The Gas and Electricity Markets Authority (GEMA) v R (SSE Generation Limited & Ors) (SSE) [2022] EWCA Civ 1472

[2022] EWCA Civ 1472Case No CA-2022-000852
IN THE COURT OF APPEAL (CIVIL DIVISION)
[2022] EWHC 865 (Admin)
ON APPEAL FROM THE HIGH COURT OF JUSTICE (KING’S BENCH DIVISION)
MR JUSTICE SWIFT
Venue Royal Courts of Justice, Strand, London, WC2A 2LLDate 08/11/2022
Before
SIR JULIAN FLAUX, CHANCELLOR OF THE HIGH COURTLORD JUSTICE GREENLORD JUSTICE SNOWDEN
Between
R (SSE GENERATION LIMITED & OTHERS) (SSE)Claimant / Respondent
Kassie Smith KC, Amy Rogers, George Molyneaux (instructed by GEMA)Tom Hickman KC, Christopher Brown (instructed by the CMA)Kieron Beal KC, instructed by Addleshaw Goddard LLP for SSE for National Grid Electricity System Operator Ltd, Centrica Plc and British Gas Trading Ltd did not appear and were not representedHearing Hearing dates : 18-19 July & 22 September 2022
Approved JudgmentThis judgment was handed down remotely at 10.30am on 8 November 2022 by circulation to the parties or their representatives by e-mail and by release to the National Archives..............................

The appeals

[1]There are two appeals before the Court. They raise issues of importance to the operation of the system for the setting of charges for the transmission of electricity but also to wider public law.[2]They concern a decision of the Gas and Electricity Markets Authority (“GEMA”) dated 17th December 2020 (“the GEMA Decision”). The decision relates to a modification to the methodology for setting transmission charges payable to network operators by electricity generators using the network in Great Britain. In the decision GEMA (inter alia):(i) required a methodology to be adopted for the setting of electricity transmission charges which was not compliant with the relevant law but which was intended to constitute an interim position pending a fresh decision which would be legally compliant; and(ii) adopted a definition of the expression “congestion management” which meant that the costs attributable to the management of “congestion” on the transmission system within Great Britain were not taken into account when calculating limits on the transmission charges which could be levied This had the effect of limiting the costs taken into account when transmission charges were set to those incurred by the network operators attributable to the management of congestion on interconnectors governing international trade between EU member states only, and not also the costs of managing congestion elsewhere within the transmission system. For reasons explained below it is in the commercial interests of generators for all the costs attributable to congestion management to be taken into account because, under the relevant rules, there is a cap on the transmission charges that can then be levied on generators by network operators.[3]One affected generator, SSE Generation Limited (“SSE”), brought a statutory appeal to the Competition and Markets Authority (“CMA”) which in a decision dated 30th March 2021 upheld the GEMA Decision (“The CMA Decision” SSE Code Modification Appeal 2021- An appeal under section 173 of the Energy Act 2004: SSE Generation Limited v Gas and Electricity Markets Authority and National Grid Electricity System Operator Limited (Intervener) and Centrica plc/British Gas Trading Limited (Intervener) ). SSE sought judicial review of the CMA Decision on a variety of grounds, all but one of which were rejected. As to the one successful ground, the High Court held that it was unlawful for the CMA to approve the GEMA Decision endorsing the imposition of unlawful charges, even as an interim measure. The Court came to this conclusion notwithstanding that implementation of the decision would give rise to a significant improvement upon the pre-existing position and its setting aside would resurrect a situation that was less compliant with the law than the interim position set out in the decision. The High Court however upheld the conclusion of the CMA, and therefore of GEMA, that the only congestion management costs falling outside of the exclusion for ancillary services were those relating to the management of congestion on interconnectors between EU member states.[4]There are two appeals against the judgment of the High Court. The first is brought by GEMA in relation to its power to adopt interim measures which include non-legally compliant components. The CMA supports GEMA as an interested party to the appeal. The respondent to the appeal is SSE. The interested parties are National Grid Electricity System Operator Limited (“NGESO”) and, Centrica PLC/ British Gas Trading Limited, who were also parties to the CMA appeal proceedings and were interested parties in the judicial review (Issue I). The second appeal is brought by SSE. GEMA are the respondent party. CMA appeared as an interested party only in this proceedings. This concerns the treatment of “congestion management” costs (Issue II). Outline of Issue I: The duty of regulators to ensure compliance with the law - how to transition from non-compliance into compliance[5]Issue I concerns the powers and duties of a relevant regulator (GEMA) when confronted with a situation whereby a system (here a charging regime) is not compliant with the law. The regulator then has to devise a system to bring the system into compliance. But this is not a state of affairs that can be brought about easily or immediately; it will be a complex exercise which will take time. The expert judgment of the regulator was that the transition to a state of compliance is best achieved in two stages with the first being a “stop gap” applicable in the short term pending a full resolution of the issue. A corollary is that because the stop gap is a staging post it will, unavoidably, not yet be fully compliant with the law. It will, though, be a significant improvement upon the status quo ante and it will lead in due course to a state of full compliance. The issue arising is whether, in adopting such a non-compliant stop gap, the regulator was acting lawfully? The Judge below held that it was not. The question for this Court is whether this is correct. Outline of Issue II: The meaning of the expression “congestion management” – purposive interpretation - EU retained law – correcting errors in legislation[6]Issue II is about the meaning of the phrase “congestion management” and, more specifically, whether the costs of managing certain aspects of network congestion must be taken into account in the calculation of transmission charges. We were told that very substantial sums of money rested upon the correct meaning of this expression. At the risk of over simplification congestion arises on a network where there is some physical blockage or bottleneck which means that the flow of electricity moving in one direction gets held up. There are many costs associated with managing this problem. Congestion can arise anywhere on a network. One point where it can be especially problematic is at the interconnections between different national networks.[7]If such costs of congestion management amount to costs for an “ancillaryservice” provided to generators then, under the relevant law, they must be excluded from the calculation of transmission charges that are subject to annual average regulatory limits. The issue then concerns the meaning of the expression “congestion”. A definition used in a 2009 Regulation made clear that it was congestion at interconnectors between different states which impeded international trade that the legislation was aimed at. However, a decade later, in the 2019 Recast Regulation, the definition of “congestion” was changed and all references to interconnectors and to international trade between states were removed from the definition. The earlier definition was expressly repealed. At the same time, “congestion management” was explicitly carved out of the exclusion for “ancillary services”. This had the consequence, prima facie, that charges in respect of such costs fell to be included in the calculation of the core transmission charge that was subject to annual average limits. The dispute now is whether “congestion” is still to be taken as referring to the costs of congestion management only on interconnectors governing international trade, or, on the entirety of a transmission system wherever it arises.[8]The position has been complicated by the fact that when the relevant (2019) EU measures were translated into EU retained law, upon the expiry of the post-Brexit implementation period, for reasons which are very difficult to discern, the drafter of the EU retained law, whilst purporting not to change anything, instead of just copying and pasting the 2019 definition of “congestion” into the domestic retained EU law, instead reverted back to the 2009 version and used the repealed definition of “congestion”, but at the same time suggested that this reversion was not intended to bear a meaning differing from the new 2019 version. To complicate matters even further, the version of the law applicable under the Northern Ireland Protocol (“the NI Protocol”) left intact the applicable 2019 EU definition.[9]The domestic electricity market is interconnected to wider EU markets. It is no part of current domestic energy policy to create a set of rules diverging from those governing the operation of transmission systems elsewhere in the EU. It is considered important that everyone in the market works from the same overarching charging principles. If, however, Great Britain adopts a definition of “congestion management” differing from that being used elsewhere (whether in Northern Ireland or in the EU) this risks creating market distortions. Divergence in the pricing signals being communicated to the market arising from inconsistent charging regimes can cause distortions in the free flow of energy, and this in turn risks distorting investment decisions. This court has no information about how the expression “congestion management” is construed and applied in the EU or by the relevant regulatory authorities in Europe. We have no jurisdiction to make a reference to the CJEU which can, using its normal procedures, receive submissions from all affected states and EU Institutions and thereby, when delivering judgment, achieve consistency. For this reason, the court mooted with the parties the possibility that the court might, in some suitable manner, seek the assistance of the European Commission in providing us with evidence of how charges were set elsewhere and as to their views on the correct construction of the law. In the event, for reasons explained below (see paragraphs [132] – [133]), this course was not adopted.

The legislative and regulatory system

[10]The law and regulatory practices governing these appeals is of Byzantine complexity. It includes a plethora of regulations, directives, decisions and rules of the EU Council, Parliament and Commission which have been repeatedly amended, modified and recast, and certain of which have become EU retained law and continue to have effect in Great Britain pursuant to section 3 of the European Union (Withdrawal) Act 2018 (“EU(W)A 2018”). The position in Northern Ireland is different, being governed by the NI Protocol.[11]In domestic law, powers and duties exist in relation to regulators under the Electricity Act 1989 (“EA 1989”) which includes extensive licensing powers for (inter alia) transmission. The regulatory practices consist of of a series of statutory decisions of GEMA and the CMA. There is also a Code adopted by NGESO, and given statutory approval by GEMA, called the Connection and Use of System Code (“CUSC” or “the Code”) which takes legal effect as a contract between NGESO, electricity generators and others. This sets out the relevant charging methodology and the rules governing modifications to such methodologies. There are also multiple proposals for modification to the Code charging principles made pursuant to mechanisms laid down in the Code submitted by electricity generators and others, all subject to approval procedures adopted by GEMA.[12]Then there are binding guidelines, Explanatory Memoranda, impact assessments and other working documents said to amount to travaux préparatoires to EU and UK retained measures to which reference has been made. And finally there is an array of informally stated views and opinions expressed by officials at the national and international level to which we have also been referred.[13]For the purposes of the comprehensibility of this judgment I have endeavoured to be as selective as possible.

