Wales & West Utilities Limited, R (on the application of) v Competition and Markets Authority

The proceedings before the CMA

The “cost of equity” ground, namely that the cost of equity chosen by GEMA was too low and therefore wrong (all appellants, including WWU);

The “outperformance wedge” ground, namely that it was wrong for GEMA to apply a reduction to the allowed cost of equity, reflecting its view that companies would be expected to achieve operational outperformance during the price control period (all appellants, including WWU);

The “ongoing efficiency” ground, namely that GEMA had erred in its approach to setting ongoing efficiency (i.e. the cost reduction applied by GEMA to account for expected productivity improvements in the energy sector) (five appellants, including WWU); and

The "licence modification" process ground (three appellants, including WWU).

The final determination of the CMA

found in favour of all appellants (including WWU) that GEMA was wrong to impose the outperformance wedge (Order at §§3-5);

partially found in favour of WWU (and other appellants) that GEMA was wrong to impose the innovation uplift (Order at §§6-13);

partially found in favour of Scottish Hydro Electric Transmission Plc (“SSEN-T”) and one other appellant that GEMA acted ultra vires in the manner in which it sought to use a directions process to modify certain licence conditions (Order at §§14-18), but dismissed WWU’s appeal under joined Ground D (Licence modification process) (Order at §19); and

in all other respects, dismissed WWU’s appeal, including on the cost of debt and tax clawback (Order at §29).

The principal objective of the Secretary of State and the Gas and Electricity Markets Authority (in this Act referred to as “the Authority”) in carrying out their respective functions under this Part is to protect the interests of existing and future consumers in relation to gas conveyed through pipes.

In performing the duties under subsections (1B) . . . the Secretary of State or the Authority shall have regard to -

Subject to [not material] the Secretary of State and the Authority shall carry out their respective functions under this Part in the manner which he or it considers is best calculated -

An appeal lies to the CMA against a decision by the Authority to proceed with the modification of a condition of a licence under section 23.

An appeal may be brought under this section only by –

The permission of the CMA is required for the bringing of an appeal under this section.

The appeal is heard by three members of the “CMA panel” (paragraph 4(2)) which is comprised of individuals appointed for their expertise in (broadly) competition economics, law, finance, and business (see paragraph 35 of Schedule 4 to the Enterprise and Regulatory Reform Act 2013).

The CMA may require the production of documents and information (paragraph 6), and may also require oral evidence and cross-examination (paragraphs 7(2) and (4)).

The CMA may commission expert advice on any matter a party raises (paragraph 9).

This section applies to every appeal brought under section 23B.

In determining an appeal the CMA must have regard, to the same extent as is required of the Authority, to the matters to which the Authority must have regard—

In determining the appeal the CMA -

The CMA may allow the appeal only to the extent that it is satisfied that the decision appealed against was wrong on one or more of the following grounds—

that the Authority failed properly to have regard to any matter mentioned in subsection (2);

that the Authority failed to give the appropriate weight to any matter mentioned in subsection (2);

that the decision was based, wholly or partly, on an error of fact;

that the modifications fail to achieve, in whole or in part, the effect stated by the Authority by virtue of section 23(7)(b);

that the decision was wrong in law.

To the extent that the CMA does not allow the appeal, it must confirm the decision appealed against.

quash the decision (to the extent that the appeal is allowed);

remit the matter back to the Authority for reconsideration and determination in accordance with any directions given by the CMA;

substitute the CMA's decision for that of the Authority (to the extent that the appeal is allowed) and give any directions to the Authority or any other party to the appeal.

For other appeals, the CMA has the powers under subsections 23E(2)(a) and (b), but not (c).

At the point of privatisation, a new form of price control regulation was developed and implemented in the UK. This was known as RPI-X and was designed to incentivise improvements in efficiency, particularly as the networks were seen as being inefficient with high operating costs and spare capacity.

Under RPI-X the regulator limits average network charges from rising at the rate of inflation (measured by the retail price index, or RPI) less an efficiency factor (called ‘X’). Since price limits are set before expenditure has been incurred by companies, RPI-X is a form of ex-ante regulation. Once price caps are set, companies face incentives to reduce expenditure as much as possible to maximise their profits. In doing so, they reveal the potential for efficiency gains to the regulator. The regulator periodically resets the caps, and ‘squeezes out the remaining efficiency potential’ by increasing the X factor and reducing prices.

While RPI-X was generally considered to have worked well initially, some potential issues arose, particularly around the incentives to reduce customer service (in order to reduce costs), limited levels of innovation, and a bias towards capital expenditure in order for companies to expand their asset bases.

In 2008, GEMA decided to review its approach to setting price controls, in particular to try and address some of the issues it saw arising from RPI-X. This resulted in the creation of the ‘RIIO’ framework (where RIIO stands for: Revenues = Incentives + Innovation + Outputs). This aims to build on the previous RPI-X regime, but places more emphasis on incentives to drive the outputs and innovation needed to deliver a sustainable energy system in a way that provides value for money to existing and future consumers. GEMA’s first RIIO price control (RIIO-1) ran from 1 April 2013 to 31 March 2021 for the gas transmission and distribution networks as well as the electricity transmission networks.

The latest iteration of GEMA’s updated regulatory framework (RIIO-2) is the price control that is subject to this appeal. When needing to differentiate between the controls for different sectors these are often referred to as RIIO-GT2 (for the gas transmission network), RIIO-GD2 (for the gas distribution networks), and RIIO-ET2 (for the electricity transmission networks).

Structure of RIIO and components of allowed revenue

The RIIO model is an incentive-based model under which GEMA sets both the amount that regulated companies can earn over the price control period and what the companies must deliver in return for those revenues. Although it is referred to as a price control, GEMA actually controls the total revenues each company can recover rather than specific prices. GEMA explained the total allowed revenue as comprising:

Baseline revenue is the revenue that a company needs to cover the efficient operating and financing costs of delivering outputs and long-term value for money, including allowances for maintenance of, and investment in, capital assets and taxation. GEMA assesses the efficiency of these costs.

GEMA describes baseline revenue as comprising four different categories:

RIIO adopts a totex approach to ensure companies make balanced decisions between different types of solution. Totex includes capital expenditure (capex), operating expenditure (opex), and replacement expenditure (repex). Totex is remunerated by a combination of ‘fast money’ and ‘slow money’:

In order to better balance these requirements, GEMA made a number of changes from RIIO-1, in particular:

This is the second time that an appeal has been brought under section 11C of EA89. In the appeals of the RIIO-1 price control decision (ED1 Determinations) the CMA published its conclusions on the standard of review in its two decisions of 29 September 2015. Neither decision was challenged by any party in the ED1 Determinations, and the applicable legislation remains unchanged. We have therefore applied the standard of review employed in the ED1 Determinations in the current appeals.

Furthermore, in considering the submissions we have received on the statutory framework in the course of these appeals, we have also drawn on the approach taken in other regulatory appeals involving substantially the same statutory framework, where relevant. Notably, as in the ED1 Determinations, we have had regard to cases under the Energy Act 2004 (EA04) on the basis that the grounds for allowing an appeal under EA04 are very similar to the grounds for allowing an appeal under EA89 and GA86. As such, these provide helpful guidance as to the level of scrutiny which an appellate body with particular expertise such as the CMA should adopt in reviewing GEMA’s decisions in the present case. We have also taken guidance from the approaches taken in more recent appeals brought under the Gas (Northern Ireland) Order 1996 and the Electricity (Northern Ireland) Order 1992, given that the statutory grounds for allowing an appeal are essentially the same as the grounds for allowing an appeal under EA89 and GA86. Although we are not bound by these decisions, which concern a different statutory appeal mechanism under a different legislative scheme, we consider that they are instructive for the approach which the CMA should take in the present appeals.

Similarly, we have also drawn guidance from appeals brought under other regimes such as the Communications Act 2003 (CA03). Whilst we acknowledge that there is no direct analogy with the present appeals given some differences in the statutory provisions, we consider the approach adopted in these cases by the Competition Appeal Tribunal (the CAT) to be broadly analogous to the approach taken by the CMA in the ED1 Determinations and that they therefore also provide some guidance as to the level of scrutiny which an appellate body with particular expertise such as the CMA should adopt in reviewing GEMA’s decision in the present case.

Both ED1 Determinations concluded that the CMA’s role was not limited to reviewing the decision on traditional judicial review grounds and that the CMA is not only able, but required by EA89, to consider the merits of the decision under appeal, albeit by reference to the specific grounds of appeal laid down in the statute. In this context, the appeals took note of the fact that, when this appeal mechanism was inserted into the provisions of EA89, the preceding consultation document referred to the grounds of appeal that would be available to appellants, stating:

In the ED1 Determinations, the CMA found that a useful analogy could be drawn between the appeal on the basis of EA89 and the approach taken by the CAT in the context of appeals against Ofcom price controls in the telecommunications sector. In BT v Ofcom, the CAT stated that:

We agree with and adopt the CMA’s position outlined in the ED1 Determinations, where the CMA decided that an appeal under EA89 does not involve a rehearing where it is open to it to decide matters afresh untrammelled by GEMA’s decision. Furthermore, we also agree with the ED1 Determinations that consideration of the merits of the decision under appeal must be within the confines of the specific grounds of appeal laid down in the statute.

We note that the CAT confirmed this approach in two recent cases brought under CA03:

In the present case, whilst there was common ground between the appellants and GEMA as to the fact that the CMA’s role is not limited to reviewing the Decision on traditional judicial review grounds, there was some disagreement on the extent of the merits review:

In line with the CMA’s position outlined in the ED1 Determinations, we reject WWU’s submission that we should ‘re-consider the case as if [we] were the primary decision-maker’ and agree with GEMA’s submission that the standard of review falls short of a full rehearing. We are required to consider the merits of the Decision but only through the prism of the specific errors alleged by the appellants. The appeals do not entitle the CMA to proceed with a re-run of the original investigation or have a de novo re-hearing of all the evidence. The key question is whether GEMA made a decision that was wrong (on one of the prescribed statutory grounds). Only to that extent must the merits of the Decision be taken into account and we have done so in the present appeals.

Our final determination in this appeal reflects the application of a standard of review that is in line with the approach set out above. We consider that this approach is consistent with the approach taken by the CMA in the ED1 Determinations and the more recent decisions in appeals against decisions by the Northern Ireland Authority for Utility Regulation, as well as the CC in Energy Code appeals, and by the courts in relation to appeals under CA03. It also reflects the government’s intention in implementing the relevant appeal provisions.

On the point of deciding errors of fact, in 2003 the Court of Appeal in Assicurazioni Generali Spa v Arab Insurance Group issued a Practice Note where the Court held that:

Both ED1 Determinations were based on the reasoning in Assicurazioni and also referred to the CC’s approach in the energy code modification appeal of E.ON UK plc v GEMA(E.ON), which we discuss below. While the CMA recognised that it was not bound by the decision in E.ON (which concerned a different statutory appeal mechanism under a different legislative scheme), it considered that the E.ON decision accurately characterised the approach which the CMA should take in the RIIO-1 appeals.

