HCA International Ltd v The Competition And Markets Authority

This case raises the question of the circumstances in which a court or tribunal quashing a decision by an administrative body should remit that decision to be remade by a freshly constituted decision making body.

HCA International Limited (“HCA”), the UK subsidiary of a US healthcare company, is a major provider of private healthcare services in the UK and particularly in London. It has been one of the main subjects of an investigation under Part IV of the Enterprise Act 2002 (the “2002 Act”) into the supply of privately funded healthcare services commissioned by the Office of Fair Trading (“OFT”) from the Competition Commission, which later became the Competition and Markets Authority; I shall use the abbreviation “CMA” to refer to both. As a result of the decisions reached by the CMA in its final report dated 2^nd April 2014, HCA applied on 30^th May 2014 to the Competition Appeal Tribunal (the “CAT”) for a review under section 179 of the 2002 Act.

Ultimately, the CMA agreed that two of its main decisions affecting HCA (including the decision that HCA should divest itself of two hospitals that it owns in central London) should be quashed by the CAT, but HCA asked the CAT to remit those decisions to be remade by a freshly constituted inquiry group of the CMA. The CMA strongly resisted that course, and the CAT decided to remit the decisions to be remade by the original inquiry group.

The main question before this court is whether the CAT was right. In the broadest of outline, HCA contends that the inquiry group of the CMA that handled the original investigation was incompetent, treated HCA unfairly over an extended period, and would be affected by either or both of apparent bias and “confirmation bias” (the difficulty of persuading a decision maker that has once made up its mind to change its view). The appeal has been expedited because it is important that the CMA’s investigation should reach a timely conclusion.

For reasons that will shortly appear, the disposal of this appeal has required a more than usually detailed treatment of the facts. That is not because those facts are especially complicated, but because HCA’s attack on the CAT’s decision was founded on a series of contentions concerning its understanding of the CMA’s conduct of the investigation.

The proceedings have been the subject of confidentiality orders and arrangements, but this judgment does not contain confidential information.

In the course of its investigation, the CMA had gathered a large amount of invoicing information for the period between 2007 and 2011 in relation to the charges made by private hospitals to insurance companies and to private individuals for healthcare services. The CMA concluded, based on its insured prices analysis (“IPA”), that there were adverse effects on competition (“AECs”) associated with the structure of the market for privately funded healthcare services in central London, that HCA faced

weak competitive constraints from its rivals in central London and that HCA charged significantly higher prices to insurers than the London Clinic (“TLC”).

The CMA made 3 main decisions (together the “3 decisions”) as follows: (a) the decision that HCA should divest itself of two of its hospitals in central London, being either Wellington Hospital together with its Platinum Medical Centre, or the London Bridge Hospital and the Princess Grace Hospital (the “divestment decision”), (b) the decision that there was an AEC in relation to the charges made to individuals who paid for their own healthcare services (the “self-pay AEC decision”), and (c) the decision that there was an AEC in relation to the charges made to insurance companies for healthcare services provided to insured patients (the “insured AEC decision”).

HCA challenged the CMA’s decisions on 5 grounds: (i) that the CMA had not given HCA an adequate opportunity in the course of its investigation to comment on the IPA and its price concentration analysis (the “PCA”), (ii) that there were material errors in the IPA and the PCA which meant that they did not provide a sustainable basis for the CMA’s 3 decisions, (iii) that the CMA’s assessment that HCA’s hospitals face weak competitive restraints was flawed because it was based on an arbitrary definition of the relevant geographical market, (iv) that the CMA’s assessment that there were high barriers to entry and expansion in central London was irrational and unsustainable, and (v) that the divestment decision was not a proportionate remedy.

The CMA’s response to HCA’s application was to acknowledge on 14^th November 2014 that HCA had shown that there were errors in the CMA’s IPA which HCA had not been given an opportunity to identify and comment upon during the CMA’s investigation, and that the insured AEC and divestment decisions (the “2 decisions”) should be quashed. In effect, the CMA conceded the first of HCA’s five grounds of challenge, but persisted in its defences to the other four grounds (“grounds 2 to 5”). The CAT ultimately decided that the CMA’s concession was more limited than HCA had argued. It concluded that the concession was on the basis that there were (a) errors in the regression analysis and calculation of R², in particular a computer coding error which meant that the CMA had consistently overestimated the R² levels it reported (the “R² error”), and (b) errors in the CMA’s assessment of the statistical significance of price differences between HCA’s hospitals and a competitor hospital, TLC (the “statistical error”). These two errors, according to the CAT, went to the confidence intervals applicable to the CMA’s estimated price differences between HCA and TLC, but did not change those estimates themselves. HCA has always contested and still contests those conclusions.

The CMA’s concessions resulted in a case management hearing before the CAT on 15^th December 2014 at which it considered what order it should make in response to the CMA’s concessions. I shall deal in more detail with the CAT’s ruling dated 23^rd December 2014 (the “ruling”) in due course, but the essence of its decision was that it was appropriate to refer the matter back to the same CMA inquiry group and the same CMA case team as had undertaken the original investigation (the “remission decision”). The CAT held that no case of actual bias or pre-determination had been made out against the CMA. It also held that it could not be said that the CMA’s investigation was wholly flawed or had been completely mishandled, or that the CMA had knowingly consulted on a false basis, or that a fair minded and informed observer

would conclude that there was a real risk that the CMA’s existing inquiry group and case team would be biased against HCA, or that there was an inappropriate risk of confirmation bias. The CAT also held that the costs of grounds 2 to 5 should be reserved pending the outcome of the reconsideration of the 2 decisions (the “reserved costs decision”), and that the CMA should not be required to pay the costs of HCA’s data room review resulting from another of the CAT’s decisions on 25^th July 2014 (the “data room costs decision”).

HCA now appeals the remission decision on the main bases that the CAT failed to attribute sufficient weight to maintaining confidence in the CMA’s decisions, the CMA’s investigations were wholly flawed and mishandled, and that the CAT wrongly concluded that there was no appearance of bias, and wrongly took the practical difficulties of remission to a new inquiry group into account. HCA also appeals the reserved costs and the data room costs decisions.

Before turning in detail to these arguments, it is necessary to set out the factual background, including significant passages from the evidence of Mr Roger Witcomb (“Mr Witcomb”), who was the chair of the original inquiry group of the CMA. The reason why so much of that evidence is relevant to what we have to decide is because one of the central submissions made by HCA was that Mr Witcomb’s inquiry group had behaved unfairly towards HCA and that, even when this unfairness was pointed out, Mr Witcomb did not understand why what he was alleged to have done was wrong.

In April 2012, the OFT made a reference to the CMA for it to investigate the supply or acquisition of privately funded healthcare services. Had it been made today, the reference would have been made to the Chair of the CMA under what is now schedule 4 to the Enterprise and Regulatory Reform Act 2013 (the “2013 Act”).

The CMA gathered, as I have said, invoicing information for the period between 2007 and 2011. The investigation was undertaken by an inquiry group comprising Mr Witcomb, Ms Jayne Almond, Mr Anthony Morris, Mr Jeremy Peat, and Mr Jonathan Whiticar, and by a staff project or case team including Mr Thomas Wood, the inquiry manager (“Mr Wood”).

On 2^nd September 2013, the CMA published its provisional findings (the “provisional findings”) including the details of its original IPA. The provisional findings based on the original IPA included findings that HCA’s prices were overall higher by a defined percentage than those of TLC.

On 2^nd October 2013, the CAT (Mr Marcus Smith QC, Mr William Allan and Ms Margot Daly) ruled on an application by BMI Healthcare Ltd (“BMI”), HCA and Spire Healthcare Group (“Spire”), who contended that the CMA had acted irrationally in the judicial review sense of that word. The CAT held that the CMA’s operation of its disclosure room regime was unfair and in breach of its statutory duty under section 169 of the 2002 Act and in breach of the rules of natural justice in comprehensively failing to give BMI, HCA and Spire a fair opportunity to correct or contradict the CMA’s provisional findings or to make worthwhile representations. The CAT made clear that, given the technical nature of the material disclosed, the CMA was required

to provide a high degree of disclosure and transparency, and that HCA, BMI and Spire needed to see the material that was hitherto undisclosed in order to prepare its response to the CMA’s provisional findings.

Between 28^th October and 11^th November 2013, the CMA opened its new data room which related in part to the IPA. On 11^th November 2013, HCA responded to the CMA’s consultation on the IPA. On 21^st January 2014, the inquiry group was presented with a revised version of the IPA taking into account representations made by HCA and others (the “revised IPA”), and the CMA published its provisional decision on remedies (the “provisional decision on remedies”), which included the divestment decision, but was based on the original IPA rather than the revised one. The revised IPA had the effect of increasing significantly the CMA’s previous assessment (based on the original IPA) of the percentage by which HCA’s prices were said to be higher overall than TLC’s prices. The revised IPA also, therefore, had the effect of substantially increasing the CMA’s assessment of the net present value (the “NPV”) of the price benefits of the proposed divestment of HCA’s hospitals.

