Npower Direct Ltd & Ors v The Gas and Electricity Markets Authority & Ors

In June 2014, Ofgem made a reference to the CMA for an investigation into the energy market. On 24 June 2016, the CMA reported as to its investigation of the energy market. It identified a number of adverse effects on competition (“AECs”). One AEC was a lack of customer engagement within the market resulting in a significant financial detriment of £1.4 billion. One of the CMA’s recommended remedies was that Ofgem carry out a programme of identifying, testing (through, for example, market trials) and implementing measuresto increase customer engagement with the domestic retail energy markets and introduce a new Standard Licence Condition (“SLC”) requiring suppliers to comply with directions to carry out such market trials: see Ch 13 of CMA report, at paragraph 13.20.

Ofgem consulted by a letter on 19 October 2016 on the introduction of a new SLC and the proposed selection criteria for trials. It stated in its October consultation that “In light of the CMA’s compelling AEC findings, Ofgem considers that the introduction of proposed SLC 32A would be both a necessary and proportionate regulatory intervention. We also consider that this intervention is consistent with Ofgem's principal objective and general duties. In particular we note that the power of direction afforded by proposed SLC 32A licence condition will ultimately facilitate Ofgem decisions on the most appropriate ways of protecting the interests of existing and future consumers, including in terms of balance between promotion of effective competition and more direct methods of consumer protection."

Ofgem received responses including from npower. Ofgem then used its powers under s.11A EA89 and s.23B GA86 to implement the recommended modification on 30 January 2017, SLC 32A. It decided that selection for trial would be based on (1) whether the supplier had enough of the types of customer of relevance to the specific research question(s), and (2) whether the burden of the trial was “proportionate” to the particular supplier, taking into account its existing capabilities including its participation in previous Ofgem-led trials and its own testing and trialling work. It expressed the view that its revised criteria would provide “a fair and proportionate” method of selecting participants.

A letter of 30 January 2017 of Ofgem considered the impact of trials on suppliers quoting from the June 2016 report of the CMA. It was called ‘Implementation of Standard Licence Condition 32A: Power to direct suppliers to test consumer engagement measures – decision to make licence modifications’. A further letter of the same date was headed ‘Decision: Selection criteria for mandatory supplier testing of measures to promote domestic consumer engagement’. This included the following:

“Impact of trials on suppliers

A concern was raised by a large supplier that the selection criteria’s impact would be disproportionate because they would cause an unequal impact on certain suppliers, based upon potential customer losses and damage to their commercial interests. A mid-tier supplier raised a concern over how costs of trials could be spread over a smaller customer base, particularly if the characteristics of that customer base meant it could be selected multiple times. We have considered these points in the context of the CMA’s recommendation of the trials programme as part of the remedies to address the domestic weak consumer engagement adverse effect on competition (“AEC”). In its summary of the remedies, the CMA stated:

The above overarching feature of weak customer response, in turn, gives suppliers a position of unilateral market power concerning their inactive customer base. In relation to unilateral market power, our finding is that suppliers in such a position have the ability to exploit such a position…

The full report states:

The gains available to customers from promoting engagement are potentially high….We recognise that there will also be costs to suppliers of complying with the requirement to participate in the Ofgem-led programme (including, where used, RCTs) and implementing the resulting interventions. However, we note that our starting point for the proportionality of the Ofgem-led programme, and any individual decision subsequently taken by Ofgem in the context of the programme, is the scale of the detriment, which is very substantial.

The CMA recognises that Ofgem is already obligated to consider the cost and proportionality of all its activities.

In designing the programme, Ofgem will be required to assess the proportionality of the various stages involved in the programme, including the testing involved in each specific proposed measure.

To ensure proportionality, we will choose research questions for trials that we believe stand a good chance of seeing positive consumer outcomes. We will also ensure that the scale of the trial is no larger than necessary to obtain robust results. We have revised criterion 2, to clarify that the assessment of proportionality is linked to capability, which will include size of supplier in relation to the scale of the trial.

……..

3.

Burden of the trial is proportionate to a particular supplier

A few respondents called for the costs of each trial to be shared between Ofgem and the supplier(s) involved. We do not consider it appropriate for Ofgem to fund the trials in part or in full. Our role in this area will be to supply the research questions for trials, based upon desired consumer outcomes, and to provide guidance to suppliers on how to undertake robust trials. We would also monitor the progress and outputs of trials. None of this would be a substitute for suppliers using their own resource and expertise to carry out the specified trial.

In respect of the feedback on proportionality, we have already stated that we will take account of a supplier’s participation in previous Ofgem-led trials and its activities in the area of the research question. This is to encourage suppliers to take proactive steps to engage consumers who are not currently taking advantage of the competitive market. To take a supplier’s actions into account, there would need to be considerable alignment with the measures planned for Ofgem-led trial and any supplier-led testing would need to have been undertaken in a robust manner.”

A further letter of Ofgem of the same date to suppliers set out the reasons for adopting SLC 32A and summarises the consultation responses. It stated:

“As part of the remedies to address this AEC, the CMA recommended that Ofgem establish a programme of work to provide customers with measures to prompt them to engage. It considered that suppliers have the ability to engage their existing customers through the regular communications they send to them, but they are likely to face limited incentives to do so.

As such, the CMA recommended that this Ofgem-led programme should identify, test (through randomised controlled trials (RCTs), where appropriate) and implement measures to provide domestic consumers with different or additional information with the aim of promoting engagement in the domestic retail energy market. The CMA further recommended that we introduce a licence condition (SLC 32A) to require suppliers to participate in this programme, to ensure the programme’s effective implementation, for which it provided the draft licence condition wording.

SLC 32A has two key features: (i) the power to direct licence holders to undertake testing and trialling; and (ii) information gathering powers. Both are with regards to us being able to use RCTs or other forms of testing to identify effective domestic consumer engagement measures.”

SLC 32A is time limited and will expire on 31 December 2022: 32A.9.”

This letter then went on to consider commercial impact and referred to a concern of respondents

“that an intervention being trialled might prove to be highly effective at prompting engagement and, as a result, a supplier loses a significant number of consumers either to another supplier or from a particular tariff type.The impact of this can be managed by ensuring that the burden of trials is fairly spread across suppliers, that the size of a trial is appropriate and proportionate – ie a trial with 200,000 customers should not be pursued when a trial with 40,000 customers is sufficient, and that by selecting suppliers of a sufficient size proportionate to the scale of the trial. However, we do recognise the potential for some commercial impact and would not consider this was a valid reason for a supplier to avoid or be excluded from a trial. Of course, if a proposal was proven to be successful in promoting engagement, it is highly likely that we would seek to put this in place for all suppliers to facilitate competition.”

Ofgem published these selection criteria, following consultation with stakeholders in October 2016, also on 30 January 2017. These were not challenged at the time, and nor were the terms of SLC 32A themselves nor Ofgem’s power to modify the licence to include the condition. Ofgem considered the proportionality of adopting both the condition and the selection criteria, based on the consultation (inter alia), before doing so in accordance with its duties under e.g. s.3A(5A) EA89.

The relevant background in relation to the CMA’s energy market investigation (“EMI”), the Final EMI Report and the views and experience of the CMA regarding randomised controlled trials (“RCTs”) such as the Autumn trial is set out in a witness statement of Mr Prevett, which has been adduced by the Intervener, the CMA. As noted

In the skeleton argument on behalf of npower, it is stated that there is no challenge to (i) the history and content of the CMA Report (and ancillary documents) following the EMI (2014-2016), (ii) the finding of an AEC of weak customer response by the CMA, and (iii) the implementation of SLC32A by Ofgem. The skeleton of npower at paragraph 3 says as follows:

“This is agreed context. npower adopt the CMA’s description, and has made clear that it does not challenge the finding of an AEC in the terms set out by the CMA or the introduction of SLC32A by Ofgem as an important part of a package of remedies. npower agrees with the expressed importance of Ofgem using SLC32A to address the AEC. SLC32A is an important and necessary provision to help address an identified consumer detriment (an AEC identified by the CMA) and is not in question.”

The details of the above matters have been set out in some detail above. They show an important history before any of the events of which complaint is made. It is the identification of an AEC by the CMA in its report of 24 June 2016. That was the existence of numerous inactive customers who did not consider alternative packages which might lead to a substantial reduction in the amounts paid for gas and electricity. Although issue is taken with the amount of £1.4 billion of annual customer detriment caused by customers remaining on the SVT, no issue is taken to the AEC and to its being a problem from a competition standpoint.