In the decision of SONI Limited v Northern Ireland Authority for Utility Regulation (SONI), having referred to Assicurazioni, the CMA concluded that:

In the present appeals, the parties made submissions on what ‘wrong’ means, variously arguing that the Decision may be wrong if:

We consider each of these issues below.

Alternative approaches

In the present appeals, GEMA submitted that:

Consistent with the ED1 Determinations, we note and adopt the finding made in E.ON and various other more recent cases that our starting point should not be to determine whether there is an alternative approach and then decide whether it is better. We should only determine whether there is an error in the approach chosen by GEMA, as alleged by the appellants.

In SONI (an appeal under The Electricity (Northern Ireland) Order 1992) the CMA also adopted this position, stating that:

In their response to the provisional determination, NGET/NGG submitted that the CMA should have considered the appellants’ alternative approaches alongside the GEMA’s approach and evaluated the relative merit of the different approaches to determine whether GEMA’s approach was wrong. Similarly, Cadent submitted in its reply to GEMA’s Response that: ‘[w]here an alternative approach would have been clearly better, GEMA’s decision is wrong.’

As noted at paragraph 3.40 above, our starting point will be to consider the adequacy of GEMA’s chosen approach rather than considering which approach we ourselves might have chosen had we been in GEMA’s position. However, we agree that, in considering whether GEMA’s chosen approach discloses an error, we will consider its inherent merits including by comparing its merits with those of any reasonable alternatives advanced by the appellants. If, out of the alternatives available, we conclude that some alternatives clearly had greater merit than the solution chosen by GEMA, then we are more likely to be persuaded that GEMA has erred. On the other hand, where the alternative options each have competing pros and cons, and none is clearly superior, it will be more difficult to persuade us that GEMA has erred.

GEMA’s assessment of the evidence

In the present appeals, NGN submitted ‘that a number of GEMA’s decisions are extreme and not supported by a reasonable consideration of the evidence’ and that GEMA is required to ’[take] into account all relevant evidence, [give] it appropriate weight and [adopt] a decision which is consistent with such weighing of evidence.’

SSEN-T submitted that the CMA should find an error if ‘GEMA has not taken account of relevant evidence or data, and placed weight on unreliable data’.

In Northern Powergrid, the CMA overturned GEMA’s decision to impose a cost adjustment relating to Smart Grid Benefits (SGB) on the basis that:

We thus agree with the appellants that we may find an error if the Decision is based on unreliable data or fails to take account of the relevant evidence.

GEMA’s reasoning

In the present appeals, SGN argued that ‘GEMA acts unlawfully where its decision is unreasonable…’ and that ‘A decision can be unreasonable … where there is a demonstrable flaw in the reasoning’.

GEMA argued in relation to cost of equity, aiming up and sufficiency of return that ‘it is not enough for an appellant to identify some error or absence of reasoning, since an appeal can only succeed if the decision itself cannot stand.’

In Firmus Energy (Distribution) Limited v Northern Ireland Authority for Utility Regulation (Firmus), the CMA found that:

In line with Firmus, we agree with GEMA that the fact that its reasoning underpinning its Decision is wrong does not mean that the Decision itself will necessarily be wrong if it can be supported on a basis other than that on which GEMA relied.

Procedural deficiencies

In the present appeals, several appellants invited us to take into account procedural deficiencies in GEMA’s decision-making process when determining whether its Decision was wrong. In particular:

In Firmus, the CMA found that:

Consistent with Firmus, our analysis should only take into account procedural deficiencies (including a flawed consultation process) if they are so serious that we cannot be assured that the Decision was not wrong.

The margin of appreciation afforded to GEMA in matters of regulatory judgement is important to bear in mind in reaching a finding as to whether a decision is ‘wrong’, as recognised by the CMA and courts in previous regulatory appeals.3.56 In the present appeals, we were invited by GEMA to find that:

Most of the appellants took issue with the way GEMA described its regulatory discretion in its submission. They submitted that GEMA had overstated the extent of its margin of appreciation.

Cadent argued that:

Further, Cadent contended that it was not correct for GEMA to say that ‘[T]he question before the CMA … is whether GEMA’s decision “lay outside the bounds within which reasonable disagreement is possible”’ and that ‘… the standard of review in relation to matters of judgement is not limited to Wednesbury unreasonableness’.

NGET/NGG said that GEMA had ‘taken the point on regulatory discretion and judgement too far’ and that ‘there is no principle that an appeal must fail if it relates to a judgement on which “reasonable people may disagree”.’

Similarly, SGN maintained that ‘GEMA seeks to distort and subvert the relevant legal framework by replacing the applicable statutory test, which entails considering whether GEMA’s decision is wrong on one of the grounds set out in section 23D(4) GA86, with an alternative test, which entails considering whether the Decision fell outside reasonable bounds…a decision may be “wrong” even if it is not “unreasonable”.’

Quoting the CMA’s decision in Northern Powergrid, NGN submitted that ‘there has to be … a limit to the discretion of regulators to make adjustments’ and ultimately ‘[t]he exercise of regulatory discretion remains bounded and subject to legal principles.’ NGN also argued that ‘GEMA’s margin of discretion does not allow GEMA to take decisions based on extremely selective, inconsistent and/or inadequate evidence and assumptions.’

SSEN-T argued that ‘GEMA’s attempt to hide behind discretion betrays a lack of confidence in the substance of its decisions and must be rejected.’

SPT submitted that ‘Whilst of course it is true that GEMA is to be allowed some margin of appreciation as an “expert regulator”, this does not give GEMA carte blanche. It must not be allowed to denude the CMA of its own expert regulatory function.’

General principles

The CMA’s starting point when considering an appeal is as set out in SONI, where the CMA considered that:

In Virgin Media, which was a case brought under CA03, the CAT held that ‘[p]roper respect must be accorded to Ofcom’s role as a specialist regulator, and the expertise of Ofcom’s staff.’

We consider that these dicta in SONI and Virgin Media are equally apposite for the margin of appreciation afforded to GEMA in the present context.

In essence, while we agree that GEMA should be afforded a margin of appreciation as an expert regulator, we also agree with the appellants that its margin of appreciation is not unbounded. As explained below, we consider that the approach we should take is different depending on whether the appellants are alleging an error of primary fact (or inferences from such primary facts) or an error of law, or whether they are alleging that GEMA’s weighting of factors or other exercise of judgement is wrong.

Errors of primary fact and errors of law

In E.ON, in relation to an appeal under EA04, the CC found in relation to errors of fact that:

As regards errors of law, in its recent decision in the SSE code modifications appeal, the CMA agreed with the parties that ‘in the context of challenges relying on an alleged error of law, … there [is] no role for ‘regulatory judgement’ or discretion on the question of what is the correct construction of legislation’ and also that ‘on that question, the concept of reasonable judgement, as embodied in the Wednesbury test, has no part to play.’

In the present appeals, GEMA itself recognised that ‘where an error of law is alleged, the CMA must make its own decision as to what was the correct conclusion, without showing deference to GEMA’s reasoning or regulatory discretion.’

Thus, if we find that the Decision is based on a plain error of fact or is wrong in law, there is no room for a margin of appreciation.

Regulatory judgement

In E.ON, the CC considered the appellate body’s function. The CC noted that the appellate body will exercise a degree of restraint when considering matters of regulatory judgement:

The CC further took the view that the statutory test under section 175 of EA04:

In BT v Ofcom the Court of Appeal considered the CAT’s approach to its jurisdiction to review Ofcom’s exercise of judgement. While that appeal concerned a different statutory scheme, the court confirmed the general principle underlying the approach, saying:

In line with E.ON and BT v Ofcom, we find that where the exercise of regulatory judgement is involved, GEMA will have a margin of appreciation. GEMA’s margin of appreciation will be at its greatest where all that is impugned is an overall value judgement based upon competing considerations in the context of a public policy decision. We will apply appropriate restraint and, in principle, not question issues of judgement on unchallenged primary findings and inferences determined by GEMA unless we are satisfied that GEMA’s decision is wrong.

Similarly, where GEMA has exercised regulatory judgement in selecting amongst various alternative solutions to a regulatory problem, we will not substitute GEMA’s assessment or weighting of the evidence or reasoning with our own unless we are satisfied that GEMA’s approach was wrong – for example, because there was a clearly superior alternative approach (see discussion at paragraph 3.42 above). In their responses to the CMA’s provisional determination, a number of the appellants argued that the CMA had shown undue deference to GEMA’s expertise in economic matters, which they contended was unwarranted. SSEN-T noted that the CMA considered issues relevant to the Capital Asset Pricing Model (CAPM) and cost of equity across a range of regulated sectors and that, by contrast, GEMA’s experience was confined to the energy sector. It went on to contend that we should therefore show no deference to GEMA’s decision-making on these points.

We accept that the CMA is itself an expert body, in particular in relation to CAPM and cost of equity, and that it benefits from cross-sector experience. We also agree that we must not uncritically accept GEMA’s assessment and weighting of the considerations before it simply because GEMA is an expert body. However, we do not accept that these points mean GEMA has no margin of appreciation when these matters are appealed to us. Nor do we accept that allowing GEMA a margin of appreciation equates to the adoption of a judicial review standard. We note that even in a full merits appeal of an infringement decision imposing quasi-criminal penalties under the Competition Act 1998, it can be consistent with such an appeal to allow the decision-maker a margin of appreciation. We shall carefully scrutinise the substance of GEMA’s decision-making in line with the grounds of appeal advanced before us and we shall intervene where we are persuaded GEMA has erred.

We note that in their replies WWU, NGET/NGG, and Cadent all objected to GEMA including the following quote from the Cellcom case: ‘the court should be very slow to impugn decisions of fact made by an expert and experienced decision-maker, it must surely be even slower to impugn his educated prophesies and predictions for the future.’ The appellants noted that these observations were made in the context of judicial review proceedings and so are not applicable in the present appeals. While we agree with the appellants that Cellcom is a judicial review case, as explained above, we will apply appropriate restraint when reviewing issues which entail the exercise of GEMA’s judgement, which may include GEMA’s predictions for the future.

Past decisions taken by the CMA in other regulatory appeals are not binding. This means that we are not required to adopt an identical approach to determining the issues, nor to ensure that our decision in the present appeals mirrors assessments made and conclusions reached by the CMA in other regulatory appeals. While the CMA may have regard to such past decisions in considering the present appeals, it will also take into account the fact that other regulatory appeals may be subject to different statutory regimes and relate to different sectors.

For example, the CMA may decline to adopt an approach taken in the recent CMA PR19 Redetermination which was a full redetermination of the PR19 price control by the CMA. By contrast, our role in these appeals is limited to finding whether GEMA was wrong on any of the specific grounds raised by the appellants. As indicated above (see paragraph 3.22), the appeals do not entitle the CMA to proceed with a re-run of the original investigation or have a de novo re-hearing of all the evidence.

This ground covers errors alleged by WWU relating to GEMA’s cost of debt allowance within the RIIO-2 price control.

Background

The cost of debt component of the WACC estimate reflects the return required to compensate debt investors for lending to a business. Unlike the forward-looking cost of equity, the majority of debt costs are accounted for by interest costs on historic (embedded) debt already held by businesses.