On 29^th January 2014, the CMA wrote to HCA’s solicitors about the arrangements for an oral hearing between HCA and the inquiry group and the case team to provide HCA an additional opportunity to supplement any of its written submissions on the provisional findings or the provisional decision on remedies.

On 30^th January 2014, the CMA wrote to Nuffield Health (“Nuffield”), one of the other healthcare providers that was the subject of the inquiry, saying that it had made changes to the original IPA following comments received on its provisional findings. The CMA indicated in this regard that it had dropped the use of the average revenue per admission, because it was not a reliable measure, and it had changed the methodology used to calculate the price index to control for any differences in the operator’s mix of patients’ ages, gender and length of stay. The CMA said that the revised results showed that the price differences between Nuffield, BMI and Spire were smaller than those previously found so that it was the CMA’s current view that it could not distinguish between BMI, Spire and Nuffield in assessing the level of their market power in negotiating with insurers. The CMA also said that the new IPA had not caused it to change its provisional decision on remedies, but that since Nuffield had had a limited opportunity to review the IPA “it should be given the opportunity to review these matters, given the change in [the CMA’s] current thinking”.

On 6^th February 2014, the CMA wrote to each of 3 healthcare providers and 8 healthcare insurers saying that it proposed to disclose certain confidential information relating to that participant, which had been used in its IPA, to Nuffield’s external advisers. It said it had revised the IPA following comments received on its provisional findings, and the revised results had had a significant impact on the CMA’s assessment of Nuffield’s position. The CMA said that the key revisions it had made to the IPA’s methodology were that it had dropped the use of the average revenue per admission, because it was not a reliable measure, and it had changed the methodology used to calculate the price index to control for any differences in the operator’s mix of patients’ ages, gender and length of stay. The CMA said that it was allowing Nuffield’s external advisers (but not the other providers) access to the revised IPA in a disclosure room. HCA was not sent any letter similar to these 11

letters to these other participants in the inquiry.

On 11 February 2014, HCA’s solicitors emailed Mr Wood noting that the CMA had refused HCA’s request that its advisers be allowed further access to the disclosure room to prepare its response, and asked for confirmation that no other party was being granted such access. On 12^th February 2014, Mr Wood responded to HCA’s solicitors confirming that no other party had been given access to a disclosure room in order to prepare a response, but that Nuffield’s advisers had been given such access “to review our [IPA] and the extracts from the national bargaining analysis which are relevant to Nuffield”. He continued by saying that “[y]ou will recall that Nuffield’s advisers did not have access to the Data Room when the advisers to HCA and others reviewed this information”.

On 18^th February 2014, the inquiry group undertook an oral hearing that was attended by HCA. HCA’s experts made submissions about various aspects of the methodology and efficacy of the CMA’s original IPA. The CMA’s representatives did not inform HCA that the IPA had been revised.

On 2^nd April 2014, the CMA published the non-confidential version of its final report, from which HCA was able to ascertain that the IPA had been revised.

On 16^th April 2014, HCA wrote to the CMA asking for disclosure of its revised IPA, which the CMA refused to provide on 24^th April 2014. On 30^th May 2014, HCA issued its Notice of Application to the CAT seeking the quashing of the 3 decisions on broadly the 5 grounds that I have already indicated.

On 18^th June 2014, HCA’s solicitors asked the CMA either to agree to a preliminary issue hearing on the procedural fairness of the CMA’s failure to disclose the revised IPA to HCA or to allow HCA to comment upon it, or to concede the application on that ground. The CMA refused that invitation on 20^th June 2014.

On 25^th July 2014, the CAT (composed as for the ruling) ruled that the CMA should give disclosure of the revised IPA, after a contested hearing that took place on 25^th June 2014 and 8^th July 2014. It concluded that fairness between the parties required HCA to have access to the revised IPA, since the IPA was “absolutely critical as the basis for the CMA’s findings of relevant AECs and thus for its decision to impose the divestment remedy”.

In August 2014, HCA’s advisers, KPMG and Professor Michael Waterson (“Professor Waterson”) were given access to the data room and identified the R² error and the statistical error in the revised IPA, and a number of other errors including one allegedly simple and purely arithmetical error. Professor Waterson said in respect of one of the coding errors that “a rather clear mistake has been made”, concluding that there was a “clear error in the coding routine employed by the CMA which has the effect of greatly exaggerating the precision of the estimates … [h]ence it is much more difficult to reject the hypothesis that the mean prices charged by HCA and TLC are the same”.

On 8^th October 2014, HCA’s solicitors wrote to the CMA’s solicitors identifying the R² error and the statistical error based on its expert evidence, and enclosing the reports by KPMG and Professor Waterson, inviting the CMA to concede that the 3 decisions should be quashed to avoid further costs. The CMA’s solicitors declined that invitation on 9^th October 2014.

On 17 October 2014, HCA filed its amended Notice of Application dealing with the R² error and the statistical error and the other errors identified by HCA’s experts.

On 13^th November 2014, the CMA’s solicitors wrote to HCA’s solicitors saying that it was minded to concede the quashing of the divestiture decision and the insured AEC decision on the limited basis that it accepted its procedure had been unfair. On 25^th November 2014, the CMA filed an amended defence admitting the R² error and the statistical error (though it had previously declined to file such a document before it was ordered to do so). It accepted that the proper course was for the divestment decision and the insured AEC decision to be remitted to the CMA so that HCA could be properly heard. It also accepted in broad terms that the other errors identified by HCA’s experts would be considered afresh by the CMA on the remission of the 2 decisions. The arithmetic error was also broadly accepted in the amended defence.

On 2^nd December 2014, Mr Witcomb and other members of the original case team met with a representative of the UK Trade and Investment division of the Foreign and Commonwealth Office (“UKTI”), and gave an account of the current state of these proceedings.

On 4^th December 2014, HCA’s solicitors proposed that the 2 decisions should be remitted for reconsideration by a new inquiry group and a new case team. The letter was regarded by the CAT as being, in effect, HCA’s application for that relief. HCA relied on the following main points:-

i)

Procedural defects including the following: (a) The CMA knowingly consulted HCA on a false basis, by reference to reasoning and analysis to which it no longer adhered, having already decided to adopt a new IPA, (b) the CMA inquiry group received HCA’s oral and written submissions on the original IPA without informing HCA that the original IPA had been abandoned. The CMA thereby conducted a sham consultation which seriously misled HCA and its advisers, (c) the CMA discriminated against HCA in the conduct of the consultation process, by failing to consult it about or inform it of the existence of the revised IPA, whilst allowing Nuffield’s advisers to have access to it, and (d) the CMA informed all the main parties except HCA of the existence of the revised IPA and that Nuffield was being consulted on it, but when HCA asked whether any other party was being given access to the data room, the CMA concealed the existence of the revised IPA from HCA, falsely suggesting that disclosure to Nuffield was in respect of the original IPA. At first sight, allegations (a) and (d) appear to be allegations involving bad faith.

ii)

The CMA’s procedure was affected by serious and pervasive unfairness, so that it would not be appropriate for the same inquiry group or the same case team to conduct any further investigations.

iii)

Remission to the original inquiry group and case team would give rise to an appearance of bias and a significant danger of confirmation bias.

iv)

Remission to a new inquiry group and a new case team would be proportionate in all the circumstances, given the significance of the matters in issue. The extra time and cost associated with the appointment of a new inquiry group and a new case team would be justified.

On 8 December 2014, UKTI emailed HCA to inform it of the meeting on 2^nd December 2014 with Mr Witcomb and the case team. The email reported that “[t]he CMA stated that the modelling errors which were identified by HCA advisors do not change the actual results or conclusions of the modelling. The errors concerned the confidence intervals for the results of the modelling, which were lower than thought by the CMA (they noted they looked at over 1000 treatments across 6 hospitals for 6 years and the results were unchanged by the errors that had been found)”. On 9^th December 2014, HCA’s solicitors wrote to the CMA’s solicitors passing on UKTI’s email and expressing concern that the CMA had told UKTI that its conclusions remained valid and unchanged notwithstanding the errors identified.

On 9^th December 2014, the CMA’s solicitors wrote to HCA’s solicitors rejecting the suggestion of remission to a new inquiry group and a new case team, and stating that the CMA would approach the reconsideration of the 2 decisions with an open mind.

On 10^th December 2014, the CMA served Mr Witcomb’s 3^rd statement addressing the suggestion that the remission should be to a new inquiry group and a new case team. He referred back to his 2^nd statement dated 7^th August 2014 which had included the following passages:-

“66.

