Competition and Markets Authority v Flynn Pharma Ltd & Ors

Rule 104 of the Competition Appeal Tribunal Rules 2015 (“the CAT rules”) provides that the Competition Appeals Tribunal (“the CAT”) may “at its discretion … make any order it thinks fit in relation to the payment of costs in respect of the whole or part of the proceedings”. It is common ground that this rule confers a discretion on the CAT. The question on this appeal is whether there is a starting point for the exercise of that discretion; and, if so, what it is.

The proceedings with which we are concerned arose out of an investigation by the Competition and Markets Authority (“the CMA”) into the pricing of an epilepsy drug called phenytoin sodium. By its decision of 7 December 2016, after an investigation lasting over three years, it found that Flynn and Pfizer had abused their dominant positions in the UK market for phenytoin sodium capsules under both domestic and EU competition law by charging excessive prices. It imposed a fine of £84.2 million on Pfizer and £5.2 million on Flynn.

Both Flynn and Pfizer appealed to the CAT. The CAT recognised in its substantive decision the “importance of this case for the public interest”: [2018] CAT 11 at [5]. After a four week trial, the CAT found that although Flynn and Pfizer held dominant positions in the market in question, the CMA had made errors in deciding that they had abused their positions. The CAT therefore set aside the penalties and remitted the question of abuse to the CMA for redetermination. The CMA’s subsequent appeal to this court in large part failed; but that court also recognised that the issues “were of considerable public importance”: [2020] EWCA Civ 339 at [15].

Following its decision on the substantive appeal, the CAT considered the question of costs. In its ruling of 29 March 2019 ([2019] CAT 9) it held that the established practice of the CAT in appeals of this nature was that the starting point for the exercise of discretion was that the unsuccessful party should pay the successful party’s costs. I refer to this starting point as “costs follow the event”. It considered the decision of this court in British Telecommunications plc v Office of Communications (“BT v Ofcom”) [2018] EWCA Civ 2542, [2019] Bus LR 592, but decided that the decision in that case did not require a different starting point. From that starting point the CAT considered who had won and who had lost on the various main issues. Its ultimate ruling was that the CMA should pay a proportion of the costs of both Flynn and Pfizer.

The CMA challenge that conclusion. In the skeleton argument filed on their behalf, it was argued that the starting point is that no order for costs should be made against a public body performing its functions in the public interest unless it has acted unreasonably. They described this as “the Principle” (with a capital P). The Principle may be displaced where rules of court or equivalent lay down a different starting point (such as the general rule in CPR Part 44.2 (2) (a) that the unsuccessful party pays the successful party’s costs). But there is nothing in the CAT rules which displaces the Principle. In the oral submissions presented by Sir James Eadie QC, the CMA adopted a more nuanced position. The principle for which the CMA now argues is

Flynn and Pfizer, on the other hand, say that it was open to the CAT to adopt “costs follow the event” as its starting point; and that, as the CAT held, the decision of this court in BT v Ofcom does not require a different approach.

The CMA was established by section 25 of the Enterprise and Regulatory Reform Act 2013. Section 25 (3) provides:

In very broad terms, the CMA has succeeded to the powers and duties of the Office of Fair Trading (and before that the Director General of Fair Trading) in investigating and dealing with anti-competitive practices.

The CMA has a range of powers which enable it to carry out an investigation where it has reasonable grounds to suspect an infringement of competition law. As Mr Brealey QC pointed out on behalf of Pfizer, those powers are extensive. They include power to enter business premises without a warrant, power to require the provision of documents or information, and power to interview persons. An investigation by the CMA into a suspected breach of competition law may be, and often is, protracted. As mentioned, in the present case the investigation took over three years. If, as a result of its investigations, the CMA is minded to find a breach of competition law, it must give notice to the persons likely to be affected, and give them an opportunity to make representations. Those representations may be (and usually are) followed up by a hearing which, Mr Brealey said, is to all intents and purposes akin to a trial (although this latter point was disputed by Sir James). This stage (referred to as the administrative phase) can be very expensive for the subjects of the investigation, requiring the engagement of both lawyers and expert economists, as well as the expenditure of management time; and any expenses they incur in the administrative phase cannot be recouped from the CMA, whatever the outcome of the investigation or an appeal to the CAT.

Once the CMA has decided that there has been a breach of competition law, it has a range of powers open to it. These include the making of directions designed to bring the anti-competitive conduct to an end. The CMA has power to apply to the court to enforce such directions. Section 36 of the Competition Act 1998 enables the CMA to impose financial penalties on undertakings if it is satisfied that the infringement of competition law has been committed intentionally or negligently by the undertaking. A decision by the CMA to this effect is an appealable decision, giving rise to a right of appeal to the CAT: Competition Act 1998 s 46.

The CAT must determine the appeal on the merits by reference to the grounds of appeal set out in the notice of appeal: Competition Act 1998 Sched 8 para 3 (1). Because the appeal is an appeal “on the merits” rather than akin to judicial review, the CAT will often hear detailed evidence, both factual and expert. This may in some cases include evidence that was not considered by the CMA in arriving at its contested decision. The CAT may confirm or set aside the decision which is the subject of the appeal, or any part of it, and may:

remit the matter to the CMA;

impose or revoke, or vary the amount of, a penalty;

give such directions, or take such other steps, as the CMA could itself have given or taken; or

make any other decision which the CMA could itself have made.

