when drafting that provision, its wording by replacing ‘the congestion between two bidding zones observed’ with ‘the congestion between and within bidding zones observed’, cannot succeed either.

141    In that regard, irrespective of whether the Federal Republic of Germany may rely on documents relating to the legislative trilogue in order to demonstrate the legislature’s intention when adopting Article 16(13) of Regulation 2019/943, it is sufficient to note that any refusal expressly to include congestion within a zone is not decisive for the question of which elements should be taken into account in assessing contributions to congestion ‘between two … zones’.

142    As is apparent from paragraphs 81 to 84 and 135 above, Article 16(13) of Regulation 2019/943 does not define the network elements on which the costs of remedial actions intended to relieve congestion between zones must be shared, which does not mean that the legislature intended to exclude a group of network elements from cost sharing.

143    Accordingly, the applicants’ arguments that internal network elements that are not ‘critical’ should be excluded from the contested cost sharing methodology on the basis of Article 16(13) of Regulation 2019/943 and Article 74 of Regulation 2015/1222 must be rejected.

–       The incentives derived from including non-critical network elements within the scope of the contested cost sharing methodology

144    As regards the applicants’ argument relating to the allegedly incorrect incentives that stem from including network elements with a voltage level higher than or equal to 220 kV within the scope of the contested cost sharing methodology, it should be recalled that Article 16(13) of Regulation 2019/943 provides for the obligation for TSOs to bear the costs of remedial actions relating to flows resulting from internal transactions that contribute to the congestion between two bidding zones observed and that do not exceed the threshold. That incentivises TSOs to develop their network in order to be able to accommodate such flows, when they do not exceed the threshold, which corresponds to the level of loop flows that could be expected without structural congestion in a bidding zone.

145    By contrast, the relevant legal framework does not impose an obligation on TSOs to develop their networks in order to accommodate loop flows above the threshold referred to in paragraph 144 above. Those flows are, by their nature, unforeseen and unforeseeable and the TSO that accommodates those flows has no influence on them.

146    It should be noted, in that regard, that, as stated in recital 27 of Regulation 2019/943, reducing the effect of loop flows and internal congestions on cross-zonal trade was one of the main objectives of the EU legislature when establishing that legislative framework.

147    Similarly, it is apparent from page 59 of the Commission’s impact assessment of 30 November 2016, carried out as part of the preparation of the legislative ‘Energy’ package (SWD(2016) 410), that loop flows can reduce cross-zonal capacity and lead to expensive out-of-market redispatching and significant distortions to prices and investment signals in neighbouring bidding zones. That results in a significant loss of prosperity.

148    In addition, the effect of excluding non-critical network elements would be that the TSO that causes loop flows would not have an incentive to invest sufficiently in its network, since it would not bear the full cost of the remedial actions necessary to resolve the congestion it caused. It is only that TSO, which has knowledge of and responsibility for its network, which may adopt other necessary measures, such as reconfiguring the zone or investing in its network. In that context, it should be recalled that, according to Article 35(5) of Regulation 2019/943, each TSO is responsible for its network and, accordingly, its internal problems should not be borne by neighbouring TSOs.

149    Furthermore, if each of the two responsible neighbouring TSOs had an incentive to invest in its own network, either to accommodate or to reduce loop flows, that would lead, in the absence of coordination between those two TSOs, to overinvestment and poor allocation of resources, as ACER rightly mentioned in paragraphs 110 to 115 of its defence.

150    In the light of the foregoing assessments, it must be observed that the applicants seek, in essence, to be exempted from the costs that they cause to the other TSOs on the non-critical elements of their networks with their loop flows that exceed the threshold, even if the remedial actions on those elements contribute to ensuring cross-zonal trade.

151    Such an approach would, moreover, be contrary to the principle of energy solidarity, as interpreted by the Court of Justice.

152    The principle of energy solidarity entails rights and obligations both for the European Union and for the Member States, the European Union being bound by an obligation of solidarity towards the Member States and the Member States being bound by an obligation of solidarity between themselves and with regard to the common interest of the European Union and the policies pursued by it (judgment of 15 July 2021, Germany v Poland, C‑848/19 P, EU:C:2021:598, paragraph 49).

153    In those circumstances, it would be contrary to the principle of energy solidarity to allow TSOs that cause loop flows that pass through other bidding zones to avoid the costs of remedial actions that have been activated in the common interest in order to optimise cross-zonal capacity, while at the same time ensuring network security as efficiently as possible, to the benefit of all TSOs and, accordingly, electricity consumers.

154    In the light of the foregoing considerations, the first plea must be rejected.

 The second plea, concerning flow decomposition

155    By their second plea, the applicants claim that the flow decomposition method used in the contested cost sharing methodology, as confirmed by the contested decision, is unlawful.

156    The second plea is divided into two parts.

157    By the first part, the applicants submit that the flow decomposition method used, namely the Power Flow Colouring method (‘the PFC method’), is not an appropriate method and that applying it therefore leads to incorrect results.

158    By the second part, the applicants criticise, in particular, the treatment of network elements relating to the transmission of direct current.

 The first part, concerning the inadequacy of the PFC method

159    By the first part of the second plea, the applicants submit that the PFC method is not an appropriate flow decomposition method, in that it produces incorrect results.

160    The applicants raise numerous objections, divided into four groups.

161    By the first group of objections, the applicants claim that the PFC method, because of its design, is not able to decompose flows properly.

162    By the second group of objections, the applicants complain that the Board of Appeal did not effectively examine the arguments which they had raised against Decision No 30/2020.

163    By the third group of objections, the applicants claim that the PFC method leads, in several respects, to unrealistic results.

164    By the fourth group of objections, the applicants submit that the PFC method, because of its design, does not meet the relevant legal criteria.

165    ACER disputes the applicants’ arguments.

–       Preliminary observations

166    Flow decomposition is a necessary step in sharing the costs of remedial actions.

167    The importance of flow decomposition stems from the fact that it provides input data necessary for the subsequent sharing of costs. Thus, if flow decomposition is carried out incorrectly, it necessarily affects the results of the sharing of costs.

168    The purpose of flow decomposition is to identify the types of flow which cause congestion on network elements.

169    In accordance with Article 16(13) of Regulation 2019/943, it is necessary to determine which flows contribute to the congestion between bidding zones.

170    The different flows, for example internal flows or loop flows, which should be identified for cost sharing purposes, embody the different origins and destinations of the ‘components’ determined on the basis of a total physical flow, which alone is measurable.

171    Thus, flow decomposition consists of a process, first, of dividing the total flow into components and, second, of determining the share of those components.

172    For the purposes of that flow decomposition, the contested cost sharing methodology is based on the PFC method. As is apparent from the explanatory document of 22 February 2019 accompanying the proposal for the cost sharing methodology of 27 March 2019, that method was developed with the main goal to stay consistent with the European zonal market model and, at the same time, to allow for a complete partitioning of the power flow for each network element of the power system.