The transmission network and the relevant Codes

[14]To place the legislative and regulatory systems into context it is helpful to provide a brief summary of the GB transmission system. NGESO is the operator of the transmission network. Chapter 2 of the CMA Decision provides a concise overview of the ownership and operation of the GB electricity generation, transmission and supply sector:
“NGESO operates the transmission system in GB 2.10 The TOs “TOs” are Transmission Owners. and the OFTOs “OFTOs” are Offshore Transmission Owners. own and maintain the physical transmission infrastructure but they do not operate it. The system as a whole is operated by a single system operator: the Transmission System Operator (TSO). The TSO functions for the whole of GB, including offshore transmission, are performed by a separate system operator entity, NGESO. 2.11 The TSO is also responsible for keeping the electricity system balanced in real-time by coordinating the output of generating stations and ensuring that supply and demand are exactly matched, and all equipment is being operated within safe physical limits at all times. The tools and actions that NGESO use to do this are known as ‘Balancing Services’, and include, for example, paying particular generators to reduce or increase their output. Balancing Services are essential to the safe operation of the GB transmission system. 2.12 The nature of Balancing Services is that one balancing action can be used to serve multiple purposes (for instance imbalance, network constraints, voltage issues). Where a constraint on the capacity of a piece of infrastructure is the initial driver of a balancing action, the costs of that action will be categorised as a system ‘constraint’ cost in NGESO’s reporting of charges for Balancing Services, even in instances where taking that action has also resolved multiple issues that are not constraint-related. 2.13 The characteristics of electricity mean that quantities of energy generated and consumed are very likely to differ from the quantities for which contracts have been struck in advance. Consequently, central arrangements are required to: (a) meter the quantities produced and consumed; (b) compare these with the quantities covered by bilateral contracts between generators, interconnectors and suppliers; and (c) provide financial settlement for the differences (known as ‘imbalances’).”
[15]The EA 1989 sets out the relevant licensing regime. Pursuant to sections 6-8 GEMA is empowered to grant licences including for generation and transmission and set conditions. Section 11A empowers GEMA to modify licence conditions. NGESO’s transmission licence requires it at all times to have in force a “Balancing and Settlement Code” (“theBSC”) setting out the governance arrangements for electricity balancing and settlement in GB. Under the BSC, NGESO can increase or decrease the quantities of electricity to be delivered to, or taken off, the total system at any time in order to ensure that the system is in balance. It also provides for the settlement of consequential financial obligations between Code parties. Paragraphs [3.11] – [3.13] of the CMA Decision state:
“3.11 The balancing mechanism provides a means by which NGESO can buy or sell additional energy close to real-time to maintain energy balance, and also to deal with other operational constraints of the transmission system. The energy balancing aspect of the BSC specifies how parties can make submissions to NGESO to either buy or sell electricity into/out of the market at close to real time through the balancing mechanism. 3.12 The settlement aspect of the BSC specifies how the actual positions of generators and suppliers (and interconnectors) against their contracted positions are monitored and metered and how imbalances are settled when actual delivery or offtake does not match contractual positions. 3.13 The BSC is administered by Elexon Limited (Elexon), a not-for-profit company which carries out this financial settlement process. Elexon’s administrative costs are recovered through BSC charges. The method of calculation of these is set out in Section D of the BSC.”
[16]There is also a requirement fora CodeCondition C4 of NGESO’s transmission licence sets out the basis for “Use of System”’ charges. These comprise(i) the Transmission Network Use of System (“TNUoS”) and(ii) Balancing Services Use of System (“BSUoS”) charges. The CMA Decision states:
“3.15 The licence conditions include the following: Condition C4.1(a) requires NGESO to set a use of system charging methodology which will be approved by GEMA; and C4.1(b) requires NGESO to conform to the methodology (as modified in accordance with standard condition C5 (Use of system charging methodology) and standard condition C10 (Connection and Use of System Code)). Standard condition C4 paragraph 7 states that for these purposes ‘charges do not include references to: (a) connection charges’. 3.16 By way of standard condition C10.2(d), the charging methodology is to be set out in the CUSC which NGESO is required by standard condition C10 to establish (see paragraph 3.21 below). The methodology for calculating TNUoS and BSUoS charges is in section 14 of the CUSC.”
[17]The Code is an important instrument. Under a separate agreement known as the CUSC Framework Agreement, the Code is contractually binding as between NGESO and those (including generators) it is applied to. GEMA is not a party to the Code.NGESO’s licence requires it to keep the use of system charging methodology (in the Code) under review at all times.[18]The impugned GEMA Decision concerns a modification proposal. The Code contains provisions governing how modifications come about and it has been prayed in aid in a variety of ways by the parties in argument. Proposed modifications must be assessed against specified objectives (“Applicable CUSC objectives” or “ACOs”) which are set out in standard condition C5.5. One of these, as summarised in the CMA Decision, was to ensure: “… compliance with the EU Electricity Regulation 2009 on conditions for access to the network for cross-border exchanges in electricity (the Electricity Regulation) and any relevant legally binding decisions of the European Commission”.[19]Section 8 sets out a detailed process whereby modifications to the charging methodology can be made. There are various ways in which the Code can be modified. Under paragraph 8.16.1 the process is started by a proposal from any of the persons identified in Section 14, paragraph 8.16, which excludes GEMA:
“CUSC MODIFICATION PROPOSALS 8.16.1 … (b) A proposal to modify the Charging Methodologies may be made: (i) by a CUSC Party, by the Citizens Advice, by the Citizens Advice Scotland or by a BSC Party; or (ii) under Paragraph 8.28.5, by the CUSC Modifications Panel; or (iii) by a Relevant Transmission Licensee in relation to Exhibit O Part IB, Exhibit O Part IIB, Exhibit O Part IC and Exhibit O Part IIC only; or (iv) by a Materially Affected Party, unless otherwise permitted by the Authority. (v) by the Authority, or by The Company under the direction of the Authority, pursuant to Paragraph 8.17A.1.” (i) by a CUSC Party, by the Citizens Advice, by the Citizens Advice Scotland or by a BSC Party; or (ii) under Paragraph 8.28.5, by the CUSC Modifications Panel; or (iii) by a Relevant Transmission Licensee in relation to Exhibit O Part IB, Exhibit O Part IIB, Exhibit O Part IC and Exhibit O Part IIC only; or (iv) by a Materially Affected Party, unless otherwise permitted by the Authority. (v) by the Authority, or by The Company under the direction of the Authority, pursuant to Paragraph 8.17A.1.”
[20]A proposal is evaluated by the “CUSC Modification Panel” which may refer a proposal to a Workgroup for consideration. Where a reference is made the group evaluates the proposal and alternatives and reports back to the CUSC Modification Panel. A process of consultation then follows on the proposals and any alternatives following which the CUSC Modification Panel reports to GEMA. The final decision is taken by GEMA.[21]The options open to GEMA are, however, circumscribed under paragraph 8.23.7:
“… in accordance with the Transmission Licence, the Authority may approve the CUSC Modification Proposal or a Workgroup Alternative CUSC Modification(s) contained in the CUSC Modification Report (which shall then be an "Approved CUSC Modification" until implemented). If the Authority believes that neither the CUSC Modification Proposal (nor any Workgroup Alternative CUSC Modification(s)) would better facilitate achievement of the Applicable CUSC Objectives, then there will be no approval. In such a case, the Code Administrator will notify CUSC Parties and will raise the issue at the next CUSC Modifications Panel meeting.”
[22]When deciding whether to approve or reject proposals submitted to it, GEMA must consider whether those proposals would “better facilitate” achieving the ACOs (see paragraph [18] above). The benchmark status quo ante is referred to as the “baseline”. GEMA cannot approve a proposal unless it considers that the proposal would better facilitate achievement of the ACOs.[23]Under paragraph 8.23.12, GEMA has the power to remit proposals for revision and resubmission where GEMA “cannot properly form an opinion” on the proposals in question. GEMA may approve the original proposal or one of the alternatives, but if it believes that none would better facilitate achievement of the ACO “…there will be no approval".[24]GEMA does have a power to adopt its own proposal under the Code to direct a change to its terms to cure a breach of EU law under paragraph 8.17A(1)(a):
“8.17A AUTHORITY RAISED OR DIRECTED MODIFICATION 8.17A.1 The Authority may: (a) itself; or (b) direct The Company to raise a CUSC Modification Proposal where the Authority reasonably considers that such CUSC Modification Proposal is necessary to comply with or implement the Electricity Regulation and/or any relevant Legally Binding Decisions of the European Commission and/or The Agency or in respect of Significant Code Review.”
[25]Even where GEMA can raise such a modification proposal it must still however use the contractual procedure laid down in the Code. It cannot, unilaterally, take its own decision and then impose it.

Regulation (EC) No 714/2009 - The 2009 Regulation

[26]I turn now to the principal legislative measures. The starting point is Regulation (EC) No 714/2009 of the European Parliament and of the Council of 13th July 2009 (“the 2009 Regulation”). This concerns “… conditions for access to the network for cross-border exchanges in electricity and repealing Regulation (EC) No 1228/2003”. It has now been repealed by the 2019 Recast Regulation (below). Recitals (1) – (3) set out the context:
“(1) The internal market in electricity, which has been progressively implemented since 1999, aims to deliver real choice for all consumers in the Community, be they citizens or businesses, new business opportunities and more cross border trade, so as to achieve efficiency gains, competitive prices and higher standards of service, and to contribute to security of supply and sustainability. (2) Directive 2003/54/EC of the European Parliament and of the Council of 26 June 2003 concerning common rules for the internal market in electricity and Regulation (EC) No 1228/2003 of the European Parliament and of the Council of 26 June 2003 on conditions for access to the network for cross-border exchanges in electricity have made significant contributions towards the creation of such an internal market in electricity. (3) However, at present, there are obstacles to the sale of electricity on equal terms, without discrimination or disadvantage in the Community. In particular, non-discriminatory network access and an equally effective level of regulatory supervision do not yet exist in each Member State, and isolated markets persist.”
[27]Recital 29 stated:
“… the Commission should be empowered to establish or adopt the Guidelines necessary for providing the minimum degree of harmonisation required to achieve the aims of this Regulation.”
Article 18(2) made clear that “Guidelines” were “rules” which set out the overarching principles governing the setting of charges. Article 19 imposed a duty upon national regulatory authorities to “ensure compliance” with the Regulation and Guidelines adopted pursuant to Article 18.[28]Article 2(2)(c) is central to Issue II. It defined “congestion” in terms of hindrances to international trade:
“…a situation in which an interconnection linking national transmission networks cannot accommodate all physical flows resulting from international trade requested by market participants, because of a lack of capacity of the interconnectors and/or the national transmission systems concerned.”
[29]The rules on the management and allocation of available transmission capacity of interconnections between national systems were laid down in a lengthy Annex. This is also relevant to Issue II, concerning the meaning of “congestion”. Two paragraphs ([1.2] and [2.1]) are particularly relied upon by GEMA as highlighting the nexus between congestion and international trade:
“1. General Provisions 1.1. Transmission system operators (TSOs) shall endeavour to accept all commercial transactions, including those involving cross-border-trade. 1.2. When there is no congestion, there shall be no restriction of access to the interconnection. Where this is usually the case, there need be no permanent general allocation procedure for access to a cross-border transmission service. … 2. Congestion-management methods 2.1. Congestion-management methods shall be market-based in order to facilitate efficient cross-border trade. For that purpose, capacity shall be allocated only by means of explicit (capacity) or implicit (capacity and energy) auctions. Both methods may coexist on the same interconnection. For intra-day trade continuous trading may be used.”
(emphasis added) (emphasis added)

Directive 2009/72/EC -The 2009 Directive

[30]The 2009 Regulation was adopted on 13th July 2009 and was by its nature directly applicable and did not need measures of transposition into the law of the Member States to be legally effective. On the same day the Parliament and Council also adopted Directive 2009/72/EC of the European Parliament and of the Council “concerning common rules for the internal market in electricity and repealing Directive 2003/54/EC” (“the 2009 Directive”). This supplemented the 2009 Regulation by laying down rules to be implemented into domestic law.[31]Article 2 contains important definitions including that of an “ancillary service”, but it did not address whether the costs associated with congestion management were ancillary:
“17. ‘ancillary service’ means a service necessary for the operation of a transmission or distribution system…”
[32]Article 37(1) listed the duties and powers of national authorities. It supplemented the duty imposed by Article 19 of the 2009 Regulation to ensure compliance with the relevant laws including those in Guidelines. Under Article 36 regulatory authorities were instructed to “take all reasonable measures” to achieve the objectives set out in the article “within the framework of [the] duties and powers” laid down in Article 37. One such duty in Article 37 compelled regulators to fix or approve tariffs, to “ensure” compliance with the law on the part of commercial operators, and to respect “contractual freedom” but only insofar as consistent with the law (see subparagraph (l) below). The provisions of relevance to this appeal are as follows:
“Article 37 Duties and powers of the regulatory authority 1. The regulatory authority shall have the following duties: (a) fixing or approving, in accordance with transparent criteria, transmission or distribution tariffs or their methodologies; (b) ensuring compliance of transmission and distribution system operators and, where relevant, system owners, as well as of any electricity undertakings, with their obligations under this Directive and other relevant Community legislation, including as regards cross-border issues; … (d) complying with, and implementing, any relevant legally binding decisions of the Agency and of the Commission; … (l) respecting contractual freedom with regard to interruptible supply contracts and with regard to long-term contracts provided that they are compatible with Community law and consistent with Community policies; … (q) monitoring the implementation of rules relating to the roles and responsibilities of transmission system operators, distribution system operators, suppliers and customers and other market parties pursuant to Regulation (EC) No 714/2009…”
.