Elements of WWU’s appeal involve GEMA’s treatment of derivative debt instruments when measuring the cost of debt. For further background on such instruments, please see paragraphs 16.6 to 16.8 below.

The RIIO-2 Decision

GEMA based its cost of debt allowance on an average of the yield on the iBoxx GBP 10yr+ Utilities index. Under this approach, companies receive an allowance based on the 10-year average of this index at the start of the control, with the average used increasing by a year throughout each year of the control, so that in the final year companies receive a 14-year average of the index.

In changing from the mix of A and BBB indices used in previous controls, GEMA stated that it considered that broadly matching the average borrowing costs of networks, by using an investment grade index that was expected to be more representative of network borrowing costs, was more important than precisely matching a theoretical notional company rating.

GEMA noted evidence that indicated that the average rating of the constituents of the Utilities index had fallen over time and suggested that would be prudent to monitor the average rating over time before reassessing the approach for RIIO-3.

GEMA tested its chosen approach against a range of rate and inflation scenarios and was comfortable that its 10 to 14 year trailing average approach was an appropriate calibration.

GEMA allowed 0.25% for additional costs of borrowing (outside of explicit interest costs). This allowance comprised:

GEMA noted that it remained of the view that it was appropriate to exclude intercompany loans and derivatives from its calibration exercises. However, it also noted that its modelling suggested that if both these products were included in modelled expected costs for the Gas Distribution and Transmission (GD&T) companies, the 10 to 14-year trailing average plus 25 bps of additional allowance (GEMA’s RIIO-2 cost of debt approach) would be expected to be sufficient to cover the combined expected GD&T company debt and derivative costs.

GEMA assessed SSEN-T’s Regulatory Asset value (RAV) growth profile over the RIIO-1 and RIIO-2 period as materially different from other networks, justifying a RAV-weighted approach to its debt allowance.

GEMA noted that SGN Scotland and NGN had provided estimates of the additional costs faced by smaller networks as being 6 bps. GEMA considered this estimate to be reasonable and decided to ‘err on the conservative side’ in allowing this additional provision for notional licensees expected to issue debt at smaller size or less frequently than other networks due to their lower RAV size and RAV growth for RIIO-2. GEMA defined less frequently issuing notional networks as those that are expected to issue less than £150 million per annum on average, namely SGN Scotland, NGN and WWU.

GEMA considered these adjustments for RAV profile differences to be appropriate for notional company allowances because RAV was not solely determined by management decisions and was heavily influenced by regulatory decisions and regulatory requirements for investment. GEMA distinguished these notional company adjustments from other requests for actual company-specific adjustments (for example, from WWU for actual company debt costs), which were to a much greater extent driven by management or shareholder decisions on capital structure, M&A activity, dividend policies, and type and timing of debt.

WWU submitted that GEMA had erred in both law and policy in that it had failed to provide an adequate cost of debt allowance for WWU. WWU submitted that GEMA had misdirected itself in law and failed to give effect to its financing duty, and, as a result, had determined an allowance for all GDNs [ie. Gas distribution networks] based on an average of the actual cost of debt of a group of companies in circumstances where the effect of that approach was to over-remunerate some companies while penalising others. WWU submitted that GEMA’s approach discriminated against WWU. WWU submitted that, in addition, GEMA had adopted an irrational and inconsistent policy of not taking account of derivatives in assessing the cost of debt.

WWU divided its appeal on this ground into three main sections, relating to alleged errors stemming from:

In the paragraphs below we briefly describe GEMA’s approach in RIIO-2, summarise the evidence that has been presented to us and present our assessment of whether GEMA’s cost of debt allowance for WWU was wrong.

WWU’s case is that GEMA has a statutory duty to secure the actual financeability of any particular licence-holder. It contends that this duty is embedded in GEMA’s principal objective.

For the reasons given below, we do not accept WWU’s submission that the duty to have regard to financeability requires GEMA to secure the actual financeability of particular licence-holders.

In our assessment, we consider first the structure of section 4AA GA86 before analysing the nature of the financeability duty itself.

When looking at the wording of section 4AA:

The Principal Objective is clearly set out in subsection (1) as being ‘… to protect the interests of existing and future consumers in relation to gas conveyed through pipes’.

Subsection (1A) sets out certain specific interests to be included in the definition of ‘interests of existing and future consumers’.

Subsection (1B) introduces an element of discretion on GEMA’s part in exercising its functions, in that it shall do so ‘… in the manner which the Secretary of State or the Authority (as the case may be) considers is best calculated to further the principal objective …’, factoring in the need to promote effective competition ‘wherever appropriate’.

Subsection (1C) directs GEMA to consider certain factors when deciding how to carry out its functions under Part 1 GA86 (also with a view to promoting competition).

Subsection (2) refers to the duties under (1B) and (1C) and directs GEMA, in performing those duties, to have regard to the need to secure that reasonable demands for gas are met and the need to secure that licence holders are able to finance the activities which are the subject of obligations imposed by or under (inter alia) the GA86 itself.

Subsection (3) refers to the duties under subsections (1B), (1C) and (2), and directs GEMA, in performing those duties, to have regard to the interests of a range of consumers.

Subsection (4) confers on GEMA the discretion, in carrying out its functions, to have regard to the interests of other utility consumers.

Subsection (5), which is explicitly subject to subsections (1B) and (2), directs GEMA to perform its functions in the manner best calculated to promote efficiency on the part of (inter alia) licence-holders and protect the public from danger.

It seems clear that the financeability duty does not have the same status as the Principal Objective in subsection (1). In our view, this is not so much because of the positioning of subsection (2) following the amendments made to the GA86 by the Utilities Act 2000, but rather because of its wording.

Subsection (2) contains matters to which GEMA must have regard in carrying out its functions in furtherance of the Principal Objective. Unlike the Principal Objective, however, the duty is to ‘have regard’.

We note that WWU focused on the words ‘need to secure’ in subsection (2)(b) and sought to minimise the effect of the words ‘have regard to’. We do not agree that that is the correct focus. The words ‘have regard to’ are contained in the main body of subsection (2) and describe the nature of the obligation on GEMA to consider the needs described in (2)(a) and (2)(b). Subsection (2)(b) does not say, either expressly or by necessary implication, that GEMA has a duty to secure the actual financeability of any particular licence-holder.

We therefore accept GEMA’s submission (see paragraph 14.44(a) above) that a duty to ‘have regard’ is not a duty slavishly to follow and that the statutory wording does not impose an obligation of result. We do not accept WWU’s argument that the words ‘need to secure’ in section 4AA(2)(b) lead to a different conclusion.

Contrary to WWU’s submissions, we consider that the use of a notional company approach does properly have regard to the need to secure that licensees are able to finance their activities, bearing always in mind GEMA’s principal objective of protecting the consumer interest. In short, we agree with GEMA that it creates strong incentives on the part of licensees to manage company debt prudently and efficiently. GEMA explained its thinking on the placement of risk as being:

As Ofwat explained (see 14.72 above), a notional company approach incentivises companies to secure efficient costs of finance and protects customers from the risks of companies’ financing decisions; the notional approach reflects the principle that companies and their investors are best placed to bear the risks associated with their borrowing choices. It is therefore unsurprising that the notional company approach has been used in a variety of regulatory contexts.

We asked WWU in its hearing to reconcile the elements of its case set out at paragraphs 14.20(a) and 14.24(a) above (that GEMA is not required to indemnify licence holders but that its financeability obligations are mandatory). WWU initially said that GEMA had enforcement sanctions available to it and would be expected to take action under the relevant condition. We did not find that a convincing explanation because adopting WWU’s proposed interpretation of the financing requirement would mean that GEMA would in fact be in breach of its duty if a situation were to arise in which enforcement sanctions became necessary.

WWU later appeared to accept that GEMA could be said to have discharged its financeability duty even in a situation where a licence holder ultimately could not finance its activities:

In its response to the provisional determination, WWU said that the two positions could be reconciled on the basis that GEMA was only required to put licence holders in a position whereby they were able to finance their activities. WWU then described the position of ‘being able’ as meaning ‘… by allowing them to recover all reasonable costs’. In paragraphs 14.182 to 14.183 we assess WWU’s notion of reasonably incurred costs and do not repeat that here.

We do not agree that the financeability duty requires GEMA to ensure that each licensee can recover all of the costs which it has reasonably incurred. Furthermore, as we have explained, there is in our view a sound reason for avoiding an approach which focuses on market rates, in that such an approach would not provide sufficient incentives to licensees to manage their debt costs efficiently (see 14.81 above).

We are not persuaded by WWU’s argument that GEMA must have regard to the individual circumstances of each company regulated by the RIIO regime. WWU contended that the reference to ‘licence holders’ in section 4AA of GA86 must be read to mean individual companies when setting the parameters of the price controls. Section 4AA sets GEMA’s Principal Objective and duties ‘…in carrying out [its] respective functions under this Part…’, and Part 1 of GA86 covers many aspects of GEMA’s functions in regulating the entire gas transmission and distribution system across England, Scotland and Wales. When section 4AA is read in the wider context of Part 1, it does not lend itself to the interpretation that section 4AA focuses on individual companies. On the contrary, we consider section 4AA to be a provision that sets the priorities and considerations for GEMA when governing the whole regulatory system.

WWU contended that if GEMA did not investigate licence holders’ individual circumstances it must be in breach of its public law duty. We do not agree with that characterisation. WWU’s argument is premised on its anterior argument that GEMA is required to have regard to the individual circumstances of each licence holder; but we have rejected that argument immediately above.

WWU pointed to the fact that GEMA used the actual circumstances of a company in setting the cost of debt when dealing with OFTOs. When asked about that point in GEMA’s hearing, GEMA described OFTOs as being ‘…single asset, fixed term contracts that are subject to competition’. GEMA had previously described OFTOs as being for a fixed period of 20 or 25 years. We agree with GEMA that the competitive nature of the OFTO auction process means that bids based on inefficient debt costs would be likely to be uncompetitive and so would likely fail. As a result, the efficiency of debt costs within the OFTO sector is ensured by a market process, and so does not require regulatory judgement from GEMA in order to protect consumer interests and ensure efficiency. That is very different to the regulation of monopoly energy networks with an ongoing licence to operate.

It is therefore clear that OFTOs are in a different position, where because of the competitive nature of the tender process, GEMA can legitimately take the view that the tendering companies’ costs of debt are set at efficient levels. That same logic does not apply in the case of the licensees subject to RIIO-2.

. . .

We note that section 4AA(1B) confers on GEMA a wide scope in decision making, ‘…the Authority shall carry out their respective functions under this Part in the manner which the Secretary of State or the Authority (as the case may be) considers is best calculated to further the principal objective, wherever appropriate by promoting effective competition…’.

In our view, to displace that scope and impose specific and mandatory outcomes in terms of licensees’ financeability would require clear and unambiguous wording. We do not think subsection (2) is worded in such a fashion as either to restrict GEMA’s decision making scope by compelling it to assess the individual circumstances of companies or, specifically, to compel GEMA to design a system that automatically passes through licence holders’ debt costs to customers and consumers.