The CMA is not the only body with power to enforce competition law. The same scheme applies to regulators in the communications and postal services, electricity, gas, water and sewerage, railways, air traffic and air operation services, payment systems, healthcare services in England and financial services sectors.

In addition to dealing with appeals against infringement decisions, the CAT also deals with challenges to decisions of the CMA or the Secretary of State in connection with a market investigations or merger situations. In that kind of case, the appeal is not an appeal on the merits, but the CAT applies the principles of judicial review. The principles of judicial review have applied to an appeal to the CAT against a decision of Ofcom under the Communications Act 2003 since July 2017 (when s.194A of the Act was enacted).

Rule 4 of the CAT rules is headed “Governing principles”. It provides, so far as relevant:

As noted, rule 104 of the CAT rules gives the CAT the power to make such costs order as it thinks fit. Rule 104 (4) lists a number of factors that the CAT may take into account in making an order. They are:

These factors are similar to those in the CPR; but there is no equivalent to the general rule under the CPR that “the unsuccessful party will be ordered to pay the costs of the successful party”.

Rule 104 (5) provides:

Rule 104 (5) does not prescribe the basis on which costs will be assessed; unlike CPR Part 44.3 which states that the court “will” assess costs either on the standard basis or the indemnity basis. But it is to be expected that where the CAT refers the assessment of costs to a costs officer of the Senior Court that officer will apply one or other of the two bases for which the CPR provides.

The main building blocks in the CMA’s argument are four decided cases: Bradford MDC v Booth (2000) 164 JP 485; Baxendale-Walker v Law Society [2007] EWCA

Bradford is a decision of the Divisional Court, presided over by Lord Bingham CJ.

The argument for the Council on appeal, recorded at [10], was that “it can never be just and reasonable to order a local authority to pay costs in the absence of bad faith or unreasonable behaviour.” At [22] Lord Bingham rejected that submission, because

What the court will think just and reasonable will depend on all the relevant facts and circumstances of the case before the court. The court may think it just and reasonable that costs should follow the event, but need not think so in all cases covered by the subsection.

Where a complainant has successfully challenged before justices an administrative decision made by a police or regulatory authority acting honestly, reasonably, properly and on grounds that reasonably appeared to be sound, in exercise of its public duty, the court should consider, in addition to any other relevant fact or circumstances, both (i) the financial prejudice to the particular complainant in the particular circumstances if an order for costs is not made in his favour; and (ii) the need to encourage public authorities to make and stand by honest, reasonable and apparently sound administrative decisions made in the public interest without fear of exposure to undue financial prejudice if the decision is successfully challenged.”

This case, therefore, does not explicitly support the proposition that the starting point is that no order for costs should be made. The need to encourage public authorities to make and stand by decisions is no more than a factor to be considered. Sir James submitted, however, that subsequent decisions of this court have indeed elevated that factor into a starting point or default position.

In Baxendale-Walker the Law Society instigated disciplinary proceedings against a solicitor in the Solicitors Disciplinary Tribunal (“the SDT”). One of the two allegations of conduct unbefitting a solicitor was not proved, but the second was admitted and the Tribunal found the solicitor guilty of unbefitting conduct and suspended him from practice for three years. The SDT ordered the Law Society to pay 30 per cent of the solicitor’s costs. The Law Society’s appeal to the Divisional Court succeeded; and this court dismissed the solicitor’s appeal against the decision of the Divisional Court. The SDT had statutory power “to make such order as it may think fit, and any such order may in particular include provision for any of the following matters … the payment by any party of costs or a contribution towards costs.” In the Divisional Court Moses LJ referred to Bradford and said:

But in another case (Law Society v Adcock [2007] 1 WLR 1096) Waller LJ had said that Moses LJ had put the matter too highly in favour of a regulator. He said that:

In this court in Baxendale-Walker Sir Igor Judge PQBD said that the apparent difference of opinion needed to be resolved.

He went on to say at [34]:

He concluded at [40]:

I take this last sentence to be explicit approval of what Moses LJ had said. Thus the difference of opinion between Moses LJ and Waller LJ was resolved in favour of the former. I find it difficult to see this as anything other than approval of a starting point or default position. I note, also, that Sir Igor described the bringing of disciplinary proceedings as the Law Society’s “exercise of … regulatory responsibility” rather than a regulatory obligation.

In Perinpanathan Mrs Perinpanathan's daughter, who was 15, was stopped at Heathrow Airport. She was carrying some £150,000 in cash. The cash was detained by the police on the basis that there were reasonable grounds to suspect it was intended for use in unlawful conduct, namely terrorism. On 13th December 2007, following a two-day hearing at Westminster Magistrates' Court, the justices declined to order forfeiture of the cash. The police had failed to prove, on the balance of probabilities, that the intention was to use the cash for unlawful purposes. The magistrates refused to make an order for costs against the police. It was common ground in this court that when they seized the cash the police had reasonable grounds for their suspicion that it had been intended for use in unlawful conduct; and that they had reasonable grounds to believe that the cash had been intended for such use when they made their application for its forfeiture. The power to order costs in that case was the same power that the Divisional Court had considered in Bradford. Stanley Burnton LJ said at [19]:

Among the many cases that Stanley Burnton LJ considered was the decision of the

Commenting on that decision, Stanley Burnton LJ said at [31]:

He summarised his conclusions as follows:

This, too, is clear recognition that where the principle applies it is the starting point or default position.