Commission Regulation (EU) No 838/2010 – the Annex B Guidelines

[33]The relevant rules (the Guidelines) are in Commission Regulation (EU) No 838/2010 of 23rd September 2010 “laying down guidelines relating to the inter-transmission system operator compensation mechanism and a common regulatory approach to transmission charging” (“Regulation 838/2010”). The text below is that operative following the end of the post-Brexit implementation period, and it forms part of EU retained law. It is the only version that needs to be considered for the purpose of this appeal. Article 1 concerns the hosting of cross-border electricity flows:
“Transmission system operators shall receive compensation for costs incurred as a result of hosting cross-border flows of electricity on their networks on the basis of the guidelines set out in Part A of the Annex.”
[34]Article 2 concerns access to transmission systems. It is broader than Article 1 which is limited to cross-border payments:
“Charges applied by network operators for access to the transmission system shall be in accordance with guidelines set out in Part B of the Annex.”
[35]Recital 10 addresses Article 2 and proceeds from a recognition that absent a common system for determining aggregate, average charges there might be variations in charges applied at the local level which could distort international trade. Article 2 states that “for this reason” there is a requirement that variations should not “undermine” the internal market and should therefore be within a range (“the Permitted Range”):
“(10) Variations in charges faced by producers of electricity for access to the transmission system should not undermine the internal market. For this reason average charges for access to the network in Member States should be kept within a range which helps to ensure that the benefits of harmonisation are realised.”
[36]Part B of the Annex (“Annex B”) is central to both Issues I and II. It is entitled “Guidelines for A Common Regulatory approach to Transmission Charging” and provides the framework for the Permitted Range: “1. Annual average transmission charges paid by producers in Great Britain shall be within the range set out in point 3. 2. Annual average transmission charges paid by producers is annual total transmission tariff charges paid by producers divided by the total measured energy injected annually by producers to the transmission system of Great Britain. For the calculation set out at Point 3, transmission charges shall exclude:(1) charges paid by producers for physical assets required for connection to the system or the upgrade of the connection;(2) charges paid by producers related to ancillary services;(3) specific system loss charges paid by producers in Ireland, Great Britain and Northern Ireland. 3. Annual average transmission charges paid by producers in Great Britain shall be within a range of 0 to 2.5 euros per megawatt hour.” (1) charges paid by producers for physical assets required for connection to the system or the upgrade of the connection; (3) specific system loss charges paid by producers in Ireland, Great Britain and Northern Ireland.[37]Annex B thus sets a floor and ceiling for average annual transmission charges and identifies categories of charges which must be excluded from the calculation of those charges. In short:(i) The Permitted Range: Annual average transmission charges paid by generators shall be within the range of €0.00 to €2.5 per MWh.(ii) The Connection Exclusion: The charges included in the calculation of the charge shall not include “Charges paid by producers for physical assets required for connection to the system or the upgrade of the connection” (“the Connection Exclusion”); and,(iii) The Ancillary Services Exclusion: The charges included in the calculation shall also exclude “Charges paid by producers related to ancillary services” (“theAncillary Services Exclusion” or “the ASE”).[38]Although, as already observed, there were definitions of “congestion” and “ancillary service” in the 2009 legislation there was no indication in Annex B as to how costs incurred by network operators relating to congestion management were to be allocated or recovered from producers in the transmission charging regime. Whether related costs were to be treated as an ancillary service and therefore to be excluded was a matter left to the discretion of individual national authorities, reflecting the fact that these were measures of minimum harmonisation which, by their nature, left a good deal to be worked out at the national level.

Regulation (EU) 2019/943 - The 2019 Recast Regulation

[39]In 2019 the 2009 Regulation was repealed and replaced by Regulation (EU) 2019/943 on the internal market for electricity with effect from 1st January 2020 (“the 2019 Recast Regulation”). This brought together the many amendments to the 2009 Regulation but also introduced significant updating alterations which reflected the fact that in the intervening decade there had been deep changes in the market including in relation to the degree of integration between the transmission networks of the different states.[40]For the first time the EU legislature addressed whether congestion management should be treated as an ancillary service, and stated that it was not.

(i) Article 2(1) defines ‘interconnector’ as:

“a transmission line which crosses or spans a border between Member States and which connects the national transmission systems of the Member States” (ii) Article 2(4) defines “congestion” 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”; (iii) Article 2(6) defines “structural congestion” 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.” (iv) Article 2(60) provides that: “‘ancillary service’ means ancillary service as defined in point (48) of Article 2 of Directive (EU) 2019/944.”
Article 2(48) of that directive states that ‘ancillary service’ means:
“a service necessary for the operation of a transmission or distribution system, including balancing and non-frequency ancillary services, but not including congestion management”
(emphasis added). (emphasis added).[41]No changes were made to Annex B as a result of the coming into effect of the 2019 Recast Regulation.

The basic facts

[42]I turn to the facts giving rise to the GEMA Decision. A useful summary is set out in the CMA Decision at paragraphs [4.28] – [4.34].[43]In August 2017 GEMA commenced a targeted charging review (“theTCR”) of the Code as it applied to electricity network charging. In February 2018 the CMA took a decision (“the CMA 2018 Decision”) on a Code modification proposal called “CMP 261” in which it concluded that the charging methodology being used and as set out in the Code was inconsistent with Annex B. This was pursuant to an appeal by EDF Energy (Thermal Generation) Limited and SSE Generation Limited against a decision of GEMA dated 16th November 2017 which had rejected an industry proposal to modify the TNUoS charges paid by generators in the charging year 2015/16 ostensibly to bring them within the Permitted Range in Annex B. In that decision GEMA had addressed the correct interpretation of Annex B and had concluded that the charging methodology provided for under the Code was inconsistent with Annex B. The CMA upheld the decision of GEMA that some local charges fell within the Connection Exclusion (see paragraph [37(ii)] above) and were therefore wrongly included in the basic transmission charge. EDF Energy (Thermal Generation) Ltd and SSE v GEMA (26th February 2018)[44]During the course of the TCR consultation GEMA circulated an open letter (dated 4th May 2018) in which it stated that:(i) there was no need for an immediate change to the Code calculation subject to there being, in the interim, no breach of the Permitted Range; and(ii) it would make sense to consider the possibility of any definitive change to the calculation alongside the TCR. On 28th November 2018 GEMA published a “minded to” decision and draft Impact Assessment. This proposed a decision to:
“Set the Transmission Generation Residual to zero, subject to maintaining compliance with [Annex B].... The ESO is developing a modification which would enact the post CMP261 definition of the [Permitted Range], and would allow us to direct that our policy position of no residuals charged to generation is met.”
[45]On 21st November 2019, GEMA published its decision entitled “Targeted charging review: decision and impact”. It acknowledged that its decision would not be compliant with Annex B, in that whilst charges would be within the Permitted Range they would not take full account of the Connection Exclusion: “4.79 For the avoidance of doubt, we consider that the CUSC is compliant with EU Regulation 838/2010 except for the interpretation of the ‘exclusion connection’ which needs to have the correct interpretation, in accordance with the CMA appeal regarding CMP261. We think that generators should face transmission charges for:• off-shore local charges,• on-shore local charges (less those which fall into the ‘Connection Exclusion’), and• wider locational charges. For compliance with the EU Regulation 838/2010 we expect these annual average transmission charges paid by producers … not to exceed €2.50/MWh or fall below €0/MWh. We accept that an ‘adjustment charge’ may be necessary to rectify this."[46]GEMA directed NGESO to formulate proposals to modify the Code to give effect to its conclusion that the charging methodology was non-compliant with the law:
“45. The Proposal(s) must set out proposals to modify the Use of System Charging Methodology, Section 14 CUSC to set the TGR to £0, subject to ensuring ongoing compliance with EU Regulation No 838/2010 (in particular, the requirement that average transmission charges paid by producers in each Member State must be within prescribed ranges – which for Ireland, Great Britain and Northern Ireland is 0 to 2.50 EUR/MWh). This should be achieved by charging generators all applicable charges (having factored in the correct interpretation of the connection exclusion as set out in EU Regulation 838/2010) and adjusted if needed to ensure compliance with the 0 to 2.50 EUR/MWh range. 46. …the Proposal(s) must set out proposals for an appropriate adjustment charge to ensure compliance with the EU Regulation 838/2010, if NGESO considers it necessary…”
[47]NGESO then raised a composite proposal termed “CMP317/327” (“the NGESO Proposal”). There was an industry wide consultation on this which resulted in 83 proposals (or Workgroup Alternative Code Modifications (“WACMs”)) being advanced in accordance with the procedure laid down under the Code, in addition to the NGESO proposal. GEMA decided however that none of these, including the NGESO proposal, correctly applied the Connection Exclusion in Annex B. However, the NGESO Proposal came closest and amounted to a material improvement on the status quo ante. The difference between the NGESO Proposal and a legally compliant proposal would be c.£1.7m in the 2021-2022 charging year. In contrast, the difference between the pre-existing position and a legally compliant proposal would be c.£423m with the financial burden being shifted from generators to suppliers.[48]The CMA Decision summarised the position as follows: “4.70 Under the status quo, no Local Charges are considered to fall within the Connection Exclusion. Therefore, all locational TNUoS charges are included within the ex ante CUSC Calculation of the downward adjustment required to achieve compliance with the upper limit of the Permitted Range. As we set out in Table 4.3, this would result in a much larger downward adjustment to TNUoS charges to achieve compliance with the upper limit of the Permitted Range (-£451 million) than would be the case under the Original Proposal (-£2 million) or WACM7 (-£5 million). Also, as a consequence of this the amount of TNUoS charges that are forecast to be paid by Generators in 2021/22 under the Original Proposal and WACM7 was much higher (£814 million and £811m) than under the status quo (£365 million). 4.71 The calculation of the adjustment to TNUoS charges required to achieve compliance with the upper limit of the Permitted Range is separate to the calculation of likely compliance with the Permitted Range. At this stage of calculating likely compliance with the Permitted Range, GEMA applied its view of the correct definition of the Connection Exclusion (all but £1.7 million of Local Charges are considered by GEMA to fall under the Connection Exclusion, which is different from that set out in the status quo, Original Proposal and WACM7. 4.72 We note that because under the status quo no Local Charges are considered to fall within the Connection Exclusion, when calculating the downward adjustment to achieve compliance with the Permitted Range, this results in a large downward adjustment to TNUoS charges which leads to a seemingly contradictory outcome. Under the status quo, in 2021/22 the downward adjustment to Generator TNUoS charges aimed at bringing them below the upper limit of the Permitted Range would in practice result in a likely breach of the lower limit. In December 2020, GEMA concluded in its Decision that there was a serious and imminent risk of a breach under the status quo. 4.73 On page 20 of its Decision, GEMA set out some figures that underpinned its view that ‘Under the status quo, there is a serious and imminent risk of a breach of the lower limit of the Permitted Range.’ In particular, it explained that a large amount of offshore Local Charges that should be deducted from the transmission charges for the purposes of calculating compliance with the Permitted Range, under the status quo would not be deducted. This meant that the total level of charges would be £423 million higher than it should have been. Consequently, the TGR reduction would reduce transmission charges by €2.09/MWh more than it should. This led GEMA to conclude that there was a serious risk that charges would fall below the lower limit of the Permitted Range.” 49. The issue – the competing contentions[49]I turn now to the ground of appeal advanced by GEMA, supported by the CMA. The issue is narrow. The dispute is over the method chosen by GEMA to bring the charging methodology into compliance. In this respect it is common ground that the GEMA Decision incorporated a stop gap methodology based upon the temporary adherence to the NGESO proposal which was based upon an erroneous interpretation of Annex B.[50]The judge concluded that it was, in principle, unlawful to embed a temporary, stop-gap, non-compliant methodology into a modification decision. In his judgment refusing permission to appeal the Judge pithily observed:
“What the Code requires … is compliance with the legal standard. Something that amounts to a nearer miss than the pre-existing state of affairs is not good enough” ([2022] EWHC 987 at paragraph [16]). In the judgment under appeal at paragraph [45] the Judge said: “ … I accept the factual premises of the proposition. The modification proposal that GEMA approved made things better insofar as the amendment made to the CUSC prevented transmission charges outside the specified range. It was, in that regard, an improvement on the unamended provisions of paragraph 14.14.5 of the CUSC. I also accept that the criterion at paragraph 8.23.7 of the CUSC, for approval of a modification, is whether GEMA is satisfied that what is proposed "would better facilitate achievement of the Applicable CUSC Objectives". In many circumstances that criterion will bring with it a practical margin of judgement for GEMA as expert regulator. This may be particularly significant when assessment of a modification proposal requires a balance to be struck between overlapping or competing practical considerations. But that was not this case. The relevant objective was compliance with a legal standard and in that context a miss is as good as a mile. In the circumstances of this case, it is immaterial that the amendment to the CUSC improved the situation, making for a situation that better complied with Regulation 838/2010. Actual compliance with Regulation 838/2010 was what was required to permit GEMA to approve this aspect of modification proposal CMP 317/327.”
[51]GEMA argues that the judge erred. In its formal grounds of appeal GEMA argued that the duty under Annex B was as to the “result” ie that any charge should be within the Permitted Range and the NGESO proposal, which it had endorsed, did just that. The CMA puts the argument in a slightly more nuanced way. It contends that the judge failed properly to apply the Code under which GEMA was bound to accept any proposal which “better facilitated achievement” of the applicable objectives which include compliance with the law. As to this GEMA did form just such a conclusion which indeed was one which, in substance at least, the Judge agreed with because he accepted that the NGESO proposal was nearer to full compliance than any other proposal and was an improvement on the status quo ante. To this extent the GEMA decision did “better facilitate achievement” of the law, albeit that there remained some way to go before full compliance could be attained. The judge failed to understand the progressive and relative nature of the legal test. In their skeleton argument the CMA set out an important rider:
“In circumstances where this was an explicitly short term solution, with NGESO expected to bring forward a further CUSC modification proposal to align the correct methodology with GEMA’s own (in the CMA’s view correct) interpretation of the Connection Exclusion, the CMA’s view was that GEMA had not fallen into legal error in approving CMP 317/327”
[52]SSE argues that the Judge was right in emphasising that:(i) the rules in Annex B emanate from an EU regulation which is directly applicable and bound GEMA;(ii) the obligation in Annex B was clear and unequivocal and required charges relating to certain costs to be excluded from the relevant calculation of transmission charges;(iii) GEMA acknowledged that the charging methodology was based upon an incorrect interpretation of Annex B;(iv) GEMA nonetheless embedded within a formal decision a methodology that it knew was inconsistent with Annex B; and (v), it followed that the decision was unlawful. As to GEMA’s argument that its decision was better than alternatives:
“GEMA is wrong to suggest that some constructions are more erroneous than others. A construction of a legal provision is either right or wrong…”
. The effect of the GEMA Decision was that “…an unlawful legal definition” was “… hardwired into a legally binding Code document.”[53]In written submissions SSE argued that GEMA’s approach subverted the rule of law:
“GEMA is wrong … to suggest that if any of the proposed amendments was assessed to be better than the status quo, it could lawfully be chosen as a “stop gap” measure. That would suggest that an unlawful construction of legislation could be adopted and applied if it represented the “least bad” option. That is not a lawfully relevant consideration. The Court does not permit a public body to maintain an error of law for the sake of expediency, until it can be redressed in a more lasting fashion: R v. Paddington Valuation Officer ex p Peachey [1966] 1 QB 380, CA, per Danckwerts LJ at p. 418 and Salmon LJ at p. 419. To do so would subvert the rule of law.”