The [CMA] published its Provisional Decision on the Remedies (“the PDR”) on 21 January 2014 … the review of the parties’ comments to the IPA was a lengthy process. Given the statutory deadline for the Investigation was 3 April 2014, the [CMA] could not have delayed publication of the PDR, while waiting for the conclusion of the review of the parties’ comments on the IPA. Accordingly, both work streams had to run in parallel. Therefore, the Group decided not to reflect the revised IPA results in the PDR.

…

Further consultation on the IPA

79.

… HCA, BMI and Spire had been given access, in the October Data Room, to the IPA and the underlying cleaned data. Nuffield was not provided with access to the October Data Room, as in the [provisional findings] there had been no provisional finding that Nuffield had market power via-a-vis the [private medical insurers].

80.

Having considered the impact upon Nuffield of the revised IPA … , and its duties under the 2002 Act, the [inquiry group] decided that consultation with Nuffield was necessary. The [inquiry group] considered whether disclosure to HCA and the other hospital operators was also required, but decided it was not, noting that the position of those hospital operators had not materially changed as a result of the revised IPA results. In addition, the [inquiry group] had to consider the restrictions imposed by the statutory deadline for publication of the Final Report and, in particular, (i) the time required to organise a data room with four parties, and (ii) the time required to consider properly further representations on the methodology and underlying data made by all four hospitals. …

83.

Following the [inquiry group’s] decisions, the [CMA] held the data room between 10 and 14 February 2014, providing Nuffield’s external advisers with access to (i) the main changes to the methodology used in the revised IPA; and (ii) the main results of the IPA relating to the four hospital operators located outside central London (BMI, Spire, Nuffield and Ramsay) … The [CMA] did not disclose the underlying data and analysis, or the more detailed description of the methodology and results equivalent to that contained in the relevant appendix to the [provisional findings]. Nor did the [CMA] disclose any of the IPA results relating to central London.

84.

The [inquiry group] considered that those parties whose confidential data was to be disclosed to Nuffield should be notified of that disclosure in writing. Accordingly, on 6 February 2014, the [CMA] wrote to three hospital operators (BMI, Spire and Ramsay) and eight insurers (Bupa, AXA, Aviva, PruHealth, Simplyhealth, WPA, Exeter Family Friendly & Cigna). The [CMA] did not notify HCA at this time as its confidential data was not being disclosed in the data room, given that the information disclosed to Nuffield did not relate to central London (and HCA’s data related only to central London)”.

Mr Witcomb’s 3^rd statement explained the CMA’s position as follows:-

“11.

HCA claims that the CMA consulted on the [provisional decision on remedies] on a “false basis, by reference to a reasoning and analysis to which [the CMA] no longer adhered”, as the CMA decided not to include the revised IPA in the [provisional decision on remedies].

12.

For the reasons given below, I do not consider that, when the [CMA] consulted on the [provisional decision on remedies] it was consulting on a reasoning to which it no longer adhered. The role of the IPA in the remedies assessment is explained below. The fact that the [CMA] amended the IPA methodology – and its results – did not materially affect the reasoning in the [provisional decision on remedies] or prejudice HCA’s ability to comment on the [provisional decision on remedies].

…

The role of the IPA in the [provisional decision on remedies]

18.

I shall now explain the IPA’s contribution to the remedies’ analysis in the [provisional decision on remedies].

19.

The numerical results of the IPA in central London were only used in the proportionality assessment of the divestiture remedy. As part of this assessment, the [CMA] estimated the net present value (“NPV”) of the proposed package of divestiture remedies to assess the extent to which the benefits of divestiture might outweigh its costs. In this context, the [CMA] considered the extent to which the proposed divestitures of HCA’s hospitals would affect prices and sought to quantify the price benefits that are likely to accrue to customers as a result of the proposed divestiture, for self-pay, insured and international patients. The [CMA] set out a detailed analysis of the likely impact of divestiture on the prices charged to both self-pay and insured parties …

20.

In quantifying the price benefits for insured patients in central London (where HCA operates), the [CMA] thought that it was appropriate to estimate the likely reduction in prices (and hence revenues) by applying the difference in insured price between HCA and its closest London competitor [TLC] to HCA’s private revenues. To this end, the [CMA] relied on the numerical results of the IPA in central London, namely the average difference in the prices charged by HCA and TLC to [private medical insurers] for 2011.

21.

It follows from the above that only one of the results of the IPA – the average difference in the prices charged by HCA and TLC to [private medical insurers] – was relevant for the proportionality assessment of the divestiture remedy. The methodology of the IPA and the other IPA results were not a factor in our consideration of the proposed remedies.

22.

As previously explained, the revised IPA methodology and revised results were considered by the [inquiry group] on 21 January 2014. The [inquiry group] concluded that, in respect of central London, the revised results confirmed the [CMA’s] findings in [the provisional findings]. In relation to the average difference in the prices charged by HCA and TLC to [private medical insurers], this figure was [increased] in the revised version of the IPA presented in the Final Report. This higher result did not undermine the [CMA’s] proportionality assessment of the divestiture remedy in the [provisional decision on remedies]. On the contrary, it made the price benefit of the divestment remedy higher, and hence, the justification for the remedy clearer.

23.

In due course, the CMA received comments on the [provisional decision on remedies] from parties including HCA. HCA did not make any comments on the implications of the average difference in prices charged by HCA and TLC to [private medical insurers], but did make a number of comments concerning the way the [CMA] sought to quantify the price benefits of divestiture, which were not related to the IPA methodology or results.

24.

… The CMA accepted these criticisms, and updated the remedies analysis accordingly.

25.

Given the way in which HCA responded to the [provisional decision on remedies], it was not prejudiced in the comments it made or could have made by the changes made to the IPA. Had its consultation response depended on the average difference in price between HCA and TLC as determined by the IPA in the [provisional findings], then that is something we would have recognised and taken steps to address. However, that was not the case.

26.

In fact, as explained in paragraph 23 above, the change to the average price difference did not alter out overall conclusion on the proportionality assessment on the proposed divestitures between the [provisional decision on remedies] and the Final Report. …

27.

HCA was therefore able to make effective submissions to the CMA about the approach used to quantify the price benefits of divestiture and its views were taken into account in the Final Report. These submissions were not affected by the revised IPA results, and did not render consultation on the [provisional decision on remedies] a “sham”.

…

HEARING OF 18 FEBRUARY 2014

39.

In its reply, HCA also criticises the CMA for allowing it to make oral representations on the IPA at the hearing on 18 February 2014, without explaining that the CMA had already revised the IPA.

40.

I shall start by explaining the context and purpose of the hearing of 18 February 2014.

41.

In a meeting which took place on 9 January 2014, the [inquiry group] agreed that hearings would be offered to HCA, BMI and [another], and that, if Circle wanted a hearing, this should also be offered. In accordance with the thinking of the [inquiry group] at the time, these parties were likely to be the most affected by the proposed remedies in the [provisional decision on remedies]. …

42.

On 13 January 2014, the [CMA] wrote to HCA to invite it to attend a hearing in February to “discuss the provisional decision on remedies”. In response, HCA requested a list of topics to be covered at the hearing ...

43.

On 29 January 2014, the [CMA] wrote to HCA, BMI, [another] and Circle providing further detail on the arrangements for the hearings to be held in February … In particular, the letter to HCA indicated that the aim of the hearing was to provide it with an additional opportunity to supplement any of its written submissions – whether on the [provisional findings] or the [provisional decision on remedies] – with oral submissions and to raise any new issues with the members it thought necessary. It was also an opportunity for the [CMA] to clarify points before a final decision was taken on the AEC and (if required) the remedies. The [CMA] clarified that it had not identified areas in respect of HCA’s previous submissions on which the members or staff wished to ask specific questions, as it regarded HCA’s position and submissions, in particular as regards the [provisional findings], to be very clear. …

44.

On 18 February 2014, the [CMA] held a hearing with HCA. During the hearing, HCA made a number of submissions about the flaws it perceived in the CMA’s pricing analysis, which included the IPA …

45.

In relation to the IPA, HCA repeated criticisms it had raised previously with the [CMA], namely that:

a.

the IPA measured episode charges and not prices: the [CMA] had not controlled for the factors that influence that charges in its IPA such as complexity and patient mix;

b.

there was no statistically significant price difference between HCA and TLC in relation to the analysis of prices charged to BUPA; and

c.

there was no connection between prices and concentration; something else other than the level of concentration was driving prices.

46.

HCA had already submitted these criticisms to the [CMA] in its written submissions …

47.

The purpose of the hearing was therefore to give HCA – and the other parties most likely to be affected by the proposed remedies – a further opportunity to expand on their representations. The [CMA] had however clearly understood HCA’s views as regards the [provisional findings], including HCA’s main comments on the IPA, which were repeated at the hearing.

48.