Finally, we agree with WWU’s general point (paragraph 14.32 above) that, when interpreting any given statute, material relating to other statutes governing other sectors is not binding. In any event, we do not consider the references to other statutes are necessary for reaching a conclusion on this point.

For the reasons set out above, we conclude that WWU has not demonstrated that GEMA’s interpretation of its statutory duties was wrong.

In assessing WWU’s claims, we considered the impact of GEMA’s use of an index as well as the underlying analysis of actual costs that underpin GEMA’s choice of index used. We presented our analysis of this error with reference to the issues raised by WWU in its NoA, namely:

Inherent irrationality and basis in economic theory

WWU provided an extensive response to the provisional determination, noting that WWU strongly disagreed with both the assessment and the conclusions of the provisional determination on this matter. We note that WWU’s disagreement, in the main, focused on our interpretation and our weighing up of the evidence, as well as our interpretation of GEMA’s duties, rather than a failure to take material evidence into account.

As at the provisional determination stage of assessment, we have collated our assessment of the alleged irrational reliance on a cost of debt index into three main sub-categories:

Inherent irrationality and basis in economic theory

WWU raised a series of complaints in relation to the provisional determination assessment of this sub-section, which we address in the paragraphs below.

On the issue of the finance duty and the potential for competition with other duties, our assessment of the statutory duties under GA86/EA89 is set out above (see paragraphs 14.76 to 14.80).

On the issue of conflating efficiency and prudence, we continue to agree with GEMA that, in the absence of a statutory definition of efficiency, regulators can use their judgement to take a broader assessment of efficiency than simply ‘cost versus benchmark at the time of issue’. We maintain the view expressed in the provisional determination (see paragraph 14.144) that an average cost of debt within the sector (including when matched to an equivalent external benchmark) is an acceptable way to consider a number of factors when deciding an efficient and appropriate cost of debt allowance.

We disagree with WWU’s assessment that there is more scope for efficiency than acknowledged in the provisional determination. WWU highlights factors such as credit rating, capital structure and effort as ways to create debt efficiencies. It is our view that efficiency must be measured on a like-for-like basis. To claim that companies can achieve lower costs by, for example, having a higher credit rating, would seem to be ‘comparing apples and oranges’. If this was evidence of like-for-like efficiency, we would expect all energy network companies to have extremely high credit ratings to capture such efficiency benefits – this is clearly not the case. It appears evident that such decisions involve trade-offs rather than obvious net efficiencies.

Similarly, capital structures approaches, such as WBS schemes, may increase the level of debt burden that is sustainable at a given credit rating, but this appears to trade-off equity holders’ control of cashflows for higher financial returns – rather than increase the ‘efficiency’ of debt. In the case of WWU, the use of a WBS scheme does not appear to have led to more ‘efficient’ or lower debt costs . . .

We continue to view WWU’s interpretation of the ‘correct’ approach to setting debt allowances as being one where almost all financing approaches, however esoteric, would very likely be allowed and required to be funded by customers.

In its response to the provisional determination, WWU countered this view by stating that the requirement for companies to maintain a strong investment grade rating negates this risk. It is our view that WWU’s argument inappropriately conflates two issues – costs and financial resilience. Under WWU’s preferred approach, the investment grade requirement would largely constrain excess leverage rather than esoteric (or high cost) treasury approaches. It is a circular argument to suggest that all company debt approaches must be funded (at the notional level of gearing) as long as they are executed at or below a benchmark price, and then claim that credit ratings will constrain choice. In a scenario where (effectively) all approaches must be funded by customers, there is likely to be little or no impact from treasury strategy on a company’s credit rating. The more likely impact on credit ratios would be if debt costs were funded at the notional level, but total debt costs were significantly higher due to much higher levels of gearing.

Looked at a different way, a regulated company taking out 100-year fixed rate debt denominated in Yen (at or below benchmark prices) may not cause an immediate credit rating issue, especially if a regulator is required to allow the recovery of these costs. However, funding such an approach for 100 years, regardless of prevailing costs of debt over that period, may still be far from in the best interest of customers. This range of arguments is similar to WWU’s statements refuting the CMA’s view that an (effectively) cost pass through approach would reduce incentives to appropriately manage debt costs. It is precisely in taking a broader assessment of efficiency and prudence (relative to the average cost) where such incentives are likely to feature.

WWU has also disputed our comparison to the approach taken in the water sector, primarily on the basis of a higher level of company and ownership concentration in the energy network sector. While we agree with WWU that company and ownership concentration is higher in the energy network sector, this higher concentration does not appear to be a contributing factor to WWU’s higher than average costs. As WWU acknowledges, its costs are exclusively a feature of its choices in relation to debt issuance strategy. WWU’s size (either in absolute terms or relative to other companies in the sector) is not a material factor in WWU’s actual costs of debt being higher than GEMA’s allowance. GEMA told us that size and other non-controllable factors did not skew its actual cost calculation or cost of debt allowance. As a higher level of concentration versus another sector does not appear to cause material issues in GEMA’s assessment, or be the cause of WWU’s actual cost differences, we do not find it to be an error in GEMA’s approach.

WWU has questioned the consistency of GEMA’s historical approach, and what we meant by companies being in a position to ‘know the rules’. Our view is that, on the basis of the evidence submitted, GEMA’s historical approach is sufficiently stable for companies to have an appropriate and broad sense of future debt cost remuneration approaches without GEMA constraining future decisions. Perhaps more importantly, WWU has failed to persuade us that GEMA’s pre-2007 approaches to calculating its cost of debt allowance would have given any reasonable expectation that future licence modification exercises would result in company-specific cost recovery.

WWU disagreed with our assessment that, in a system where customers can ‘win’ or ‘lose’ due to non-identical debt costs at different companies, the fairest system is for everyone to be exposed to the same cost of debt allowance (as they are with the cost of equity allowance). At provisional determination, we noted that there are pros and cons to different approaches in this area, but on balance we considered that the ‘average’ approach is consistent with GEMA’s duties (see paragraph 14.149). It remains our view that this approach has the benefit of ensuring no customer is faced with extreme outcomes resulting from decisions taken by a single local monopoly that they cannot substitute away from.

In its response to the provisional determination, WWU has made several mentions of the fact that Cadent enjoys lower costs of debt which WWU argues is the result of more recent funding, and WWU specifically focuses on the fact that this means that Cadent’s customers are currently paying ‘much more than they need [to]’. WWU has noted that the total cost is the same in both the ‘aggregate cost’ and ‘company-specific’ approaches, and has argued that it would be more rational for Cadent’s customers to pay less in RIIO-2. The logical implication of WWU’s argument is that customers in other regions should pay less, funded by WWU’s customers paying more. While this is a plausible approach, on balance we continue to view it as fairer to all customers that they are not generally exposed to the specific financing choices of their regional monopoly energy network.

We take this view on the basis of financing approaches being fully in the control of management and owners of the networks. This is not the same as regulators making adjustments for facts that were largely outside of the control of management, for example due to the relative size of the company. We note that in RIIO-2 there are already differences in the allowance that companies receive, such as the lower allowance for SSEN-T and the 6 bps higher allowance received by WWU (see paragraphs 14.11 and 14.12), based on circumstances outside of the control of management.

Our conclusion here is based on a general principle and does not preclude a regulator matching allowances to specific circumstances such as these. Customers may also face different bills as the result of localised operating costs, as these again are most often the result of structural differences between regions that are outwith the control of management teams.

In practical terms, if GEMA were to follow WWU’s request for (efficiency tested ex-ante) actual costs at the notional structure, then WWU’s customers would be faced with higher bills for no other reason than their energy network company historically took a differentiated debt issuance approach that has subsequently become unprofitable. We consider that the outcome that WWU appears to be promoting would result in the imposition of higher costs for its own customers to fund lower costs for customers elsewhere in the UK, and we do not view this approach as superior to the approach taken by GEMA in RIIO-2.

Skewed results and unlawful discrimination

WWU raised a series of complaints in relation to the provisional determination assessment of this sub-section, which we address in the paragraphs below.

WWU has complained that it does not understand what the CMA meant by ‘suitably similar companies’ within our provisional determination assessment. At paragraph 14.153 above, we state that:

We are unclear as to WWU’s confusion with regard to this point. For clarity, it remains our view that regulated monopolies in the same sector and subject to the same price control regime are suitably similar for the purposes of assessing appropriate costs of debt within that sector. For the purpose of this exercise, we disagree with WWU that a regulator must consider GD, GT and ET companies separately.

WWU has also argued that all costs of debt approaches involve management decisions and should only be assessed on an ex-ante basis. WWU noted that the CMA should not judge WWU’s approach with the benefit of hindsight. As discussed extensively above, it is our view that the regulator can take a number of factors into account when assessing debt, and it is far from clear that sole reliance an ex-ante assessment of price is sufficient. In addition, the CMA is not judging WWU’s decision with hindsight – that is not the function of this appeal process. WWU has appealed against a licence modification and, having considered the evidence and arguments, we are of the view that in introducing that modification GEMA was not in error.

WWU has reiterated its argument that GEMA’s acceptance of a single debt allowance and the failure to treat WWU differently amounts to unlawful discrimination. As described above, we disagree with this assessment. GEMA has taken steps to address what might otherwise have been potentially discriminatory factors when calibrating its cost of debt allowance, as evidenced by the exclusion of SSEN-T’s lower costs as a result of its recent growth and the additional allowance awarded to WWU, SGN Scotland and NGN as a function of size. As WWU has clearly stated, its debt approach and subsequent costs are exclusively the result of its own decisions, not structural factors. As a result, we find no reason why WWU should require further company-specific adjustments to its cost of debt allowance, and conclude that GEMA has committed no error in this regard.

WWU has also stated that it is impossible for any company to issue debt in line with the GEMA allowance or the ‘prudent treasury approach that issues debt incrementally and at appropriate tenors’ as described in the provisional determination. We disagree with WWU’s assessment on three grounds.

Retrospective imposition of an interest rate policy.

Many of WWU’s points on retrospective imposition of an interest rate policy are adequately covered by the assessments above. Here, we specifically note that WWU reiterated that GEMA’s approach to the cost of debt allowance had not been as consistent as suggested and that neither the approach to the cost of debt nor the future path of interest rates could have been predicted ex-ante. We disagree, and consider that the evidence presented by GEMA showed sufficient consistency around debt allowance calculation. It may not be in customers’ best interests to have one fixed cost of debt calculation policy, as market conditions may change over time. Conversely, and as discussed above, there has been no evidence presented that would suggest that WWU could have reasonably expected the outcome of RIIO-2 to be their actually incurred and company specific costs being covered in full (at the notional level of gearing or otherwise).

We agree with WWU that the future path of interest rates was unknowable at the time. Despite this truism, WWU chose to, effectively, lock in its costs for several decades despite it being highly likely that the absolute level of debt allowances would change in each price control. Once again, we pass no judgement on this decision – it is inescapable however that WWU must take responsibility for it, as we expect that it would have if such a decision had proven to be financially advantageous.