The GEMA Decision

[54]Central to determining this appeal is an understanding of the structure of the GEMA Decision. There are six key points to make.[55]First, the CMA 2018 Decision found that under the Code certain costs were treated as falling outside the Connection Exclusion and were therefore to be included in the CUSC Calculation under Annex B. This however was in error since the costs fell within the Connection Exclusion and should therefore have been excluded from the CUSC Calculation. There was (and is) no challenge to this reasoning. In Annex 1 to the GEMA Decision, GEMA explained why it agreed with the CMA analysis. The starting point for GEMA was hence that there was an error of law to be corrected.[56]Secondly, GEMA sought to determine whether the error had led NGESO to impose incorrect charges in the past, or whether there was a real risk that unlawful charges might be levied in the future. So far as the past was concerned, NGESO sought to reassure GEMA that there had been no past over- or under-charging. GEMA was not wholly convinced:
“Historical compliance with the Limiting Regulation In addition to bringing forward future modification(s) as set out above, NGESO should evaluate whether there has been non-compliance with the Limiting Regulation in 2020-21 and/or previous charging years as a result of the i) the inclusion in the CUSC Calculation of Distributed Generator volumes and charges; and/or ii) the fact that the CUSC Calculation is based on an erroneous interpretation of the Connection Exclusion. We understand that NGESO’s initial assessments indicate that it is unlikely that there has been any non-compliance in respect of previous charging years. The position in respect of this current charging year is less clear and requires further investigation. If and insofar as any non-compliance is identified, we expect NGESO to bring forward additional modifications to address such issues. We will work together with NGESO to understand the compliance position in respect of previous years.”
[57]As for the future, GEMA examined all the modification proposals advanced to it and found that none was compliant. It conducted an analysis to decide which - relatively speaking - was he closest to the correct interpretation of Annex B and found that this was the NGESO proposal. It undertook a counterfactual analysis of how charging in the future would operate comparing the status quo ante against theNGESO proposal. If it preserved the status quo then, as Ms Smith KC for GEMA put it, the risk was that charges came “crashing through the floor”. There was a “real and imminent risk” of breach of the lower limit of the Permitted Range:
“Under the status quo, there is a serious and imminent risk of a breach of the lower limit of the Permitted Range. We understand from the ESO that the estimated value of offshore Local Charges in Charging Year 2021-2022 is c. £423m and estimated generator output is 223 TWh (or 223 million MWh). We expect that the vast majority, if not all, of these charges would fall within the Connection Exclusion, but the current CUSC Calculation assumes that they do not, and therefore takes them into account. The effect of including these charges in the CUSC Calculation is to increase calculated average charges by £1.90/MWh above what would otherwise be the case (and thus to produce a matching increase in the negative adjustment made by the TGR). On the basis of currently-forecast GBP/EUR exchange rates of around 1.1, this is equivalent to €2.09/MWh. Taking into account a c.20% ex ante error margin, there is therefore a serious risk under the status quo that annual average transmission charges (calculated in accordance with the Limiting Regulation) will fall below the lower limit of the Permitted Range. This would constitute non-compliance with the Limiting Regulation. Since the Baseline presents an immediate risk of non-compliance with the Limiting Regulation, we consider that a proposal which would secure compliance (or reduce the risk of such non-compliance) would represent an improvement on the status quo. Approval of such an option would be preferable to allowing the status quo to remain (whether by rejecting all proposals or exercising our send back powers).”
[58]Thirdly, GEMA decided that in moving towards a solution it had to act in stages. It had no power under the Code simply to impose its own solution forthwith but had to go down one of the mandated modification routes. Even with all due haste and expedition this would take time. There would therefore have to be a “stop gap” as part of the glidepath towards compliance.[59]Fourthly, GEMA acknowledged that in consequence it faced the dilemma of choosing between the lesser of two evils because it would have to include in its decision a component which was based on an incorrect interpretation of Annex B:
“The Authority therefore needs to choose between the imperfect status quo and a series of imperfect alternatives. It is open to the Authority to approve a modification proposal which is based on an incorrect interpretation of the Connection Exclusion, if that proposal is better than the (imperfect) Baseline and the other (imperfect) proposals at facilitating the achievement of the ACOs.”
The NGESO proposal was the nearest of all submitted proposals to full compliance. That was accordingly the best solution to go with, for the time being.[60]Fifthly, GEMA confirmed that its approval of the NGESO proposal was conditional upon a new proposal being put in place as soon as possible to bring about full compliance. Using the NGESO proposal as the stop gap would “better facilitate” achievement of the ACOs, which included compliance with the law.[61]Sixthly, to obviate the risk of embedding, systemically, into the charging methodology a non-compliant component, the new proposal would have to ensure that any past over or under-charging was corrected, ex post facto.[62]GEMA’s conclusion was in the following terms:
“Our decision We have considered the issues raised by the modification proposals and the FMR dated 13 August 2020, including taking into account the responses to the Workgroup Consultation and Code Administrator Consultation. We have also considered and taken into account the votes of the Workgroup and the CUSC Panel on CMP317/327. We do not consider that any of the proposals incorporate the correct interpretation of the Connection Exclusion. Notwithstanding this, we have concluded that the Original Proposal would be likely to avoid the imminent risk of a breach of the Limiting Regulation that is posed by the status quo, and better facilitate achievement of the ACOs than either the status quo or any of the WACMs. We also consider that approval of the Original Proposal would be consistent with our principal objective and statutory duties. Accordingly, our decision is to approve the Original Proposal and direct that it be implemented. Our approval of the Original Proposal is on the express basis that it is a ‘stop-gap’ measure which should avert an imminent risk of breach of the Limiting Regulation, and allow time for the formulation of a longer-term solution that properly reflects the correct interpretation of the Connection Exclusion. We expect NGESO to bring forward a further CUSC Modification Proposal that will fully give effect to the correct interpretation of the Connection Exclusion. We also expect NGESO to bring forward a CUSC Modification Proposal to remove from the CUSC Calculation the TNUoS Charges payable by Large Distributed Generators and their associated volumes (MWh). The need for this further modification is explained in more detail in the ‘Future Modifications’ section below.”