173    It is against that background that the applicants put forward various objections against the contested decision, in so far as it rejected their arguments raised before the Board of Appeal.

–       The statement of reasons for the contested decision

174    By the first part of their second plea, the applicants criticise the PFC method as regards network elements that transmit alternating current.

175    Among the objections made against the contested decision, the applicants claim, in particular, that the Board of Appeal failed to carry out a proper examination of their arguments raised against the PFC method, which was used for the flow decomposition.

176    In accordance with the case-law, as regards complex technical and economic assessments contained in a decision of ACER, the Board of Appeal cannot confine itself to a review of manifest errors of assessment, but must, relying on the expertise of its members, examine whether the arguments put forward before it are capable of demonstrating that the considerations on which that decision is based are vitiated by error (see, to that effect, judgment of 18 November 2020, Aquind v ACER, T‑735/18, EU:T:2020:542, paragraph 69).

177    In that regard, it should be recalled that the composition of the Board of Appeal meets the requirements necessary to enable it to conduct a full review of decisions adopted by ACER. If its members must have prior experience in the energy sector, this is because they have or should have the technical knowledge necessary to enable them to conduct a detailed examination of appeals. The EU legislature intended to provide the Board of Appeal with the necessary expertise to allow it itself to carry out assessments of complex technical and economic facts relating to energy (judgment of 9 March 2023, ACER v Aquind, C‑46/21 P, EU:C:2023:182, paragraphs 63 and 64).

178    However, an examination of whether the Board of Appeal carried out its review in the present case with the requisite intensity in accordance with the case-law cited in paragraph 176 above presupposes that the Court is in a position, in particular in the light of the statement of reasons provided in the contested decision, to carry out such an examination. In that context, the Court must raise of its own motion a plea alleging a failure to state reasons in the contested decision, as regards the part of the contested decision dealing with the PFC method used for the flow decomposition.

179    In that regard, it must be borne in mind that an absence of or inadequate statement of reasons constitutes an infringement of essential procedural requirements for the purposes of Article 263 TFEU and is a plea involving a matter of public policy which may, and even must, be raised by the EU judicature of its own motion (see judgment of 9 March 2023, Les Mousquetaires and ITM Entreprises v Commission, C‑682/20 P, EU:C:2023:170, paragraph 39 and the case-law cited).

180    In the light of the foregoing, in connection with the measure of organisation of procedure of 5 May 2023, the parties were expressly invited by the Court to comment on the ‘sufficiency of the statement of reasons in the part of the contested decision which deals with the objections raised by the applicants against the [PFC method]’.

181    It must also be borne in mind that, according to consistent case-law, the extent of the obligation to state reasons depends on the nature of the measure in question and on the context in which it was adopted. The statement of reasons must disclose in a clear and unequivocal fashion the reasoning followed by the institution which adopted the measure in such a way as to enable the persons concerned to ascertain the reasons for the measure, so that they can defend their rights and ascertain whether or not the decision is well founded, and the EU judicature to review the lawfulness of the measure. It is not necessary for the reasoning to go into all the relevant facts and points of law, since the question whether the statement of reasons meets the requirements of Article 296 TFEU must be assessed with regard not only to its wording but also to its context and to all the legal rules governing the matter in question (judgments of 2 April 1998, Commission v Sytraval and Brink’s France, C‑367/95 P, EU:C:1998:154, paragraph 63, and of 30 November 2011, Sniace v Commission, T‑238/09, not published, EU:T:2011:705, paragraph 37).

182    In particular, the institution which adopted the measure is not obliged to adopt a position on all the arguments relied on before it by the parties concerned. It is sufficient if it sets out the facts and the legal considerations of fundamental importance in the context of that measure (see judgment of 1 July 2008, Chronopost and La Poste v UFEX and Others, C‑341/06 P and C‑342/06 P, EU:C:2008:375, paragraph 96 and the case-law cited).

183    Similarly, the question whether the statement of reasons is sufficient must also be assessed in the light of the interest which the addressees of the measure may have in obtaining explanations. Consequently, the reasons given for a measure adversely affecting a person are sufficient if that measure was adopted in a context which was known to that person and which enables him or her to understand the scope of the measure concerning him or her (see, to that effect, judgment of 15 July 2021, Commission v Landesbank Baden-Württemberg and SRB, C‑584/20 P and C‑621/20 P, EU:C:2021:601, paragraph 104 and the case-law cited).

184    It must also be borne in mind that the obligation to state reasons is an essential procedural requirement that must be distinguished from the question whether the reasoning is well founded, which goes to the substantive legality of the measure at issue. The reasoning of a decision consists in a formal statement of the grounds on which that decision is based. If those grounds are vitiated by errors, those errors will vitiate the substantive legality of the decision, but not the statement of reasons in it, which may be adequate even though it sets out reasons which are incorrect. It follows that objections and arguments intended to establish that the measure at issue is not well founded are irrelevant in the context of a plea in law alleging an inadequate statement of reasons or a lack of such a statement (see, to that effect, judgments of 10 March 2022, Commission v Freistaat Bayern and Others, C‑167/19 P and C‑171/19 P, EU:C:2022:176, paragraph 77, and of 13 May 2020, Koenig & Bauer v EUIPO (we’re on it), T‑156/19, not published, EU:T:2020:200, paragraph 55 and the case-law cited).

185    Given that, in the present case, the Board of Appeal was to conduct, in accordance with the case-law cited in paragraphs 176 and 177 above, a full review of Decision No 30/2020, including the assessments contained in that decision concerning complex technical and economic facts relating to energy, and could not confine itself to a review of manifest errors of assessment, the statement of reasons provided by the Board of Appeal had to make it possible, on the one hand, for the addressee of the contested decision to be able to defend its rights and ascertain whether or not that decision was well founded and, on the other, for the EU judicature to exercise its power of review.

186    Without a sufficient statement of reasons, the applicants and the Court would be unable to examine whether the Board of Appeal did in fact examine the arguments raised before it in accordance with the requirements referred to in paragraphs 176 and 177 above.

187    It is in the light of those considerations that the plea raised by the Court of its own motion, concerning the statement of reasons for the contested decision as regards the PFC method used for the flow decomposition, must be examined.

188    In their appeal before the Board of Appeal, the applicants set out in detail their criticism of Decision No 30/2020. They claimed that the PFC method was vitiated by a number of design errors and that, consequently, the PFC method produced incorrect results.