Ofgem consulted on selection criteria by a letter on 19 October 2016 on the introduction of a new SLC and the proposed selection criteria for trials. Such consultation included with npower as to how to deal with the AEC and as to the creation of new SLC 32A. Its purpose was to compel suppliers to participate in a trial to test the market with a view to finding remedies to address the domestic weak consumer engagement AEC. Ofgem considered that SLC 32A would be “both a necessary and proportionate regulatory intervention. We also consider that this intervention is consistent with Ofgem's principal objective and general duties.” In its recommendation of 30 January 2017, Ofgem was alive both to the need to deal with the AEC by trials and to the fact that this was inevitably going to cause a burden on the supplier(s) chosen for the trial: Ofgem would do what it could to provide a fair and proportionate way of selecting participants.

It was clear from letters of Ofgem of 30 January 2017 that the trials were going to cause difficulties to the particular supplier(s) selected. They would not like any selection because what was proposed would inevitably wake up the customers who had been passive and cause them directly or indirectly to leave the selected supplier with a consequent loss of income. They would also be concerned that other suppliers should have been selected instead of or before them. This would be despite the emphasis throughout the creation of the response to the AEC on Ofgem acting in a fair and proportionate way.

The inevitable and realistic concession at paragraph 3 of npower’s skeleton argument forms not simply background in this case, but provides a problem to the nature of the challenge. Given that the trial would inevitably have such consequences to selected suppliers, it is necessary for any challenge to be by reference to the actions of Ofgem at the point of and immediately leading up to the challenge in the summer of 2018. The difficulty for npower is to show that that which occurred was not simply the corollary of the very difficulties that would ensue from the decision to tackle the particular AEC, the implementation of SLC 32A and the selection of a supplier which could be, and in the event, was npower.

Since then, Ofgem has run several trials in accordance with its broad powers under SLC 32A. In February 2018, Ofgem ran its first “Collective Switch Trial”. This involved disengaged customers being provided with information about an opportunity to switch providers based on an exclusive tariff which they are offered by a third-party intermediary on behalf of other suppliers who bid to provide that tariff. The first trial involved 50,000 of Scottish Power’s customers, during which 22% of the customers switched to a cheaper tariff. Mr Fisher’s third witness statement refers to the fact that the trial “achieved a consumer response eight times higher than that of the control…group and considerably greater than that achieved in other trials previously conducted under SLC32A”. Mr Fisher at paragraph 9 said the following about this trial:

“It is possible that the collective switch model as trialled in the first collective switch trial is a particularly effective way of addressing the AEC identified by the CMA (lack of customer engagement) and the resulting £1.4 billion of annual customer detriment caused by customers remaining on the SVT (Standard Variable Tariff). However, it is also possible that there was something unusual about the incumbent supplier and/or the particular group of its customers trialled, the time of year or other factors which led to the customer response rate being so high. Given the size of the detriment identified by the CMA and the number of customers affected (up to eight million customers), it is essential that Ofgem tests whether the results are replicable and whether such a service could be run at levels that would enable efficient and timely communications with affected customers.”

In the course of the contest about proportionality, internal documents have been provided to npower of the period of June and July 2018 (which were not sent to npower

The key points are that the position of Ofgem as represented to npower was replicability of the Scottish Power and to understand whether it was scalable. In addition to paragraph 9 of Mr Fisher’s statement cited above, he amplified this in his third statement where he referred (at paragraphs 5-6) to the Scottish Power test achieving a customer response which was eight times higher than that of the control (or comparator) group and considerably greater than in other trial previously conducted under SLC 32A. The concern was therefore that the Scottish Power test was an outlier (for example due to characteristics of the customers tested) or whether it was due to the design characteristics of the intervention.

There is a complaint of npower about alleged delay on the part of Ofgem in relation to the disclosure of these documents (between 19 October 2018 and 15 November 2018) and a further document to which I shall refer below, namely a document of 20 July 2018. Attention is drawn to the fact that the former document of 14 June 2018 does not refer to scalability or to larger trial being required. That is first mentioned in the paper of 4 July 2018, whereas in a previous paper (not identified and not disclosed), the possibility of scaling up was regarded as important as was ‘generalisability’ (which seems to be another term for replicability). The submission of npower is that at this stage there was no concern about the ability of EHL to deliver on a larger number of subjects: on the contrary, there was a confidence about its ability to deliver “because we have a procurement route in place with the delivery partner.” It is therefore contended that the reason of scalability is ex post facto justification for the scale of the trials.

It is suggested that the wrong costs and benefits were taken into account. As regards suppliers, there was taken into account the cost of sending out letters, but not the more substantial costs of loss of customers and consequent revenue. Indeed, it is estimated by npower that the consequence of the trial of 100,000 customers will be about £30,000,000 per annum (its earlier estimate in a letter of 10 August 2018 was potentially up to £120,000,000 per annum if all customers switched). As regards consumers, there was taken into account savings (“reduced consumer detriment”), whereas it is said that this was not relevant to the robustness of the tests. According to the skeleton argument for npower, “it is a dangerous approach to consider that a trial (the primary value which must be obtaining robust data) is to be considered, as regards quantified financial value, on the basis of a hoped for actual outcome. SLC 32A should be not policy implementation measure, it is said, but to obtain data to inform future policy interventions.”

Ofgem decided to run a second collective switch trial. It informed the “big six” energy suppliers of this on 12 July 2018. This paper refers to two suppliers, each providing 100,000 customers for the trial. It invited expressions of interest, according to the published selection criteria. npower replied on 18 July 2018 that they did not wish to be included in the trial and giving their reasons. It stated that they had participated in an earlier trial and were to be involved in the Disengaged Customer Database project.

The document to which the largest criticism is directed by npower is a document sent on 20 July 2018 and headed “Impact of the price cap on the Autumn CS trial options note”. It is claimed by npower at paragraphs 31-45 of its skeleton that this document shows “an astonishing approach to misleading customers and the supposed trade-off between honesty to customers and obtaining higher switching figures.” It states that that the key research question is to test replicability. An issue is whether the impact of a trial will vary according to the supplier. Further, the impact of the forthcoming price cap will be to reduce any saving to the customer because of the savings due to the price cap. The effect of the price cap will undermine the ability to replicate the Scottish Power test because the consumer will be looking at the impact of the price cap in addition to or instead of the impact of obtaining a better price from a different supplier.

Concern was expressed that the impact of the price cap is that there will be less people switching because of a belief that the price reduction is making it unnecessary to explore the market to find out if a further reduction can be obtained in that way. A potential con in respect of the trial would be that the “potential impact of the trial may be diminished and less people will switch as a result [of price cap]”. It was said that an improper purpose in these circumstances is that the purpose has become to procure consumer switching rather than to test robustly whether or not it is in the consumer’s interest to switch: it may not be following the cap. The aim should be that the trial should inform about a certain result, not that it should achieve a preferred result.

The gravamen is summarised at paragraph 45 of the submissions of npower as follows, namely “The only explanation for Ofgem continuing with the npower trial in these circumstances is that it is doing so solely or materially for the purpose not of testing anything (more specifically, replicability) but rather to obtain a result: consumer switching. This is clearly beyond the scope of SLC32A – it is a measure to trial consumer engagement measures to inform future policy interventions, it not a regulatory tool to achieve a (direct) result.” I shall return to this in the consideration below of the improper purpose issue.

In his third statement, Mr Fisher at paragraph 12 stated that in parallel with the announcement of 12 July 2018, enquiries were made of Energyhelpline to see if it had the capacity to deal with that quantum of two waves of 100,000 customers. By 20 July 2018, this possibility had been ruled out as it was clear that there was no capacity to deal with a trial of more than 100,000 customers at one time. This then led to proposed trials of cohorts of 100,000 customers consecutively rather than simultaneously. At paragraph 13 of his third statement, Mr Fisher identified changes in deliberations from three trials of 100,000 each with two trials to be before the price cap (pursuant to the Domestic Gas and Electricity (Tariff Cap) Act 2018 which received Royal Assent on 19 July 2018). The change was then to two waves of 100,000 customers to be conducted in the autumn trial, one to be before the price cap and one after. When it became apparent that the price cap level would not be until at the earliest early November 2018, it was concluded that the first test should be in 2018 before the price cap, and one following the introduction of the price cap in early 2019.

Ofgem subsequently selected the two suppliers, one being npower, to take part in this trial in accordance with its published selection criteria under SLC 32A. On 31 July 2018 it notified npower of its decision to select them. It explained that in fact all “big six” suppliers had run earlier trials and npower’s trial had had the smallest customer population (1200) and that work on the Database project was unlikely to coincide greatly with the trial. This was elucidated in a document of Ofgem of 18 September 2018 referred to below: it stated that the other suppliers had participated in up to two trials, with the suppliers providing data for over 160,000 customers (the largest being over 250,000). Npower had thus far been exposed to very limited customer losses through Ofgem’s trialling activity. On 31 July 2018, npower indicated by email that they “recognise the statutory nature of the process (i.e. it is not voluntary) and hence we will get on with the trial.” According to paragraph 15 of Mr Fisher’s third statement, communications took place with the second supplier who worked “collaboratively” with Ofgem and Energyhelpline in agreeing a revised timetable.