The correct question to be asked

[63]For the reasons I set out below I have reached the conclusion that the judge erred in ruling that this staged approach to compliance was unlawful. This issue does not concern the correct interpretation of the rules in Annex B. It is accepted that the basic obligation upon GEMA is to ensure that charges are within the Permitted Range and that in performing the relevant calculation there are certain required charging inputs and that certain charges relating to specified costs are required to be excluded. It is also common ground that the charges are annualised and aggregated. The limiting regime is not about setting the charges to be paid by individual market participants. An individual undertaking might therefore be charged a rate outside the Permitted Range provided that the aggregated annualised charge was within the range. It was also common ground that insofar as at the end of a charging year (April to April) the aggregate charge was not in accordance with the rules then a process of ex post facto adjustment would be applied to bring it back into line. This could entail debits or credits being applied to the accounts of market participants.[64]At base this appeal turns upon the duty of regulators, once non-compliance has come to light, to ensure observance with the law. There is a statutory duty on GEMA to both comply with the law and ensure compliance by its regulated community, which duty has not been changed by the exit of the UK from the EU. The existence of a duty does not however preclude the decision maker also having a discretion or power as to “how” to go about ensuring compliance; the two are not mutually exclusive. This flows from the proper interpretation of the legislative regime as a whole. For example, Article 37 of the 2009 Directive refers to the taking of “reasonable measures” in the framework of the duty to ensure observance of the law (see paragraph [32] above) indicating that there might be a range of different “reasonable” ways in which compliance can be secured. The conclusion of the Judge however was that the GEMA Decision was unlawful because during the glidepath to adherence - the stop gap - an unlawful methodology would temporarily subsist and be incorporated into the Code. It was implicit in the Judge’s reasoning that there was no discretion or power for GEMA to do anything more than demand immediate or instantaneous observance, even if this was impossible to achieve in any realistic and practical sense and left the state of observance with the law in a worse situation. To prohibit the interim stage upon the basis that it reflects a degree of temporary (diminishing) non-observance begs the question of what, if the Judge is correct, regulators are meant to do in a case such as the present in order to meet their statutory duty.[65]In my judgment under the relevant legislation GEMA had a power as to how it went about performing its duty to secure compliance with the law. A decision whether GEMA acted unlawfully in the exercise of this power is fact and context specific. Under EU law the test to be applied would be proportionality. It is unnecessary in this case to devote time to determining whether proportionality remains the right test or whether the test is simply one of domestic law rationality. In either case a relatively broad margin of judgment or discretion will be implied into the test and in my view both lead to the same end result.[66]On the present facts, the relevant questions to be posed include:(i) whether there was a proper basis for GEMA to conclude that there was non-compliance which needed to be cured;(ii) whether compliance could be achieved forthwith or, alternatively, had to be staged over time involving the use of a stop gap;(iii) whether the solution adopted was the best available including how it compared with other proposed solutions and what it meant in practical terms for the risk of future non-observance;(iv) whether a non-compliant stop-gap was avoidable: and(v) whether the solution finally adopted was reasonably certain and practicable. I consider each of these questions below by reference to the analysis in the GEMA Decision set out above.[67]Was there a proper basis for taking steps to impose a remedy? First, GEMA formed the view that there was a real problem to be solved. There was some debate as to whether there was a history of breach prior to the GEMA Decision. The evidence before us is not conclusive. Mr Beal KC, in his submissions for SSE, said that there were “concerns” amongst generators as to a lack of transparency on the part of NGESO in its performance of the calculations and as to the possibility of wrongful charges being levied. NGESO informed GEMA that there had been no past errors in calculation. The GEMA Decision nonetheless left open the possibility that there had been historic errors of calculation (see paragraph [56] above). The basis for GEMA’s decision was not, however, that there had been proven past breach. It was that there was an imminent and real risk of future breach. The basis for this was(i) GEMA’s agreement with the analysis in the CMA 2018 Decision that the methodology for calculating the charge was inconsistent with Annex B and(ii) its counterfactual analysis of the impact on adherence to the Permitted Range of retaining the status quo compared with adopting the NGESO proposal. If the status quo was retained then there was an imminent risk that annual average charges would fall below the floor of the Permitted Range. In my judgment GEMA was entitled to conclude that there was a serious problem to be resolved.[68]Could compliance have been achieved forthwith / the need for a “stop gap”? Secondly, GEMA was entitled to conclude that the cure could not be imposed forthwith. Whichever procedural route GEMA decided upon required some degree of consultation and evaluation which would take time. It was impossible to impose an overnight solution. That being so it inevitably followed that there would have to be an interim stage pending full compliance and that this would entail a “stop gap”.[69]Was the solution adopted the best available stop gap? Thirdly, the NGESO proposal was the best that could be achieved in the circumstances. There is no quibble on the part of any party with GEMA’s maths and, indeed, the Judge agreed that the NGESO proposal was an marked improvement upon the subsisting position.[70]Whether a non-compliant stop gap was avoidable? Fourthly, for the stop gap GEMA was compelled to choose a non-compliant solution. There are a number of points which arise here and the context is important. GEMA faced the present conundrum because, when it had earlier consulted and sought proposed solutions, it had been handed nothing but non-compliant proposals, including from SSE. Yet, SSE still argues that it should have rejected the proposal that was nearest to being compliant and adopted some other, worse, proposal. SSE seeks to hoist GEMA on a petard the making of which SSE contributed to. In my judgment on the facts GEMA was entitled to form the view that it was impossible to avoid a situation where, as a stop gap, there was a risk of non-compliance, brought about by the failure of the entire regulated community to come up with a modification proposal which was adherent to the relevant charging principles which could have been adopted earlier.[71]I deliberately use the expression “risk” of non-compliance. There was a dispute as to whether or not the GEMA Decision would or would not lead to a future unlawful state of affairs. GEMA argued that it would not. The duty upon it was one of result only. On its calculation, adoption of the NGESO proposal would lead to charges remaining within the Permitted Range for 2-3 years which, it argued, allowed ample time for a new solution to be introduced. But if, in the interim, the calculation of charges did, because of the incorrect treatment of exclusions, lead to an erroneous calculation this would be ironed out by a process of ex post facto reconciliation. That reconciliation process would cover any period of non-observance, including therefore periods prior to the last charging year (April to April). Accordingly, when looked at from the correct vantage point (after the reconciliation process) there would be no adverse “result” and no breach. Mr Beal KC for SSE said, no matter, there would still be a breach because, on GEMA’s own admission, there would be a non-adherent methodology being “hard baked” into the CUSC, albeit pending the new solution. In my view Annex B spells out how the calculation must be performed and this entailed not only respecting the Permitted Range but also applying the Connection Exclusion and ASE (see paragraph [37] above). I am inclined to accept the argument of SSE. GEMA’s solution involves the unattractive proposition that the performance of a perpetually incorrect calculation does not matter if it can be periodically cured. I therefore take as a working hypothesis that to some degree there was embedded into the charging system a methodology which could in due course give rise to an erroneous charge being calculated and levied. However, even if, technically, I do not accept GEMA’s analysis it nonetheless highlights that, because of the periodic cure, charges would be likely to remain within the Permitted Range and any breach could well be of very limited practical effect, which is a conclusion that is relevant to the rationality or proportionality of GEMA’s approach.[72]When pushed by the Court as to precisely what it was being suggested GEMA could, alternatively, have done, SSE advanced three solutions all of which entailed a higher degree of non-compliance with Annex B than GEMA’s solution. None involved GEMA being able to secure compliance forthwith. The only proposals advanced to GEMA were unlawful (see paragraph [47] above) but these still represented the source material from which GEMA was compelled to select in order to identify a suitable stop gap. If GEMA had wished to use some other solution it considered that it was required to go through the elongated and cumbersome modification procedures mandated in the CUSC. It was common ground that GEMA had to find a solution through use of the CUSC modification regime and this was indeed the view expressed by the Judge (judgment paragraph [43]). It follows that, in all practical senses, GEMA had no option but to choose a non-compliant solution as the stop gap.[73]Counsel for the CMA suggested that the obligation upon GEMA was simply to choose an option that was better than the status quo. Provided that it did this then by definition it would be meeting the regulatory objective imposed by the Code of “better facilitating” compliance with the law and, as such, it would be complying with its legal obligations qua regulator. There is some support for this analysis in paragraph [43] of the judgment. With respect I do not agree. Nothing in the Code – which is a private law contract - can override or permit a departure from an overarching obligation to secure compliance with the law as set out in Annex B. This was fully recognised in the EU regime. Thus, Article 37(1)(l)of the 2009 Directive (see paragraph [32] above) made it clear that whilst regulators had to respect private law contracts their terms could not be permitted to justify a departure from overarching legal obligations. This reflects a fundamental principle of public law that a decision maker cannot prevent itself from complying with its statutory duties by entering into, or approving, an inconsistent contract. A contract, such as the CUSC, cannot serve to dilute the obligation to “ensure” compliance with Annex B by a contractual obligation to adopt a proposal which is no more than an improvement on that which went before. The duty is not to choose the best of a pack of bad options; it is to ensure adherence to Annex B.[74]In fact I did not understand GEMA to be arguing that its duty lay only in finding a solution that was better than the status quo. The logic of the GEMA Decision assumes that its duty was to ensure ultimate compliance and the “better than the status quo” test was applied only to enable GEMA to determine the “stop gap” which was itself no more than a step on the route to that ultimate compliance. The approval of the NGESO proposal as “better facilitating” observance of the law was, in context, limited to the stop gap component of the process which was a proper interim course to take, but would not have been proper for the end result.[75]Was the solution adopted reasonably clear and certain? Fifthly, Mr Beal KC for SSE complained that the road map for the future set out in the GEMA Decision was vague, unspecific and did not set out any firm deadlines. It created no certainty of a lawful outcome and all that it did was to “expect” NGESO to come forward with a new modified proposal. I disagree. The nuts and bolts of the solution were very much matters falling within the discretion of GEMA. It is in any event not true to say that the decision did not set out its objective (namely full compliance with the CMA 2018 Decision) or that that NGESO had no indication of the timeframe. It is true that GEMA only “expected” NGESO to come forward with a solution but in pith and substance NGESO had no real option but to cooperate. In my judgment the criticisms made are unwarranted.

GEMA’s powers

[76]There is a final matter of significance concerning the powers available to GEMA to remedy defects. In this case the problem flows from the fact that when GEMA sought proposals to bring the charging methodology into line it received 84 non-compliant solutions. These were submitted even though everyone knew what the problem was, it having been set out and analysed in the GEMA decision of 2017 which led to the CMA 2018 Decision (see paragraph [43] above). As of the date of this appeal over 5 years have elapsed since the writing was first clearly inscribed on the wall. Moreover, at the outset of the appeal we were informed that the matter had still not been resolved and disagreements remained between industry and GEMA.[77]GEMA has throughout considered itself bound to seek remedies using the powers of modification under the Code. The Judge agreed. His view was that there was “…no suggestion that there was any other route to compliance with Regulation 838/2010 other than though the Code” (judgment paragraph [43]). But, as this case shows, returning the resolution of the problem into the hands of the regulated community under the Code can serve to generate dissension, delay and multi-layered litigation. This appeal is the third level of review and appeal. It seems obvious that there is a substantialrisk that the CUSC modification system permits market participants to game the system. Even the one Code mechanism which allows GEMA to initiate a modification process (Paragraph 8.17A(1)(a) – see paragraph [24] above) still involves a process governed by the Code. There is no Code power for the Authority simply to take (subject to a fair procedure) the relevant decision and then impose the required result.[78]The facts of this case suggest that the Code is capable of hindering the discharge by GEMA of its statutory duty. The Code cannot take precedence over the duty of GEMA to ensure timely compliance with the law. Insofar as there is a conflict between the Code and the statutory duties of GEMA, the latter prevail. With respect to the Judge I do not therefore agree that GEMA is bound by the Code when it comes to taking the steps necessary to ensure compliance with the law. If the Code is an impediment to proper enforcement it is either inoperative and GEMA should deploy other powers at its disposal to give effect to its duty, or it needs to amend the Code, or both.