189    In particular, the applicants criticised the PFC method in that it did not take physical reality into account. In that regard, it must be borne in mind that, according to the regulatory framework, an approach that reflected physical reality was required. Only a physical flow of electricity can give rise to congestion and the cost sharing methodology can, accordingly, share the costs incurred through remedial actions only in respect of flows that actually contribute to congestion. The applicants claimed that, as a result of ‘Kirchhoff’s law of least resistance’, exchanges of electricity take place on the shortest connection between the point of generation and the point of consumption, irrespective of the borders of the bidding zones and irrespective of agreed exchanges of electricity. However, according to the applicants, the PFC method ignores that scientific fact, and an exchange of electricity between a point of generation and a point of consumption in different zones is allowed only if, taking into account all consolidated agreed cross-border exchanges, the bidding zone in which the point of generation is located has a positive net position and if, taking into account all consolidated agreed cross-border exchanges, the bidding zone in which the point of consumption is located has a negative net position. Therefore, depending on the net position of the bidding zones concerned, the PFC method would identify internal flows and loop flows instead of allowing a cross-border flow. Accordingly, by disregarding ‘natural laws’, the PFC method tends to identify loop flows that do not actually exist. Last, the applicants submitted that the PFC method is also not justified by considerations relating to the zonal configuration of the electricity market in the European Union.

190    According to the table of contents of the contested decision, the issues relating to flow decomposition raised by the applicants before the Board of Appeal are dealt with under the ‘Third Consolidated Plea’, in paragraphs 407 to 587 (pages 61 to 87) of the contested decision. In connection with that ‘Third Consolidated Plea’, paragraphs 407 to 551 (pages 61 to 81) of the contested decision are devoted to network elements that transmit alternating current, which are the subject of the first part of the present plea, whereas paragraphs 552 to 587 (pages 82 to 87) of the contested decision concern network elements that transmit direct current, which are the subject of the second part of the present plea.

191    By contrast, paragraphs 588 to 631 of the contested decision relate, in accordance with the table of contents of that decision, to the ‘Fourth Consolidated Plea’, relating to the overestimation of loop flows and internal flows from importing bidding zones, raised, as is apparent from the relevant paragraphs of the contested decision, not by the applicants in the present case, but by the applicant in the case of Polskie sieci elektroenergetyczne v ACER (T‑484/21). In those circumstances, for the purpose of examining compliance with the obligation to state reasons, there is no need to take into account paragraphs 588 to 631 of the contested decision, unless reference is made, in those paragraphs, to paragraphs 407 to 551 of the contested decision.

192    It must be stated that the reasoning set out in paragraphs 407 to 551 of the contested decision does not, despite its length, provide a statement of reasons that complies with the requirements set out in paragraphs 181 to 183 above.

193    First, the fact that the ‘Third Consolidated Plea’ is examined over approximately 150 paragraphs in the contested decision does not, in itself, demonstrate a rigorous and detailed examination of the applicants’ arguments before the Board of Appeal. The length of that part of the contested decision is largely due to the way in which it is drafted, which consists of dealing with the same subject multiple times in the light of the various provisions of law or elements of fact relied on, which involves numerous references and repetitions.

194    Second, it should be noted that, in general, in the contested decision, the Board of Appeal confined itself on several occasions to making general statements, without actually responding to the applicants’ arguments, then relying on those same statements in the remainder of the examination. Thus, the Board of Appeal creates the appearance of detailed reasoning, without, however, genuinely responding to the applicants’ arguments.

195    Third, as regards, specifically, the applicants’ objection that the PFC method does not take natural laws into account and thus results in the identification of loop flows that do not actually exist, the core of the Board of Appeal’s reasoning is set out, as ACER stated in response to a written question put by the Court, in paragraphs 421 to 461 of the contested decision.

196    It must be stated, however, that paragraphs 421 to 432 (pages 65 and 66) of the contested decision, which appear under subheading 3.2 ‘Flow decomposition in the [contested cost sharing methodology]’, are limited, so far as responding to the applicants’ objections is concerned, to the finding, in paragraph 424 (page 65), that ‘various flow decomposition methods exist’.

197    Paragraphs 433 to 451 (pages 66 to 69) of the contested decision, which appear under subheading 3.3 ‘The PFC method ignores electrical distance, creates fictional flows and thereby obstructs any reasonable cost-attribution’, are limited, so far as responding to the applicants’ objections is concerned, to a few general and abstract statements that do not genuinely respond to the applicants’ arguments.

198    In paragraph 437 (page 67) of the contested decision, it is merely stated that ‘Sub-Pleas 3.1 and 3.2 show that, from a technical point of view, … various valid flow decomposition methods co-exist’.

199    However, it should be noted that, in the examination of ‘Sub-Pleas 3.1 and 3.2’ in the contested decision, it was concluded only that various flow decomposition methods exist. By contrast, the Board of Appeal did not draw any conclusions as to the ‘validity’ of those various methods.

200    It follows from paragraph 439 (page 67) of the contested decision that the Board of Appeal took the view that the use of generation shift keys did not render the PFC method, or any part of it, fictitious. It is added that the use of generation shift keys is a standard feature of capacity calculation methods.

201    However, first, it should be noted that the applicants, as part of another objection raised before the Board of Appeal, claimed that generation shift keys, precisely because they were designed for capacity calculation, were not capable of being used in connection with flow decomposition. Second, the reference to the use of generation shift keys does not respond to the applicants’ arguments that the PFC method is vitiated by a design error since it does not take account, ab initio, of electrical distance.

202    In paragraph 447 (page 68) of the contested decision, the Board of Appeal concluded that it ‘does not find that the PFC method is based on fictitious flows’.

203    However, the preceding paragraphs do not make it possible to understand the reasons why the Board of Appeal rejected the applicants’ criticism that the PFC method led to the identification of loop flows that did not actually exist.

204    In paragraph 451 (page 69) of the contested decision, the Board of Appeal concludes that ‘the [Full Line Decomposition (FLD)] method was not an appropriate flow decomposition method’.

205    However, in so far as the PFC method is at issue, the ground set out by the Board of Appeal, based on the possible inconsistency of the FLD method with the division of the market into bidding zones, does not respond to the applicants’ argument criticising the PFC method.

206    Even if it were accepted that the reasoning relating to the FLD method must be understood as an implicit response to the objections raised by the applicants against the PFC method, that reasoning does not disclose the reasons why the Board of Appeal rejected the applicants’ objections raised before it.

207    In paragraph 449 (page 68) of the contested decision, it is stated that ‘electrical distance … can only be fully applied in a nodal market model’ and that, ‘however, if the nodal market model were to be the applicable model, there would be no [loop flows] at all, nor the need for flow decomposition’.

208    The alleged impossibility of ‘fully’ taking account of electrical distance does not mean that it is not necessary to take electrical distance into consideration in order to achieve a correct flow decomposition or that the PFC method takes due account of electrical distance.

209    Nor, moreover, is it apparent from the statement reproduced in paragraph 207 above that the non-consideration of electrical distance under the PFC method, criticised by the applicants, is inherent in the zonal model, on which the electricity market in the European Union is based, and that that zonal model does not allow electrical distance to be taken into account.

210    Paragraphs 452 to 462 (pages 69 and 70) of the contested decision, under subheading 3.4 ‘The PFC method infringes Article 16(13) [of Regulation 2019/943 and the “polluter pays” principle]’, do not contain information that makes it possible to understand the reasons why the applicants’ objections were rejected.