Lord Neuberger MR said at [59]:

He went on at [65] to say that the principles formulated by Lord Bingham in Bradford applied not just to licensing cases, but to:

At [73] he said:

I agree with Sir James that Lord Neuberger’s choice of the word “functions” means that the principle is not confined to the performance of specific duties. That echoes Sir Igor Judge’s phrase “regulatory responsibility”.

At [75] Lord Neuberger recognised that there were respectable arguments (a) for a presumption that the successful party should recover its costs and (b) for saying that there is no such presumption and that, absent other factors, a successful party should only be able to recover costs where the actions of the police were unreasonable or otherwise open to criticism. Again, I agree with Sir James that, in so formulating the rival arguments, Lord Neuberger was approaching the question not merely as a question of authority but also as a question of principle.

At [76] he reiterated:

He expressed his conclusion at [77]:

Maurice Kay LJ agreed with both judgments.

The CAT has a UK-wide jurisdiction. It is therefore necessary to look beyond the practice of the courts of England and Wales: BPP Holdings Ltd v HMRC [2017] UKSC 55, [2017] 1 WLR 2945 at [23]. We were referred to the decision of Lord Ericht, sitting in the Outer House, in Ahmed-Sheikh v Scottish Solicitors Discipline Tribunal [2019] CSOH 104, 2020 SLT 1. Ms Ahmed-Sheikh had been found guilty of professional misconduct. The Scottish Solicitors Discipline Tribunal awarded expenses (i.e. costs) against her on the agent and client basis (which I understand to be the equivalent of what used to be called solicitor and client basis under the RSC). The argument was not about whether the Tribunal was wrong in making a costs order at all: the argument turned on the scale of costs that had been awarded. The solicitor contended that costs should have been awarded on a party and party basis. The decision in that case was that the Tribunal was entitled to make the order for costs that it did. In the course of his judgment Lord Ericht drew attention to a number of differences between the practice in Scotland and the practice in England and Wales. Having referred to Baxendale-Walker Lord Ericht continued at [50]:

He also noted that he had been referred to 20 cases dating from between 1999 and 2017 in four of which expenses had been awarded against the Law Society. It is not apparent from the judgment whether the principles in Baxendale Walker have been considered and rejected by the Tribunal in Scotland (and, if so, why); or whether they have simply been overlooked. Nor did Lord Ericht himself comment on the practice of the disciplinary tribunal in Scotland: he merely noted that it did not apply Baxendale-Walker. Moreover, the obiter observations of the Outer House in AhmedSheikh cannot detract from the binding precedent of BT v Ofcom (in which BPP was quoted at [65]) which holds that those principles apply to at least some proceedings in the CAT.

The practice of the CAT, which I have summarised above, was not developed in ignorance of the principles formulated by Lord Bingham in Bradford. In the early days of the CAT Sir Christopher Bellamy QC delivered an influential costs ruling in

The rules then current did not contain a general rule that costs should follow the event. Nor did they contain a general rule to the effect that costs will be awarded against a losing party only if that party has behaved unreasonably, frivolously or vexatiously. Rules in other tribunals did contain a rule of the latter kind but they reflected a specific policy decision on the part of Parliament that the particular objectives of the legislation in question could best be met by restricting the circumstances in which costs may be awarded.

The CAT referred to Lord Bingham’s judgment in Bradford (which had not been cited to it) and also to the practice of the European Court of First Instance, now the General Court (although it noted that the rules of that court did provide for the normal rule that costs should follow the event).

It disavowed seeking to formulate rigid rules but stated that the CAT should proceed on a case-by-case basis, retaining flexibility to meet circumstances as they arise.

It then went onto consider what factors were relevant to the exercise of its discretion. They were:

The financial prejudice, by way of costs, that the successful appellant had suffered as a result of having brought the case. The fact that a successful appellant had been put to expense in exercising his rights was a factor relevant to the exercise of the discretion, even though it was not necessarily a decisive factor.

In many cases it would not be possible to identify a winner, in which case costs should lie where they fall.

A party who has succeeded in part might be awarded only part of his costs.

Different circumstances were likely to apply to different kinds of cases. Cases involving penalties would require particular consideration, and

Any analogy there might be with the rule in civil litigation that the losing party should pay the winning party’s costs, should be displaced, in the exercise of discretion, where the CAT was satisfied that such a rule would frustrate the objects of the Act.

There was force in the contention that a general or rigid rule to the effect that losing appellants should normally be liable for the Director’s costs, as well as their own, could tend to deter appeals.

As far as cases where it was the Director who is unsuccessful, the principal policy argument was that it would bear unduly heavily on the public purse if the Director was regularly faced with large bills of costs from successful appellants. The Tribunal noted Lord Bingham’s statement that there was a public interest in encouraging public authorities to make and stand by honest, reasonable and apparently sound administrative decisions made in the public interest without fear of exposure to undue financial prejudice if the decision is successfully challenged. On this point it stated that “we accept that the factors urged on us by the Director are potentially relevant to the exercise of our discretion” but they were not decisive.