WWU flagged that the CMA had noted that, due to the long nature of the interest rate ‘fix’ within WWU’s debt strategy, WWU may still benefit relative to future allowances. WWU has argued that the CMA has failed to acknowledge that the opposite scenario is also possible, and that WWU’s position could continue to deteriorate versus the average of the sector and/or future allowances. We acknowledge this, but would point out that if this is a genuine concern for WWU, then it is within WWU’s power to unwind the current treasury strategy and/or corporate structure. WWU has told us that it has already prepaid debt that was not yet due in order to improve financeability, but also that further restructuring of the derivative portfolio would not be viable as issuing new debt would ‘absorb leverage headroom’ while issuing equity ‘is not justifiable due to the low returns that such action would achieve’. As a result, we consider WWU’s decisions in this area to be influenced by the potential costs and benefits of restructuring, rather than the technical viability of such actions. We are very clear in our view that these issues and decisions sit entirely with WWU, and that individual company financing decisions are not the concern of GEMA when calculating the RIIO-2 price controls.

GEMA’s alleged irrational reliance on a cost of debt index - Overall conclusion

In the paragraphs above we have again considered WWU’s arguments and evidence relating to GEMA’s alleged irrational reliance on a cost of debt index. As at the provisional determination stage of this appeal, we do not consider that WWU has provided sufficient evidence that GEMA’s use of average actual costs or an equivalent debt index was wrong in theory or in practice. As a result, we conclude that GEMA’s reliance on a cost of debt index was not irrational and did not lead to GEMA’s cost of debt allowance being wrong.

WWU submissions on GEMA’s alleged irrational failure to take account of derivatives

WWU NoA

WWU submitted that GEMA’s approach to derivatives was essentially to treat them as company-specific management decisions which were entirely at the risk of equity investors and so could safely be disregarded by regulators. WWU submitted that derivatives should not be regarded separately from the underlying debt to which they relate, but must be viewed as intrinsic to a company’s debt financing. WWU submitted that GEMA had disregarded derivatives when they were a relevant factor which it was required in law to take into account when determining the cost of debt.

WWU’s submission on the reasons for this error are summarised as:

WWU also submitted arguments and evidence that GEMA’s objections to including derivatives were invalid. These are summarised as:

WWU’s Reply to GEMA’s Response

WWU submitted that it rejected GEMA’s implication that derivatives were too difficult or complex or variable to be subject to appropriate regulatory assessment, or that differences in the way they are used across different companies makes them unsuitable for regulatory consideration. WWU submitted that the first point was a simple error of fact, and that the second point was impossible to understand in a context in which networks adopted widely different approaches to financing strategies as a whole and that GEMA has said that these were matters it was best placed to manage.

WWU’s RFI Response

In response to questions from the CMA regarding the impact of WWU’s decision to use derivatives instead of index-linked bonds in 2007, WWU noted that one of the factors affecting WWU’s cost of debt was the 2007 decision to fix a base rate in RPI-real terms with RPI swaps with maturity dates in 2023 and 2035-39.

WWU submitted that it had estimated the (CPIH-real) cost for the RIIO-GD2 cost of debt that WWU might have been able to achieve if it had issued an index-linked bond instead of entering swaps contracts in 2007 to be 2.5% to 3.4%. WWU submitted that its expected RIIO-GD2 cost of debt and derivatives (including transaction costs) of 3.2% fell within the range, highlighting that the decision to use derivatives with nominal debt instead of natural index-linked debt was not a key driver of WWU’s GD2 efficient cost of debt at notional leverage.

GEMA submissions on its alleged irrational failure to take into account derivatives

GEMA’s response to WWU’s appeal against the irrational failure to take account of derivatives is summarised as:

GEMA submitted that although approaches to derivative use vary across licensees, the use of cross currency swaps to return foreign currency liabilities to GBP is the one area of commonality. It is for this reason that GEMA had taken into account the post-swap GBP equivalent costs of foreign currency issuance in its calibration. GEMA stated that the derivatives that have caused WWU’s costs to deviate most substantially from allowances are largely bespoke ‘bond style’ inflation derivatives, which are the least common and least well understood. GEMA submitted that they are also typically the most credit intensive for the bank counterparty because they involve an expectation of the bank being a net payer in the swap in the early years and a net receiver in the latter years. This can be viewed as ‘lending through the swap’ and is the reason these swaps typically attract much higher credit charges from banks than the other forms of swaps.

GEMA submitted that it disagreed with the efficiency assessment of WWU’s debt and derivatives conducted by Oxera. GEMA submitted that, contrary to the simplified approach taken by Oxera to assessing whether WWU’s derivatives were undertaken at market rates, it was GEMA’s view that valuing derivatives properly would require the entire trade history and all related flows (including all historical cash movements on both pay and receive legs) to be discounted at the rate indicated by the appropriate yield curve at the precise time of the derivative trade. GEMA also stated that such an exercise would also require an assessment of any subsequent restructurings, and that it was not simply a case of looking at the rate on the derivative at each five-yearly interval as suggested by WWU. In GEMA’s view, such an exercise should not be necessary if it is accepted that a trailing average of an external benchmark calibrated to average actual debt costs provides a reasonable allowance. GEMA further stated that even taking a simplified approach, its checks had highlighted that, given the bespoke nature of derivatives, it was difficult to make comparisons and assess if they had been incurred at market rates. GEMA stated that it does not consider its decision not to include most derivatives in its calculations to have removed the ability of companies to make choices. Rather, GEMA is of the view that derivatives are not intrinsic to the need to finance a company’s activities, they represent choices, the consequences of which should be borne by shareholders. In addition, GEMA stated that WWU’s argument that GEMA was informed in 2005 of its plan to use derivates is ‘not relevant’. GEMA submitted that network companies do sometimes update GEMA about financing or structural decisions they are taking but that it is not GEMA’s role to advise companies on their treasury strategies, pass judgement or approve or reject financing plans. GEMA stated that financing decisions are for companies and their boards and GEMA awareness does not convey implicit approval or otherwise.

GEMA submitted that although it stands by its position that derivatives should not be included in the calibration, it did collect detailed data on these instruments, and presented the results of the calibration testing both including and excluding derivatives (and intercompany loans). According to GEMA’s modelling, on an aggregate basis across GD&T, these costs are expected to be covered by the allowed return on debt.

[this is the gist of what was set out]

In its submissions to the CMA, GEMA relied on payments made by WWU to its shareholders and asserted that WWU had chosen to make these rather than address what GEMA considered to be WWU’s relatively high costs of debt. GEMA put forward figures regarding payments to shareholders and submitted that these payments were higher than assumed by the notional company in RIIO-1.

GEMA stated that these payments to shareholders had been made despite the company facing increasing pressure from a derivative position that was building in terms of negative mark to market (and principal accretion payments) as real rates continued to fall. GEMA submitted that breaking derivatives over time would have involved a cost (as they had a negative mark to market given rate movements since fixing date) and may have meant less would have been available to distribute to shareholders to date. In GEMA’s view, these choices had been made by WWU and had contributed to the current situation which was that the cost of these derivatives was having a negative impact on their credit quality, financial resilience and potential to distribute to shareholders in future years. GEMA stated that not providing WWU with a specific allowance to cover these costs was not only in line with its duties, but was a fair and reasonable decision of a stable and predictable regulator.

We conducted our provisional assessment of this issue in relation to two main sub-arguments – the first an assessment in principle, the second an assessment in practice. We considered the following questions:

Is it irrational to exclude derivatives in principle?

We agreed with the view expressed by WWU that derivatives are a generally accepted and widely used tool within corporate treasury departments. This is especially true if derivatives are used to replicate instruments such as index-linked debt, which are useful debt instruments in a regulatory framework. Such debt may not always be available from the markets in the quantities or calibrations required – leading companies to synthetically create them using derivatives. We also noted WWU’s evidence that GEMA does count some derivatives in its assessment of actual costs, such as those relating to currency swaps and in the pricing of OFTO cost of debt allowances. As a result, we would agree that in order to achieve the most comprehensive view of the actual cost of debt incurred in the energy sector, it would be beneficial to also count incurred derivative costs.

Conversely, GEMA contends that it has set (both in RIIO-2 and previously) an appropriate debt allowance that can reasonably be achieved using standard debt instruments. As such, if companies choose to use derivatives it must be to their advantage and should not require additional compensation. This stance is rational and defensible. WWU has not provided any evidence that this is not the case, and as discussed above, WWU’s debt (excluding derivatives) has a cost which is below GEMA’s allowance.

GEMA has also noted that companies, including WWU, have used derivatives that are specifically designed to move cashflows between periods rather than to be economically equivalent to standard debt (in terms of interest rate profile). We agreed with GEMA that the use of such instruments would potentially increase the obligation on a regulator to assess the merits and appropriateness of a large number of derivative contracts in order to count only the appropriate derivatives in the context of calculating the cost of debt. In particular, we noted that this raises legitimate concerns that companies might be in a position to report derivative costs which differ from the true underlying financing costs.

GEMA indicated that it counted derivatives related to currency swaps as these were commonly used by companies, and did not have some of the esoteric or high-cost characteristics of other types of derivatives. GEMA also argued that derivatives that fixed interest costs were acceptable in OFTO auctions as these were priced on a competitive basis (helping to ensure the ‘efficiency’ of the instruments used) and fixing costs via derivatives was appropriate in relation to contracted levels of revenues. We agreed with GEMA that there is sufficient difference between the specific scenarios highlighted by GEMA and the general assessment of the cost of debt, and that it is open to GEMA to include these derivatives in the relevant areas of its analysis, whilst focusing on the cost of bond debt when designing an index for the cost of debt.

We consider whether GEMA was wrong to include derivatives within the measure of interest for tax clawback purposes at paragraph 16.97 to paragraph 16.106

In making our assessment, we noted that all ’standard’ derivatives should have an NPV zero at the point of deployment (ie are deployed as a ‘fair bet’) and that GEMA has provided evidence that companies could, at least on average, work within its cost allowances or capital structure without derivatives. As a result, we provisionally concluded that GEMA was not wrong to apply an approach to indexation that excludes derivatives.

Does the exclusion of derivatives lead to GEMA’s cost of debt allowance being too low?

Turning to the question of whether the exclusion of derivatives leads to GEMA’s cost of debt allowance being too low, we considered the evidence provided as to what the impact of this decision has been.

As noted in paragraph 14.207, WWU has presented evidence that its actual costs of debt in this price control (3.2%) is broadly similar to the cost that would have been the case if it had instead taken out long-term index-linked debt in 2007 (an estimated range from 2.5% to 3.4%).

GEMA has presented evidence that suggests that the type of derivatives used by WWU are expensive relative to the derivatives used elsewhere in the sector, but also that WWU’s costs are higher as the result of WWU’s 2007 decision to lock-in long-term interest rates (and view that matches WWU’s evidence at paragraph 14.206). In addition, GEMA has provided evidence which demonstrates that, when measured at an aggregate level, the sector’s actual cost of debt including derivatives is not significantly different to the measure excluding derivatives – and that both ‘actual’ measures sit below the index-based allowance used by GEMA.