211    In paragraph 460 (page 70) of the contested decision, the Board of Appeal concludes that ‘the PFC method … is neither arbitrary nor based on incorrect assumptions. It correctly decomposes physical flows …’.

212    However, the preceding paragraphs do not make it possible to understand the reasons why the Board of Appeal rejected the applicants’ criticism that the PFC method leads to the identification of loop flows that do not actually exist.

213    Indeed, the preceding paragraphs are confined to mere assertions devoid of reasoning and references to conclusions drawn previously that were not themselves based on any real reasoning.

214    Thus, so far as responding to the applicants’ objections is concerned, paragraph 456 (page 69) of the contested decision merely states that ‘as set out above in Sub-Plea 3.3, the PFC method correctly identifies the physical flow components [and that] it is not based on fictitious flows’. However, the other paragraphs preceding the conclusion drawn in paragraph 460 of the contested decision do not contain anything that specifically responds to the applicants’ objections.

215    Fourth, it must also be noted that the other paragraphs of the statement of reasons in the contested decision do not fill the gaps that vitiate the reasoning set out in paragraphs 421 to 461 of the contested decision. That applies, in particular, to paragraphs 513 to 551 (pages 76 to 81) of the contested decision, concerning the use of generation shift keys, which do not directly respond to the applicants’ objections that the PFC method, in disregard of ‘natural laws’, leads to the identification of loop flows that do not actually exist. Moreover, ACER, in the course of the present proceedings and, in particular, in response to a question put by the Court, has also not specified which elements of the contested decision would make it possible to fill the gaps in the statement of reasons for that decision.

216    Fifth, according to the case-law cited in paragraphs 181 to 183 above, the obligation to state reasons for a measure varies according to the context and the addressee.

217    It is true that the subject matter of the applicants’ objections is highly complex and technical. In addition, the applicants have, as they themselves state, a high level of expertise in that field.

218    However, in the present case, it is apparent that throughout the procedure that led to the contested decision the applicants put forward the same arguments as those that were rejected, first of all, by Decision No 30/2020, then by the contested decision. It is therefore the lack of a statement of reasons enabling the applicants to understand why their objections were rejected by the Board of Appeal that led them to repeat, before the Court, the same criticism of the PFC method that they had made before the Board of Appeal.

219    It is thus apparent that, even taking into account the nature of the subject matter addressed in the contested decision and the applicants’ expertise, the statement of reasons provided does not meet the requirements set out in paragraphs 181 to 183 above.

220    Sixth, according to the case-law cited in paragraph 182 above, the institution which adopted the measure is not obliged to adopt a position on all the arguments relied on before it by the parties concerned, in so far as it may confine itself to setting out the facts and the legal considerations of fundamental importance in the context of that measure.

221    In that regard, it should be noted, first, that, as stated in paragraphs 166 and 167 above, flow decomposition is a necessary and important step in sharing the costs of remedial actions, in that it provides input data for the subsequent sharing of costs. Furthermore, the applicants’ criticism relates to the very design of the PFC method. That criticism is central to the applicants’ arguments against the flow decomposition method adopted in Decision No 30/2020 and, if it were well founded, would lead to the conclusion that the contested decision must be annulled, in so far as it confirmed Decision No 30/2020 in that regard.

222    In those circumstances, the applicants’ arguments concerned essential aspects of the contested decision and the Board of Appeal was therefore required to respond to them in order to discharge its obligation to state reasons, in accordance with the requirements set out in paragraphs 181 to 183 above.

223    It follows that the statement of reasons for the contested decision, as regards the PFC method used for flow decomposition, is incomplete and insufficient and that it does not therefore enable the applicants to ascertain the reasons for that decision or the Court to exercise its power of judicial review of that decision.

224    Accordingly, the contested decision infringes essential procedural requirements within the meaning of Article 263 TFEU, so that the contested decision must be annulled in so far as, in that decision, the Board of Appeal confirmed, without stating sufficient reasons, the flow decomposition method used by ACER in Decision No 30/2020.

 The second part, concerning the treatment of network elements that transmit direct current

225    Since the contested decision must be annulled in so far as, in that decision, the Board of Appeal confirmed, without stating sufficient reasons, the flow decomposition method used by ACER in Decision No 30/2020 (see paragraph 224 above), and in so far as the treatment of electricity network elements that transmit direct current is linked to that flow decomposition method, there is no need, for reasons of the sound administration of justice, also to examine the second part of the second plea.

 The third plea, concerning the determination of the threshold

226    By their third plea, the applicants, supported by the Federal Republic of Germany, claim that the determination of a common loop flow threshold, set at 10%, as confirmed by the contested decision, is unlawful.

227    In the first place, the applicants claim that ACER failed to recognise that the ‘polluter pays’ principle applied only to loop flows above the threshold, such that that principle could not justify a high threshold.

228    In the second place, the applicants submit that the loop flow threshold is not based on a ‘robust technical analysis’.

229    In the third place, the applicants claim that ACER was not entitled to set its ‘own’ temporary loop flow threshold.

230    In the fourth place, the applicants claim that ACER, by setting the loop flow threshold at 10%, infringed Article 16(8) of Regulation 2019/943.

231    In the fifth place, the applicants submit that ACER erred in applying the principle of proportionality.

232    In the sixth place, the applicants argue that the threshold was not determined per bidding zone border, as required by Article 16(13) of Regulation 2019/943.

233    The Federal Republic of Germany supports the applicants’ arguments.

234    ACER disputes the applicants’ arguments, as supported by the Federal Republic of Germany.

235    It is common ground that, even without structural congestion, loop flows are inevitable in a highly meshed interconnected electricity network operating according to a zonal model.

236    For that reason, the purpose of determining a loop flow threshold is to exclude such flows from the sharing of costs incurred as a result of redispatching and countertrading.

237    As regards the determination of the threshold, Article 16(13) of Regulation 2019/943 provides as follows:

‘when allocating costs of remedial actions between [TSOs], [NRAs] shall … allocate the costs … except for costs induced by flows resulting from transactions internal to bidding zones that are below the level that could be expected without structural congestion in a bidding zone.

That level shall be jointly analysed and defined by all [TSOs] in a [trading] capacity calculation region for each individual bidding zone border, and shall be subject to the approval of all [NRAs] in the capacity calculation region.’

238    First, it is apparent from Article 16(13) of Regulation 2019/943 that the threshold presupposes that the level of loop flows that could be expected without structural congestion is simulated.

239    ‘Structural congestion’ is defined in point 6 of Article 2 of Regulation 2019/943 as ‘congestion in the transmission system that is capable of being unambiguously defined, is predictable, is geographically stable over time, and frequently reoccurs under normal electricity system conditions’.

240    Second, it is apparent from the second subparagraph of Article 16(13) of Regulation 2019/943 that the determination of the threshold must be preceded by an analysis and that that analysis must be carried out by the TSOs.