It is in the nature of the market investigation process that the [CMA] may receive further comments from parties at a time where it has already developed its thinking. The fact that HCA was given a further opportunity to make its case orally even though the [CMA] had in fact already considered and addressed the points which it made does not mean that the consultation was a sham. The position is rather that, because HCA’s previous representations had been clear, its oral representations did not materially add to the points it had made in writing. However, we did not want to shut HCA out from making oral representations if it wished to do so. As explained in Witcomb 2, the [inquiry group] had by time of the oral hearing already reached a view that re-consulting HCA on the IPA was not required, because the position of HCA (as well as certain other hospital operators) had not materially changed as a result of the revised IPA results. The fact that we had a hearing on 18 February 2014 at which HCA was given the opportunity to address the [inquiry group] did not change that assessment, although the position has now changed as a result of developments in this litigation.

OTHER ALLEGATIONS OF UNFAIRNESS

49.

In the Re-Amended [Notice of Application], HCA submits that the CMA informed all other main parties except HCA of the existence of the revised analysis and of the fact that Nuffield was being consulted in relation to it. … The [CMA] did not notify HCA at this time as its confidential data was not being disclosed in the data room, given that the information disclosed to Nuffield did not relate to central London (and HCA’s data related only to central London).

…

THE COMPOSITION OF THE INQUIRY GROUP AND THE STAFF TEAM

…

54.

I am of the view that the composition of the staff team and Group should be a matter for the CMA to determine. This would provide the CMA with the necessary flexibility to manage its resources in an efficient manner, considering not only the requirements of the remittal, but also the overall workload of the CMA and the specific availability of members and staff.”

On 11^th December 2014, the CMA served Mr Witcomb’s 4^th statement dealing with his meeting with UKTI, exhibiting UKTI’s representative’s manuscript notes of the meeting recording “[c]ouple of modelling errors … R2 error … Precision of stats relating to insurance prices paid by their rivals -> Within 1 of 1000 index codes”. Mr Witcomb also explained the meeting as follows:-

“8.

The meeting with UKTI took place at the CMA’s offices on 2 December 2014. I do not have any written notes of the meeting …. However, I have been provided with Mr Thompson’s handwritten notes of the meeting … and my recollection of the meeting as set out below broadly accords with those notes.

…

11.

Mr Thompson then asked about the transparency of the modelling work relied upon by the CMA in the private healthcare market investigation. I explained that the modelling work was based on data which were commercially sensitive and which it would not be appropriate to share directly with HCA. However, a data room had been set up to which HCA’s advisers had been given access. I explained that a similar process would be likely to apply in the context of any future review of the sector.

12.

The discussion then moved on to the current proceedings before the [CAT]. I explained that HCA’s advisers had identified two main errors in the CMA’s modelling of insured prices, namely an error in the CMA’s statistical significance testing and an error in the calculation of R squared. I noted that both these errors went to the confidence intervals for the CMA’s parameter estimates (i.e. the robustness of the CMA’s estimates), but did not change the parameter estimates themselves (i.e. the estimated price difference between HCA and TLC). In particular, I explained that the confidence intervals were lower than had originally been found by the CMA. Neither I, nor my colleagues at the meeting, made any comment as to the impact of these errors on the CMA’s conclusions. The explanation I

gave was purely a factual explanation as to the nature of the errors.

…

16.

In particular, at point 3 of the email, it is suggested that “The CMA stated that the modelling errors which were identified do not change the actual results or conclusions of the modelling. The errors concern the confidence intervals for the results of the modelling, which were lower than thought by the CMA”. Whilst the second sentence is an accurate reflection of what I said at the meeting, the first sentence is not (and, moreover, I note that this is not reflected in Mr Thompson’s handwritten notes of the meeting). I did not say that the errors identified by HCA’s advisers do not change the CMA’s actual results or conclusions of the modelling. As explained above, what I said was that the errors do not change the parameter estimates in the modelling (i.e. the estimate price difference between HCA and TLC). Importantly, I did not make any statement as to the consequences of the errors in terms of the CMA’s conclusions, including its AEC finding and the appropriateness and proportionality of the divestment remedy.”

The hearing before the CAT took place on 15^th December 2014, and on 23^rd December 2014, it published its ruling. On 12^th January 2015, the CAT made an order quashing the insured AEC decision and the divestment decision, and remitting them to be reconsidered by the CMA. It stayed HCA’s application insofar as it related to the self-pay AEC decision, and ordered the CMA to pay HCA’s reasonable costs in respect of ground 1 on the standard basis, not including the costs of the data room review conducted by HCA’s advisers as a result of the CAT’s disclosure ruling dated 25^th July 2014. The CAT also made the reserved costs decision already described.

On 11^th February 2015, the CAT granted HCA permission to appeal on the remission issue, but refused permission on the costs issues. It declined to order a stay of the remitted investigation. On 5^th March 2015, I granted HCA permission to appeal on the costs issues and directed that the hearing of the appeal be expedited. I also declined to order a stay, expressing the hope that the appeal would come on quickly enough to render the question academic.

Section 134(1) and (2) of the 2002 Act requires the CMA to decide whether “any feature, or combination of features, of each relevant market prevents, restricts or distorts competition in connection with the supply or acquisition of any goods or services in the United Kingdom or a part of the United Kingdom”. Section 134(4) provides that, if the CMA has decided that there is an AEC it shall decide whether action should be taken under section 138 for the purpose of remedying or preventing the AEC concerned or any detrimental effect on customers.

Section 133A of the 2002 Act provides that, where a reference is made to the chair of the CMA for the constitution of a group under schedule 4 of the 2013 Act, the functions of the CMA “are to be carried out on behalf of the CMA by the group so constituted”. Thus, for the purposes of this case, the inquiry group is the relevant decision maker.

Section 169(1) and (2) provides that where the CMA is proposing to make a relevant decision in a way which it considers is likely to have an impact on the interests of any person, it “shall, so far as practicable, consult that person about what is proposed before making that decision”.

Section 179(1) of the 2002 Act allows a person aggrieved by a decision of the CMA in connection with a reference to apply to the CAT for a review of that decision. Section 179(4) provides that the CAT shall apply the same principles as would be applied by a court on an application for judicial review. Section 179(5), however, provides that the CAT may: “(a) dismiss the application or quash the whole or part of the decision to which it relates; and (b) where it quashes the whole or part of that decision, refer the matter back to the original decision maker with a direction to reconsider and make a new decision in accordance with the ruling of the [CAT]”. Section 179(6) provides that an appeal lies on any point of law arising from a decision of the CAT. The “original decision maker” in section 179(5)(b) appears to be a reference to the CMA itself in section 179(1), so there is no presumption built in to the section, one way or another, that the decision will be referred back to the same inquiry group within the CMA.

Paragraphs 34-38 of schedule 4 to the 2013 Act provide that the members of a CMA group are to be selected by the chair of the CMA from the CMA panel, and that each group must consist of at least 3 members of the CMA panel. Paragraph 41 of schedule 4 to the 2013 Act allows the chair of the CMA to remove a member of a CMA group if the member will be unable to perform his duties or if, because of a particular interest of that member “it is inappropriate for him or her to remain a member of the group”. Paragraph 49 of the same schedule expressly provides that “[i]n making decisions that they are required or permitted to make … CMA groups must act independently of the CMA board”.

Rule 55(2) of the Competition Appeal Tribunal Rules 2003 (SI No. 1372 of 2003 as amended by SI No. 2068 of 2004) (the “CAT rules”) provides as follows in relation to costs: “[t]he [CAT] may at its discretion … at any stage of the proceedings make any order it thinks fit in relation to the payment of costs by one party to another in respect of the whole or part of the proceedings and in determining how much the party is required to pay, the [CAT] may take account of the conduct of all parties in relation to the proceedings”.

The CAT dealt at some length in paragraphs 5 to 54 with the factual background to the challenge. These paragraphs also referred in some detail to the parts of Mr Witcomb’s evidence upon which HCA had relied for its contentions that the CMA had behaved either unfairly or in bad faith, and had still not understood the nature of that unfairness. At various points in the summary, the CAT expressed itself in positive terms about Mr Witcomb’s evidence or the CMA’s approach. For example, it described the CMA’s view that it was not appropriate to trespass onto the precise extent, impact and effect of the “errors” identified in KPMG’s report as a “responsible approach for the CMA to adopt”; it accepted Mr Witcomb’s evidence of what had been said at the meeting with UKTI on 2^nd December 2014 (albeit concluding that it had not been a good idea for him to have gone along in the first place); and it concluded that Mr Witcomb had made no statement to suggest that he or the inquiry group had pre-determined views on HCA’s case on the matters to be remitted to the CMA. The CAT also rejected HCA’s contention that, by deleting large parts of its defence by amendment, it was retracting those parts of its factual case, rather than narrowing the areas of active dispute between the parties in the light of the CMA’s concessions. It made clear that Mr Witcomb’s 3^rd statement had given largely the same explanations as had been deleted in the amended defence.