As a result, we provisionally concluded that the practical impact of GEMA’s decision to exclude derivatives from its calculation is immaterial. It is clear that WWU’s costs are the result of the type and intended consequences of the derivatives used by WWU (specially to fix real interest rates for a multi-decade period), not the use of derivatives in principle. As above, we concluded that WWU is not being discriminated against through GEMA’s decision not to include derivatives when setting the cost of debt in this price control. Rather, WWU’s higher costs are the result of risk explicitly taken as part of its financing strategy. As in relation to the use of an index, there is no evidence that GEMA was historically willing to count derivatives in its assessment of actual costs, and thus WWU was aware of GEMA’s approach when it decided to follow its strategy.

We provisionally concluded that WWU should not be entitled to relief on appeal simply because its chosen strategy has not proven to be profitable. GEMA has presented evidence that the inclusion of all derivative costs would not have had a material impact on its estimate of the sectors actual debt costs, and both approaches (including and excluding derivatives) give a result that is lower than the cost of debt allowed by GEMA. As a result, we disagree with WWU’s assessment that GEMA’s decision not to include derivatives led to a cost of debt allowance that was too low.

GEMA’s alleged irrational failure to take into account derivatives – Our provisional conclusion

In our provisional determination we provisionally concluded, for the reasons set out above, that WWU had not demonstrated that GEMA had irrationally failed to take account of derivatives.

In making our final assessment of this cost of debt sub-issue, we note that GEMA has broadly agreed with the assessment conducted for the provisional determination and supported its conclusions, other than the recommendation that it may be useful to count derivative costs in the future. WWU provided an extensive response to the provisional determination, noting that WWU strongly disagreed with both the assessment and the conclusions of the provisional determination on this matter. We note that WWU’s disagreement, in the main, focused on our interpretation and balance of the evidence, in particular the CMA’s view of WWU’s derivative portfolio, rather than a failure to take material evidence into account.

As at the provisional determination stage of assessment, we have collated our assessment of the alleged irrational reliance on a cost of debt index into two main sub-categories:

Is it irrational to exclude derivatives in principle?

WWU has disagreed with the assessment of its derivative portfolios implied in the provisional determination, and has variously asked the CMA to clarify what it means by ‘standard’ swaps, ‘esoteric’ instruments and ‘underlying financial costs’. We view much of this semantic debate as irrelevant to the basis of WWU’s appeal. As in the previous sub-section, our role in this appeal is not to judge or opine on the nature of WWU’s particular financing strategy. Rather, it is to assess whether GEMA’s cost of debt allowance was wrong when considering the issues raised in WWU’s appeal.

Conversely, GEMA has reiterated its view that it should not be required to include derivative costs in any future assessment, as it believes that a ‘snapshot’ of derivative costs does not provide the same clarity on overall funding costs as would be expected from debt instruments held on balance sheets.

It remains our view that, in theory, it may be useful to count some derivative instruments when calculating an average actual cost of debt for a regulated sector. The most useful derivative instruments to count would seem to be those that are used to synthetically replicate debt instruments, such as index-linked debt – particularly when such approaches are used when useful debt instruments such as index-linked debt are not readily available in the size or tenor required. However, this approach may not be practical or desirable if companies use either more complex derivatives or use derivatives for other purposes – and assessment of each derivative may place an inappropriate burden on regulators.

WWU’s own description of using derivatives to ‘defer income from GD1 to GD2 and accelerate costs from GD2 to GD1 by altering certain payment and receipt legs on a relatively small number of RPI swaps’ is clearly not an example of synthetically replicating a debt instrument that was not otherwise available. As WWU stated, these actions were ‘taken to improve financeability for GD2’, and as a result the cost of these instruments cannot readily be seen as a proxy for debt costs. We disagree with WWU’s view that such actions are equivalent to WWU deciding to ‘buy back bond debt before its due date’ as both ‘had the effect of accelerating interest costs’.

In our view, buying back debt uses shareholder funds to reduce the future cost of debt, while WWU’s description of ‘altering certain payment and receipt legs’ is more likely to result in the movement elements of the debt costs between periods for the benefit of cashflow management, with no overall reduction in debt burden and no reduction in leverage.

WWU also argued that GEMA’s exclusion of derivatives in general leads to one uniform interest rate and inflation profile of the notional efficient operator which is not followed in practice by any network. We disagree with WWU that this is either strictly the case or, looked at more broadly, is an irrational approach.

On the basis of this assessment, we agree with GEMA that there are complications to consider when choosing whether and how to count derivatives within an actual cost of debt calculation – particularly if such instruments are used for reasons other than to synthetically replicate standard debt instruments such as index-linked debt. The challenges in interpreting this data mean that regulators must use judgement when choosing whether and how to apply this information within the estimation of an appropriate cost of debt allowance. In this case, there is evidence that derivative costs within the energy network sector may not be representative of the costs of debt we are looking to assess (as opposed to other costs, such as risk management) and so we do not consider their exclusion from GEMA’s calculations to be irrational. In the paragraphs below we consider whether their exclusion led to a cost of debt that was too low in practice.

Does the exclusion of derivatives lead to GEMA’s cost of debt allowance being too low?

While in the section immediately above we conclude that exclusion of derivatives is not wrong in principle, we must also make a final assessment of the more practical question of whether such exclusion led to GEMA’s cost of debt allowance being too low in this price control.

In response to the provisional determination, WWU have argued that derivatives are vital risk management tools and that GEMA’s allowance cannot be achieved (in terms of following the financing strategy implied by the index) without derivatives. This argument appears flawed.

WWU has also stated that the CMA has mischaracterised the nature of its appeal in this area by assessing that ‘WWU is not being discriminated against through GEMA’s decision not to include derivatives’. While we consider our provisional determination assessment of whether excluding derivatives was discriminatory to be a valid consideration in determining whether GEMA’s decision was wrong, we concur with WWU that ‘it is not WWU’s case that Ofgem’s [GEMA] decision not to include derivatives led to a cost of debt allowance that was too low’.

As noted in the provisional determination assessment, GEMA has provided evidence which demonstrates that, when measured at an aggregate level, the sector’s actual cost of debt including derivatives is not significantly different to the measure excluding derivatives – and that both ‘actual’ measures sit below the index-based allowance used by GEMA. WWU has not disputed this fact with new evidence (subject to the procedural comments below). As a result, we continue to conclude that the exclusion of derivatives did not lead to GEMA’s cost of debt allowance being too low.

The alleged irrational failure to take into account derivatives – Overall conclusion

In the paragraphs above we have again considered WWU’s arguments and evidence relating to GEMA’s alleged irrational failure to take into account derivatives. As at the provisional determination stage of this appeal, we do not consider that WWU has demonstrated that GEMA’s exclusion of derivatives was wrong in theory or in practice. As a result, GEMA’s exclusion of derivative costs was not irrational and did not lead to GEMA’s cost of debt allowance being wrong.

As a result of the assessments described in the paragraphs above, we determine that GEMA was not wrong in its approach to, or estimate of, its cost of debt allowance. We have received insufficient evidence to demonstrate, and are not persuaded, that GEMA has failed correctly to interpret or to give effect to its financing duty, that GEMA was irrationally reliant on a cost of debt index nor that GEMA irrationally failed to take account of derivatives in its cost of debt allowance.

WWU has taken out two sorts of derivatives, namely:

Interest rate swaps

Inflation swaps

Interest rate swaps

We understand WWU has entered into interest rate swaps where it exchanges one series of interest payments on a defined principal amount with another series of interest payments. The most common type of such swaps is to exchange fixed interest for floating interest. These swaps will be specified as per paragraph 16.6.

Inflation swaps

WWU has entered into a number of swaps which, when combined with its bond debt, have the effect of creating an obligation to pay interest on an index-linked basis, ie interest that increases and decreases with the level of underlying inflation. This form of inflation-linked financing is commonly used by network companies, as price-controlled revenues also increase and decrease in line with inflation.

This form of derivative is comprised of three elements of ‘interest’ on the notional principal sum:

One of the characteristics of index-linked bond debt is that the inflation-linked part of interest payable is added to the principal during the life of the bond, and therefore is only paid out to holders of that debt when the principal is repaid at the end of the debt term. We understand that the kind of inflation swaps used by WWU to a certain extent replicate this feature. The element of interest payable related to inflation, as per 16.12(b) above, is only settled every few years, typically every three or five years. Although the inflation-linked element of interest on these bonds and derivatives is not necessarily paid (ie settled) each year, they still represent a cost in those years. As a result, normal reporting conventions apply as set out in paragraph 16.16 below with this interest being reported as a finance charge within the relevant periods and the sum owed shown as an accrual in the balance sheet. This process of rolling up interest is called interest accretion.

In summary, firms can use inflation swaps, in conjunction with debt they have issued with a nominal coupon rate, to synthetically create index-linked debt. Network operators can use this index-linked debt to match the nature of the revenue stream they are awarded under their price control, which is also linked to inflation, and therefore better match revenues received under the price control with payments out to holders of debt.

Up until relatively recently most firms only reported interest relating to the current period on the derivatives they held within their accounts. Both interest payable and receivable, including any elements linked to an inflation index, that was incurred or earned in respect of the period would be reported as a financing charge within the profit and loss account. Likewise any amounts due or payable in respect of interest charges at the balance sheet date would be reported as a debtor (accrued income) or accruals (an accrued expense).

For firms that report under International Financial Reporting Standards derivatives came on balance sheet in 2005. For other UK firms that requirement came in with the introduction of Financial Reporting Standard 102 (FRS102) in 2016. Bringing derivatives on balance sheet has implications for reporting of the finance charge within the profit and loss account. As a result, the finance charge may need to include an adjustment, so that changes in market value of the derivatives from one balance date to the next are appropriately reflected. This adjustment is called the fair value adjustment and relates to interest payable / receivable in relation to future periods (ie beyond the balance sheet date) in respect of derivatives held by the firm.

In the price control period before RIIO-1, GEMA identified the need to put in place a tax clawback policy. The concept of a tax clawback policy has also been applied by other regulators. This section sets out the process by which and the legal instruments through which GEMA designed and implemented its tax clawback policy pre RIIO-1 and throughout RIIO-1 and RIIO-2.

The process by which GEMA re-calculates annually each licensee’s allowed revenue by updating the forecast values set at FD for actual outturn data is called the Annual Iteration Process (AIP). This process is not specific to tax clawback.

GEMA implements its tax clawback policy via two documents which are incorporated in the licence by reference, and which it collectively terms ‘Price Control Financial Instruments’:

Pre RIIO-1 (GDPCR1)

Following a consultation process, GEMA set out its policy on tax clawback by way of an Open Letter in 2009 (2009 Open Letter).

According to the 2009 Open Letter, the intention was to claw back from licensees the revenue benefit they obtained from lower tax costs as a result of high gearing. The clawback would be triggered when in any year:

The 2009 Open Letter explained that, when both of these conditions were satisfied, GEMA would clawback the tax benefit which resulted from the difference between actual and modelled interest costs in that year.

The 2009 Open Letter noted that the clawback adjustments for the then current price control period would be made at the next Price Review, ie RIIO-1.