241    Third, it is apparent from that provision that the threshold must be analysed and determined ‘for each individual bidding zone border’.

242    In the present case, it is common ground that the TSOs did not carry out the required analysis.

243    It is also common ground that ACER did not carry out that analysis either.

244    In that regard, it is apparent from paragraph 112 of Decision No 30/2020 that, in the absence of a threshold analysed and defined by the TSOs and approved by the NRAs in accordance with Article 16(13) of Regulation 2019/943, ACER examined whether it was in a position to perform that analysis itself and concluded that that was not the case, due to constraints on resources, the time available and the necessary expertise.

245    In addition, it is apparent from recital 8 of the contested cost sharing methodology and from paragraphs 110 to 114 of Decision No 30/2020 that ACER, in such a situation, considered itself authorised to set a temporary threshold itself.

246    Moreover, it is apparent from paragraphs 115 to 122 of Decision No 30/2020 that ACER set the temporary threshold for the whole of the Core region uniformly at 10% of the maximum capacity of each network element concerned and then divided that threshold equally between all of the bidding zones that create loop flows on the network element concerned.

247    In the contested decision, the Board of Appeal rejected as unfounded the criticism relating to ACER’s determination of a temporary threshold for the reasons set out in paragraphs 909 to 1077 (pages 137 to 164), 1210 to 1221 (pages 187 to 189) and 1192 to 1226 (pages 212 to 218) of the contested decision.

248    In that regard, it is apparent in particular from paragraphs 924 to 946 (pages 140 to 144), 1217 to 1221 (pages 187 to 189) and 1199 to 1226 (pages 213 to 218) of the contested decision that ACER considered that it had to set the threshold itself, given that the setting of such a threshold was, in its view, essential in order to be able to adopt the cost sharing methodology.

249    Furthermore, it is apparent from those same paragraphs of the contested decision that ACER’s determination of the threshold was based on a ‘rigorous analysis’.

250    It is also apparent from those parts of the contested decision that the Board of Appeal considered that ACER was authorised, or even obliged, to set a temporary threshold itself without having at its disposal the analysis prescribed in Article 16(13) of Regulation 2019/943, in order to avoid a deadlock situation.

251    In that regard, the temporary nature of the threshold thus set by ACER is emphasised in the contested decision. It is apparent from paragraph 943 (page 143) of that decision that the TSOs may still carry out the analysis required at any time and that NRAs may at any time replace the temporary threshold laid down in the contested cost sharing methodology with a definitive threshold.

252    In the light of the foregoing, it must be examined whether the Board of Appeal was entitled, without erring in law, to find, in the contested decision, that ACER’s determination of the threshold in the contested cost sharing methodology complied with the requirements stemming from Article 16(13) of Regulation 2019/943. If that is not the case, it will be necessary to examine whether, as the Board of Appeal noted, ACER, in the particular circumstances of the present case, could nevertheless rely on an implicit competence authorising it to determine a threshold in a different way than that established by that provision.

 Compliance with the requirements of Article 16(13) of Regulation 2019/943

253    In accordance with Article 16(13) of Regulation 2019/943, the threshold is to be analysed and defined ‘for each individual bidding zone border’ and must correspond to the ‘level that could be expected without structural congestion’.

254    In the first place, it must be examined whether the method for determining the threshold followed by ACER in the contested cost sharing methodology, confirmed by the contested decision, complies with the requirement for that threshold to be analysed and defined ‘for each individual bidding zone border’.

255    In that regard, it is apparent from Article 7(3) and (4) of the contested cost sharing methodology that the threshold is determined in two stages.

256    In the first stage, a common threshold for the whole of the Core region is applied to each cross-border relevant network element. That common threshold is set at 10% of the maximum capacity of the network element concerned.

257    In the second stage, the common threshold is divided equally, for each network element concerned, by the number of bidding zones within the Core region from which the loop flows that pass through that network element originate. If a bidding zone does not fully use the share of the threshold allocated to it, the unused part is then divided equally between the remaining bidding zones.

258    It follows that the threshold of each cross-border relevant network element corresponds to 10% of its maximum capacity, divided equally by the number of bidding zones within the Core region from which the loop flows that pass through that network element originate.

259    It is true, as ACER submits, that such an allocation results in an individual threshold for each bidding zone and, therefore, that that determination of the threshold entails a certain ‘individualisation’ of the threshold, in that it is determined on the basis of the individual maximum capacity of each network element concerned and according to the number of bidding zones from which loop flows that pass through those network elements originate.

260    However, it must be stated that the ‘individualisation’ referred to in paragraph 259 above is not the individualisation required, in the second subparagraph of Article 16(13) of Regulation 2019/943, by the words ‘for each individual bidding zone border’.

261    It is common ground between the parties that the level of loop flows changes according to the characteristics of the bidding zones, such as the size, the degree of meshing, the proportion of electricity injected that comes from renewable energies and the number of borders of the bidding zone concerned. Thus, the level of loop flows may vary from one bidding zone to another, on a ‘border’ within the meaning of Article 16(13) of Regulation 2019/943, or even on the elements of relevance to congestion between two zones. That is why the second subparagraph of Article 16(13) of Regulation 2019/943 requires the threshold to be determined on the basis of the characteristics of the bidding zones in question and the different borders between them.

262    However, ACER’s determination of the threshold is based, in the first stage, on a single threshold for all bidding zones in the Core region, so that no account is taken of the specific characteristics of those zones or of the borders between them.

263    Furthermore, the ‘individualisation’ carried out in the second stage also does not take account of the characteristics of the different bidding zones, but depends solely on the number of bidding zones from which the loop flows that pass through the relevant network elements originate. The same reasoning applies in the event of a subsequent allocation, among the other bidding zones, of the share of the threshold not used by a bidding zone.

264    It follows that the threshold set by ACER does not comply with the requirement laid down in the second subparagraph of Article 16(13) of Regulation 2019/943, according to which the threshold must be defined ‘for each individual bidding zone border’.

265    In the second place, as regards the requirement for the threshold to correspond to the ‘level that could be expected without structural congestion’, it is common ground that the analysis normally required to determine the level of loop flows that could be expected without structural congestion, required by the first subparagraph of Article 16(13) of Regulation 2019/943, was not carried out in the present case.

266    In the absence of such an analysis, the threshold set by ACER cannot comply with the requirement for that threshold to correspond to the level of loop flows that could be expected without structural congestion.

267    In that regard, it is apparent from paragraphs 958 (page 145) and 1221 (page 189) of the contested decision that ACER considered that the threshold which it had set corresponded to a situation without structural congestion.

268    It is apparent from paragraph 115 of Decision No 30/2020 and from paragraph 223 of the defence that ACER consulted with TSOs regarding the level of loop flows that could be expected without structural congestion. While some TSOs indicated values varying between 3, 5 and 10%, others did not respond or indicated thresholds higher than 10%. In those circumstances, and assuming that the TSOs’ responses had been influenced by their own interests, ACER set the threshold at 10% of the maximum capacity of the network element concerned as an ‘average’ of the opinions provided, as is also apparent from paragraph 115 of Decision No 30/2020.