The CAT then cited extensively from Heard v. Sinclair Roche & Temperley [2004] IRLR 763 (Burton J, Mr D Bleiman and Mr B M Warman) (the “Sinclair Roche case”), and drew attention to the fact that the applicable standards of independence and impartiality were not necessarily the same for the CMA, as a specialist administrative body, as they would be for a judicial tribunal.

Where the CMA was taking decisions that were determinative of civil rights and obligations, the CAT held that the requirements of article 6 of the ECHR may be satisfied where the area is a specialised one and there is a combination of fair treatment by the administrative body and the availability of judicial review, and it may be appropriate to allow the administrative body a substantial margin of appreciation.

The CAT also held that it was inherent in the requirement of fairness that there should be no bias or pre-determination or appearance of bias or pre-determination by the CMA in its consideration. In the circumstances, the Sinclair Roche case provided relevant guidance in respect of remission.

Remission to a new inquiry group and a new case team was not appropriate in this case, according to the CAT, because (i) no case of actual bias or pre-determination had been made out, (ii) there were no objective grounds to consider that the inquiry group might fail to approach the redetermination professionally and with a genuine open mind (as Mr Witcomb had said in evidence it would), (iii) the CMA was composed of persons of high repute and professionalism, who could be expected to understand their obligations of open-mindedness and be sincere and genuine in their efforts to achieve it, (iv) the CMA investigation was neither wholly flawed nor completely mishandled; the CMA’s explanations for their contested judgments could not be discounted, (v) on the evidence, HCA’s allegations that the CMA had knowingly consulted on a false basis, conducted a sham consultation, deliberately misled HCA, discriminated against HCA, and behaved dishonestly and in bad faith, were not accepted; the CMA explanations appeared reasonable, (vi) a fair-minded and informed observer would not conclude either that there was a real risk or a “real possibility” that the inquiry group and the case team would be biased against HCA, or that the risk of confirmation bias was such as to infringe the relevant standard, (vii) HCA had not given proper weight to the professionalism that could be expected of a body like the CMA and the inquiry group, (viii) it could not be said that the inquiry group were actually or apparently so thoroughly committed that a rethink appeared impracticable, (ix) the CMA had not conducted itself with conspicuous unfairness or completely mishandled things, and (x) neither the CMA’s failure to give disclosure before the disclosure ruling nor its failure to concede ground 2 of HCA’s challenge provided objective grounds for thinking it was biased, pre-judging or acting on a wholly flawed basis.

There were persuasive reasons why, where no bias or pre-determination was made out, the matter should be remitted to the same inquiry group, including the greater practical difficulties and risks of error of remission to another inquiry group, and the additional costs and delay involved in doing so.

At paragraph 68 of its ruling, the CAT explained why it had decided to reserve the costs of HCA’s grounds 2 to 5. It regarded that as the just, appropriate and fair course. It said that it was likely that the CAT would in due course closely examine the merits of HCA’s arguments on grounds 2 to 5 in relation to the self-pay AEC decision, which remained alive, and in relation to a new challenge to any remade insured AEC and divestment decisions. The CAT was more likely to be better informed in the future as to who ought fairly to bear these costs. Accordingly, the CAT stayed HCA’s application for costs of HCA’s grounds 2 to 5 pending the CMA’s redetermination of the 2 decisions.

At paragraph 76 of its ruling, the CAT explained why it was refusing HCA’s application for the costs of the data room exercise to be included in its costs of the application. The costs of the data room exercise were not incurred as a result of any unlawful action by the CMA. The CAT held that the effect of its disclosure ruling was to give HCA the opportunity to review the IPA data in the way that HCA should have been able to do during the administrative investigation phase of the enquiry. Had that occurred, it was common ground that HCA’s costs of the data room exercise would not have been recoverable from the HCA. In these circumstances, awarding HCA its costs would give it an unmerited windfall.

HCA’s grounds of appeal can be briefly summarised as follows:-

i)

The CAT misconstrued the guidance given in the Sinclair Roche case, failing to give proper regard to the public interest in maintaining confidence in the CMA’s decisions (the “public confidence ground”).

ii)

The CAT ought to have held that, viewed objectively, the CMA’s investigations had been wholly flawed and mishandled, and should not have placed reliance on the fact that the CMA had not behaved dishonestly so as to justify sending the matter back to the original inquiry group (the “mishandling ground”).

iii)

The CAT conflated the issues of appearance of bias and the risk of confirmation bias, failing to consider the latter separately, and improperly concluding that there was no appearance of bias or confirmation bias (the “bias ground”).

iv)

The CAT wrongly took into account the practical difficulties that the CMA would face if required to establish a new inquiry group and a new case team (the “practicalities ground”).

In oral argument, Ms Dinah Rose QC, leading counsel for HCA, adjusted the emphasis of her submissions. She relied on 3 central matters that she submitted demonstrated inappropriate unfairness and a lack of understanding by the original inquiry group that it had acted unfairly. When pressed about whether HCA pressed its allegations of bad faith, she replied that HCA could not say whether the inquiry group had acted in bad faith or not, but the answer to that question could not affect the fact that there had been pervasive unfairness. The 3 matters were as follows:-

i)

First, Ms Rose submitted that the CMA had told all the other participants about the revisions it had made to the IPA in early 2014, yet it had not informed HCA, and had actually misled HCA in its email of 12^th February 2014. Moreover, the CMA had proceeded with an oral hearing on 18^th February 2014 at which HCA’s experts were making submissions, and at which the IPA had been debated and the CMA had remained silent about the substantial changes it had made to the modelling (the “revised IPA concealment ground”).

ii)

Secondly, Ms Rose drew attention to Mr Witcomb’s explanation of the alleged unfairness in his 3^rd statement, which had demonstrated that neither he nor the CMA had any proper understanding of why the procedure that had been adopted was unfair and inappropriate (the “realisation of unfairness ground”).

iii)

Thirdly, the UKTI incident demonstrated that the CMA was maintaining that the effect of the errors in the revised IPA did not affect the conclusions and decisions that the CMA had reached. This was inconsistent with its stance that all the errors would be looked at again with a view to remaking the decision once all the errors had been corrected and accurate data was available (the “UKTI ground”).

Ms Rose’s basic submission was that neither HCA nor the public could have confidence that the existing inquiry group and project team would behave competently and fairly in the future, when they had repeatedly failed to do so in the past. She relied on the whole history of the investigation and the CMA’s repeated refusal to accept that it had behaved wrongly or unfairly. It was, as she put it, an issue of “confidence and competence”.

The CMA submitted that HCA’s factual approach was partial and misleading. In essence, the CMA contended that Mr Witcomb had provided proper explanations for all the alleged unfairness in the procedure so that the CAT had been right to conclude that HCA had not been improperly or unfairly singled out for special treatment.

In relation to HCA’s main complaints, the CMA relied on Mr Witcomb’s explanations for the CMA’s approach: (a) it was a coincidence occasioned by the statutory imperative of promulgating its final decision in April 2014 that the provisional decision on remedies was published on the same day, 21^st January 2014, that the revised IPA was approved, (b) the other participants had only been told of the revised IPA in the context of their confidential information concerning hospitals outside London going across to Nuffield, whereas HCA with its London-based hospitals was not affected by that issue, (c) there was no need to re-consult on the revised IPA since the revision had itself taken into account HCA’s submissions and the CMA was not aware of any errors in the revised IPA, (d) the oral hearing of 18^th February 2014 was simply to allow HCA to elaborate on its written submissions on the provisional findings and the provisional decision on remedies, and was not affected by the revisions to the IPA, (e) an overlap between different phases of the investigative process was inevitable given the tight timescales, (f) Mr Witcomb’s explanations for what he had said to UKTI and what had been written to HCA in relation to the revised IPA were entirely appropriate and credible, and (g) the CAT’s findings on the evidence had been well-grounded and correct.

The CMA drew attention to the fact that its evidence had been filed in answer to the detailed allegations in HCA’s solicitors’ letter of 4^th December 2014 and not the later refinements to HCA’s case that were advanced before the CAT and in this court. Those allegations suggested that the CMA had acted deliberately to disadvantage HCA, which was without foundation.

Ms Kassie Smith QC, leading counsel for the CMA, submitted that the overarching principle was that the court should not remit to the same tribunal unless it is satisfied that the parties would not enjoy a fair hearing of their case on remission. In practical terms, Ms Smith said that that would be if either the original decision maker no longer had an open mind, if there was bias or an appearance of bias, or if the original decision maker had so completely mishandled the case as to show itself to be incompetent to handle the matter fairly on remission. She submitted that the touchstone was the public law duty of fairness as was reflected in paragraphs 80-82 of the CAT’s ruling. The normal rule was to remit to the original decision maker, and there was no basis here for the court to interfere with the CAT’s decision, since it was neither perverse nor exercised on a basis that was wrong in law.