Considerations of public expenditure could not be decisive in cases where considerations of fairness point in the opposite direction.

The CAT thus concluded:

As I read this, the CAT rejected the argument based on Bradford (i.e. Booth’s case) that the public interest element bore on the question whether to make a costs order at all. Rather, the concerns could be dealt with by controlling the level of costs. It is also fair to say that GISC was decided before it was appreciated that Lord Bingham’s principles had been extended into other areas of litigation, and before they had been characterised by later decisions of this court as a starting point or default position, rather than being no more than a factor to be taken into account.

In GISC the CAT then went onto consider the facts of the particular case. Of particular note is the CAT’s reasoning at [67] where it said that it was doubtful that the decisions were taken on grounds that reasonably appeared to be sound; and that even on the Director’s submission that costs should only be awarded in cases of manifest error, such an error had been made.

GISC was followed by the CAT (Rimer J presiding) in The Racecourse Association v Office of Fair Trading [2006] CAT 1. Having referred to the first of the factors mentioned by the CAT in GISC, the CAT went on to say at [8]:

At [10] it repeated that:

In the same paragraph it confirmed that the effect of GISC was that:

The difference between GISC and The Racecourse Association, as I see it, is that whereas in GISC the fact that a successful appellant had been put to expense in exercising his rights was no more than a factor relevant to the exercise of the discretion; in The Racecourse Association it was elevated to a starting point. The difference between the two, in practical terms, is that the starting point is the default position. That must mean that there is some sort of burden cast on the unsuccessful party to move the CAT away from that starting point. Lord Bingham’s principle, in so far as it relates to the making of a costs order at all, has got lost. Subsequent cases in the CAT have also rejected the application of Lord Bingham’s principle in this category of case, and have adopted “costs follow the event” as a starting point. It is also pertinent, in my judgment, to observe that the suggested meeting of the regulator’s concerns by control of costs takes place (for the most part) as a matter of case management, which necessarily takes place before the outcome of the appeal is known. Costs control at that stage (which is now a feature of all litigation) is evenhanded as between the parties. It is difficult to see how that bears on the subsequent decision whether or not to make a costs order once the outcome of the appeal is known. In addition, the subsequent adoption by the CAT of “costs follow the event” as a starting point loses sight of Sir Christopher Bellamy’s observation that a general rule to the effect that losing appellants should normally be liable for the regulator’s costs, as well as their own, could tend to deter appeals.

In Tesco plc v Competition Commission [2009] CAT 26 the CAT (presided over by Barling J) distinguished between two types of case. Having referred to Bradford, the CAT went on to say at [32]:

It concluded at [34]:

In Eden Brown Ltd v Office of Fair Trading [2011] CAT 29 the CAT (presided over by Roth J) again pointed out that it had adopted different starting points for different types of case. Where a regulator had determined a dispute between undertakings, and defended its decision, the starting point was that if unsuccessful it should not be liable to pay costs. On the other hand, where the contested decision was a finding of infringement, or the imposition of a penalty, the starting point was that costs would follow the event. That was the position established in The Racecourse Association and followed in subsequent cases. Delivering its ruling Roth J said at [10]:

Having referred to Bradford, Baxendale-Walker and Perinpanathan, he continued at [16]:

Eden Brown has been followed in a number of cases in the CAT, to which Ms Bacon QC, on behalf of Flynn Pharma, referred us. I do not think that it is necessary to refer to them. Suffice it to say that on each occasion the CAT agreed with the analysis in Eden Brown.

The CAT considered these cases (including Eden Brown) again in British Sky Broadcasting Ltd v Office of Communications [2013] CAT 9 (commonly known as

In deciding as it did in Pay TV the CAT distinguished the cases to which I have referred, following its previous decision in Tesco plc v Competition Commission [2009] CAT 26.

Finally, in this review of the cases, I must refer to the decision of this court in Quarmby Construction Co Ltd v Office of Fair Trading [2012] EWCA Civ 1552. The OFT had found that Quarmby had infringed the Chapter l prohibition by rigging bids for tenders in the construction industry; and imposed a substantial penalty. Quarmby appealed to the CAT. The appeal failed as regards the infringement; but the CAT reduced the penalty. It made no order as to costs. Quarmby’s appeal to this court failed. It argued that the CAT should adopt the general principle that costs follow the event; and referred to previous decisions of the CAT which supported that approach. At [23] Lloyd LJ said:

At [31] he said that there was much to be said for the CAT’s view that neither side was the winner. He said that the CAT’s balance between success on the penalty issue but loss on the infringement issue was a view that the CAT were entitled to take. At [32] he concluded:

Lloyd LJ’s judgment does not record what was argued on behalf of the OFT. Significantly, for present purposes, none of the authorities on which the CMA relies in our case appears to have been cited in Quarmby. And the outcome of the case itself, i.e. no order for costs, is consistent with the outcome for which the CMA argues. Moreover, Quarmby was considered and applied in the decision under appeal in BT v