Conclusion on Issue I

[79]The issue in this case concerns the exercise of a power conferred upon GEMA as to the methodology it selected to bring a non-observant charging methodology into a state of observance. The decision is susceptible to public law challenge. GEMA enjoys a relatively broad margin of discretion and judgment in identifying a solution. On the facts of this case the GEMA Decision was lawful, even though it contained as part of the transition to observance the temporary adoption of a methodology which might have created a transient risk of non-observance. I would therefore allow the appeal of GEMA. Issue II: The allocation of costs attributable to “congestion management”

The issue

[80]I turn now to the cross-appealadvanced by SSE. The context is immensely convoluted. The underlying issue can however be shortly expressed: does the expression “congestion” in the legislative regime apply only to congestion on an interconnector between different states, or to something wider? Both GEMA and the CMA take the view that it applies only to the interconnector and international trade and the Judge agreed. SSE takes the view that this is too narrow an interpretation of the law which, properly construed, applies to congestion on any part of the overall transmission system.[81]The significance of the debate lies in how the costs associated with congestion management are to be attributed and whether, or not, they must be included as part of the relevant transmission charge for the purpose of calculating the Permitted Range under Annex B. The underlying commercial rationale is that if charges attributable to the costs of congestion management are excluded from the calculation of transmission charges then they are not subject to the limitations inherent by having to adhere to the Permitted Range in Annex B (see paragraphs [35] – [37] above). But, conversely, if such charges must be included within the calculation they will be subject to the limits imposed by Annex B, and this may be of commercial benefit to generators. In paragraph [8.1] of its Decision the CMA thus said:
“In this section we address Ground 3 of the Appellants’ NoA, namely that GEMA erred in law in its construction of the Ancillary Services Exclusion (ASE) and its corresponding treatment of (i) the relevant BSUoS charges; and (ii) the relevant BSC charges. It was wrong, the Appellants said, to treat those charges as falling within the ASE and accordingly not limited by the Permitted Range for the average annual transmission charges Generators pay.”
(emphasis added) (emphasis added)[82]It was common ground before GEMA, the CMA and the Judge that the relevant law was that found in the EU law prior to the end of the post-Brexit implementation period: this included Articles 2(4) and 2(60) of the 2019 Recast Regulation. It was further assumed that the EU retained law did not affect any change to the substance of the law. Indeed, the regulators have argued that the EU retained version could be used as a form of after the event guide to the construction of Article 2(4) (eg paragraph [8.60] of the CMA Decision (below)). These are conclusions I have real difficulty with and which I address below.[83]The Judge endorsed the analysis of the CMA. In its decision the CMA held that charges based upon costs attributable to congestion management generally were excluded because they fell within the Ancillary Service Exclusion. It concluded that congestion management referred specifically to congestion on interconnectors. . The CMA held as follows: “Congestion management 8.51 If, as we conclude, the services to which both the BSUoS and BSC charges relate are services necessary for the operation of the system, they will only fall outside the scope of the ASE if they relate to congestion management. This consideration arises only if(a) the relevant definition of an ancillary service is that contained in the Recast Electricity Directive (and the Electricity Regulation) or(b) the definition in the Electricity Directive (and the Electricity Regulation) applies but is taken to exclude from its scope services relating to congestion management within a network area. 8.52 There is no contention that the BSC charges relate to congestion management, however broadly construed. Accordingly, on the basis of the analysis set out above they will fall within the scope of the ASE, and GEMA did not err in concluding that they do. 8.53 In so far as the BSUoS charges are concerned, the starting-point is the definition of ‘congestion’ in Article 2(4) of the Recast Electricity Regulation: “congestion” means 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; 8.54 Although ‘congestion management’ is not defined, it cannot sensibly be interpreted as being anything other than the management of ‘congestion’ within the terms of that definition in Article 2(4). A relevant service will therefore be one which is directed to managing congestion arising from a situation involving trade between ‘network areas’. 8.55 As GEMA pointed out, the reference to network areas reflects changes in the European electricity market between 2009 and 2019 during which some network areas comprised more than one Member State and could exchange electricity without the need for capacity allocation. 8.56 It thus seems to us that the definition of congestion is an updating of essentially the same concept as contained within the definition of congestion in Article 2(2)(c) of the Electricity Regulation (from 2009): “congestion” means a situation in which an interconnection linking national transmission networks cannot accommodate all physical flows resulting from international trade requested by market participants, because of a lack of capacity of the interconnectors and/or the national transmission systems concerned; 8.57 The words ‘because of’ are important. The definition draws a distinction between cause (i.e. lack of capacity of interconnectors and/or the national transmission systems) and effect (i.e. the inability of the interconnector to accommodate all physical flows). It is the effect which constitutes ‘congestion’ not the cause. Therefore, even applying the definition in the Recast Electricity Regulation, congestion is limited to congestion on interconnectors. 8.58 It therefore appears that the EU legislation has taken a consistent approach that congestion applies to a cross-border situation which is managed by a process of capacity allocation. It does not refer to purely internal constraints within a network area, even if those internal constraints may have the consequence that there is congestion and a need for capacity allocation at the boundary between, formerly, two interconnectors and, since 2019, two network areas. 8.59 We also note that the definition of ancillary services in the Recast Electricity Directive and the Recast Electricity Regulation distinguishes between (i) services for congestion management (charges for which are expressly excluded from the ASE); and (ii) balancing and non-frequency allocation services (charges for which are expressly included). The latter, which are defined via the Recast Electricity Directive and the Recast Electricity Regulation, relate to the management of internal constraints within a network area. Such internal constraints and those relating to cross-border congestion are managed differently in practice: internal constraints by way of separate services and charges (in GB, by the balancing services NGESO provides and the settlement services administered by Elexon) and cross border congestion by the capacity allocation process described in paragraph 8.23 above. 8.60 The interpretation in paragraph 8.58 is also consistent with the amendments made to the Recast Electricity Regulation by the Amendment Regulations to deal with the consequences of the UK’s withdrawal from the EU. The reinstatement of language relating to interconnection reflects the reality that the UK is no longer part of an EU single market. In so far as that amendment did not come into force until after the Decision, the relevant legislation was enacted well before the Decision was taken. In any event, it did not have any practical impact because it does not change the inherent meaning of congestion but merely reflects the changed situation of the UK being outside the EU. 8.61 We therefore consider that the concept of congestion, and hence congestion management and related charges, is, in the definition of ancillary services, concerned with the issue of capacity allocation across interconnectors, not with congestion management internal to a single network area. We also agree that this is made clear by the definition of ancillary services in the Recast Electricity Directive and the Recast Electricity Regulation (in specifying that balancing and non-frequency services are within the definition, but not congestion management). 8.62 On this basis, we do not consider that the relevant BSUoS charges relate to congestion management. They therefore fall within the scope of the ASE, as GEMA concluded.” “congestion” means 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; “congestion” means a situation in which an interconnection linking national transmission networks cannot accommodate all physical flows resulting from international trade requested by market participants, because of a lack of capacity of the interconnectors and/or the national transmission systems concerned;[84]The Judge agreed with this analysis. He said:
“71. … I consider the reasoning in the CMA's decision is correct. As with most arguments of statutory instruction there are points that can be made for both sides of the argument. SSEGL places reliance on several linguistic points for example, that in some other provisions in the 2019 Regulation specific reference is made to congestion arising from use of interconnectors (see article 19), and that in other provisions congestion must include reference to congestion within a transmission network. SSEGL also submits that it is not possible to divorce congestion that affects the use of interconnectors from congestion within a network. However, I consider there are two matters within the CMA reasoning that clearly tip the balance in favour of the conclusion it reached. The first is that the notion of congestion charging is defined as congestion arising from requests to trade between network areas. I accept the point made at paragraph 8.56 of the CMA's decision that this is synonymous with a request to trade across an interconnector. The second is the point at paragraph 8.59 of the CMA's decision, the distinction drawn between congestion management services and balancing and non-frequency allocation services. I am satisfied that the CMA's conclusion was correct, and for this reason, the second part of SSEGL's Ground 3 also fails.”
The legislative framework: What is “congestion management” and which part of the network does it apply to?[85]The EU measure in force when the GEMA Decision was taken was Article 2(4) of the 2019 Recast Regulation. All parties invited the Court to adopt a purposive or teleological interpretation of Article 2(4) of that Recast Regulation, though the governmental agencies (GEMA and the CMA) arrived at a different end result from SSE. When adopting a purposive approach to interpretation, courts will consider a range of tools including:(i) an analysis of the natural meaning of the particular statutory language used;(ii) an analysis of legislative drafting techniques including how the language has changed over time and the inferences which are to be drawn from such changes;(iii) an analysis of the substantive provisions as a whole which will enable the language of the disputed measure to be placed into context and which can also (especially in an EU context) indicate whether a measure is intended to be a measure of full or only partial harmonisation; and(iv) an analysis of the recitals as instrumental in identifying the legislative history, relevant travaux préparatoires and the purposes said to have influenced adoption of various provisions of the measure in question, and in due turn, any implementing measure. An illustration of the use of these sorts of techniques in an EU context is Kocur v Angard Staffing Solutions Limited and Royal Mail Group Limited [2022] EWCA Civ 189. These techniques are also common to the interpretation of domestic law as the citations below demonstrate.[86]I start with the natural language used in the context of the changes to the relevant language over time.

(c) of the 2009 Regulation defined “congestion” as:

“…a situation in which an interconnection linking national transmission networks cannot accommodate all physical flows resulting from international trade requested by market participants, because of a lack of capacity of the interconnectors and/or the national transmission systems concerned.”
[87]It is evident from this that “congestion” was a problem relating to trade across international interconnectors and concerned impediments to international trade between the transmission systems of different states. The article differentiated between the reasons for congestion (causes) and the congestion itself (the effect). The effect was that “… an interconnection linkingnational transmission networks cannot accommodate all physical flows resulting from international trade requested by market participants”. The causes were two-fold:(i) lack of capacity of the interconnectors and/or(ii) lack of capacity elsewhere on the national transmission systems concerned. The “and/or” in the definition makes it clear that a cause of congestion could arise from the system itself.[88]As already observed, in Annex B as originally formulated in 2009 an exclusion to the charges to be taken into account when setting transmission charges was created for “charges paid by producers related to ancillary services”. There was however no definition of “ancillary services”. The law was thus silent as to whether charges related to the costs of congestion management were to be regarded as charges for ancillary services (and to be excluded) or core transmission charges (and to be included). How such congestion management costs were allocated between network operators and generators was for the different national regulatory authorities.[89]All of that changed on 1st January 2020 when the 2019 Recast Regulation formally repealed the 2009 Regulation. By a cross-reference to Article 2(48) of Directive (EU) 2019/944, Article 2(60) expressly provided that ‘ancillary service’ means:
“…a service necessary for the operation of a transmission or distribution system, including balancing and non-frequency ancillary services, but not including congestion management.”
There is no reference in this definition to congestion management services being limited to the management of congestion on interconnectors or to international trade.[90]Read naturally in conjunction with Annex B, charges for costs attributable to balancing and non-frequency ancillary services were ancillary and were therefore excluded from the calculation of annual average transmission charges for the purpose of compliance with the Permitted Range; but charges for costs attributable to “congestion management” were not excluded - ie they were to be included in the calculation for the purposes of the Permitted Range.[91]The definition of “congestion” was also altered in Article 2(4) of the 2019 Recast Regulation. Under the new definition:
“congestion” means 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;”
[92]For the purpose of interpretation, importance attaches to the way in which the definition of “congestion” changed over time (see the case law at paragraphs [112] and [113] below). The new definition removed all references to:(a) interconnectors (newly defined in Article 2(1) of the 2019 Recast Regulation in terms of international infrastructure crossing or spanning a border between Member States);(b) links (as in linking of national networks);(c) international trade; and(d) national transmission systems.[93]The contrast between the use of very specific language in the 2009 Regulation and the much broader and looser language in the 2019 Recast Regulation is stark. In my view these changes were deliberate. They involved a policy shift away from congestion being exclusively an issue connected to problems at national borders and affecting international trade. In the Recast Regulation the expression, “trade betweennetwork areas”, which has replaced the earlier reference to “international trade”, is nowhere defined. It seems clear however that a “network area” is not limited to a political area, ie a state. There are, as SSE points out in its evidence, some unitary “networkareas” in the EU which embrace more than one state but there are also different intra-state “areas” for example, as between England and Scotland, where there are limited capacity linkages. In similar vein the expression “network elements” is also undefined. Again given a natural language meaning it refers to all of the “elements” of a network of which one (but only one) would be an interconnector.[94]If the legislator had wished to limit “congestion” to interconnections between states then it could easily have done so. It could for example have retained the language of the old Article 2(2)(c); but it did not do so. Equally, it could have inserted into Article 2(4) terms which it used elsewhere in that same 2019 regulation which did incorporate language limiting their operation to international trade such as “interconnector” in Article 2(1) (which refers to linkages which cross or span a border between Member States) or “cross-border flow” in Article 2(3) (defined in terms of physical flows of electricity on a network resulting from the impact of activity outside of that Member State). But, again, it did not do so. The salient point is that in the 2019 Recast Regulation the legislator used specific terminology to connote international trade but avoided using these very terms in the new 2019 definition of “congestion” and instead used much broader language, and this must be taken as a deliberate policy choice.[95]It follows that the natural meaning of “congestion” in Article 2(4), when read in the light of both its legislative history and the deliberate use and non-use of terminology, suggests that the legislature did not intend to define “congestion” in narrow terms limited to congestion on international interconnectors.[96]Next, there are the substantive provisions of the regulation when taken as a whole. Not surprisingly the issue of congestion is addressed at length for the simple reason that it is a commonly occurring technical and commercial issue that arises and for which harmonised regulatory solutions were needed. The costs associated with the management of such congestion, and how they are allocated as between market participants, is but one subset of this broader topic. Save for the inference to be drawn from the definitions of “ancillary services” and “congestion” (discussed above) there is nothing explicit in the regulation on the allocation of costs.[97]Mr Beal KC drew our attention to various provisions of the 2019 Recast Regulation to highlight, in his submission, that the legislature recognised that as a phenomenon congestion could occur at any point in a network and not only at or on cross-border interconnectors. Ms Amy Rogers, for GEMA, argued that it was not surprising that the measure should address congestion and that it was not of course disputed that congestion could and did arise in all sorts of places on networks. But she argued, in effect, that the treatment of the costs incurred in managing that congestion was a quite different issue and that not a great deal of assistance could therefore be taken from the legislation as a whole. Further, that the split between costs associated with managing cross border congestion and other congestion costs was a legitimate “compromise” policy choice made at the EU level.[98]In my view, the 2019 Recast Regulation read as a whole, albeit only in modest degree, supports the analysis of SSE. This is because if the legislation had implemented the compromise that Ms Rogers contended for then one would have expected to see it clearly articulated in the recitals and/or in the substantive provisions, whereas it is not. Further, when the expression “congestion management” is used in the regulation, for instance in Article 16 on “General Principles of capacity allocation and congestion management”, the notion is described in broad and generic terms, and not in a way suggesting that it embodies a compromise or is limited only to the costs of the management of congestion on interconnectors hindering international trade. Moreover, Article 16(1) emphasises that when dealing with congestion management, non-discriminatory and market-based principles should apply which rather undermines the idea that some “compromise”, based upon differentiated principles, applies:
“Network congestion problems shall be addressed with non-discriminatory market-based solutions which give efficient economic signals to the market participants and transmission system operators involved. Network congestion problems shall be solved by means of non-transaction-based methods, namely methods that do not involve a selection between the contracts of individual market participants. When taking operational measures to ensure that its transmission system remains in the normal state, the transmission system operator shall take into account the effect of those measures on neighbouring control areas and coordinate such measures with other affected transmission system operators as provided for in Regulation (EU) 2015/1222.”
[99]Article 18 is also relevant. It is contained in section 2 of the Recast Regulation which is headed “Network Charges and congestion income”. Article 18 itself is headed “Charges for access to network, use of networks and reinforcement”. It underscores that network charges should be non-discriminatory and transparent and it expressly provides, in sub-paragraph (5), that the setting of charges under Article 18 “is without prejudice to charges resulting from congestion management under Article 16.”[100]Neither Articles 16 nor 18 use the language of international trade in the context of congestion management. For example, instead of referring in Article 16 to the networks of different member states it refers to “neighbouring control areas”, which is a much broader concept than one limited to political, state, entities.[101]Next, there are the recitals and any relevant travaux. I draw guidance from the recitals to the 2019 Recast Regulation which reflect an energy sector “… that was in the middle of its most profound changes in decades”. Technological developments allowed “…for new forms of consumer participation and cross-border cooperation” (recital [3]). It is proper to infer that when the legislature modified the definition of “congestion” this reflected this evolution in the wider EU market which was no longer based entirely on national boundaries. Ms Rogers for GEMA, in her attractive submissions, argued that the picture painted by the background travaux again supported her analysis of compromise. She for instance referred us to documents preceding the 2009 Regulation which drew a clear linkage between congestion management and international trade and to the Annex to the 2009 Regulation (see paragraph [29] above). But, with respect, these do not assist. It is not in dispute that prior to 2009 the legislature did link congestion management to international trade but it is difficult to see how that rationale necessarily held true in 2019 when, as the recitals to the 2019 Recast Regulation make plain, markets had undergone “profound” change and had become less nationalistic and were now characterised by increased cooperation and integration.[102]The Judge held that the repealed Article 2(2)(c) and the new Article 2(4) were “synonymous”. With respect I disagree. Article 2(4) was a clear change to the concept of “congestion” and was quite different to the 2009 predecessor and was not therefore “synonymous” with it. Article 2(4) is not limited to the costs associated with managing congestion across international interconnectors. In my judgment SSE is correct in its analysis.[103]I would add an important caveat. Whilst I am clear as to what Article 2(4) is not limited to I am not seeking to express any view as to how it would be applied in the future, in practice. It contains a variety of new and broad concepts which may well, when GEMA comes to construe and apply them, necessitate it forming a view as to novel, complex and technical issues of law and fact. For instance, GEMA must not only work out what it considers the terms actually mean but must then, for example, evaluate when “requests” cannot be “accommodated”, or what “significantly affects” means in relation to aflow, or what amounts to a “network element”, etc.[104]Subject to the analysis of EU retained law below I would therefore allow SSE’s appeal.

EU retained law: The conundrum

[105]The next question is whether the position in EU retained law alters the above conclusion. Prima facie, the EU retained law definition of “congestion” is markedly different to the definition in Article 2(4), and if applied according to its terms would lead to the opposite conclusion to that I have arrived at above.[106]Mr Beal KC made the point that the GEMA Decision was taken at a point in time when the EU retained law measure was not yet in force and that therefore the EU retained law is not relevant to the analysis. Whilst I ultimately agree with this point there is a good deal of unravelling of the law to undertake first. The court cannot simply ignore the EU retained law version. It is expressly relied upon by the CMA and GEMA as a guide to the interpretation of Article 2(4). GEMA argues that Article 2(4) and the EU retained version mean one and the same thing. The court should consider the EU retained version as a guide to the construction of Article 2(4) albeit that there is an admitted element of the tail wagging the dog in such an approach. Moreover, during the proceedings below all of GEMA, the CMA and the Judge assumed that there was nothing in the EU retained law definition of “congestion” which altered the substance of the 2019 Recast Regulation definition of “congestion”. If correct this would guide the law going forward. The issue is thus live before this court and we should not shrink from the duty to seek to provide legal certainty and guidance.[107]I turn then to the position under EU retained law. As to this for reasons that are very hard to fathom, the definition in the EU retained law differs from the definition in Article 2(4) of the 2019 Recast Regulation and is premised upon the repealed predecessor in Article 2(2)(c) of the 2009 Regulation. Yet, at the same time, it is said that this resurrected version still means the same as Article 2(4) of the 2019 Recast Regulation, even though as set out above it is drafted in very different language.[108]The EU retained law version says:
“‘congestion’ means a situation in which an interconnection linking the Great Britain transmission network with the transmission network of another country or territory cannot accommodate all physical flows resulting from international trade required by market participants, because of a lack of capacity of the interconnectors or the transmission systems concerned.”
[109]The EU retained law also includes a modified version of Article 2(1) of the 2019 Recast Regulation, defining an “interconnector” as:
“… a transmission line which crosses or spans a border between Great Britain and another country or territory and which connects the national transmission system of Great Britain with the transmission system of that other country or territory.”
[110]The re-insertion of interconnection as a concept into the definition of “congestion”, coupled with the reinforcement that an interconnector means a connection between Great Britain and another country or territory, plainly reintroduces into the EU retained law the limiting condition of international trade that had been deliberately removed from Article 2(4).[111]An agreed note prepared by all the parties as to the position under EU retained law refers to correspondence between BEIS and the parties in which BEIS explains that in its view the new definition of “congestion” was not intended to effect any change to either the 2009 Regulation or the Recast 2019 Regulation. The view was taken that there was no difference in substance between trade between “network areas” (in the 2019 Regulation) and trade across an (international) interconnector (in the 2009 Regulation). For this reason when the 2019 Recast Regulation came to be modified (under the process set out in section 8 EU(W)A 2018) to make it Brexit compliant, the draftsperson decided to work from the old, repealed, Article 2(2)(c) and not the new, replacement, Article 2(4).[112]With respect I find the legislative approach baffling. It is generally considered a cardinal rule of legislative drafting that a measure should not be amended unless it is intended to convey a change of meaning. It was for this reason that in R (Quintavalle) v Secretary of State for Health [2003] 2 AC 687 at page [695] Lord Bingham identified as relevant the "…historical context of the situation which led to its enactment". The task of the Court was to work out what had changed since it was this that then elucidated the purpose of the legislature in the measure being construed: "The basic task of the court is to ascertain and give effect to the true meaning of what Parliament has said in the enactment to be construed. But that is not to say that attention should be confined and a literal interpretation given to every particular provision which gives rise to difficulty… Every statute other than a pure consolidating statute is, after all, enacted to make some change, or address some problem, or remove some blemish, or effect some improvement in the national life. The court's task, within the permissible bounds of interpretation, is to give effect to Parliament's purpose. So the controversial provisions should be read in the context of the statute as a whole, and the statute as a whole should be read in the historical context of the situation which led to its enactment".[113]In Solar Century Holdings Ltd v Secretary of State for Energy & Climate Change [2014] EWHC 3677 (Admin) (“Solar Century”) the High Court (at paragraph [52] (ii)), having cited the leading appellate judgments, also highlighted the importance of identifying what had changed in order to identify the relevant purpose of the new measure:
“i) When construing an enactment, including the exercise of power under an enactment, it is relevant to identify the intention or purpose of the measure, i.e. the mischief to which it is directed. ii) In all cases (save with regard to consolidating enactments) the purpose or mischief may be identified by the posing of questions (cf Lord Bingham in Quintavalle) such as: If the legislation has changed, what has changed? If there is a problem which had to be resolved, what was the problem? If there was a blemish in the legislation, what was that blemish? If there was an improvement which was sought to be achieved, what was that improvement?”
[114]Legislative drafters change language to bring about change, but it follows that they do not change language not to bring about change. Yet, here, not only is the change said not to be intended to convey a change of meaning but:(i) in order to ensure, so it is said, conformity with Article 2(4) of the 2019 Recast Regulation the EU retained version has been changed to conform to the repealed predecessor version in the 2009 Regulation; and moreover(ii) the version now applicable in GB is materially different to the version applicable in Northern Ireland such that there are now two linguistically different texts applicable in the UK which are nonetheless said to represent the same law.[115]We were taken at length to the Explanatory Notes which accompanied the EU retained version. There is no need to set them out. They support the conclusion, explicitly advanced by BEIS, that the Government did not intend to bring about any policy change when introducing a definition of “congestion” into EU retained law differing from the EU counterpart. No impact assessment was produced to accompany the adoption of the EU retained law measure, which itself is an indication that the retained measure was not considered to bring about substantive change.[116]For the reasons I have given I am clear that the definition of “congestion” in Article 2(4) of the 2019 Recast Regulation is quite different to that in Article 2(2)(c) of the 2009 Regulation. The rendition of “congestion” in EU retained law thus reflects an error which lies in the assumption made that the new definition in Article 2(4) of the 2019 Recast Regulation was not intended to bring about any substantive change to that in the old, repealed, 2009 Regulation.[117]However, and notwithstanding this error, the fact remains that the governing law in Great Britain is now that in the EU retained law. What is the court to do? Do we apply, faithfully, the new EU retained version upon the basis that it is clear on its face? Or do we apply some hyper-elongated interpretative approach to construction in order, somehow, to make the EU retained version say that which, palpably, it does not say?[118]To answer this the starting point is that the court has the power to correct errors in legislation and that this applies to primary and to secondary measures. However, there are limits to this jurisdiction which sit at the intersection between a legitimate exercise in purposive interpretation and judicial legislation. We can do the former but not the latter. Which side of the line does the present case fall on?