Scalability was not referred to in the document of 14 June 2018, but it was mentioned by 4 July 2018 (and apparently in an earlier paper). In any event, the document of 14 June 2018 was much less detailed than subsequent documents, and so there is little, if any, significance as to how little there was in the document. That there were discussions within the team at this stage is apparent from the third witness statement of Mr Fisher referred to above. Hence, the reasons provided in August 2018 were not an ex post facto rationalisation, but reflected the discussions and decisions which preceded the notice of intention of the Decision provided on 1 August 2018.

Following that, on 1 August 2018, Ofgem, giving the requisite one month’s notice (SLC 32A.2), issued the Direction to npower pursuant to SLC 32A, along with a covering letter explaining the basis of the trial and the reasons for the Direction. It provided the relevant background to the trial in terms of the CMA’s report and recommendations to Ofgem in terms of testing customer engagement measures. It is accepted by npower, and correctly, that the letter of 1 August 2018 was a draft Direction: see npower’s skeleton argument, paragraph 51 and the judgment of Zacaroli J at paragraph 4, to which more reference will be made below. On 31 August 2018, that is a month later, a letter in substantially the same terms was sent. That was the Direction.

The reasons included the following:

The letter went on to explain that npower would be required to identify “100,000 domestic customers” and take the various specified steps in relation to the trial outlined. The Direction itself was then enclosed, along with an indicative timetable with the first steps to take place on 3 September 2018.

The early concerns of npower were said to be because Ofgem chose to have a trial of 100,000 customers in respect of npower (one month’s notice of a Direction given on 1 August 2018) which was treated by npower as being disproportionate. There was a willingness to have a trial of 10,000 to 30,000 customers (response of 10 August 2018), even 50,000 customers (19 September 2018), but a trial of 100,000 customers was regarded as disproportionate. I shall now consider the correspondence in this regard.

On 10 August 2018, npower made a commitment to comply with the Direction, but requested that the number of participants on trial should be limited to 30,000 maximum (not 100,000). On 17 August 2018, Ofgem responded by email as to the rationale for 100,000 rather than 30,000, following a call with npower. The call on 16 August 2018 was between Ofgem (particularly Ed Charlish) and Alan Hannaway and Nicholas Murphy of npower “regarding the collective switch trial and the plan for database. We did feed back to them through the meeting the rationale for performing the trial with 100k customers rather than 30k.” (referred to in the email of 17 August 2018). The email stated that it was discussed in the call that “The previous trial yielded encouraging results. One of the main objectives of the second trial is to prove that we can scale this up and service provider can organise a collective switch for a larger group of customers. Should a collective switch be rolled out as a steady policy intervention, you can appreciate that we deem it absolutely necessary to test it for this amount of customers.”

In response to a request for more detail, Ofgem replied on 20 August 2018 by email explaining further “why 100,000 was needed rather than something smaller.” Because of the “extremely positive” results of the earlier trial (with a switch rate of over 20%), Ofgem wanted to understand whether it was “scalable”, and needed “to understand two things”: (1) whether call centres can deal with the increase in the volume of customers they will need to interact with, (2) the market appetite for bidders on the collective switch auction at larger volumes.

The Direction took effect on 31 August 2018 (with the first steps of an indicative timetable to be taken on 3 September 2018) and Ofgem sent npower a copy of the Direction on that date, along with the covering letter setting out the reasons for it. The Direction was issued in finalised form materially identical to the 1 August 2018 Direction.

Npower responded with their representations on the Direction in email at 17.02 on 31

In the email of npower of 31 August 2018, as regards the legitimate aim of the trial, it was suggested that insofar as the rationale of Ofgem was commercial appetite for suppliers taking on large volumes on Commercial Switches, this fell outside the scope of the scope of the Energy Market Investigation. It was stated that this was confirmed in the SLC 32A consultation decision letter of Ofgem on 30 January 2017.

In early September 2018, npower complied with the first step required under the Direction. On 7 September 2018, Ofgem by Mr Fisher responded to the representations of npower including on validity, aims and proportionality. It responded as regards legitimate aim saying that the purpose was consistent with the scope and purposes provided for in SLC 32A: that SLC was to gather evidence in order to identify appropriate means to test effective ways of prompting consumer engagement in the domestic energy market. The “legitimate aim of this trial, is therefore in line with the aim of the licence condition under which the trial is launched i.e. to test (i) large scale measures, that will prompt market-wide consumer engagement and (ii) the capacity and capability of third parties, such as price comparison websites (eg Energyhelpline) to deal with large numbers of switching customers and potential implications from there.”

In this context, Ofgem stated that “we consider that npower therefore may have misunderstood the exchange of emails on purpose of trial and we would therefore note for the avoidance of doubt that the purpose of the autumn CST was not to test the commercial appetite for suppliers to take on large volumes in Collective Switches, or the operational capability of gaining and losing suppliers to maintain service levels for high gains and losses during Collective Switches (although we note that such matters may be relevant factors for consideration in the implementation of such measures.) The legitimate aim of the trial is to test the sustainability of a collective switch of a large number of inactive customers while prompting further engagement via informed choices i.e. by providing recommendations on even cheaper tariffs through Energy Helpline beyond the collective switch tariff.” Thus, it was not accepted that Ofgem acted outside its powers under SLC 32A.

Under the heading proportionality, Ofgem said on size “in order for the measure to be effective and appropriate to prompt consumer engagement through switching and informed choices, a trial population of 100,000 customers would be a reasonably sizable trial population to measure a large scale intervention. In this way, Ofgem will be able to measure the effectiveness of such an intervention on a wider basis than the first collective switch trial as well as confirming what is the market appetite for bidders on the collective switch auction to undertake larger volumes of customers. We will also be able to trial what is the industry’s capacity to deal with large numbers of potentially switching customers, for example, whether call-centres could deal with the increase in the volume of customers they will need to interact with.”

There was a meeting on 12 September 2018 between Ofgem’s officials and npower to discuss mainly the number of the trial population. On 14 September 2018, there was a further meeting between Mr Fisher, Ofgem’s Deputy Director, Consumer Engagement and npower: see Fisher (1) paragraphs 30-32.

Despite this, npower was not willing to proceed with issuing 100,000 letters (wanting, by its letter of 19 September 2018, to limit trial to 50,000 instead, stating that it was not satisfied regarding proportionality). Absent agreement, Ofgem indicated that the Direction would be enforced if 100,000 letters were not issued by 20 September 2018 as per Direction.

On 18 September 2018, Ofgem sent a paper on proportionality to npower. It states that having run the first collective switch trial with one incumbent supplier and at a small sample size relative to the relevant customer population “it was not clear whether this was replicable or deliverable at greater scale. Hence the need to run a second trial.”

It referred to matters relevant to impacts / alternatives. It bore in mind the fact that npower had 737,549 non-prepayment customers who had been on the SVT for 3 years or more. In selecting npower, it took into account the fact that npower had conducted a trial (on 1200 customers) and therefore been subject to limited customer losses whereas other suppliers had participated in up to two trials with some of the suppliers having provided data on over 160,000 customers. It was explained why a wider intervention than the Scottish Power trial was required consistently with the information provided in August 2018, namely (1) to measure the appetite by suppliers to provide a competitively priced exclusive tariff (the appetite might not be there if the number were not large enough), and (2) to measure the feasibility of a service provider to scale up its operations from operational and cost-related perspectives. Calculations are identified which had been undertaken to ensure that the trial would be sufficient to detect the impact of intervention. These calculations have not been subjected to detailed analysis or criticism by npower in its submissions or evidence. The conclusion of Ofgem was that a trial population of 100,000 customers would be an appropriate and reasonably sizable group to measure the effectiveness of a larger scale intervention.

On 19 September 2018, npower confirmed that they would not comply with a trial of 100,000 and would only comply with 50,000, and set out potential grounds of legal challenge for Direction. By 20 September 2018, npower had not complied with the stage of the Direction of sending the necessary letters to the selected customers for the trial required by that date.

This led to the Provisional Order made on 24 September 2018 (“PO”) under s.25 EA89 and s.28 GA86 to secure compliance npower with the Direction forthwith and require any non-compliance to be remedied by 26 September 2018. On 26 September 2018, Ofgem published its Statement of Reasons for making the PO pursuant to s.38A GA86 and s.49A EA89. By 27 September 2018, npower failed to comply with the PO.

Ofgem then issued and served a claim and N244 in the Chancery Division for an urgent injunction under s.27(7) EA89 (and s.30(8) GA86) (Claim No FS-2018-000014) supported by the first witness statement of Mr Paul Fisher setting out the background and exhibiting the key documents.

On 1 October 2018, Paul Fisher of Ofgem in the injunction proceedings commented about the value of sequential trials on account of price cap coming into effect in January 2019 (inter alia). He reiterated that “Ofgem wishes to establish whether the results of the first collective switch trial …are replicable as well as scalable.”