BT v Ofcom involved a market review under which Ofcom was required (a) to identify the relevant product and geographic market(s) in accordance with EU law and guidelines; (b) to carry out an analysis as to whether there was a lack of effective competition in the defined market by reason of an entity's dominance; and (c) to decide what remedies to impose. BT appealed to the CAT which quashed certain determinations made by Ofcom concerning market definition in its Final Statement. The appeal was brought under section 192 of the Communications Act 2003. At that time section 195(2) of that Act provided that:

The CAT followed its own previous decisions in drawing a distinction between different types of appeal. It said that, although there was no express starting point in rule 104, for certain categories of case the CAT had an established practice in relation to costs. It gave an appeal against Ofcom's decision resolving a price dispute under as an example of a case where the starting point was that there should be no order for costs against Ofcom if it had acted reasonably and in good faith. It said, however, that in the case of infringement decisions and applications for judicial review of merger decisions and market investigations the CAT had taken the view that the starting point should be that costs should follow the event. In the decision under appeal in that case, the CAT described the practice which had grown up (summarised in this court at [16]):

The question for this court was whether that was correct. It is to be particularly noted that the practice as described included “infringement decisions”.

In its submissions to this court BT emphasised the “important distinction” between dispute resolution appeals and regulatory appeals. In deciding a dispute resolution appeal, Ofcom was performing a unique quasi-judicial role. In the case under consideration Ofcom was imposing conditions on BT. In voluntarily defending a decision it had made pursuant to its public functions, it must face the possibility of an adverse costs order just as any other public authority defending an appeal or judicial

This court considered all the English cases to which I have referred with the exception of Eden Brown. But since Eden Brown was quoted in Pay TV, I do not regard that omission as significant. At [66] the court said of Bradford, Baxendale-Walker and Perinpanathan:

At [69] it noted that in Perinpanathan Lord Neuberger MR had “expressed himself in terms wide enough to be of relevance to the competition arena.”

The court concluded as follows:

Turning to the previous practice of the CAT, the court said at [74]:

Since the CAT’s description of its established practice had expressly included infringement decisions, it seems to me that this court’s reference to “merits appeals” must at least have included infringement decisions. The court went on to say at [75] that, although there are distinctions between different kinds of decision, in many cases “this may be a distinction without a difference”:

At [78] the court said:

At [81] the court “particularly” endorsed the statement of the CAT in Number (UK) Ltd v Office of Communications (Costs) [2009] CAT 5 that:

It concluded at [83]:

In my judgment, it is clear from [74] that the court intended its decision to apply to “merits appeals” such as the case with which we are concerned. The case before it was, indeed, a “merits appeal”. It is all of a piece with its evaluation at [66] of the trilogy of Bradford, Baxendale-Walker and Perinpanathan where the observations were intended to be widely applied. That is reinforced by the entirely general statement at [83], which deals compendiously with “a claim before the CAT” in which Ofcom (or the CMA) is acting purely in its regulatory capacity. The court had explained what it meant by “in its regulatory capacity” at [72]; namely that the regulator was acting solely in pursuit of its public duty and in the public interest. In other words, it was not acting in its own commercial interest. Both Tesco and Pay TV, which followed Eden Brown in distinguishing different kinds of case brought before the CAT, were specifically disapproved. Even in GISC the CAT held that the Director’s concerns were best addressed by control over the level of costs, rather than the prior question whether costs should be awarded at all.

In my judgment, therefore, this court has comprehensively rejected the proposition that the starting point, even in a merits or judicial review appeal in the CAT, is that costs follow the event. It is true that in BT v Ofcom this court does not appear to have been referred to the practice in disciplinary proceedings in Scotland, where the Baxendale-Walker principle is apparently not applied. But, as I have said, I do not consider that that detracts from the authority of BT v Ofcom. It could, of course, be said that the case before the court in BT v Ofcom was not an infringement appeal or an appeal under the Competition Act 1998, with the consequence that what it said on that topic was, strictly speaking, obiter. But that, to my mind, is an over-technical approach. In a previous dispute between BT and Ofcom this court had noted that the Tribunal’s function under both the Competition Act 1998 and the Communications Act 2003 was similar: British Telecommunications plc v Office of Communications [2011] EWCA Civ 245, [2011] 4 All ER 372 at [63]. The judgment in BT v Ofcom was a judgment of the court which is intended to carry special authority. It was designed to give general guidance to the CAT, especially in view of the court’s observations at [74] and [78].

Accordingly, I consider that even if we are not technically bound by BT v Ofcom we ought to follow it.

Nevertheless, although the court rejected “costs follow the event” as the starting point, I do not consider that it unequivocally endorsed “the Principle” as originally framed by the CMA. That, I think, emerges from the final paragraphs of the judgment in BT v Ofcom in which the court said:

Although in these paragraphs this court recognised that the CAT could develop its own guidelines, that was subject to the proviso that they were based on “applicable legal principle”. In Perinpanathan Lord Neuberger MR stressed the need for consistency of approach; and in particular the need to avoid inconsistency “in the approach to the question of costs in regulatory matters”: see [74]. That justifies the adoption of the same principles in cases involving competition regulators.