RIIO-1 price control

GEMA enacted its tax clawback policy within price control licence conditions for RIIO-1, which ran from 2013 to 2021. GEMA referred to the 2009 Open Letter in the Handbook for GD1, one of the Price Control Financial Instruments that has the same status as a licence condition. It did so by the following means:

2015 correspondence

WWU wrote to GEMA in June 2015 seeking guidance, amongst other things, as to the correct treatment of the inflation expense on RPI derivatives for the determination of actual interest for the purposes of tax clawback.

GEMA replied in July 2015 (2015 Letter) advising that:

Developments in reporting of financial performance under RIIO price control

GEMA has a long-established regulatory reporting mechanism for network operators which are referred to as the Regulatory Instructions and Guidance (RIGs).

On 2 August 2018, GEMA announced by way of an open letter that, following a review, it wanted to see a step change in how financial performance under RIIO price controls was reported by network operators. A key proposal was to require additional reporting to provide insight into the impacts of each network operator’s level of gearing, cost of debt, and actual tax payments on its returns under the price controls which, GEMA explained, current reporting did not easily allow. GEMA labelled this reporting initiative Regulatory Financial Performance Reporting (RFPR). Guidance and templates it was currently developing, GEMA explained, would include the accounting standards that it expected companies to adhere to, and establish common and consistent methodologies in how key financial data (such as the actual cost of debt, gearing and tax liabilities) was reported.

GEMA set out that it considered a key reporting requirement for each network operator was its Return on Regulatory Equity (RoRE)– including and excluding finance and tax; and based on actual as well as notional gearing. Network operators would also be required to reconcile their RFPR reports to their own:

On 13 March 2019 GEMA sent an open letter to gas distribution network operators stating that it was now proposing for comment modifications to the existing reporting regime (the RIGs) for gas network operators to include the RFPR requirement. GEMA explained that if it implemented the changes, they would take effect in April 2019, meaning that data submitted in July 2019 (in respect of the 2018/19 reporting year) would use the RIGs as modified.

On 30 April 2019 GEMA issued a direction to licensees, including WWU, which provided a link to the following documents:

We refer to the first of these documents as the RFPR Decision.

One instruction for completion within the RFPR Decision was, in respect of Tab R7 Financing of the accompanying template, the following:

RIIO-2 price control

GEMA re-enacted its tax clawback policy within price control licence conditions for RIIO-2. GEMA referred to the 2009 Open Letter in the Handbook for GD2, one of the price control financial instruments.

PCFM Guidance

On 12 April 2021 GEMA consulted on the introduction of a new Associated Document, namely the RIIO-2 PCFM Guidance.

The purpose of that guidance was as follows:

This PCFM Guidance was intended, amongst other things, to cover the annual adjustment to allowances for tax clawback.

The RIIO-2 Decision

As part of the decision to implement price control licence conditions for RIIO-2, GEMA imposed the PCFM and the Handbook for GD2 (which are an integral part of Special Condition 8.1 of WWU’s licence). Taken together, these two documents, along with the PCFM Guidance, aim to set out both GEMA’s tax clawback policy and the mechanism through which tax allowances are clawed back.

We find it helpful to articulate the errors alleged by WWU for the purposes of our assessment as follows:

The principal argument of WWU’s appeal, as set out at paragraph 16.79, is that it is logically incoherent for GEMA to adopt one position regarding derivatives for the purposes of setting cost of debt allowances and another position for the purpose of tax clawback. These are, however, two distinct purposes. The appropriate treatment of derivatives in each case are separate questions where different objectives come into play; it follows that it may be justified for GEMA to adopt differing positions in relation to these separate questions. We set out our assessment for the purpose of setting WWU’s cost of debt allowance in WWW Head A: Cost of Debt (see Chapter 14). We set out our assessment for the purpose of tax clawback below. We refer to the different objectives of the cost of debt calculation and tax clawback calculation at paragraph 16.98 below.

As set out in paragraph 16.77, we frame our assessment of this alleged error under the question ‘Was the tax clawback policy wrong because it included derivatives within the measure of interest for tax clawback purposes?’.

In considering whether GEMA’s policy to include derivatives within the measure of interest for tax clawback purposes was wrong, we consider the following three questions, which, in our view, are relevant steps in the logic as to why GEMA included derivatives in tax clawback:

Whether it is normal business practice to include derivatives within financing/ financing costs

Tax clawback calculations rely on a measure of the financing costs of debt ie interest. In our view, interest on debt derivatives are as much a part of the measure of interest as the interest on primary debt instruments for the following reasons:

As a starting point when considering what comprises interest of a regulated firm such as WWU, there is therefore a solid foundation for including interest related to both debt and any associated derivatives. This is in practice also WWU’s position: it accepts that derivatives could be included but says that GEMA is wrong to do so while excluding the same derivatives when setting the cost of debt allowance. We therefore consider it is normal business practice to use derivatives to alter the expected interest rate profile on debt taken out, and for the resulting amounts owed (or owing) at period ends to be reflected within the balance sheet and the resulting interest payable in financing costs within the profit and loss account.

Whether it is wrong to seek to assess the position of the individual firm rather than a notional firm

We consider that, for the purposes of a calculation to adjust the tax allowances of an individual firm to reflect the actions of that firm, it is appropriate to seek to assess the position of that individual firm. That is the whole point of the tax clawback policy.

We recognise that there may be questions on how ‘actual’ firm and ‘notional’ firm are defined, particularly for firms with complex financing structures. However, in our view it is clear and not in dispute that the tax clawback mechanism is based on assessing the financing costs of an ‘actual’ firm.

We therefore consider it not wrong for GEMA to seek to assess the actual financial position of the individual firm rather than the costs of a ‘notional’ firm.

Whether it is wrong to seek to measure interest based on the firm’s actual financing strategy rather than on an alternative notional financing strategy

If the above two points are accepted, and given that we do not consider that WWU is seeking to dispute those points, then the alleged error is whether GEMA was wrong in how it calculated the interest incurred by the ‘actual’ firm for tax clawback purposes. In particular, GEMA calculated interest based on the ‘actual’ financing strategy, including derivatives, whereas WWU said GEMA should calculate interest on a basis which excluded derivatives, which we consider could be described as a ‘notional’ financing strategy. WWU’s position is that GEMA’s approach is wrong, as it is inconsistent with the position on the cost of debt.

For the purposes of measuring interest payable by the actual firm for the purposes of tax clawback policy, it is not necessary to consider some of the broader questions around the definition of interest, such as whether it was efficiently incurred. This contrasts with the position for measuring interest for the purposes of preparing the evidence base used to set cost of debt allowances, which is intended to provide an allowance for a notional company. The objective of the calculation of tax clawback is to calculate and claw back for consumers the tax benefits that the actual firm has gained from the financing actions it has taken to the extent that they differ from the notional company, whether taken efficiently or inefficiently.

As described above, it is therefore evident that the calculation of tax clawback needs to be based on actual financing, in that it is based on the consequences for tax payable of the differences between actual and notional financing decisions. Given that we also consider that derivatives are a normal part of financing strategy, and indeed that WWU has not disputed this, this would imply that the tax clawback policy would also include derivatives.

The tax benefit to a firm from deducting interest is a function of its actual interest charge, which will reflect actual financing, including derivatives, rather than the notional financing used in setting the cost of debt. As a result, the tax benefit will also be affected by the firm’s choice of actual debt instruments, including the mix of floating and fixed rate debt and the use of bank and bond debt, as well as any related derivative instruments. There are therefore a number of differences between the financing assumptions used in measuring the ‘actual’ interest deductible for tax purposes and the measure of interest used to set the cost of debt allowances. Other than derivatives, WWU has not indicated that the other differences in financing policy between the actual and notional firm should be excluded when calculating interest for the purposes of tax clawback.

We recognise that there was evidence received during this appeal that the ‘actual’ interest under GEMA’s definition for tax clawback may still differ from the ‘actual’ interest applied for tax purposes, in particular for WWU’s derivatives. However, this does not change the question of whether it is right as a general principle to include or exclude derivatives in calculating the interest used for tax clawback, which was WWU’s ground of appeal.

WWU’s arguments in response to our provisional determination

In its response to our provisional determination, WWU pointed out that, whilst obligations arising from interest rate (real or nominal) derivative contracts might be presented as a financial liability within financial statements under FRS102, that did not mean that such derivatives provided debt finance. Rather such derivatives were used by WWU to achieve a target interest rate profile. WWU went on to note that GEMA agreed that interest rate derivatives were neither intrinsic to debt financing nor required for capital raising purposes.

WWU further submitted that the sole purpose of the tax clawback policy was to deter excess leverage. As its derivatives were not debt instruments, derivatives were, WWU argued, not relevant to assessing the extent of a firm’s leverage. It was therefore wrong for us to state in our provisional determination that, because derivatives were part of normal financing strategies, derivatives must be included in clawback. The fact that, WWU contended, derivatives were Net Present Value (NPV) zero contracts ab initio fundamentally distinguished them from debt. That, WWU argued, necessarily meant that the tax clawback policy could not apply any incentive power (whether negative or positive) to derivative actions ab initio.

We agree with WWU that derivatives are generally not a means by which external investors provide finance to businesses. Derivatives such as those held by WWU are instead used to alter the interest rate profile of the primary financial instruments (ie debt) that WWU holds. As such they are part of the financial toolkit with which network firms are able to service their debt obligations.

However, the evidence does not support WWU’s contention that the sole purpose of the tax clawback policy was to deter excess leverage. As stated in the 2009 Open Letter the tax clawback policy ‘claws back from licensees the revenue benefits they obtain from lower costs as a result of higher gearing.’ We note that including interest relating to derivatives would, all other things being the same, result in a calculation for the tax liability that more closely reflects what the individual firm would be required to pay. Therefore the treatment actually adopted by GEMA more closely reflects the implementation approach set out in the 2009 Open Letter than that proposed by WWU. We therefore do not consider that GEMA erred in adopting this approach.

In summary, we conclude that GEMA was seeking, in applying tax clawback, to identify a measure of interest for an actual financing structure. We consider that, for companies that choose to enter derivative contracts, the ‘actual’ financing structure would normally include derivatives as well as primary debt instruments. On that basis, we conclude that, rather than being incoherent by including derivatives, GEMA was in practice taking into account the different objectives it faced when defining a measure of interest for the purpose of, on the one hand, the cost of debt and, on the other, tax clawback. We therefore conclude that GEMA was not wrong to take a measure of actual interest which includes derivatives for tax clawback purposes.

A legitimate expectation may arise in circumstances where a public decision-maker changes, or proposes to change, an existing policy or practice. A legitimate expectation will only be created if a public authority has given a promise or adopted a practice which represents how it is going to act in a given matter or area. The promise or practice relied upon must give rise to a representation which is clear, unambiguous and devoid of any relevant qualification.