EU retained law: judicial corrections to legislation and the limits of purposive interpretation

[119]With some hesitation I am of the view that it would not be right for the court to remedy the error by reformulating the EU retained version.[120]The authors of Bennion, Statutory Interpretation (7th ed, 2017), take as a starting presumption that courts will apply a “rectifying construction” (ibid section [15.1] page [425]). This is described in the following way:
“It is presumed that the legislator intends the court to apply a construction which rectifies any error in the drafting of the enactment, where required in order to give effect to the legislative intention”
. This formulation has been approved of in a series of cases (ibid footnote 1 at page [430] and cases cited thereat). This is an important presumption especially in the case of EU retained law where the legislature pushed through literally thousands of complex and highly technical measures, overwhelmingly in subordinate legislation, in record time and invariably with but scant Parliamentary scrutiny of the detail, in order to meet the deadline for the end of the post-Brexit implementation period. It will therefore hardly be a surprise when courts encounter a range of lacunae, ambiguities and other infelicities in the drafting of such legislation.[121]The approach of the courts will always be to seek to ensure, to the greatest extent possible, that the legislative intent is fulfilled. This does not represent an exercise in judicial activism but reflects the comity that exists between the courts and Parliament when it comes to the process of construction and, as the quotation from Bennion explains, entails the court living up to Parliament’s intention that the courts will perform their role in making due sense of legislation.[122]But there are limits. These were articulated in Inco Europe Ltd v First Choice Distribution [2000] UKHL 15 (“Inco”). There, Lord Nicholls observed that the function of the court was not confined to the resolving of ambiguities but that the court could, where the intention was clear, “… add words, or omit words, or substitute words”. As to limitations, the power to correct was limited to plain cases and the courts should abstain from any course which “might have the appearance of judicial legislation”:
“This power is confined to plain cases of drafting mistakes. The courts are ever mindful that their constitutional role in this field is interpretative. They must abstain from any course which might have the appearance of judicial legislation. A statute is expressed in language approved and enacted by the legislature. So the courts exercise considerable caution before adding or omitting or substituting words. Before interpreting a statute in this way the court must be abundantly sure of three matters: (1) the intended purpose of the statute or provision in question; (2) that by inadvertence the draftsman and Parliament failed to give effect to that purpose in the provision in question; and (3) the substance of the provision Parliament would have made, although not necessarily the precise words Parliament would have used, had the error in the Bill been noticed. The third of these conditions is of crucial importance. Otherwise any attempt to determine the meaning of the enactment would cross the boundary between construction and legislation: see Lord Diplock in Jones v. Wrotham Park Settled Estates [1980] A.C. 74, 105. In the present case these three conditions are fulfilled. Sometimes, even when these conditions are met, the court may find itself inhibited from interpreting the statutory provision in accordance with what it is satisfied was the underlying intention of Parliament. The alteration in language may be too far-reaching. In Western Bank Ltd v. Schindler [1977] Ch 1, 18, Scarman L.J. observed that the insertion must not be too big, or too much at variance with the language used by the legislature. Or the subject matter may call for a strict interpretation of the statutory language, as in penal legislation. None of these considerations apply in the present case. Here, the court is able to give effect to a construction of the statute which accords with the intention of the legislature.”
[123]If this court were to correct the EU retained law so as to bring it back into line with the Article 2(4) definition then we would therefore have to be sure:(i) of the intended purpose of the provision in question;(ii) that the drafter and Parliament inadvertently failed to give effect to the requisite purpose in the provision in question; and (iii), the substance of the provision that Parliament would have made (even if not the precise words).[124]I consider the position to be marginal. The intended purpose of the EU retained measure was to implement Article 2(4) as to which I have set out my conclusions as to its scope and effect. The error (or inadvertence) is also clear in that a false assumption was made that the 2009 Regulation and the 2019 Recast Regulation both evinced the same substantive meaning of “congestion”. The third condition from Inco (above) is more problematic. I am clear that we could, at least at the technical drafting level, having sought assistance from the parties, perform these tasks and come up with a workable draft.[125]However, I fear that in so doing we would give the impression of engaging in judicial legislation. As to this:(i) neither BEIS, in the correspondence before the court, nor GEMA, in written and oral submissions, accept that there is an error;(ii) legally and constitutionally it was and remains available in principle for the Executive to depart from the text of Article 2(4) of the 2019 Recast Regulation, if it sees fit;(iii) for this court to repair the draft we would have to understand why it was thought appropriate to adopt different wording for Great Britain than Northern Ireland after the end of the post-Brexit implementation period, and (iv), we would have to ensure that the new, judicially amended (ie substantially rewritten), language was fit for purpose for Great Britain by in effect exercising a power conferred upon the Secretary of State under section 8 EU(W)A 2018. A review of the nature of the test set in section 8 highlights just how technical and value laden such an exercise is capable of being. For my part the sheer extent of this task, which would include the court seeking to replicate the exercise of a statutory power conferred upon the Minister in a complex and uncertain regulatory environment and in an area of relations between Great Britain, Northern Ireland and the EU which is currently beset with political complications, pushes the exercise over the line.[126]For these reasons it seems to me that the proper course is simply to set out, as I have, my conclusions on the law but then to leave it to the Secretary of State to make any changes that are considered necessary to achieve the stated present purpose of the measure of EU retained law (which is not to depart from the substantive meaning of congestion in Article 2(4)). The power to make regulations under section 8 EU(W)A 2018 expires at the end of the 2-year period beginning with IP completion day (section 8(8) EU(W)A which was 31st December 2020 (section 39 EU (Withdrawal Agreement) Act 2020).[127]I record that in submissions Mr Beal KC argued that if the Government did nothing then, in view of the clear statement of the Government that it was not intended to effect any substantive change to the law, this would leave the EU retained measure as ultra vires. But he also accepted that this was a battle for another day.

Conclusion

[128]I would allow the appeal. When the GEMA Decision was taken the governing definitions were to be found in the 2019 Recast Regulation. That provided that congestion was not limited to problems arising across international interconnectors, and that ancillary services excluded congestion management. That had the result that, under Annex B, charges for congestion management fell outside the Ancillary Services Exclusion, and hence were required to be included in the calculation of the average annual transmission charge for the purposes of compliance with the Permitted Range. The Ancillary Services Exclusion in Annex B is not, on its face and in accordance with the usual methods of construction, limited to charges for the costs attributable to managing congestion on interconnectors between states. Procedural issues: (i) the admissibility of the views of officials as a guide to interpretation; and (ii), the utility of seeking the views of and evidence from the EU Commission.[129]There are two procedural issues that arose during the appeal that I would briefly deal with.[130]The first concerned the admissibility of the views of officials both in London and in Brussels on the meaning of Article 2(4) of the 2019 Recast regulation. SSE sought disclosure from GEMA covering correspondence on the various definitions of congestion passing between itself and BEIS. We received various submissions about the views of officials. I am of the view that these are inadmissible. They do not fall within the class of documents and materials traditionally considered to be admissible as a guide to legislative intent. In Solar Century (ibid) the Court summarised (at paragraph [52(iii)-(vii)]) the categories of document that might be admissible but also, importantly, the limits of the use of such documents:
“iii) To identify the purpose or mischief and to answer these questions it is permissible to examine Explanatory Notes, White and Green Papers, Ministerial statements (Bradley) and Law Commission Reports, all of which may be admissible forms of evidence. iv) However, not all such admissible sources are of equal weight. Those sources (such as Explanatory Notes) whose "shape" was closely connected to the "shape of the proposed legislation" may be more informative as guides (Westminster City Council) than other sources which are more remote from the final language selected by Parliament. v) In addition, a court may draw inferences from the statutory words actually used in the scheme of the legislation as a whole and from any case law on the underlying subject matter and a court might ask whether it may be inferred that Parliament intended to act consistently with the standard set out in case law (Chetnik). vi) Material that is admissible will reflect the views of their authors. And the views of authors, including the Government of the day, do not necessarily reflect the will of Parliament (Westminster City Council). If there is an inconsistency between the statutory language and the pre-legislative, admissible, material it cannot, without more, therefore be assumed that the statutory purpose must reflect the purpose set out in pre-existing admissible material. vii) However, if there is a collision between a literal interpretation of an enactment and the contextual material with the consequence that the literal interpretation "is manifestly contrary to the intention which one may readily impute to Parliament, when having regard to the historical context and the mischief…", then the enactment should be construed in the light of the purpose as evident from the historical context and mischief (R v Z per Lord Carswell).”
[131]The documents in issue do not fall within the class of documents traditionally treated as admissible. They are simply the views of officials formed after the event. They do not necessarily represent the views of responsible Ministers or the institutions that employ them. They played no part in the legislative process. They have never hitherto been made public, and the context to the observations is unstated and obscure. To take them into account would be inconsistent with the principles upon which purposive interpretation operates: See for an example of a comparable approach and conclusion Re Agriculture Sector (Wales) Bill [2014] UKSC 43 at paragraph [39].[132]The second procedural issue concerned whether the Court should have sought the views of the EU Commission. The meaning of “congestion”, albeit technical, is important in the context of the charging regime. The legislation emphasises that consistency in charging principles is important, not least because the rules send signals to the markets which can influence investment decisions. There is minimal evidence before the Court as to how the expression “congestion” is understood and applied elsewhere.[133]Prior to the end of the post-Brexit implementation period the court might have considered a reference to the CJEU. That is not now an option. The Court mooted, tentatively, an alternative with the parties which was that the Court would invite the EU Commission to(i) provide information as to practices elsewhere and(ii) express its own (necessarily expert) opinion on interpretation. Again prior the end of the post-Brexit implementation period the Court would have had the power to request the Commission to offer assistance to the Court in this way and, as part of the duty of cooperation existing between the Member states and the EU, the Commission would no doubt have felt constrained to assist: See (Case C-2/88 Imm.,) JJ Zwartveld and others [1990] ECR I-4406. There was some debate as to whether any equivalent doctrine existed following the end of the post-Brexit implementation period which would require the Commission to cooperate under the NI Protocol or the Trade and Cooperation Agreement upon the basis that the construction placed on the law by this court could affect the interpretation of the position under the NI Protocol pursuant to which the 2019 Recast regulation remained operative. It is unnecessary to express any view on this. Had I been of the view that seeking the assistance of the Commission was prudent I very much doubt that the Court would have done any more than politely invite the Commission to provide such assistance as it saw fit, without there being any hint of an obligation upon it to do so. In the event I was persuaded by the submissions of Ms Rogers for GEMA that it would not be a good idea to make any such request. The reasons can be summarised shortly. First, SSE had at an earlier stage asked the Commission to assist and it had declined and there was no discernible reason why it should change its mind now. Secondly, as the Commission had observed any conclusion of this court would not bind the Commission or any other EU state, which undermined the utility of it expressing its view. Thirdly, the process, quite irrespective of the outcome, could take many months, and possibly well over 6-9 months especially if the Commission had to collect evidence about regulatory practices from the Member States. I found these reasons persuasive and concluded that there would be no real utility in seeking the assistance of the EU Commission.

F. Disposal

[134]It follows from the above that I would allow the appeal by GEMA on Issue I. I would allow the cross appeal of SSE on Issue II.[135]I would finally express my thanks to the parties for their detailed and helpful written and oral submissions and for the way in which all counsel responded to the very many questions from the Court. Although I would reverse the judge on all of his findings which have been challenged in this appeal we benefited from having received much more detailed arguments on the points arising than were before the High Court.[136]I agree.

Sir Julian Flaux, Chancellor of the High Court:

[137]I also agree.

Cited in 4 later judgments