The applicable legal principles to be derived from these cases are, in my judgment, as follows:

Where a power to make an order about costs does not include an express general rule or default position, an important factor in the exercise of discretion is the fact that one of the parties is a regulator exercising functions in the public interest.

That leads to the conclusion that in such cases the starting point or default position is that no order for costs should be made against a regulator who has brought or defended proceedings in the CAT acting purely in its regulatory capacity.

The default position may be departed from for good reason. iv) The mere fact that the regulator has been unsuccessful is not, without more, a good reason. I do not consider that it is necessary to find “exceptional circumstances” as opposed to a good reason.

A good reason will include unreasonable conduct on the part of the regulator, or substantial financial hardship likely to be suffered by the successful party if a costs order is not made.

There may be additional factors, specific to a particular case, which might also permit a departure from the starting point.

It is not entirely clear why the court remitted the question of costs to the CAT; but as

In our case the CAT considered the decision of this court in BT v Ofcom. The question that the CAT posed for itself was as follows:

At [33] it said that it was not clear whether the court had questions of the wider competition enforcement regime in mind. At [34] it said:

I respectfully disagree. The categorisation of decisions described in the ruling under appeal in that case distinguished between two types of case, one of which included infringement decisions. Rather than expecting the court to have expressly included infringement decisions in the scope of its general remarks, I would have expected the court to have expressly excluded them if that is what it meant. In short, I consider that the CAT was looking at BT v Ofcom through the wrong end of the telescope. It should have started with the principles as developed in the case law to which I have referred.

The CAT went on to cite extensively from Eden Brown (which had of course been cited in Pay TV). At [39] it said:

Next the CAT considered particular features of competition infringement cases. It noted that:

In a regulatory situation, the authority is obliged by its regulatory duties to take action against a particular person and, if it cannot itself impose any sanction, to apply to a further body, whether a court or a tribunal, to obtain that sanction. In the present case, by contrast, the CMA has a discretion whether to take action against a particular company, and is not obliged to do so in any particular case. Although the CAT did not expressly say so, it appears to have regarded this factor as preventing the case from being a regulatory situation at all.

The CMA's powers are extensive, as shown by the imposition in this case of a very substantial financial penalty of a quasi-criminal nature. These powers are exercised by the CMA through an administrative procedure in which objections are put to the parties accused of infringement, their responses considered, and a decision taken. The parties bear the entire cost of their participation in that process, whatever its outcome.

The appeal to the CAT is the parties' first opportunity to put their case to an independent and impartial appeal body and for the CMA to defend its decision. It is an appeal “on the merits”. It is thus an essential part of the system by which competition authorities, in return for receiving extensive enforcement powers, are held to account by the courts.

Having identified these features of competition infringement cases, the CAT concluded at [46]:

Eden Brown, it will be recalled, held that the principles established in Bradford, Baxendale-Walker and Perinpanathan had “no application” in infringement cases.

I do not consider that the fact that the CMA has a discretion whether or not to adopt an infringement decision or to impose penalties is of critical importance. Ms Bacon placed particular reliance on paragraph [75] of BT v Ofcom in which the court discussed “obligations”. But it is clear that the court was using the word “obligation” in a very loose sense. As Sir Igor Judge PQBD pointed out in Baxendale-Walker at

It is true that the CMA’s powers are extensive. The imposition of a financial penalty is a serious matter for an undertaking, as is the reputational damage likely to be occasioned by a finding of infringement. But I question whether, in context, that is of greater significance than the loss to a sole trader of a licence to carry on a trade at all, such as the refusal of the private hire licence to Mr Booth in Bradford. That, too, is a decision or sanction imposed by the regulator itself which must be challenged by the person aggrieved by the sanction. The SDT, which was under consideration in Baxendale-Walker, has power to strike a solicitor’s name from the roll and thereby permanently to deprive him or her of their professional livelihood. In that situation the sanction is imposed by the tribunal rather than the Law Society, which acts as prosecutor; but I cannot see that that is relevant distinction. I accept that an infringement of competition law is treated for many purposes as a breach of the criminal law, as this court held in the appeal against the CAT’s substantive decision: [2020] EWCA Civ 339 at [136]. But that, in my judgment, is only of tangential relevance to the question of costs. It must not be forgotten that a corporate defendant (unlike an individual defendant) has no entitlement to costs in successfully defending a prosecution in the criminal courts: see Prosecution of Offences Act 1985 s 16A which limits the making of a defendant’s costs order to cases in which the defendant is an individual. So I do not find the quasi-criminal nature of a penalty a compelling analogy.

As far as the CAT being the first opportunity the undertaking has to put its case to an independent tribunal is concerned, I do not consider that that is very different to Bradford, in which Mr Booth’s first opportunity to put his case to an independent tribunal was his complaint to the magistrates. That, too, was an appeal on the merits. Nor is it any different to Baxendale-Walker where the SDT hearing was Mr Baxendale-Walker’s first opportunity to defend himself. So too was Perinpanathan,

In short, therefore, I consider (a) that the CAT misinterpreted the decision of this court in BT v Ofcom and (b) that the reasons given for singling out competition infringement cases are not compelling. What is lacking in the present ruling is that the CAT appears to me to have given no weight at all to the position of the CMA as a public authority carrying out its functions in the public interest. As Stanley Burnton LJ said in Perinpanathan, commenting on British Telecommunications plc v Office of Communications, that is the feature that is relevant in considering what (if any) order for costs should be made. It was not the nature of the dispute. The costs ruling in our case contains the same flaw as caused this court to set aside the ruling in BT v Ofcom.