The CAT relied, as I have said, on the Sinclair Roche case, where the Employment Appeal Tribunal set out the principles that it was applying in deciding to remit a discrimination case back to the same Employment Tribunal as had heard the original application. The guidance has been much debated in argument before us, so I shall set out the pertinent parts of the decision at paragraphs 45-46 as follows:-

“[45] We are satisfied that the Tribunal's conclusions that there was direct discrimination against SF and SH (the referrals issue) and indirect discrimination of SF (the part-time working issue) cannot stand. Mr Gatt QC submits that the case should be remitted for hearing before a different Tribunal. Mr Bean QC submits that this would be catastrophic for his clients, and in any event unnecessary. He refers to the somewhat different procedure which we have been adopting at the preliminary hearing, and indeed sift, stages, by reference to [English v Emery Reimbold and Strick Ltd [2002] 1 WLR 2409], as explained in Burns v Consignia (No 2) [2004] IRLR 425, by way of what he referred to as a carefully controlled remission. That however is a practice which is adopted at the interlocutory stage where, inter alia, there is a case alleged of inadequacy of reasoning, or absence of a finding, and the case is sent back to the same tribunal simply to answer specific questions, based on its existing notes of evidence. That is not the case here, where we have concluded that the Tribunal has in fact not done, or at any rate finished, its job. This is not a question of what has been described in Burns as a referral back, but of a straightforward remission. The issue nevertheless remains as to whether it should be remitted back to the same Tribunal, which, subject to our guidance, will be able to make use of its existing knowledge of the case and notes of evidence, or a fresh tribunal to start again.

[46] There is no authority which has been cited to us, or of which we ourselves know, which would assist us in such a situation, and we set out what appear to us to be relevant factors:

46.1

Proportionality must always be a relevant consideration. Here the award was for £900,000, and although we are conscious that ordering a fresh hearing in front of a different Tribunal would add considerably to the cost to parties on both sides who have already invested in solicitors and Counsel, both at the Tribunal and on appeal (in the case of the Applicants, two Counsel for the appeal), sufficient money is at stake that the question of costs would from the one point of view not offend on the grounds of proportionality and from the other not be a decisive, or even an important, factor. Similarly the distress and inconvenience of the parties in reliving a hearing must be weighed up, but (a) are rendered necessary in any event by the decision to set aside the original decision and (b) will not be greatly less by virtue of the extra time taken by a fully, rather than partially remitted, hearing, the main distress and inconvenience being caused by the matter being reopened at all.

46.2

Passage of Time. The appellate tribunal must be careful not to send a matter back to the same tribunal if there is a real risk that it will have forgotten about the case. …

46.3

Bias or Partiality. It would not be appropriate to send the matter back to the same Tribunal where there was a question of bias or the risk of prejudgment or partiality. This would obviously be so where the basis of the appeal had depended upon bias or misconduct, but is not limited to such a case.

46.4

Totally flawed Decision. It would not ordinarily be appropriate to send the matter back to a tribunal where, in the conclusion of the appellate tribunal, the first hearing was wholly flawed or there has been a complete mishandling of it. This of course may come about without any personal blame on the part of the tribunal. There could be complexities which had not been appreciated, authorities which had been overlooked or the adoption erroneously of an incorrect approach. The appellate tribunal must have confidence that, with guidance, the tribunal can get it right second time.

46.5

Second Bite. There must be a very careful consideration of what Lord Phillips in English (at para 24) called "A second bite at the cherry". If the tribunal has already made up its mind, on the face of it, in relation to all the matters before it, it may well be a difficult if not impossible task to change it: and in any event there must be the very real risk of an appearance of prejudgment or bias if that is what a tribunal is asked to do. There must be a very real and very human desire to attempt to reach the same result, if only on the basis of the natural wish to say "I told you so". Once again the appellate tribunal would only send the matter back if it had confidence that, with guidance, the tribunal, because there were matters which it had not, or had not yet, considered at the time it apparently reached a conclusion,

would be prepared to look fully at such further matters, and thus be willing or enabled to come to a different conclusion, if so advised.

46.6

Tribunal Professionalism. In the balance with all the above factors, the appellate tribunal will, in our view, ordinarily consider that, in the absence of clear indications to the contrary, it should be assumed that the tribunal below is capable of a professional approach to dealing with the matter on remission. By professionalism, we mean not only the general competence and integrity of the members as they go about their business, but also their experience and ability in doing that business in accordance with the statutory framework and the guidance of the higher courts. … where a tribunal is corrected on an honest misunderstanding or misapplication of the legally required approach (not amounting to a "totally flawed" decision described at 46.4), then, unless it appears that the tribunal has so thoroughly committed itself that a rethink appears impracticable, there can be the presumption that it will go about the tasks set them on remission in a professional way, paying careful attention to the guidance given to it by the appellate tribunal.”

We were also referred to the decision of the Court of Appeal (Chadwick and Buxton LJJ) in Loftus-Brigham v. Ealing London Borough Council [2003] EWCA Civ 1490; 103 ConLR 102, where a case was sent back to a different judge once errors had been identified in the original judge’s causation analysis. Chadwick LJ did not, as it seems to me, make any statement of principle when he said at paragraph 26 that the case should be remitted to another judge “in view of the judge’s strong expression of view, when the issue was ventilated with him after the trial, that the test that he had applied was in fact the same as that which the law requires”. He said that “in view of the history of the matter, it would be better for the matter to be approached afresh”. The report provides little detail about that history.

The relevant authorities on bias are not in dispute between the parties. In Porter v. Magill [2002] 2 AC 357, the House of Lords approved the test for apparent bias as being that the court must ascertain all the circumstances which have a bearing on the suggestion that the judge was biased, and “the question is [then] whether the fairminded and informed observer, having considered the facts, would conclude that there was a real possibility that the tribunal was biased” (see the speech of Lord Hope at paragraphs 102-3).

The first question relates to the appropriate principles to be applied. The parties agreed that, apart from the Sinclair Roche case, there was very little guidance to be obtained from the authorities.

It seems to me that it is necessary to understand the basis on which the court will act, having quashed a decision by a decision maker, whether that decision maker is an administrative body such as the CMA or a tribunal such as the employment tribunal. The principles are unlikely to be different as between remission to an administrative body and remission to a tribunal, although there may be different relevant factors arising, for example, from the size and composition of the body or tribunal in question.

Different considerations may apply to political bodies such as local authorities, since they have a legitimate policy making function, so I shall say nothing about that situation having heard no argument on the point. Moreover, different considerations may also apply to administrative bodies where statute provides for a single decision maker – that is not the case here under section 179(1) and (5) of the 2002 Act, as I have already explained.

The principle as it seems to me must be that remission will be made to the same decision maker unless that would cause reasonably perceived unfairness to the affected parties or would damage public confidence in the decision making process. The basis on which the court will approach these two interlocking concepts of “reasonably perceived unfairness to the affected parties” and “damage to the public confidence in the decision making process” may depend heavily on the circumstances of the remission.

A variety of factors will undoubtedly be relevant to the application of these principles. I would not want to limit those factors by setting out anything that could be regarded as an exhaustive list as the Employment Appeal Tribunal seems to have attempted to do in the Sinclair Roche case. There will be many different situations which cannot be predicted from a single case.

It is, however, always the case that the presence of actual bias, apparent bias or confirmation bias would make remission to the original decision maker undesirable, because any such bias would amount both to reasonably perceived unfairness to an affected party and potentially serious damage to public confidence in the decision making process.

It is important also to understand the kind of unfairness that is relevant to the question of whether the decision should be remitted to the original decision maker. The unfairness concerned is such as contravenes the public law duty of fairness, not some abstract concept of unfairness based on a colloquial usage of that word. It is wellestablished that what fairness requires will vary with the factual circumstances, but what is required in order to achieve fairness is a matter of law, and not a matter of discretion for the decision maker.

In these circumstances, in my judgment, it was incumbent on the CAT in this case to consider the factual circumstances in order to decide first whether remission to the CMA’s original inquiry group would give rise to a reasonably perceived unfairness to HCA, by reason of that inquiry group’s conduct of the original investigation. That process would also have entailed a consideration of whether remission to the original inquiry group would damage public confidence in the CMA’s decision making process. I emphasise that these were questions of law. I have no doubt that the CAT did seek to apply these principles. The question is whether it did so correctly. In this context, I turn now to deal with HCA’s attacks on the CAT’s reasoning. I will start with the 3 matters that Ms Rose emphasised in oral argument.

The revised IPA concealment ground

The burden of this complaint is that the CMA behaved unfairly in deliberately concealing or not informing HCA about the revised IPA either in the email of 12^th February 2014 or at the oral hearing on 18^th February 2014. Mr Witcomb’s explanation is that there were good reasons for telling the other participants, but no good reason to tell HCA, whose criticisms of the original IPA had already been taken on board and catered for in the revised IPA. The inquiry group had consulted once and had no need to re-consult.