On 2 October 2018, npower issued an application to quash the provisional order under s.27 EA89 and s.30 GA86. The matter was heard by Zacaroli J on 4 October 2018 and judgment was given on the next day 5 October 2018, who granted an injunction, ordering npower to comply with the PO. On 8 October 2018, npower complied with the injunction by issuing customer letters.

On 11 October 2018, an application was issued by npower for determination of a “preliminary procedural issue”, on jurisdiction on the papers, namely whether legality of the underlying Direction could be challenged in the statutory review of PO under

On 19 October 2018, npower sought an explanation, and disclosure of documents, in relation to various matters (consecutive trials, effects of price-cap, and various detailed matters including the disclosure of a contract with a third-party, Energyhelpline).

There were then various communications with Lang J between 23 and 25 October 2018 as follows:

On 31 October 2018, npower issued the Judicial Review Claim. An order of Murray J was made on the same date linking the Judicial Review to the section 27 proceedings and ordering a linked-up hearing for 28-29 November 2018 with expedited consequential directions. Ofgem sought to set aside that order on the basis that after 12 November 2018, there was no longer urgency. On 2 November 2018, an order of King J was made with a revised expedited timetable with npower to file grounds in the

On 8 November 2018, npower renewed their request made on 19 October 2018 raising four detailed questions and asking for various documents.

On 9 November 2018, the CMA applied to intervene in the linked claims. On 14 November 2018, Yip J ordered that the application to intervene be determined by the trial judge, and making directions that CMA’s submissions and evidence be served by 21 Nov 2018.

As regards the trial, pursuant to the PO, savings letters were sent to trial customers between 12 and 21 November 2018 and reminder letters were due to be sent between 22 November and 3 December 2018. There were further steps for 11 December 2018 and 14 December 2018 and the completion date for the trial is due to be 7 January 2019.

The procedural history has become complicated and it has given rise to a number of issues which will be considered later in this Judgment.

As noted above, one of npower’s criticisms of Ofgem’s proportionality assessment is that “Ofgem has not ever assessed the costs to npower” of requiring its participation in the Autumn trial (paragraph 56(iii) of SFG. It appears, from Dr Harris’ evidence

As pointed out when the CMA assessed the proportionality of the Ofgem-led programme it recognised that participation in the programme would entail costs to suppliers; hence the statement in the Final EMI Report that these costs should be taken into account in considering the proportionality of any testing and supplier participation: see Prevett (paragraph 33). Notwithstanding those costs, the CMA, noting that the starting point for the proportionality both of the Ofgem-led programme as a whole and for any individual decision subsequently taken by Ofgem is the scale of the detriment, found the Ofgem-led programme to be proportionate.

As noted above, the introduction of SLC 32A was itself the subject of a further assessment that included consideration of proportionality issues. This can be seen from the two documents dated 30 January 2017 referred to above, both accepting the potential loss of customers. that accompanied the decision in question, namely, (i) a letter headed ‘Implementation of Standard Licence Condition 32A: Power to direct suppliers to test consumer engagement measures – decision to make licence modifications’ and (ii) a further letter headed ‘Decision: Selection criteria for mandatory supplier testing of measures to promote domestic consumer engagement’. In both of these documents, the potential loss of customers by suppliers was accepted. In particular, the ‘Implementation’ letter stated, in Ofgem’s discussion of consultation responses:

“A second area of concern was that an intervention being trialled might prove to be highly effective at prompting engagement and, as a result, a supplier loses a significant number of consumers either to another supplier or from a particular tariff type.

The impact of this can be managed by ensuring that the burden of trials is fairly spread across suppliers, that the size of a trial is appropriate and proportionate – ie a trial with 200,000 customers should not be pursued when a trial with 40,000 customers is sufficient, and that by selecting suppliers of a sufficient size proportionate to the scale of the trial. However we do recognise the potential forsome commercial impact and would not consider this a valid reason for a supplierto avoid or be excluded from a trial. Of course, if a proposal was proven to be successful in promoting engagement, it is highly likely that we would seek to put this in place for all suppliers to facilitate competition.”

[emphasis added]

Thus, Ofgem specifically acknowledged the cost of which npower complains in its decision to introduce SLC 32A and explained that it did not consider the cost in question to be a valid reason to avoid or exclude any particular supplier from a trial. Rather, Ofgem decided that the way to address this issue was to ensure that the burden of trials was spread fairly across suppliers and their size was proportionate.

The reasons why Ofgem selected npower for the Autumn trial and the considerations that led it to reach the views that it did in relation to trial size are addressed in Ofgem’s evidence: see Fisher 1 paragraphs 12-28 and Fisher 3 paragraphs 4-15 of Fisher 3 (in which reference is made to a ‘Collective Switch options paper’ concerning the trial design) and Ms Moon’s evidence. It is to be noted that:

One of the aspects of the Autumn trial that npower alleges Ofgem has not evaluated adequately is its timing: see SFG, paragraph 56(ii). The complaints in relation to timing appear to relate primarily to the imminent price cap, on which the CMA comments separately below. However, before turning to that issue, the observations of CMA on timing more generally should be considered.

Mr Prevett (paragraph 53) expressed the view that progressing the testing of measures to promote consumer engagement is a matter of “critical importance”. The reasons for this included:

In short, where a potentially effective intervention has been identified (such as the collective switching that was the subject of the Scottish Power trial) then the testing of that intervention should progress expeditiously: the sooner there is a robust evidence basis on which to conclude that the intervention is likely to prove effective, the sooner that intervention can begin to be implemented and consequently, the sooner the features that have given rise to the CMA’s competition concerns in domestic retail energy markets can be addressed.

This is reflected in the timescales for the implementation of the Ofgem-led programme that the CMA expected in the Final EMI Report. As Mr Prevett remarks (paragraph 47), “The Final EMI Report made clear…that the CMA’s expectation was that Ofgem would begin work on the Ofgem-led programme and the Database remedy promptly.”

Finally on timing, Dr Harris comments in his second statement (paragraph 70) that npower agrees with the CMA “on the fact that the market (and all relevant aspects) change over time.” As is clear from the evidence of Mr Prevett and from the Final EMI Report itself (see e.g. paragraphs 11.57-11.61 and 15.84), the CMA designed the Ofgem-led programme – in particular, the requirement that it be ongoing, with interventions being monitored for effectiveness and updated – in a way that was intended to take this into account.

Drawing the above together, the CMA’s view is that the prospect of a change in market conditions should not, by itself, be a sufficient reason for the Court to suspend Ofgem’s testing of potentially effective interventions. The factors in favour of timely implementation of the remedy weigh as heavily as ever and the inevitability of changes in the relevant markets has been factored into its design and assessment.

The general considerations in favour of timeliness outlined above lead to consideration of the final issue on which the CMA comments, namely, the price cap. Npower’s position on this is, as set out in Dr Harris (2) at paragraph 61 that “its imminent introduction makes a nonsense of the reasons that Ofgem had given for the need to carry out the Autumn Trial”. The suggestion is then made that this is invalidated by the price cap, the effect of which would invalidate the whole basis of the further trial. The case is that it would remove or reduce the incentive of customers to change suppliers. It would also remove any ability to replicate the Scottish Power test because the conditions would have changed. It would at a stroke reduce the customer detriment said to have been £1.4 billion by up to £1 billion.

Ofgem disagrees: Mr Fisher considered that the imminent price cap rendered the Autumn trial “not only helpful, but necessary” (see Fisher 3, paragraphs 29-35). The answers to the criticisms include that the AEC would not be addressed by the price cap in that large savings could still be made by switching which far exceeded those delivered by the price cap (paragraph 29). Further, the price cap was to be temporary (up to the end of 2020 with a possibility of extension to the end of 2023) during which it would be reviewed whether conditions were in place for effective competition for domestic supply contracts such that the cap could be lifted. In short, Mr Fisher stated that the price cap was a temporary measure to protect default tariff customers until the conditions for effective competition were in place (paragraph 30). The test was required before the introduction of the price cap (paragraph 33). Ofgem did not introduce the trial “days” before the price cap was introduced: the proximity occurred because of the delay in the process caused by npower which led to the PO and the injunction being obtained on 5 October 2018 (paragraph 35).

Reference here should be made to Mr Prevett’s statement at paragraphs 49-53 of his statement: the price cap (which is a temporary measure) is different in nature from the Ofgem-led programme. It will not address the demand-side market failings identified by the CMA. Thus, the price cap does not diminish the importance of making progress in the testing of interventions to increase consumer engagement. Indeed, if such testing does not progress then there must be a risk that the same market failings will remain after the price cap is lifted, i.e. that domestic retail energy markets will be no closer to a competitive state than they are now.