That is not to say that success or failure (either overall or on discrete issues) is irrelevant. As Sir Igor Judge PQBD said in Baxendale-Walker it is a factor to be considered; and in any event success or failure is expressly referred to in rule 104. Also relevant is any financial hardship which might be caused to the successful appellant (although none was alleged in this particular case).

Pfizer and Flynn drew attention to other features of an infringement decision by the CMA. In short they are:

The potential impact of penalties on the profitability and viability of the business of an undertaking. But the withdrawal or non-renewal of a licence to trade may have similar effect.

An infringement decision may lead to the disqualification of a director; and thus the CAT’s supervisory role over the CMA protects individuals against flawed decisions. That is undoubtedly so, although the court would still have to consider the critical issue of whether the person in question is unfit to be a director. In addition, an infringement decision will be addressed to an undertaking, rather than to an individual director; and any appeal would be made by the undertaking rather than by the director. But if a director were to be a party to an appeal, the CAT would, in my judgment, be entitled to take this into account in considering what order for costs to make in any particular case, particularly where the director concerned would suffer financial hardship if no costs order were made.

An infringement decision may lead to findings of fact binding on other courts or tribunals if follow-on proceedings for damages are brought. Accordingly, in order to defend itself against potential liability in civil proceedings, an undertaking may be compelled to appeal to the CAT. There is, therefore, a

Parties are required to incur considerable costs in the administrative stage of an investigation. Those costs are irrecoverable. It would be doubly unfair for a party to have to bear its own costs both in the administrative stage and in the CAT in a case in which it turns out that the CMA made a flawed decision. This seems to me to be a feature of many forms of regulatory processes. I note that in RBS Backhaul at [60] the CAT referred to the “constant regulatory dialogue” and to the “costs of maintaining specialised regulatory and compliance departments, and taking specialised advice, [which] will not ordinarily be recoverable prior to proceedings.” Yet even so, the CAT’s approach in that case was that no order for costs should be made. Ms Bacon pointed out that those remarks were directed to a case involving a regulated industry. Competition infringement cases, by contrast, do not (or do not necessarily) involve regulated industries; and undertakings operating in unregulated industries (such as the construction industry) are not in such dialogue and do not maintain specialised regulatory departments. That is a fair point, as far as it goes; but it is accommodated by the possibility of departure from the starting point in cases of financial hardship.

If an infringement case is brought in the EU context against a decision of the European Commission, the Commission is required to pay the costs of a successful appellant if its decision is annulled. This is a red herring. Article 134 of the Rules of Procedure of the General Court provides that “The unsuccessful party shall be ordered to pay the costs if they have been applied for in the successful party's pleadings”. There is, therefore, a starting point mandated by the rules (although there are some exceptions). That is not the case under rule 104.

Pfizer and Flynn also emphasised that a starting point that “costs follow the event” is no more than that: the CAT may make a different order. I do not regard this as a weighty argument. CPR Part 44.2 says much the same:

whether costs are payable by one party to another;

the amount of those costs; and

when they are to be paid.

the general rule is that the unsuccessful party will be ordered to pay the costs of the successful party; but

the court may make a different order.”

Yet in Quarmby this court rejected the application by analogy of the approach under the CPR. In addition, in BT v Ofcom this court disapproved PayTV which had also sought to draw an analogy between proceedings in the CAT and civil litigation governed either by the CPR or by the former Rules of the Supreme Court which contained the same general rule (RSC Order 62 rule 3 (3)). For the same reason I do not consider that analogies with litigation conducted under the CPR (whether judicial review or otherwise) are helpful. To the contrary, the general rule in the tribunal system (which is not governed by the CPR) is that in the first tier tribunal each party bears its own costs, absent unreasonable conduct: see, for example Tribunal Procedure (First-tier Tribunal) (Property Chamber) Rules 2013 rule 13. That is so even in the First Tier Tax Tribunal: Tribunal Procedure (First-tier Tribunal) (Tax

Likewise in the case of a planning inquiry, which may equally involve complex evidence both factual and expert, the general rule is that the parties bear their own costs. So the mere fact that two well-resourced parties have a dispute which is judicially or quasi-judicially resolved does not necessarily engage “costs follow the event” as the starting point.

Nor do I find persuasive the point that in previous cases before the CAT the CMA (or its predecessor the OFT) either accepted or in some cases positively relied upon the principle that “costs follow the event”. Until the law was clarified by BT v Ofcom, there was no reason for them not to fall in line with the established practice of the CAT.