269    It follows that the setting of 10% of the maximum capacity of the network element concerned as a common threshold for all bidding zones of the Core region and as the first stage in the determination of the threshold per network element concerned is not based on any analysis of the threshold without structural congestion, as required by the first subparagraph of Article 16(13) of Regulation 2019/943, but is the result of a compromise with regard to the divergent opinions provided by the TSOs concerned.

270    As ACER acknowledges in paragraph 930 (page 141) of the contested decision, the analysis required by the first subparagraph of Article 16(13) of Regulation 2019/943 presupposes, inter alia, an analysis of network investments and any alternative configurations of bidding zones that would remove structural congestions. ACER admits that it did not carry out such an analysis.

271    In those circumstances, ACER’s argument that its determination of the threshold is based on a ‘rigorous analysis’ is ineffective, since, in any event, that analysis was not the analysis required by Article 16(13) of Regulation 2019/943.

272    It follows that the threshold set by ACER does not comply with the requirements laid down in Article 16(13) of Regulation 2019/943, according to which the threshold must correspond to the ‘level that could be expected without structural congestion’ and must be defined ‘for each individual bidding zone border’.

 The conditions for recognising implicit competence

273    It follows from the foregoing assessments that ACER’s determination of the threshold in the contested cost sharing methodology does not comply with Article 16(13) of Regulation 2019/943.

274    In those circumstances, the question whether, in principle, ACER was competent to determine a threshold itself on the basis of point (a) of the second subparagraph of Article 6(10) of Regulation 2019/942, as follows from paragraph 924 (page 140) of the contested decision, is irrelevant. That provision cannot, in any event, allow ACER to set a threshold that does not comply with the requirements of Article 16(13) of Regulation 2019/943.

275    It must therefore be examined whether, despite the fact that ACER’s determination of the threshold did not comply with the requirements of Article 16(13) of Regulation 2019/943, ACER had, in the specific situation in which it found itself, an implicit competence authorising it to determine a threshold in a different way than that prescribed by that provision.

276    ACER contends that its competence to determine a threshold is justified by the need for it to act. In the absence of the analysis of the level that could be expected without structural congestion that was to be carried out by the TSOs, ACER was authorised, in order to avoid a deadlock situation, to set a temporary threshold itself in the contested cost sharing methodology.

277    In the first place, it cannot in principle be accepted, in the light of the principle of legality, that an agency of the European Union, such as ACER, may derogate from the applicable legal framework. It follows that ACER could not, in principle, derogate from Article 16(13) of Regulation 2019/943.

278    In the second place, it should be noted that, in accordance with Article 6(12)(b) of Regulation 2019/942, ACER may ‘provide an interim decision to ensure that … operational security is protected’. It must be stated that, in the present case, ACER did not rely on that provision to establish the threshold. Furthermore, and in any event, the adoption of the cost sharing methodology cannot be regarded as necessary in order to ‘ensure that security of supply or operational security is protected’, within the meaning of Article 6(12)(b) of Regulation 2019/942. The purpose of that methodology is to share the costs of remedial actions and it is not intended to determine the remedial actions to be taken in order to ensure security of supply or operating security.

279    Moreover, the existence of that provision and, accordingly, of the power to adopt, in clearly defined circumstances, interim decisions militates against recognising that ACER has an implicit competence to set, even temporarily, the threshold in the cost sharing methodology in a different way than that prescribed by Article 16(13) of Regulation 2019/943.

280    In the third place, in accordance with the case-law, mere reliance on an interest linked to effectiveness is insufficient to create a competence on the part of an agency of the European Union (see, to that effect, judgment of 24 October 2019, E-Control v ACER, T‑332/17, not published, EU:T:2019:761, paragraph 69). Mere reliance on an interest linked to effectiveness is therefore insufficient to allow an agency of the European Union to derogate from the applicable legal framework.

281    However, it cannot be ruled out that an interest linked to effectiveness, provided that it corresponds to a real need to ensure the practical effect of the provisions of the Treaties or the regulation concerned, may justify the existence of an implicit decision-making power (see, to that effect, judgment of 24 October 2019, E-Control v ACER, T‑332/17, not published, EU:T:2019:761, paragraph 69).

282    It must therefore be examined whether, in the present case, the conditions for recognising that ACER has implicit competence, in accordance with that case-law, were satisfied.

 Recognition that ACER has implicit competence

283    In accordance with the case-law cited in paragraph 281 above, in order to determine whether ACER could avail itself of implicit competence, it must be examined whether recognising that ACER has such competence corresponded to a real need to ensure the practical effect of the provisions at issue.

284    In that regard, it is apparent from paragraphs 924 to 946 (pages 140 to 144) and 1206 to 1220 (pages 214 to 216) of the contested decision that ACER claimed that the need to adopt a cost sharing methodology within the prescribed time limit had forced it to set the threshold itself, despite the absence of the analysis normally required.

285    First, as regards the need relied on with regard to the timetable, it is true that ACER was, in principle, required to adopt a cost sharing methodology within the time limit laid down in Article 6(12)(a) of Regulation 2019/942 and Article 9(11) of Regulation 2015/1222, namely a time limit of six months from 27 March 2020.

286    However, EU law does not attach any penalty to exceeding the six-month time limit laid down in Article 6(12)(a) of Regulation 2019/942 and Article 9(11) of Regulation 2015/1222. That time limit is therefore indicative rather than mandatory.

287    In accordance with the case-law, where there is such an indicative time limit, even if the EU body to which that time limit applies must endeavour to comply with it, it may be necessary for it, in particular because of the complexity of the task and provided that the interests of a Member State are not impaired, to have more time (see, to that effect, judgment of 15 January 2013, Spain v Commission, T‑54/11, EU:T:2013:10, paragraph 27).

288    Consequently, the starting point of ACER’s reasoning is incorrect, since it was not mandatory for it to adopt a cost sharing methodology within the prescribed time limit, that is to say, before 28 September 2020.

289    Furthermore, it is common ground that the analysis of the ‘level that could be expected without structural congestion’ is complex and requires a considerable amount of time.

290    It was therefore, in principle, open to ACER to allow the TSOs sufficient time to carry out the analysis required without being criticised for disregarding the time limit laid down in Article 6(12)(a) of Regulation 2019/942 and Article 9(11) of Regulation 2015/1222.

291    That is all the more so since the obligation to determine a threshold by carrying out the analysis provided for in Article 16(13) of Regulation 2019/943 did not enter into force until 1 January 2020.

292    At the time when the TSOs of the Core region submitted their proposal for a cost sharing methodology to all the NRAs of that region for approval, namely on 27 March 2019, Regulation 2019/943 had not yet been adopted.