Dr Harris acknowledges that the price cap and the Ofgem-led programme are not substitutes (Harris 2, paragraph 61). That having been said, he does not address the fact that the need remains to tackle the demand-side failings that have given rise to the Domestic Weak Customer Response AEC, nor that it is Ofgem-led measures to promote consumer engagement that are expected by the CMA to perform this task.

Dr Harris alludes to the fact that the majority of the CMA’s Inquiry Group did not consider that it would be proportionate to intervene through a price cap in the period before its remedies package would take full effect (see Harris 2, paragraphs 62 and 6768). He seems to suggest that because the price cap has the potential to alter the dynamics in the relevant markets, this is a further reason to doubt the utility of an RCT now. In the CMA’s view, it is unwarranted and premature to draw the conclusion from the Inquiry Group’s concerns about the potential consequences of a price cap that no useful steps towards implementing a collective switching measure such as that tested in the Scottish Power trial could take place now.

In the light of the foregoing, it is necessary to consider the particular grounds of challenge made by npower.

In its skeleton argument, npower at paragraph 10 contend that Ofgem was unable to give reasons (either adequate or consistent reasons) to the extent of irrationality. At the heart of this analysis appears to be the words in the email of 7 September 2018 which have been selected and emboldened “for the avoidance of doubt that the purpose of the autumn CST was not to test the commercial appetite for suppliers to take on large volumes in Collective Switches”. This is said to be in contradiction with other correspondence or Mr Fisher’s statements. However, this ignores context and by selection of a part of one sentence, the terms of the email of 7 September 2018 itself.

As for context, Mr Fisher on behalf of Ofgem was responding to the suggestion in the email of npower of 31 August 2018 that the purpose of the trial was not within SLC 32A. Mr Fisher identified how the purpose of the trial was within SLC 32A which was not about testing market appetite.

As for being selective, npower has omitted to say that market appetite was stated to be a relevant factor for consideration in the implementation of the measures (as is set out

Thus, the email of 7 September did not contradict the Summary Grounds of Defence at paragraph 26 which stated that 100,000 customers were needed to understand whether call centres could deal with the increase in the volume of customers and the market appetite for bidders referred to in the email of 17 August 2018. The understanding of market appetite, albeit not the purpose of the trial was a relevant factor for consideration in the implementation of the measures.

It is suggested that there is no consistency about the reasons of replicability and scalability, thus giving rise to serious issues about whether the true reasons were being put forward. In my judgment, it is apparent that there was a desire to replicate the Scottish Power test: this is apparent from the evidence of Mr Fisher and from the internal documents of 4 July 2018 and 20 July 2018 as well as from the subsequent documents. The desire for scalability too is apparent from the same documents. The two were not mutually inconsistent. Replication did not necessarily mean that the two trials would be identical: by definition they could not be involving different suppliers and different characteristics of customers including in particular location. Far from finding that there was confusion and contradiction, there has been detailed consideration of the complex factors relating to the trials, and the sequence of correspondence internal and external shows a detailed consideration of the issues and the provision of full answers to the questions asked by npower.

I also reject the suggestions that the documents of 14 June 2018, 4 July 2018 and 20 July 2018 shed some negative light on the actual reasons and motivations of Ofgem. There was no obligation to provide the same at an earlier stage. They were provided further to disclosure requests dated 19 October 2018 and 8 November 2018. I do not accept that there is some sinister inference to be drawn from the fact that the documents were not produced more promptly at the time. There was a lot going on at the relevant time, and the gap in time from the request on 19 October 2018 to the provision of the letters as exhibits to Mr Fisher’s third witness statement does not indicate concealment or a lack of a duty of candour. Further, analysed in the way in which I have done above, the content of the letters was not, in my judgment, “deeply disturbing”. I accept the explanations provided by Mr Fisher in his third witness statement.

Further, I do not accept the criticisms made of these letters in the skeleton on behalf of npower. The criticised omissions from the letter of 14 June 2018 are by reference to the limited nature of the document and the team discussions since that time. The concern about scalability is apparent from the letter of 4 July 2018 where there was consideration of testing at greater levels. The analysis of this document in npower’s skeleton argument at paragraphs 23ff seems to me to ignore plain words in the documents about testing scalability. In the first page, referring back to an earlier paper, there are references to among other things to “generalisability” and to “possibility to scale up”. In the second page, it is stated that “the first trial arm is designed to gain insight into the possibility to scale up and whether the results can be generalisable.” There is not an obvious contradiction with the fact that they were confident that they could deliver the trials in the timelines because of a procurement route in place with the delivery partner.

Similarly, I do not accept the analysis of the document of 20 July 2018 (npower’s skeleton argument at paragraphs 31-45), to which npower attached their greatest criticisms. It was said to show “an astonishing approach to misleading customers and the supposed trade-off between honesty to customers and obtaining higher switching figures” (npower skeleton paragraph 31). An immediate difficulty of this submission is that when the allegations of improper purpose were made by npower in its skeleton argument and in the terms quoted above, Ofgem asked whether there was a suggestion of bad faith on the part of the officers of Ofgem. The response in the negative removes some of what has been alleged of any force. Npower says to this that Ofgem may have been mistaken due whether due to ineptitude or arrogance. Even if this is potentially true, the reality is that the concession about bad faith renders the criticisms of Ofgem more confined in the circumstances in which the criticisms attach. In fact, since the natural meaning of parts of the skeleton including the quotation in this paragraph is of bad faith, the concession is very significant.

In my judgment, the concession about bad faith is properly and responsibly made. Seen with this concession, the criticisms of the 20 July 2018 document do not, in my judgment, bear the scrutiny of npower for the following reasons, namely

It is suggested also that the price cap rendered otiose the need for the Autumn trial. I do not agree for the reasons given by Mr Fisher in his third witness statement especially at paragraphs 29-35. I also accept his points about the timing of the trial which was proximate in time to the commencement of the price cap, and to the continuing justification for the same.

In the light of the above analysis, I reject the criticism about the alleged breach of duty of candour.

I now consider the criticisms of npower concerning failure to give reasons, A1P1 (failure to assess proportionality), irrationality and acting for an improper purpose. Ofgem objects to the ground of acting for an improper purpose. I intend to give permission for the application to rely on this additional ground. I accept that the challenge is based in large part on the documents recently disclosed of 14 June, 4 July and 20 July. In view of the speed with which this case has had to be tried, it was not possible to have the matter dealt with in two stages of deciding preliminary matters such as addition of improper purpose to the grounds. I shall allow the addition of the ground for two principal reasons. First, it is so closely connected to the other grounds in particular of failure to state reasons and irrationality that it would be artificial to exclude it. Secondly, it could have been raised earlier, I accept that the argument as developed was in large part by reference to documents recently disclosed of 14 June, 4 July and 20 July. Having considered the full merits of the argument, it is artificial in these circumstances to consider whether if I had been considering only the question of an amendment as to whether I should have held that it was arguable. It is very possible that I should have held that it was not arguable.

I should add that as the arguments have developed the improper purpose argument has rather subsumed both the grounds of failure to state reasons and irrationality, but I shall deal with all of the grounds. I shall deal with improper purpose after the first ground of failure to state reasons.

The pleaded ground of challenge (SFG paragraphs 29-49) is that Ofgem has failed to give any, or any adequate, reasons for its decision to issue the Direction. Npower say that the duty arises under Article 37 Dir 2009/72 (“the Electricity Directive”), Article

It follows from the detailed narrative above that Ofgem has given reasons for its decision. It provided a letter dated 31 August 2018, enclosing the Direction, and summarised its reasoning for making the Direction (against the background of SLC 32A and the selection criteria). It had provided the same letter on 1 August 2018 when giving a month’s notice of the Direction, and information about the trial before that on 12 July 2018 when it was announced that the trial would be about “gauging customer responses to potential future services…”

Ofgem accepts that the correct approach to the sufficiency of reasons was stated in South Bucks v Porter [2004] UKHL 33; [2004] 1 WLR 1953 per Lord Brown at [36], and in particular:

Regardless of the origin of the duty, “what is needed is an adequate explanation of the ultimate decision”. The issue for the Court “is, in the words of Bingham MR in the

Reference should also be made to the case of R v City of Westminster ex parte Andrei Ermakov [1995] EWCA Civ 42, where under the statute (section 64 of the Housing Act 1985 there was an express duty to give reasons at the same time as giving notice of the decision) it was stated per Hutchison LJ that:

“It is well established that an obligation, whether statutory or otherwise, to give reasons for a decision is imposed so that the persons affected by the decision may know why they have won or lost and, in particular, may be able to judge whether the decision is valid and therefore unchallengeable or invalid and therefore open to challenge.