Pfizer and Flynn also emphasised the contribution of “costs follow the event” to discipline in litigation. Mr Brealey in particular emphasised that the CAT applied the “costs follow the event” principle on an issue by issue basis. If the starting point or default position were to be that no order for costs should be made against a regulator without good reason, the regulator would be encouraged to take every conceivable point: what was referred to as “kitchen sink” litigation. There is no doubt that the encouragement of discipline justifies “costs follow the event” as a starting point or general rule. Many distinguished judges have said so. But that applies to all forms of litigation, including litigation in which the starting point is that no order for costs should be made against a regulator. To the extent that the courts have balanced the twin objectives of discipline in litigation on the one hand, and the public interest in encouraging regulators to make and stand by reasonable and sound decisions without

Both sides also raised arguments before us on the potential “chilling effect” of adopting a particular starting point. In the ruling under appeal the CAT did not find it necessary to consider that question. Mr Brealey in particular drew attention to the statement in Eden Brown that the OFT (as predecessor of the CMA) did not assert that a potential liability for costs was a deterrent to the taking of robust decisions. He also drew attention to the fact that (as revealed in its accounts) the CMA was able to use monies collected by way of penalties in offsetting its legal costs. On the other hand, in RBS Backhaul at [63] the CAT said:

I would not regard the “chilling effect” on the CMA as self-evident. But in so far as it has potential to exist, I consider that it is already accommodated within the principles developed by the cases in this court.

Mr Brealey also relied on rule 4, and in particular on rule 4 (1)(a) which, he said, militated against the asymmetric approach for which the CMA contended. In the first place it is not at all clear that the CMA was contending for an asymmetric position. In its supplemental skeleton argument the CMA stated:

It did not advocate an asymmetric approach in GISC (see paragraph [25]) or in Eden Brown (see paragraph [8]). All we are asked to decide is whether, on the one hand, the starting point is that an order for costs should be made against an unsuccessful regulator, or on the other, the starting point is that no order should be made against an unsuccessful regulator except for a good reason. Accordingly, the effect of our decision in this case does not preclude the CAT from adopting a symmetric approach, if it chooses to. In deciding whether or not to do so, it will no doubt consider whether the potential exposure to a liability for costs would deter appeals by some or all undertakings. Secondly, the existence of rule 4 did not deter this court in BT v Ofcom from ruling as it did.

As I have said, although “costs follow the event” is not the starting point in an infringement appeal at first instance, I do not consider that BT v Ofcom endorses the Principle as originally formulated by the CMA. It is not the case that the only circumstances in which an order for costs can be made against the CMA is a case in which it has acted unreasonably or in bad faith. That proposition was expressly rejected in Bradford; and none of the later decisions have overruled it in that respect.

Rather, in my judgment, the starting point or default position is that no order for costs should be made against a regulator who has brought or defended proceedings in the CAT acting purely in its regulatory capacity. That starting point may be departed from for good reason; but the mere fact that the regulator has been unsuccessful is not enough.

As this court recognised in BT v Ofcom at [86], it is for the CAT to develop its own approach to an award of costs. That approach can, no doubt, include the degree of success or failure achieved by a party as one of the relevant factors as envisaged by rule 104 itself. It may also include consideration of what (if any) hardship would be suffered by a successful appellant if an order for costs was not made (which would not be a relevant consideration under the CPR). In considering hardship, the CAT could take into account, in an appropriate case, the level of irrecoverable costs that the successful party had borne in the administrative stage of the investigation. The conduct of the parties would also be a relevant consideration as the rule again envisages. It may be that if the CMA were to pursue a small or medium-sized enterprise as a test case, that would justify a departure from the starting point. But whatever approach is adopted the CAT must also put into the scales the fact that the CMA is a public body carrying out functions in the public interest; and that there is a public interest in encouraging public bodies to exercise their public function of making reasonable and sound decisions without fear of exposure to undue financial prejudice, if the decision is successfully challenged.

In BT v Ofcom this court remitted the question of costs to the CAT. The question arose whether we should do the same. But in our case the CAT considered a number of reasons for not adopting the starting point which I consider to be correct; and I have concluded that those reasons are inadequate. I have also considered the further considerations urged by both Mr Brealey and Ms Bacon; and I consider that they, too, are inadequate. The CAT found that the CMA had not acted unreasonably and could not be criticised for not accepting points made by Pfizer and Flynn. It also found that neither Pfizer nor Flynn had suffered financial hardship. In those circumstances I do not consider that any useful purpose would be served by remitting the case to the CAT.

For these reasons, I would allow the appeal and make no order for the costs of the proceedings before the CAT.

I agree.

I agree that this appeal should be allowed for the reasons given by Lewison LJ. As he notes at [38], Lord Neuberger MR recognised in Perinpanathan at [75] that there were respectable arguments (a) for a presumption that the successful party should recover costs and (b) for saying that there is no such presumption and that, absent other factors, a successful party should only be able to recover costs where the actions of the public body (in that case, the police) were unreasonable or otherwise open to criticism. The courts of England and Wales have resolved this dilemma by the classic common law method of making a series of decisions, initially on a case-by-case basis, which have come to be recognised as establishing a general principle. Even if that process was not already complete by the time of Perinpanathan, the generality of the principle was clearly established by BT v Ofcom; and our decision reinforces it. In proceeding in this way, the courts have relied, as the common law always does, upon the experience and judgment of the judges involved, some of whom were very distinguished. It is fair to say, however, that the decisions have not been evidencebased, nor have the courts been able to take into account any wider considerations of policy than those discussed in the cases. In those circumstances there may be merit in the issue being considered by the Law Commission.