It is first necessary to consider the status of allegations of bad faith made against the CMA. They have not, somewhat to my surprise, been disavowed even at the hearing before us. Yet Ms Rose did not, as far as we are aware, apply to the CAT to crossexamine Mr Witcomb. In these circumstances, it seems to me that the CAT was entirely justified in accepting at face value Mr Witcomb’s explanations of his conduct. The acceptance of Mr Witcomb’s evidence did not, of course, mean that the actions of the inquiry group were automatically to be regarded as satisfying the public law duty of fairness.

The CAT did not deal in any detail with the substance of these allegations but concluded generally that HCA had not been improperly or unfairly singled out for special treatment.

In my judgment, the CMA made a deliberate decision not to tell HCA (and certain other hospital operators) about the revised IPA. That much is confirmed by paragraph 80 of Mr Witcomb’s 2^nd statement and by parts of his 3^rd statement. The inquiry group decided that it was not necessary to re-consult on the revised IPA for understandable reasons.

I have considered carefully whether Mr Wood’s email of the 12^th February 2014 was misleading. In my judgment, it was not. It informed HCA that Nuffield was being given access to the disclosure room to review the IPA. It did not condescend to any particularity as to the detail of the IPA that Nuffield was to review. The CMA’s focus is emphasised by the last part of the email that confirms that the material disclosed to Nuffield did not contain information confidential to HCA. I accept that the email may have been carefully drafted to avoid having to inform HCA about the revised IPA, but that does not make it misleading.

In my view, the key allegation here concerns the oral hearing. It is, of course, correct to point out that only a relatively minor part of the hearing concerned the detail of the IPA. But that does not mean that Mr Witcomb and his colleagues would not have been conscious of the fact that the inquiry group had just told 11 other participants about the revised IPA and had not told HCA. They would have been fully aware that HCA was making submissions about the details of the IPA which, unknown to HCA, the CMA had just revised. It would have been the easiest thing in the world for the CMA to tell HCA at the hearing that its criticisms of the original IPA had been accepted. The CMA could have told HCA exactly what it had explained to the 11 other participants in its letters of 6^th February 2014. It could have gone on to say no further details were being made available because the inquiry group had decided that detailed re-consultation was not required, and there was no time for it. Nothing in Mr Witcomb’s detailed explanation explains why this was not the fair and transparent course to adopt. It is undoubted that in the colloquial sense that I have referred to, HCA could properly have regarded the inquiry group’s conduct in keeping quiet in a sphinx-like manner at that hearing as unfair to HCA. Moreover, in my judgment, the conduct was also a breach of the CMA’s public duty of fairness, since it meant that

The proper question for the remitting tribunal is a different one, as I have indicated, namely whether remission to the CMA’s original inquiry group would give rise to a reasonably perceived unfairness to HCA, by reason of that conduct. That question and the question of whether remission to the original inquiry group would damage public confidence in the CMA’s decision making process can only be answered against the background of all the facts and not just on the basis of a single instance of unfair conduct.

The realisation of unfairness ground

Ms Rose’s point here was that Mr Witcomb’s evidence gives the impression that he still does not understand that it was unfair and inappropriate to withhold from HCA the information that the IPA had been revised. Accordingly, there cannot be any confidence that, in the future, this inquiry group will behave fairly towards HCA. Ms Rose relies also in this connection on the entire history of the CMA’s approach to the investigation, and in particular its failure to take heed of the CAT’s strictures in its 2^nd October 2013 judgment that a high degree of disclosure and transparency on the part of the CMA was required, given the technical nature of the material.

I accept that Mr Witcomb’s evidence continued to advance justifications for the inquiry group’s approach. I also do not accept Mr Witcomb’s purported justification of his failure to tell HCA at the oral hearing about the revised IPA. In paragraph 47 of his 3^rd statement, Mr Witcomb suggests that the inquiry group was going through a mechanical process in allowing HCA’s experts to repeat the points they had made in writing. But that sees the matter entirely from the CMA’s end of the telescope. Fairness is a two-way street. HCA would have been justifiably aggrieved by the later discovery that the basis of the IPA it was addressing had been changed, albeit that those changes were largely responsive to the submissions it and others had made.

Mr Witcomb’s failure to accept that the hearing had been undertaken unfairly may add something, but perhaps not a huge amount, to the act of unfairness itself. As the CAT said, Mr Witcomb offered reasonable explanations (perhaps justifications) for what had occurred, even if in my view they were not wholly adequate ones. I shall return to this aspect in due course.

The UKTI ground

In my judgment, this ground takes HCA nowhere. Mr Witcomb has made it clear that he did not say to UKTI that the modelling errors that had been identified did not change the CMA’s actual results or conclusions. In other words, Mr Witcomb’s evidence was that he said nothing about the consequences of the errors in terms of the CMA’s conclusions, including the appropriateness of the divestment decision. Taken against the backdrop of Mr Witcomb’s assurance that he retained an open mind, it seems to me that HCA could not have hoped to persuade the CAT that Mr Witcomb

had said something he denied having said without cross-examining him, which HCA did not do.

In my view, the CAT was fully entitled to accept Mr Witcomb’s version of events. This allegation does not, I think, take HCA’s case on remission any further forward.

I turn now to consider HCA’s main grounds of appeal on the remission issue, before considering whether the CAT was mistaken in remitting the matter to the original inquiry group.

The public confidence ground and the mishandling ground

I can take these two grounds together as they are so intimately connected. HCA contends that, having accepted the guidance in the Sinclair Roche case, the CAT ought to have concluded that the CMA’s conduct meant that there could be no public confidence in the inquiry group’s future decision making. That lack of public confidence is said to arise from the mishandling of the investigation in the numerous ways relied upon by HCA. These include, again, failing to heed the CAT’s earlier warnings, and also failing to accept HCA’s repeated invitations to concede that the CMA had behaved unfairly, and the alleged catalogue of errors in the IPA and the revised IPA. HCA criticises the CAT for having confused the allegations of dishonesty with the allegations of wholesale mishandling. HCA also relies in this regard on the allegations of bias that it has made. I prefer, however, to deal with all the allegations of bias separately under a discrete heading. I should say also that I do not accept that the CAT confused the allegations of dishonesty with the allegations of mishandling. They are dealt with separately albeit concisely, in paragraph 84 of the ruling.

I entirely accept that public confidence in the decision making of an administrative body like the CMA is of central importance. But I do not accept that the CAT failed to give adequate weight to this factor. The CAT’s core reasoning is contained in paragraph 85 of the ruling where it makes clear that HCA’s submissions failed to give proper weight to the professionalism that can be expected of a body like the CMA and its inquiry group. It would be a very serious matter indeed if the CAT had taken the view, as HCA urged, that the public could not have confidence in the future decision making of this inquiry group, comprising as it does a number of highly experienced professional people. In my judgment, the catalogue of factors that HCA advanced did not lead to that inevitable conclusion.

I do not see how the CMA defending legal proceedings on legal advice can be regarded as evidence of unfairness or unreasonableness. Many of the criticisms levelled at the CMA relate to its legal handling of the investigation when challenged by BMI, HCA and Spire. We too have been treated to a blow-by-blow account of the history of the investigation and the litigation over many months, and I can say that, in common with the CAT, I did not form the view that there had been a complete mishandling as had been alleged. The errors in the IPA and the revised IPA do not amount to mishandling, even if the ones that are not now admitted are all ultimately made out. These conclusions do not detract from the specific issues that I have already indicated did amount to unfairness, but I do not accept the CMA’s approach was wholly flawed from start to finish. It has been a long and complex process and it is easy for a participant advised by the best experts to make it look as if specific identified errors invalidate the whole exercise. The approach is not necessarily correct. Of course, the CMA must now go back and look again with an open mind at its calculations and its conclusions, but it has assured HCA and the CAT that it will do so.

The CAT had made clear that there was a particular need for transparency, but the CMA was still handicapped by its need to reach a timeous conclusion. It had to balance these two factors, and it is impossible, in my judgment, to say that it mishandled the disclosure or the process of consultation in its entirety.

The bias ground

I turn then to the question of bias, which I have already explained would, if made out, lead to the conclusion that there existed reasonably perceived unfairness to HCA and potential damage to public confidence in the CMA’s decision making process.

Before this court, Ms Rose restricted her submission to that of apparent bias and confirmation bias. She did not allege actual bias. In relation to apparent bias, Ms Rose relied on all the allegations that she had advanced under other heads. The question, as I have said, is whether the fair-minded and informed observer, having considered the facts, would conclude that there was a real possibility that the CMA’s inquiry group was biased. It seems to me, largely for the reasons I have already adumbrated that the only sensible basis for this allegation is the CMA’s conduct of the oral hearing. In the circumstances, I propose to consider that aspect in the round when I come to assess whether, in all the circumstances, the CAT was right to remit the 2 decisions to the existing inquiry group.