There are numerous authoritative statements to this effect - see, for example, Thornton v Kirklees Borough Council [1979] 1 QB 626 at 638H, in the judgment of Megaw LJ; R v London Borough of Croydon, ex parte Graham (1993) 26 HLR 286 (a case to which further reference will be made) at page 291, where Sir Thomas Bingham MR said:

"I readily accept that these difficult decisions are decisions for the housing authority and certainly a pedantic exegesis of letters of this kind would be inappropriate. There is, nonetheless, an obligation under the Act to give reasons and that must impose on the council a duty to give reasons which are intelligible and which convey to the applicant the reasons why the application has been rejected in such a way that if they disclose an error of reasoning the applicant may take such steps as may be indicated."

He went on to state:

“(2)

The court can and, in appropriate cases, should admit evidence to elucidate or, exceptionally, correct or add to the reasons; but should, consistently with Steyn LJ's observations in ex parte Graham, be very cautious about doing so. I have in mind cases where, for example, an error has been made in transcription or expression, or a word or words inadvertently omitted, or where the language used may be in some way lacking in clarity. These examples are not intended to be exhaustive, but rather to reflect my view that the function of such evidence should generally be elucidation not fundamental alteration, confirmation not contradiction. Certainly there seems to me to be no warrant for receiving and relying on as validating the decision evidence - as in this case - which indicates that the real reasons were wholly different from the stated reasons. It is not in my view permissible to say, merely because the applicant does not feel able to challenge the bona fides of the decision-maker's explanation as to the real reasons, that the applicant is therefore not prejudiced and the evidence as to the real reasons can be relied upon. This is because, first, I do not accept that it is necessarily the case that in that situation he is not prejudiced; and, secondly, because, in this class of case, I do not consider that it is necessary for the applicant to show prejudice before he can obtain relief. Section 64 requires a decision and at the same time reasons; and if no reasons (which is the reality of a case such as the present) or wholly deficient reasons are given, he is prima facie entitled to have the decision quashed as unlawful.”

Npower contend that the explanation in the letters of 1 August 2018 and 31 August 2018 are not good enough: see paragraph 41 of the Grounds. In my judgment, the letters set out the grounds adequately and sufficiently: see paragraphs 46-48 above. They are not just reasons in themselves, but are to be seen in the context of what was said in the communications which led up to SLC 32A as to which there was consultation with the suppliers including npower. Reference is made to the sections above headed “Ofgem reference to the CMA” (paragraphs 21-22) and “Ofgem’s letters of 30 January 2017” (paragraphs 23-27). At paragraphs 29-33 above, there is a section headed “the absence of challenge”. As noted there, npower’s contention that these matters are simply factual background overlooks the fact that report of the CMA, the finding of an AEC and the implementation of SLC 32A set in train compulsory trials of which the Autumn trial was a part.

When the matters set out at paragraphs 46-48 above are considered against the background of the extensive communications leading up to SLC 32A and communicated to a supplier who had been consulted, npower would have been in “no genuine doubt” as to why Ofgem had decided to issue this Direction and proceeded with this trial.

In an attempt to overcome this difficulty, npower say that Ofgem failed to explain its decision to run a trial with 100,000 customers each from two different suppliers, and why it decided to run such a trial consecutively, that is with 100,000 customers being tested before the price cap and 100,000 afterwards: SFG, paragraph 42. The focus in npower’s Skeleton is on “the need for a trial of 100,000” (paragraph 10a)) and the contention that the reasons were insufficient or inadequate: (paragraphs 9-10). I take the view that the reasons given above were adequate and sufficient.

In any event, prior to the Direction, which was 31 August 2018, the issue of the numbers which would be the subject of the trial was discussed at length between the parties in conversations and written communications. This can be seen from the communications referred to above including those of 17 and 20 August 2018 as set out at paragraphs 49-51 above, that is before npower received the letter of 31 August 2018 when the Direction took effect. I therefore reject the submission in paragraph 42 of the SFG that no explanation was given as to why after a successful test of 50,000 it was considered necessary to run a larger test. Explanation was given, and such explanation was adequate and sufficient.

Npower took issue with the position by email on 31 August 2018 in particular as to legitimate aim and as to proportionality. Reference has been made above to the communications of Ofgem of 7 September 2018 and 18 September 2018 at paragraph 55-61 above. This is not a case of some retrospective provision of reasons of the kind referred to in ex parte Ermakov, where the reasons are provided for the first time in the judicial review hearing. The following points should be noted:

There was therefore a statement of reasons for the trial and for the size of the trial. It is suggested by npower that reasons given which were either misleading or inconsistent. I find that the reasons were neither misleading nor inconsistent. Reference is made to the criticisms of npower of the internal communications of June and July 2018, and to the findings there made: see paragraphs 35-45 above.

In the above mentioned correspondence repudiating any allegation of bad faith, npower in their note of 20 November 2018 explained the allegation of bad faith as follows: “Ofgem had as a material purpose (if not the only remaining purpose given the circumstances) to achieve consumer switching and that achieving an outcome is an improper purpose for a power to carry out trials” (emphasis added). This is said to be evidenced by the internal documents referred to above of 14 June 2018, 4 July 2018 and 20 July 2018, which I have considered above at paragraphs 35-45. I also refer to my findings in respect of the duty of candour at paragraphs 92-104.

It is necessary to analyse carefully what was the purpose of the trials. The submission made on behalf of npower is that the purpose of the trials ought to have been to obtain rigorous and robust data: if the data indicated that consumer switching could be achieved, so be it: equally, if the data indicated that it could not be achieved, then so be it. There is nothing to show that the robustness and rigour of the data would be overlooked. The rigour of the trials was important to give confidence in the results.

However, as was helpfully pointed out by Ms Love on behalf of CMA, the obtaining of rigorous and robust data was not a purpose in its own right. The purpose of the exercise was to address the lack of consumer engagement which had been identified as an adverse effect of competition. A purpose of the data was to find out whether a proposed intervention was likely to address the lack of consumer engagement. In this context, consumer switching to obtain a more competitive product elsewhere was addressing the lack of engagement. Ultimately, if there was customer switching and the data were sufficiently rigorous to have confidence in the results, Ofgem would then be able to scale up the intervention to millions of customers. I accept the submission of the CMA that provided that it was in the context of addressing the lack of consumer engagement as set out here, it was not an improper purpose to wish to address it by consumer switching.

More generally, insofar as it was contended that there was misleading information provided to consumers or an intent to provide such information for the purpose of achieving switching, I reject that allegation. In any event, it is difficult to understand how this allegation would not involve acting in bad faith, a case which npower has disavowed specifically.

There has been considered in detail above the alleged lack of candour, particularly by reference to the internal documents of 14 June 2018, 3 July 2018 and 20 July 2018.

The nature of the complaint 121.Npower’s complaint is that:

Ofgem submits that this argument is fallacious because npower is only entitled to supply energy pursuant to a licence which is subject to conditions. The scope of any property right is therefore limited by the terms of their licence. That is a right to supply electricity subject to compliance with licence including the obligation to comply with Directions given by Ofgem pursuant to SLC 32A. Npower did not contend that the Direction was outside SLC 32A, and conceded before Zacaroli J., that Ofgem had satisfied s. 25(2)(a) of the EA89. Ofgem contends that npower cannot establish an interference within the meaning of Article 1 of Protocol No. 1 in that the requirement to comply with a licence condition cannot constitute an interference in npower’s property rights, since the property right is contingent on such compliance.

Further, insofar as npower argue that they had property rights in respect of contracts with their customers, Ofgem submits that is correct only insofar as those contracts have not been lawfully terminated. All of the contracts held by customers involved in the trial were (and are) contracts that can be terminated by customers without notice. The trial in no way alters the contractual rights of the parties. The only effect of the trial is to provide information to a customer on which basis the customer could decide to act or not to act. On this basis, Ofgem submits that the provision of information to consumers to address an AEC does not affect any property right, namely the right to supply energy to consumers on the basis of contractual agreements subject to licence conditions.

Despite the arguments of Ofgem, I shall assume for the purpose of this argument that there is a property which attaches to the relationships with the customers notwithstanding the relationships might be terminable without notice. Thus far, the main authority referred to was Breyer Group plc v Department of Energy and Climate Change [2015] 1 WLR 4559 at paragraph [49] was to say that the relevant distinction was between goodwill and future contracts which separated existing enforceable contracts and possible future contracts. It may be that even a contract terminable at will amounts to an existing enforceable contract. However, without further citation of authority on the subject and despite the wide interpretation given to “possessions” under A1P1, I prefer to assume (without deciding the matter) that the contracts are possessions within A1P1, and to decide the matter by reference to the proportionality arguments. In the course of argument, Ofgem has conceded that the proportionality of the intervention must be satisfied: “Ofgem accepts that irrespective of whether the claimant has a A1P1 right, we have an obligation to act proportionately” (Transcript Day 2/page 99/lines 8-10).