293    In that regard, it must be stated that neither Decision No 30/2020 nor the contested decision explores the possible consequences, for the timetable for adopting the cost sharing methodology, of the fact that the obligation to determine a threshold and, therefore, to carry out the corresponding analysis did not enter into force until 1 January 2020.

294    Therefore, the Board of Appeal’s assumption in the contested decision that it was necessary for ACER to adopt the contested cost sharing methodology, without being able to wait for the analysis prescribed by Article 16(13) of Regulation 2019/943, owing to the time limit set for it to do so, is based on a reading of the regulatory framework that does not take into account either the indicative nature of the time limit to be observed by ACER or the change in the legal framework in force.

295    Accordingly, mere reliance on an indicative time limit for ACER to adopt the cost sharing methodology is insufficient to demonstrate a real need to ensure the practical effect of the provisions at issue.

296    Second, ACER stated that the need for it to act is justified by the TSOs’ failure to act. It is apparent from paragraph 955 (page 145) of the contested decision that ACER contended that the TSOs had not been able, ‘in a period of nearly 3 years’, to carry out the analysis required by Article 16(13) of Regulation 2019/943.

297    In that regard, it should be recalled that the obligation to determine the threshold by carrying out the analysis prescribed by Article 16(13) of Regulation 2019/943 did not enter into force until 1 January 2020.

298    Moreover, it must be stated that, contrary to what ACER maintains in paragraph 926 (page 140) of the contested decision, the TSOs did not consider it mandatory, in the explanatory document of 22 February 2019 accompanying the proposal for the cost sharing methodology of 27 March 2019, to determine a threshold, but rather perceived that as an option left to them.

299    Even if it were accepted that the need to set a threshold was recognised by the TSOs before the adoption of Regulation 2019/943, the fact remains that Article 16(13) of Regulation 2019/943 specifies how the threshold should be determined, namely on the basis of an analysis of the ‘level [of loop flows] that could be expected without structural congestion’ and ‘for each individual bidding zone border’.

300    ACER does not contend that it was established, before the entry into force of Regulation 2019/943, that the threshold had to be determined in that way.

301    In those circumstances, ACER could not legitimately criticise the TSOs for not being able, ‘in a period of nearly 3 years’, to carry out the analysis required by Article 16(13) of Regulation 2019/943.

302    That conclusion is also not called into question by the documents produced by ACER at the hearing to demonstrate that the TSOs and NRAs of the Core region were still discussing how the analysis required by Article 16(13) of Regulation 2019/943 should be carried out.

303    Irrespective of whether those documents are admissible, it must be stated that the ongoing negotiations relied on by ACER are irrelevant to the present case.

304    According to the case-law, the legality of a decision must be assessed purely on the basis of the elements of fact and of law existing at the time when it was adopted (see judgment of 27 April 2022, Roos and Others v Parliament, T‑710/21, T‑722/21 and T‑723/21, EU:T:2022:262, paragraph 211 and the case-law cited). Thus, the evidence relied on by ACER, which post-dates the contested decision, cannot be taken into account for the purpose of assessing the legality of that decision.

305    Third, ACER contended that the need for it to act is justified by the TSOs’ failure to comply with the time limit that it had set for them. In that regard, it states that it set for the TSOs a time limit of four months, from 18 April to 20 August 2020, to carry out the analysis required by Article 16(13) of Regulation 2019/943 and adds that they did not comply with that time limit.

306    As is apparent from paragraphs 930 (page 141), 954 (page 145) and 1131 (page 202) of the contested decision, ACER itself considered that the analysis prescribed by Article 16(13) of Regulation 2019/943 was complex and would require a considerable amount of time.

307    Moreover, ACER has not demonstrated that, during the four-month time limit that it had set for the TSOs, it facilitated, in one way or another, the work of the TSOs in carrying out the analysis prescribed by Article 16(13) of Regulation 2019/943.

308    In accordance with Article 6(11) of Regulation 2019/942, which reflects the principle of sincere cooperation enshrined in Article 4(3) TEU, ACER is required to consult NRAs and TSOs when preparing a decision on the basis of Article 6(10) of that regulation.

309    In accordance with the principle of sincere cooperation and taking into account the clear intention of the EU legislature to make decision-making on difficult but necessary cross-border issues more efficient and faster (judgment of 7 September 2022, BNetzA v ACER, T‑631/19, EU:T:2022:509, paragraph 46), ACER was to facilitate the development, by the TSOs and NRAs, of the analysis required by Article 16(13) of Regulation 2019/943.

310    In those circumstances, ACER cannot legitimately criticise the TSOs for not being able to carry out the analysis required by Article 16(13) of Regulation 2019/943 within the prescribed time limit, namely four months.

311    Fourth, ACER further contended that the need to adopt the contested cost sharing methodology without being able to wait for the analysis required by Article 16(13) of Regulation 2019/943 is justified by two other considerations.

312    First, at the hearing, ACER contended that it would have been necessary to give the TSOs sufficient time, after the contested cost sharing methodology was adopted, to put in place the necessary arrangements properly to implement it.

313    It should be noted that, according to Article 13(2) of the contested cost sharing methodology, read in conjunction with Article 37(2) of the RDCT methodology, a first step in the implementation of the contested cost sharing methodology was scheduled for 4 June 2023, whereas the full implementation of that methodology was scheduled for 4 June 2025.

314    Thus, given that the contested cost sharing methodology was first to be partially implemented only two and a half years after it was adopted, and fully implemented four and a half years after it was adopted, mere reliance on the need for the TSOs to put in place the necessary arrangements properly to implement that methodology is not sufficient to demonstrate, in the light of the very long period laid down for implementing it, a real need to adopt that methodology without being able to wait for the analysis required by Article 16(13) of Regulation 2019/943.

315    Second, it is apparent from paragraph 946 (page 144) of the contested decision that ACER considers that the contested cost sharing methodology had to be implemented simultaneously with the RDCT methodology and the ROSC methodology.

316    In that regard, it is sufficient to note that the present case in no way involves deciding whether ACER was authorised to set the same dates for the implementation of the contested cost sharing methodology, the RDCT methodology and the ROSC methodology, but rather determining whether ACER could adopt the contested cost sharing methodology without being able to wait for the analysis required by Article 16(13) of Regulation 2019/943.

317    In those circumstances, the argument based on the alleged need to implement the contested cost sharing methodology, the RDCT methodology and the ROSC methodology simultaneously is irrelevant to the question of when the cost sharing methodology had to be adopted.

318    Accordingly, ACER also could not contend that the need for it to act was justified by considerations relating to the need to allow the TSOs sufficient time to put in place the necessary arrangements properly to implement the contested cost sharing methodology simultaneously with two other methodologies.

319    It follows that ACER has not established that it was necessary to adopt the contested cost sharing methodology without being able to wait for the analysis required by Article 16(13) of Regulation 2019/943.

320    Consequently, ACER has not demonstrated that there was a real need to ensure the practical effect of the provisions at issue that would justify recognising that it has implicit competence.