I do not, however, accept that there was or is a risk of confirmation bias. The inquiry group is, as I have said, composed of highly experienced professionals. Mr Witcomb has made clear that the inquiry group will approach the process of remaking the decisions with an open mind, and I do not accept that the errors made in the past process give rise to an inference that he or his inquiry group or the case team would be affected by confirmation bias.

The practicalities ground

Ms Rose argued strenuously that there was no basis for the CMA to have relied upon the practical difficulties of appointing a new case team and a new inquiry group, when Mr Witcomb had produced no evidence of these difficulties. She criticised with particular severity the reliance by the CAT on the supposed “significantly greater difficulties and risks of error going forward” if there were remission to a new inquiry group. It is true that Mr Witcomb only said that the composition of the inquiry group and the staff team should be left to the CMA to determine. But HCA had itself made the positive submission that the additional time and cost of remission to a new inquiry group was justified on proportionality grounds by the importance of the divestment to HCA’s economic position. I have no doubt that the CAT was entitled to have regard to the practical position, though I accept that it did not go to the main questions of unfairness and public confidence that the CAT had to consider. In my judgment, the CAT came to consider the practicalities after it had addressed the overarching questions. It expressly said that practicalities were only relevant if no case of bias or pre-judgment were made out. It is, however, worth stating that tribunals remitting decisions should be careful to address the main questions first before considering the practical issues that might arise if remission to a new decision maker were necessary.

Was the CAT right to remit the 2 decisions to the existing inquiry group?

It is necessary now to draw the threads of the case together. I have rejected two of the main planks of HCA’s arguments based on the UKTI meeting and the 12^th February 2014 email, but accepted that the CMA failed unfairly to inform HCA about the revised IPA at the oral hearing. Whilst the relationship between HCA and the inquiry group seems to have been hostile and confrontational all along, I do not think that the CMA’s conduct generally was in breach of its public law duty of fairness. Much of what it did was in response to legal challenges. There were lawyers on both sides, and, as Ms Smith emphasised and I have already said, the CMA had a tight timetable to meet and many participants in the process.

The question then boils down to whether the unfairness of the CMA in its handling of the oral hearing and in failing to disclose what it knew and what it knew HCA did not know, namely the existence of a revised IPA, taken together with Mr Witcomb’s unwillingness to accept he had done anything inappropriate, means that the remission ought to have been made to a fresh decision maker. I have given this question anxious thought. It seems to me that the issues of (a) whether these matters would lead the fair-minded and informed observer to conclude that there was a real possibility that the CMA’s inquiry group was biased, and (b) whether these matters lead to the conclusion that remission to the same inquiry group would cause reasonably perceived unfairness to HCA or would damage public confidence in the decision making process, are in the circumstances of this case intimately related.

Ultimately, I have concluded, applying these tests, that there was no apparent bias and that remission to the same inquiry group would not cause reasonably perceived unfairness to HCA or affect public confidence in the CMA’s decision making process. My reasons may be shortly expressed. It seems to me that Mr Witcomb and his team were in a difficult position at the oral hearing. Time was short and they knew that HCA would want to reopen a number of issues concerning the IPA if they had the chance. This was because HCA was resisting a significant divestment decision that the inquiry group thought, in good faith (albeit perhaps mistakenly), had increased legitimacy as a result of the application of the revised IPA. The fact that the inquiry group turned out to have made some errors of calculation does not affect the fairness of its approach to the oral hearing. Thus, whilst the inquiry group’s decision not to disclose the revised IPA to HCA was, I think, wrong and unfair to HCA, I do not think that it means that either HCA or an objective observer would be reasonably justified in a perception that the inquiry group would act unfairly in the future. Nor do I think that Mr Witcomb’s failure to accept that he had behaved unfairly makes much difference. One would have expected him to explain from his own point of view why he had done what he did. Indeed, it is aspects of that explanation that have weighed heavily in my approach to his mistaken behaviour.

The inquiry group and its case team were, as I have said, composed of experienced professionals. They had and have no personal interest in the case against HCA. They took one wrong turning but that does not make it reasonable to perceive they would do the same again. Ms Rose’s reliance on the failure to heed the CAT’s earlier warning about transparency is strong in theory, but not very persuasive in practice,

since in February 2014, the inquiry group was under pressure to produce its final report by 3^rd April 2014, and must have wanted to get over all the necessary consultation hurdles as speedily as possible. As I have already said, I do not think that the errors in the revised IPA are relevant to this consideration. Errors may have been made in the investigation process, but that did not make the process wholly flawed or fundamentally mishandled.

Finally, in this regard, Ms Rose sought to argue that the CAT had been wrong to reject her argument that the CMA’s failure to concede ground 2 of its application, namely that the material errors in the IPA meant that they did not provide a sustainable basis for the 2 decisions, was itself evidence of apparent bias. The CMA conceded the R² error and the statistical error, but not most of the other alleged errors, but did not concede the consequences of those errors. It seems to me to have been entirely sustainable for the CMA to keep open its position on what conclusions ought to be drawn in the decision making process once the errors (admitted and not admitted) had been reviewed and corrected in the context of a reconsideration of the 2 decisions. This was all the CAT was saying in paragraph 87 of its ruling.

As I have already pointed out, it is not every unfairness in a process that will lead to a conclusion that remission must be to a fresh decision maker. The past unfairness must actually lead to the reasonable conclusion that the original decision maker would or might act unfairly if the decision were remitted to it. I do not think, in this case, that such a conclusion was justified.

In the circumstances, whilst I do not accept all the CAT’s reasoning, I agree with its conclusion that the 2 decisions should be remitted to the original inquiry group, assisted by the original case team if the CMA decides that that would be convenient and appropriate. I would, therefore, dismiss the appeal on the remission issue.

The CAT has a wide discretion as to costs under rule 55(2) of the CAT rules. At first sight the decision to reserve the costs of grounds 2-5 seems odd, when one considers, as HCA submitted, that the 2 decisions were quashed and grounds 2-5 comprised further reasons why they should be quashed. In formal terms, grounds 2 to 5 are of no further relevance as regards the 2 decisions, because the two decisions have been remitted. Ms Rose relied on the dictum of Lord Neuberger MR in R(M) v. Croydon London Borough Council [2012] 1 WLR 2607 at paragraphs 55 and 59 to the effect that a successful applicant for judicial review is entitled to all his costs even if only one ground is conceded, unless there are good reasons to the contrary. She made clear that she was not seeking HCA’s costs of the challenge to the self-pay AEC decision, which remains in existence and subject to challenge under grounds 2 to 5. Finally, Ms Rose submits that the effect of the reservation of the costs is to give the CMA a personal interest in the outcome of the investigation, since if there is a renewed challenge on grounds 2 to 5, the CMA risks having to pay those costs, which, if it decides the matter differently, it may never have to do.

In my judgment, the CAT was entitled to make the order it did. The CAT may well end up determining grounds 2 to 5 either in relation to the self-pay AEC decision or in relation to the remade decisions on insured AECs and divestment. The costs already incurred will not have been wasted if that occurs. It seems therefore a practical and reasonable approach to reserve the costs of the grounds 2 to 5 challenges pending the outcome of the reconsideration of the insured AEC decision and the divestment decision. I do not accept that the reservation of the costs gives the CMA a personal interest in the litigation. If it decides in such a way that the grounds 2 to 5 challenges remain alive, it risks paying HCA’s reserved costs, but that will depend on the arguments advanced at and the outcome of an entirely new hearing. The members of the inquiry group anyway have no personal financial interest in their decision making.

In my judgment, the CAT was right to exclude the data room costs from the award of costs to HCA. Those costs would have been incurred anyway, even if no challenge to the CAT had been necessary. In those circumstances, they would not have been recoverable and there is no reason why they should be recoverable merely because they were incurred after a CAT hearing.

For the reasons I have sought to give, I have concluded as follows on the remission issue:-

i)

Save in the respects mentioned in the next sub-paragraph, the CMA did not behave unfairly towards HCA in the investigation and did not mishandle the investigation, despite the errors in the revised IPA.

ii)

The CMA behaved inappropriately and unfairly in failing to inform HCA about the revised IPA at the oral hearing on 18^th February 2014. Mr Witcomb’s evidence failed, as it should have done, to accept that the inquiry group’s conduct was inappropriate and unfair in that respect.

iii)

That inappropriate and unfair conduct in the conduct of the investigation does not mean, in all the circumstances of this case, that the 2 decisions should be remitted to a fresh decision maker, because these matters (a) would not lead a fair-minded and informed observer to conclude that there was a real possibility that the CMA’s inquiry group is or was biased, (b) do not lead to the conclusion that remission to the same inquiry group would cause reasonably perceived unfairness to HCA or would damage public confidence in the CMA’s decision making process.

I would therefore dismiss the appeal on all grounds.

I agree.

I also agree.