On the premise that there has been interference, it is contended by npower that

The judgment of Dickson CJ in Oakes provides the clearest and most influential judicial analysis of proportionality within the common law tradition of legal reasoning. Its attraction as a heuristic tool is that, by breaking down an assessment of proportionality into distinct elements, it can clarify different aspects of such an assessment, and make value judgments more explicit. The approach adopted in Oakes can be summarised by saying that it is necessary to determine (1) whether the objective of the measure is sufficiently important to justify the limitation of a protected right, (2) whether the measure is rationally connected to the objective, (3) whether a less intrusive measure could have been used without unacceptably compromising the achievement of the objective, and (4) whether, balancing the severity of the measure's effects on the rights of the persons to whom it applies against the importance of the objective, to the extent that the measure will contribute to its achievement, the former outweighs the latter. The first three of these are the criteria listed by Lord Clyde in De Freitas, and the fourth reflects the additional observation made in Huang. I have formulated the fourth criterion in greater detail than Lord Sumption, but there is no difference of substance. In essence, the question at step four is whether the impact of the rights infringement is disproportionate to the likely benefit of the impugned measure.

In relation to the fourth criterion, there is a meaningful distinction to be drawn (as was explained by McLachlin CJ in Alberta v Hutterian Brethren of Wilson Colony [2009] 2 SCR 567, para 76) between the question whether a particular objective is in principle sufficiently important to justify limiting a particular right (step one), and the question whether, having determined that no less drastic means of achieving the objective are available, the impact of the rights infringement is disproportionate to the likely benefits of the impugned measure (step four).”

It should be noted that Lord Reed’s was a dissenting speech, but Lord Sumption, who was in the majority adopted the four criteria at paragraph 20. He encapsulated the fourth criterion in simpler terms with the words “whether, having regard to these matters and to the severity of the consequences, a fair balance has been struck between the rights of the individual and the interests of the community”. Lord Reed said that there was no difference of substance. At paragraph 21, Lord Sumption stated “None of this means that the court is to take over the function of the decision-maker, least of all in a case like this one.”

The CMA submit that the principles to be applied to an assessment by a court of a challenge to the proportionality of a regulatory measure such as that in various cases of the Competition Appeal Tribunal (“CAT”), which are not inconsistent with the formulation in Bank Mellat. Ms Love referred to Tesco plc v Competition Commission [2009] CAT 6, at paragraphs 135-139 and especially 137-139, which read in material part as follows:

“137 That passage [the judgment of the ECJ in Case C-331/88 R v. Ministry of Agriculture, Fisheries and Food and Secretary of State for Health, ex parte Fedesa [1990] ECR I-4023, paragraph [13]] identifies the main aspects of the principles. These are that the measure: (1) must be effective to achieve the legitimate aim in question (appropriate), (2) must be no more onerous than is required to achieve that aim (necessary), (3) must be the least onerous, if there is a choice of equally effective measures, and (4) in any event must not produce adverse effects which are disproportionate to the aim pursued.

138

The first thing to note is that the application of these principles is not an exact science: many questions of judgment and appraisal are likely to arise at each stage of the Commission's consideration of these matters. This is perhaps most obviously the case when it comes to the balancing exercise between the (achievable) aims of the proposed measure on the one side, and any adverse effects it may produce on the other side. In resolving these questions the Commission clearly has a wide margin of appreciation, with the exercise of which a court will be very slow to interfere in an application for judicial review.

139

That margin of appreciation extends to the methodology which the Commission decides to use in order to investigate and estimate the various factors which fall to be considered in a proportionality analysis (and indeed in its determination of the statutory questions of comprehensiveness, reasonableness and practicability)…”

This formulation of the proportionality principles was endorsed by the CAT in BAA Limited v Competition Commission [2012] CAT 3; upheld on appeal ([2012] EWCA Civ 1077). The CMA particularly relied on a long passage at paragraphs 20(3)-(8). I shall refer to a part of it which reads as follows:

“(3)

….In the present context, we accept Mr Beard's primary submission that the standard to be applied in judging the steps taken by the CC in carrying forward its investigations to put itself into a position properly to decide the statutory questions is a rationality test: see R (Khatun) v Newham London Borough Council [2004] EWCA Civ 55; [2005] QB 37 at [34]-[35] and the following statement by Neill LJ in R v Royal Borough of Kensington and Chelsea, ex p. Bayani (1990) 22 HLR 406, 415, quoted with approval in Khatun:

“The court should not intervene merely because it considers that further inquiries would have been desirable or sensible. It should intervene only if no reasonable [relevant public authority – in that case, it was a housing authority] could have been satisfied on the basis of the inquiries made.”

(4)

Similarly, it is a rationality test which is properly to be applied in judging whether the CC had a sufficient basis in light of the totality of the evidence available to it for making the assessments and in reaching the decisions it did. There must be evidence available to the CC of some probative value on the basis of which the CC could rationally reach the conclusion it did: see e.g. Ashbridge Investments Ltd v Minister of Housing and Local Government [1965] 1 WLR 1320,

1325; Mahon v Air New Zealand [1984] AC 808; Office of Fair Trading v IBA

Health Ltd [2004] EWCA Civ 142; [2004] ICR 1364 at [93]; Stagecoach v Competition Commission [2010] CAT 14 at [42]-[45];

(5)

In some contexts where Convention rights are in issue and the obligation on a public authority is to act in a manner which does not involve disproportionate interference with such rights, the requirements of investigation and regarding the evidential basis for action by the public authority may be more demanding. Review by the court may not be limited to ascertaining whether the public authority exercised its discretion “reasonably, carefully and in good faith”, but will include examination “whether the reasons adduced by the national authorities to justify [the interference] are ‘relevant and sufficient’” (see, e.g., Vogt v Germany (1996) 21 EHRR 205 at para. 52(iii); also Smith and Grady v United Kingdom (1999) 29 EHRR 493, paras. 135-138). However, exactly what standard of evidence is required so that the reasons adduced qualify as “relevant and sufficient” depends on the particular context: compare R (Daly) v Secretary of State for the Home Department [2001] UKHL 26; [2001] 2 AC 532 at [26]-[28] per Lord Steyn. Where social and economic judgments regarding “the existence of a problem of public concern warranting measures of deprivation of property and of the remedial action to be taken” are called for, a wide margin of appreciation will apply, and – subject to any significant countervailing factors, which are not a feature of the present case – the standard of review to be applied will be to ask whether the judgment in question is “manifestly without reasonable foundation”: James v United Kingdom (1986) 8 EHRR 123 , para. 46 (see also para. 51). Where, as here, a divestment order is made so as to further the public interest in securing effective competition in a relevant market, a judgment turning on the evaluative assessments by an expert body of the character of the CC whether a relevant AEC exists and regarding the measures required to provide an effective remedy, it is the “manifestly without reasonable foundation” standard which applies. One may compare, in this regard, the similar standard of review of assessments of expert bodies in proportionality analysis under EU law, where a court will only check to see that an act taken by such a body “is not vitiated by a manifest error or a misuse of powers and that it did not clearly exceed the bounds of its discretion”: Case C120/97 Upjohn Ltd v Licensing Authority [1999] ECR I-223; [1999] 1 WLR 927, paras. 33–37. Accordingly, in the present context, the standard of review appropriate under Article 1P1 and section 6(1) of the HRA is essentially equivalent to that given by the ordinary domestic standard of rationality. […]”.

As to when a deficiency in the regulator’s analysis will lead to the regulatory measure being quashed by a court, the CAT in Barclays Bank plc v Competition Commission [2009] CAT 27 said as follows:

“28 ... [I]t is not every perceived failure in fact-finding or analysis by a decisionmaking body which requires or permits its finding or decision to be quashed. The relevant failing must satisfy a materiality test. Generally speaking, a relevant failing will require the finding or decision to be quashed unless the Tribunal is satisfied that a reasonable decision-maker in the position of the Commission would still have reached the same finding or decision. If a decision-making process which had not included that failing could have led a reasonable decision-maker to a different conclusion, then the relevant finding or decision will usually have to be quashed.

29 This materiality test is of particular importance where a finding or decision is based upon a number of distinct grounds, only one of which is found to have been vitiated. In that context we have found the following passage in the judgment of May LJ in R v. Broadcasting Complaints Commission ex parte Owen [1985] 1 QB 1153 at 1177A-D to be of particular assistance:

“Where the reasons given by a statutory body for taking or not taking a particular course of action are not mixed and can clearly be disentangled, but where the court is quite satisfied that even though one reason may be bad in law, nevertheless the statutory body would have reached precisely the same decision on the other valid reasons, this court will not interfere by way of judicial review. In such a case, looked at realistically and with justice, such a decision of such a body ought not to be disturbed.

…

Another approach to the same problem in such circumstances, which really reflects the same thinking is this: the grant of what may be the appropriate remedies in an application for judicial review is a matter for the discretion of this court. Where one is satisfied that although a reason relied on by a statutory body may not properly be described as insubstantial, nevertheless even without it the statutory body would have been bound to come to precisely the same conclusion on valid grounds, then it would be wrong for this court to exercise its discretion to strike down, in one way or another, that body's conclusion.”