321    In any event, it must be stated that ACER’s determination of the threshold is not capable of ensuring the practical effect of the provisions at issue.

322    It is true that that determination of the threshold enabled ACER to adopt the contested cost sharing methodology on 30 November 2020, that is to say, slightly after the time limit that had been set for it to do so, which expired on 27 September 2020. However, that did not ensure the practical effect of the substantive provisions at issue.

323    The contested cost sharing methodology must, in accordance with Article 74(6)(a) of Regulation 2015/1222, provide incentives to invest effectively to manage congestion. In addition, according to recital 34 of Regulation 2019/943, the management of congestion should provide correct economic signals to TSOs and market participants.

324    As stated in paragraph 272 above, the threshold determined by ACER, confirmed by the contested decision, does not comply with the requirements of Article 16(13) of Regulation 2019/943, according to which the threshold must correspond to the ‘level that could be expected without structural congestion’ and be defined ‘for each individual bidding zone border’.

325    In those circumstances, on account of that determination of the threshold, the contested cost sharing methodology cannot provide, by means of sharing the costs of remedial actions, ‘correct economic signals’ for network investments.

326    Furthermore, it must also be stated that the weighing-up of the interest in complying with the prescribed time limit and the interest in complying with Article 16(13) of Regulation 2019/943, apparently carried out by ACER, does not justify determining a threshold that does not comply with the relevant legislation.

327    As pointed out in paragraph 286 above, the time limit set for ACER to adopt a cost sharing methodology was merely indicative, so that, in a weighing-up of interests, the intention to comply with that time limit cannot take precedence over compliance with the requirements of Article 16(13) of Regulation 2019/943.

328    That outcome cannot be called into question by ACER’s argument that it determined the threshold only on a temporary basis. The temporary nature of that determination in no way mitigates the breach of the relevant regulatory framework by ACER and the Board of Appeal.

329    The fact that the temporary nature of the determination of the threshold exists only de jure does not make it possible to lessen ACER’s failure to comply with the relevant regulatory framework.

330    Accordingly, ACER cannot justify its approach of giving precedence, in the weighing-up of interests, to the interest in complying with the prescribed time limit over the interest in complying with the requirements of Article 16(13) of Regulation 2019/943 by relying on the temporary nature of the resulting interference with the regulatory framework, through determining a threshold in a way that does not comply with that framework. Consequently, ACER could not base its determination of the threshold on an implicit competence.

331    In the light of the foregoing assessments, it must be concluded that ACER’s determination of the threshold in the contested cost sharing methodology, as confirmed by the contested decision, infringes Article 16(13) of Regulation 2019/943, in that that threshold does not meet the criterion that the threshold must correspond to the ‘level that could be expected without structural congestion’ or the criterion that the threshold must be defined ‘for each individual bidding zone border’. Furthermore, it is apparent from the foregoing that ACER was also not entitled to determine a threshold differently in order to comply with the time limit set for it to adopt the contested cost sharing methodology.

332    In those circumstances, the third plea must be upheld, without there being any need to examine the other objections raised by the applicants in support of it.

333    In so far as it is apparent from the examination of the first part of the second plea that the contested decision is vitiated by an infringement of essential procedural requirements within the meaning of Article 263 TFEU and in so far as the third plea concerns a central element of the contested cost sharing methodology, which is the subject of the contested decision, the Court cannot annul the contested decision only in part.

334    Consequently, the applicants’ action must be upheld on the basis of the first part of the second plea and the third plea and the contested decision must be annulled in so far as it confirms Decision No 30/2020 and dismisses the applicants’ appeal in Case A-001-2021 (consolidated).

 Whether the contested decision should be maintained

335    Under the second paragraph of Article 264 TFEU, the Court may, if it considers it necessary, state which of the effects of the act which it has declared void are to be considered as definitive.

336    In response to measures of organisation of procedure ordered by the Court, the parties submitted their observations in that regard.

337    In the present case, ACER contends that the annulment of the contested decision and, consequently, of the contested cost sharing methodology would have serious consequences. TSOs would have to bear all the costs of all remedial actions activated on their network elements, even if those actions were necessary due to loop flows from other bidding zones. The effect of that situation would likely be that TSOs would have an incentive to limit cross-zonal capacity, resulting in higher electricity prices.

338    In accordance with the case-law, on grounds of legal certainty, the effects of an act may be maintained, in particular, where the immediate effects of its annulment would give rise to serious negative consequences for the persons concerned and where the lawfulness of the act in question is contested not because of its aim or content, but on grounds of lack of competence or infringement of an essential procedural requirement (see judgment of 15 July 2021, Commission v Landesbank Baden-Württemberg and SRB, C‑584/20 P and C‑621/20 P, EU:C:2021:601, paragraph 175 and the case-law cited).

339    In that regard, first, it should be noted that the annulment of the contested decision is based, inter alia, on an infringement of substantive law, namely an infringement of the second subparagraph of Article 16(13) of Regulation 2019/943, and not solely on an infringement of essential procedural requirements.

340    Second, ACER’s line of argument is based on the assumption that the contested cost sharing methodology is already applicable.

341    It is apparent from the parties’ answers to the questions put by the Court that the contested cost sharing methodology, which is the subject of the contested decision, will not be fully implemented before 4 June 2025 and that, on account of delays, its implementation may even be further postponed.

342    In those circumstances, there is no need to limit the effect of the annulment of the contested decision.

 Costs

343    Under Article 134(1) of the Rules of Procedure, the unsuccessful party is to be ordered to pay the costs if they have been applied for in the successful party’s pleadings.

344    Since ACER has been unsuccessful and the applicants have applied for costs, ACER must be ordered to pay the costs.

345    According to Article 138(1) of the Rules of Procedure, the Member States and institutions which have intervened in the proceedings are to bear their own costs. The Federal Republic of Germany must therefore bear its own costs.

On those grounds,

THE GENERAL COURT (Third Chamber, Extended Composition)

hereby:

1.      Annuls the decision of the Board of Appeal of the European Union Agency for the Cooperation of Energy Regulators (ACER) of 28 May 2021 in Case A-001-2021 (consolidated), in so far as it confirms ACER Decision No 30/2020 of 30 November 2020 on the proposal of the electricity transmission system operators of the ‘Core’ capacity calculation region, comprising Belgium, the Czech Republic, Germany, France, Croatia, Luxembourg, Hungary, the Netherlands, Austria, Poland, Romania, Slovenia and Slovakia, for the methodology for cost sharing of redispatching and countertrading, and dismisses the applicants’ appeal in that case;

2.      Orders ACER to bear its own costs and to pay those incurred by TenneT TSO GmbH and TenneT TSO BV;

3.      Orders the Federal Republic of Germany to bear its own costs.

Delivered in open court in Luxembourg on 25 September 2024.

*      Language of the case: English.

© European Union
The source of this judgment is the Europa web site. The information on this site is subject to a information found here: Important legal notice. This electronic version is not authentic and is subject to amendment.