In the seminal case Ex p Coughlan, the Court of Appeal held that:

We agree with GEMA that any expectation that would have arisen as a result of the 2015 Letter would have been defeated by the RFPR Decision. On 13 March 2019, GEMA notified its intention to consult on certain proposed changes to the RIGs, including within the consultation document the note ‘We would expect Net Interest Per Regulatory (RIIO-1) definition to include all inflation derivative payments that attract tax relief (because that is the definition used for tax clawback)…’. WWU responded to the consultation on 11 April 2019. Its response noted that the draft RIGs ‘appears to conflict with what we have previously been advised by Ofgem’, which shows it was aware of the clarification GEMA was intending to make to its approach regarding tax clawback. On 30 April 2019, GEMA published the RFPR Decision updating the RIGs, which included the language cited above. The notification of GEMA’s intention to consult on certain proposed changes to the RIGs, followed by the consultation, in which WWU participated, and the publication of the resulting Direction was sufficient to remove any legitimate expectation that would have been created by the 2015 Letter (to the extent such a legitimate expectation actually arose).

In response to our provisional determination, WWU argued that a legitimate expectation created by means of a clear and unequivocal promise given by a regulator to a licence holder could not be ‘defeated’ merely by making a public statement that was at odds with the promise. We disagree with WWU’s characterisation of the RFPR Decision (and its prior consultation process) as being merely ‘a public statement … at odds with the promise.’ Whilst GEMA might have handled the matter better than it did, we conclude that the circumstances described in paragraph 16.125 above are such that WWU ought reasonably to have understood that no reliance should be placed on the 2015 Letter following the RFPR Decision (to the extent that such a legitimate expectation actually arose).

WWU also disagreed with the way we had characterised its legitimate expectation claim in the provisional determination. WWU submitted that it had not relied on a claim of legitimate expectation with regard to the forward-looking GD2 period. WWU therefore contended that our provisional conclusions concerning its legitimate expectations ‘looking forwards’ were academic and unnecessary. We note that WWU submitted to us in its NoA that the 2015 Letter established a legitimate expectation in law as to the continuation of the policy set out in it as part of the price control arrangements. Thus, WWU clearly contended that its legitimate expectations extended into GD2. In the absence of a request by WWU to withdraw that part of its appeal, it is appropriate for us to rule upon it.

In any case, to the extent that any legitimate expectation was created by the 2015 Letter, we also agree with GEMA that it would not be unfair for it to depart from the position described in such letter. We consider that GEMA’s decision to depart from such position for RIIO-2 did not meet the threshold of being so unfair that it amounted to an abuse of power. Indeed, GEMA’s statutory duties required it to apply the approach to tax clawback which it considered best met those duties and its objectives.

For the above reasons we therefore do not find that GEMA breached any legitimate expectation on the part of WWU by including interest on derivatives within the measure of interest for tax clawback purposes in its RIIO-GD2 decision.

We agree with GEMA that the issue of recovering tax clawback from before RIIO-2 is a matter for the RIIO-1 close out process as it relates to the consequences of transactions and events occurring before the outset of RIIO-2. Therefore, the decision on this close out process, which has not yet been concluded in practice, is outside of the scope of the RIIO-2 appeal process.

We have therefore not reached a view as to whether there has been any breach of legitimate expectation in relation to the recovery of tax clawback from before RIIO-2.

We therefore determine that GEMA was, for the purposes of the RIIO-2 price control, not wrong to set the policy it did in respect of the measure of interest for tax clawback purposes. We also do not find that GEMA breached any legitimate expectation on the part of WWU by including interest on derivatives within the measure of interest for tax clawback purposes. Any breach of legitimate expectation on the part of GEMA in respect of periods prior to RIIO-2 is outside the scope of the appeal. Accordingly, we dismiss this ground of appeal.

The CMA erred as to the appropriate standard of review: the CMA misconstrued its appellate function, failing to apply the requisite scrutiny under section 23D(4) of the 1986 Act. The CMA wrongly took a supervisory approach to the appeal, akin to judicial review, and incorrectly applied an enhanced margin of appreciation to GEMA’s discretionary decision-making;

The CMA misconstrued the financing duty in section 4AA(2)(b) of the1986 Act by failing to consider the financial circumstances of individual licence holders;

The CMA erred in law in upholding GEMA’s cost of debt approach: (a) in using a “notional company” averaging approach, GEMA discriminated against WWU; and (b) GEMA unlawfully excluded derivatives (with one exception) from the cost of debt allowance;

The CMA erred in its approach to WWU’s tax clawback appeal: if (contrary to ground (iii)), GEMA was entitled to exclude derivatives in its cost of debt policy, it was irrational and inconsistent for the CMA to uphold GEMA’s inclusion of derivatives in its tax clawback policy.

Ground 1: The CMA erred as to the appropriate standard of review

Ground 2: The CMA misdirected itself with respect to the Financing Duty

Ground 3: The CMA erred in its approach to WWU’s cost of debt appeal

The breadth and difference within the sector: The “sector” included companies from different industrial sectors, for example National Grid Electricity Transmission Plc (“NGET”) and National Gas Transmission Plc (“NGT”), which undertake activities and have corporate characteristics quite different from much smaller gas distributors like WWU;

The cost of debt index was dominated by two corporate groups: The average indexed cost of debt across the “sector” was driven primarily by the actual costs of debt of National Grid which drove 47%, and Cadent which drove 23%;

The circumstances of each company were different: the “sector” included companies which incurred their debt at very different times, and this was a significant factor in the cost of debt for each company; and were factors outside of WWU’s control;

The approach adopted was arbitrary: The approach adopted by GEMA was largely based on historic outcomes for this disparate group of businesses, rather than determined prospectively, e.g. to reflect administrative guidance from GEMA. The effect of this was that it would have been impossible for WWU to have been able to predict with any reasonable degree of confidence what allowance would be given for cost of debt in the RIIO-2 price control period from April 2021 to March 2026, or to determine its financial strategy on that basis.

GEMA does count some derivatives in its assessment of actual costs, but irrationally excludes others;

it was inconsistent to exclude derivatives for the cost of debt analysis, but then include them for the tax clawback analysis: this results in an unfair asymmetry;

it was irrational for the CMA to accept that the use of derivatives constituted normal business practice -- a normal utility operator will use derivatives as part of its normal business practice: derivatives are a generally accepted and widely used tool within corporate treasury departments _-but then to place no weight on this when accepting GEMA’s submission that analysis of derivatives was so complex that they should be excluded from the cost of debt assessment;

it was irrational for the CMA to accept that it was sufficient for GEMA to have considered the impact of derivatives at a sectoral average level. CMA found that WWU’s case on derivatives made no difference when looking at matters on an average basis, but this ignored the fact that WWU’s individual position was severely prejudiced;

no assessment was made by the CMA as to whether WWU’s approach was imprudent or inefficient;

the CMA found that the “most useful derivative instruments” to include are “those that are used to synthetically replicate debt instruments, such as index-linked debt” but went on to say that WWU’s use of derivatives was “clearly not an example” of this. This was based on an error of fact: WWU’s derivatives were to a very large extent “used to synthetically replicate debt instruments” and reflected GEMA’s own contemporary guidance; and

the exclusion of “synthetic” and “esoteric” derivatives went too far. Furthermore, GEMA and the CMA were capable of analysing complexity and the CMA had the expert evidence before it from Oxera which would have enabled this - there was a section of the Oxera report on the exclusion of derivatives – and the CMA should have carried out further investigation itself. In addition, in the exercise of its regulatory role, GEMA has access to extensive information concerning individual licence holders, including annual data submissions that contained extensive information on debt and derivatives.

Ground 4: The CMA erred in its approach to WWU’s tax clawback appeal

Step 1 – Ofgem’s notional tax allowance for company X (excluding derivative costs)

Step 2 – Company X’s actual position (including costs of derivatives)

Step 3 – Clawback

Step 4 – The asymmetry problem

Ground 1: Did the CMA err in its approach to the appeal?

First, there must have been a mistake as to an existing fact, including a mistake as to the availability of evidence on a particular matter. Secondly, the fact or evidence must have been "established", in the sense that it was uncontentious and objectively verifiable. Thirdly, the appellant (or his advisers) must not been have been responsible for the mistake. Fourthly, the mistake must have played a material (not necessarily decisive) part in the tribunal's reasoning.

our starting point will be to consider the adequacy of GEMA’s chosen approach rather than considering which approach we ourselves might have chosen had we been in GEMA’s position. However, we agree that, in considering whether GEMA’s chosen approach discloses an error, we will consider its inherent merits including by comparing its merits with those of any reasonable alternatives advanced by the appellants. If, out of the alternatives available, we conclude that some alternatives clearly had greater merit than the solution chosen by GEMA, then we are more likely to be persuaded that GEMA has erred. On the other hand, where the alternative options each have competing pros and cons, and none is clearly superior, it will be more difficult to persuade us that GEMA has erred.

Ground 2: Did the CMA misconstrue the Financing Duty?

Ground 3: Did the CMA err in law in upholding GEMA’s cost of debt approach

when measured at an aggregate level, the sector’s actual cost of debt including derivatives is not significantly different to the measure excluding derivatives – and that both ‘actual’ measures sit below the index-based allowance used by GEMA. WWU has not disputed this fact with new evidence . . . As a result, we continue to conclude that the exclusion of derivatives did not lead to GEMA’s cost of debt allowance being too low.

(Emphasis added).

This was correct: WWU had not presented any new evidence as part of its representations. In the cost of debt report prepared by Oxera, on behalf of WWU, it was stated that it was “a coincidence that the estimates of the average cost of debt with and without derivatives are not too different”, but no further evidence was adduced. Further, in the representations on the provisional determination of the CMA, WWU stated that it challenged what GEMA had said and requested the CMA not to rely on it unless it was “properly interrogated”, but no further evidence was adduced.

WWU submitted that if the CMA was instead referring to the decision by WWUto defer income from GD1 to GD2 and accelerate costs from GD2 to GD1 by altering certain payment and receipt legs on a relatively small number of RPI swaps, that this was the exact opposite of lending through the swap.”

(Emphasis added).

The CMA well understood, therefore, that the deferral of income from GD1 to GD2 was not on all of WWU’s derivatives, or even a large proportion of them. Rather, it was “on a relatively small number”.

Derivatives can be used to shift financing costs and receipts from one period to another so we are not convinced that taking a snapshot of embedded derivatives at one point in time would give us an accurate picture of their costs or benefits over the long-term. An example of the ability to shift financing costs and receipts from one period into another is illustrated by WWU’s own recent swap reprofiling, . . . This particular reprofiling was undertaken to provide greater receipts in 2023-2026 (which may aid debt covenant compliance during that period) but it serves to illustrate that it would be very difficult for a regulator to accurately assess the correct rate to include in its calibration exercise. It would also be very difficult to determine whether this bespoke derivative was undertaken at market rates or whether there had been any value leakage. Valuing this derivative properly would require the entire trade history and all related flows (including all historical cash movements) to be discounted at the rate indicated by the appropriate yield curve at the precise time of the derivative trade. Banks typically have proprietary software for this task but it would be disproportionately time consuming, expensive and difficult for a regulator to replicate this, particularly months or years after the event.

Ground 4: Did the CMA err in its approach to WWU’s tax clawback appeal