Ms Love on behalf of the CMA submitted that the above decisions were consistent with the four criteria cited in Bank Mellat. In respect of the competition cases which she cited, Ms Love distilled the following principles from them:

I respectfully adopt this summary of the relevant principles relating to proportionality as are relevant in this case.

As to the first of the Four Submissions, Ofgem did carry out full consultations and analysis in determining whether to adopt licence condition 32A and how to select participants in trials, which resulted in its decisions of 30 January 2017 to adopt SLC 32A. In fact, Ofgem did assess proportionality of the original licence condition and the selection criteria. Npower never challenged either of these decisions as unlawful on the basis that either SLC 32A was disproportionate or the associated selection criteria for participation in market trials.

As to the second of the Four Submissions, as regards the autumn trial, Ofgem considered how the nature, scope and timing furthered the objective of SLC 32A. It gave notice of these matters as described above. It assessed and determined not only the proportionality of the original licence condition, but also the autumn trial. I refer to the section headed “the importance of timely progress in the testing of potential interventions” at paragraphs 80-85 above, showing how critical the testing of measures was, and how important was the timely implementation of the measures. I refer also to the section about the forthcoming price cap and how this was factored into the considerations in respect of the testing: see paragraphs 86-90 and see also paragraphs 100(3), 100(4) and 101 above. More generally in respect of the considerations of Ofgem, reference again should be made to the documents of 7 and 18 September 2018 and in the third witness statement of Mr Fisher. There is no reason not to accept this evidence.

As to the third of the Four Submissions, that Ofgem failed to assess the costs / benefits of the trial or to consider alternatives, this ignores the fact that SLC 32A is premised on a trial of this nature taking place as a way of implementing the CMA’s recommendation. This condition is not like any other condition of a licensing body, but is particularly invasive because it involves the provision of information to consumers which may well cause them to select a different supplier with concomitant loss of business to the first supplier: see paragraphs 77-79 above. Proportionality of this was considered (including in terms of commercial impacts and impacts on innovation / competition). There was further consideration of costs/benefits or alternatives in the period leading up to the Direction and thereafter. This was set out in the documents of 14 June 2018 and of 4 July 2018 above. It was referred to in documents of 1 August 2018, 17 August 2018 and 20 August 2018 which all preceded the Direction of 31 August 2018, as well as the documents of 7 and 18 September 2018 elucidating and confirming the reasons given. I am also satisfied that these documents do not contain some ex post facto reasoning, but contain information considered by Ofgem at the time of the Direction, and that conclusion is confirmed by the third witness statement of Mr Fisher, especially at paragraphs 4-15.

The fourth of the Four Submissions was whether a fair balance has been struck between the rights of the supplier and interests of the community sought to be protected. This comprises those interests identified by Mr Prevett including the significant problem posed by the customer disengagement that the CMA found in the domestic retail energy markets and the significant customer detriment. This balance is covered by the matters set out above in respect of the other Submissions.

I have borne in mind the above matters of law as regards proportionality in coming to these views. This includes the following matters:

In all the circumstances, I have come to the view that whether A1P1 is engaged or not, and especially if it is, Ofgem has undertaken a structured proportionality analysis, it has engaged in its calculations which have not been subjected to detailed analysis or criticisms in the evidence or submissions. It has therefore justified its belief that it has acted in a necessary and proportionate way, given the scale of the customer population on default tariffs, and the need through a realistic and adequate scale of intervention to investigate through trials whether there can be a market-wide implementation. In my judgment, this ground also fails. There is no reason for the Court to interfere whether by judicial review or statutory review.

In the course of argument, Mr Sinclair for npower realistically accepted that irrationality was a “wrap-up ground” from all the previous submissions on the basis that no rational body could decide to proceed with the trial on any sensible basis. I agree with Ofgem that this ground is parasitic on the previous grounds and must therefore fail for the same reasons.

It is also suggested in npower’s SFG at paragraphs 64-65 that it was unreasonable for Ofgem to carry out any trial at all when a price cap was to come in and the “policy and market dynamics are to change significantly shorty after (perhaps before) the results of the trial are available.” Ofgem refute this argument for the following reasons, namely

It follows that just as the other grounds have been dismissed, so too the irrationality argument is rejected.

I have therefore rejected each of the grounds on which npower has mounted its substantive attack, and I have accepted substantially the arguments of Ofgem. It follows that each of the bases for the application for judicial review and statutory review are rejected.

There are numerous procedural points taken by Ofgem, which become of less significance. The relevant issues include the following:

Npower submit that its Part 8 N208 Claim Form for its application for statutory review is essentially mirrored in the Judicial Review Claim. The substantive grounds are the same.

Ofgem’s submissions on this point are as follows. The statutory test in section 27(1) is whether the making of the provisional order was “not within the powers of section 25” or “that any of the requirements of section 26 had not been complied with” (i.e. the procedural requirements). That must mean that if the decision to make the provisional order was unlawful on judicial review grounds, then it could be challenged under s.27(1)(a) or (b) (compare s.288 of the Town and Country Planning Act 1990 which is in similar terms).

The power to make a provisional order has three components to which any error of law would need to attach (as Zacaroli J held at [31]): i) as to Ofgem’s decision that there had been a contravention (s.25(2)(a)), which was not here disputed on the facts; ii) as to whether the procedural requirements had been complied with, again as to which there was no dispute, and iii) as to the decision that it was requisite to make a provisional order (which encompasses the need for a provisional order urgently, as opposed to making a final order pursuant to s.25(1), which entails the procedural protection of notice, as set out in s.26).

Ofgem submits that it is impermissible to seek to raise as a basis for challenging the PO the fact that the underlying condition which is said to have been breached is itself unlawful. This is a collateral challenge to the legality of a separate decision or measure which has legal effect and in respect of which a challenge should have been made at the appropriate time, in compliance with the requisite time limits and other requirements of judicial review (including not academic and arguable).

For the reasons given by Zacaroli J [33-34], namely that a successful challenge to the PO would still leave the Direction in place, a successful challenge to a provisional or final order cannot be used as a means to challenge the underlying decision. As Ms Simor QC expressed it for Ofgem, judicial review is a challenge to a substantive decision, whereas statutory review is limited to a challenge to an enforcement decision, which would not get rid of the original Direction. That follows as a matter of logic from the purpose of enforcement orders in s. 25. If s.27(1) could be used to challenge the direction, it could also be used to challenge a longstanding licence condition which was outside the thee-month time-limit for judicial review. “It would be surprising if Parliament intended, by section 27(1) (by a side wind) to have abrogated the time limits for judicial review applications in relation to the condition or direction respectively.”

I have come to the view that Ofgem’s submissions are well made, and that the appropriate form of challenge, if any, should be by way of judicial review. In my judgment, particularly having regard to the reasoning of Zacaroli J, there is no jurisdiction under s.27(1) to consider what is, in fact, a full-frontal attack on the legality of the Direction, and the appropriate form of challenge, if any, is judicial review. If I had been inclined to hold that the challenge was effective, then it would not make any difference if the statutory review was inappropriate provided that the Judicial Review Claim could be brought.

If there are no relevant grounds for statutory review, then if a challenge can be made to the Direction, it is by way of judicial review. However, this is challenged for at least three reasons, namely (i)that it is out of time;

Delay

Ofgem submits that npower has failed to bring the application sufficiently promptly (as required by CPR 54.5(1)(a)) having regard to the nature of the administrative decision under challenge and the consequences of delaying (not least, the trial has now already substantially occurred). In view of my conclusions above, I shall deal with time issues relatively shortly. I find the following:

Ofgem contend that the consequence of the Judicial Review Claim not being brought until the end of October 2018 is that the trial had substantially finished by 12 November 2018 when letters began to be sent out to customers offering alternative tariffs [HB1/5.3/507]. The Judicial Review Claim has therefore been rendered academic ( by npower’s own delay). It is therefore contended that continuation of the Judicial Review Claim now serves no practical purpose. It is said that the claim is fact specific relating to the reasons for the particular trial and the proportionality issues associated with it.

I do not intend to consider this in detail because of the conclusions which I have reached. Indeed, having taken the decision to consider the merits of the case in view of the rolled-up nature of the hearing, it seems artificial at this stage to consider whether it is academic. In any event, I have come to the view that it is not academic because

Subject to the foregoing issues of time limits and whether it is academic, I should have two choices. The first is to find that since this has been a rolled-up hearing and I have found that the application for judicial review must fail, the permission application must also fail. However, as is apparent from the considerable argument, contemporaneous documents and witness statement evidence which this case has generated and its public interest, this is a case which would have satisfied the threshold for permission to be given.

I have come to the following conclusions:

In the circumstances, I dismiss the application to quash the PO. I also dismiss the application for judicial review.