The Competition And Markets Authority v Flynn Pharma Ltd & Anor (Rev 3)

This case raises important points of law concerning the test to be applied to determine when prices charged by dominant undertakings for goods or services amount to the abuse of a dominant position, as to the nature of the duty upon a competition authority to evaluate evidence adduced by an undertaking in its defence, and as to the powers of judicial bodies called upon to hear appeals from such authorities. The subject matter of the appeals concerns the pricing of pharmaceutical products and affects the sums paid by the public purse for drugs. But the case is of wider application because the legal issues apply to all goods and services in the economy. I have read the judgment of the Chancellor of the High Court and agree with it.

The appeals are from the judgment (“the Judgment”) of the Competition Appeal Tribunal (“the Tribunal”) dated 7^th June 2018 which set aside a decision (“the

The parties are as follows.

The CMA is the competition authority in the United Kingdom with jurisdiction to enforce the Competition Act 1998 (“the CA 1998”) and Article 102 TFEU (“Article 102”) which prohibit the abuse of a dominant position. It is the principal appellant in this case and challenges the Judgment upon the basis that the analysis therein of the test for “abuse” is wrong in law. It is a respondent to the appeal by Flynn Pharma Limited and Flynn Pharma (Holdings) Limited (see below).

Flynn Pharma Limited is a company that sells and markets pharmaceutical products. It specialises in acquiring and rescuing “tail-end” or “end-of-life” pharmaceutical products which are mature drugs with declining demand. It is a wholly owned subsidiary of Flynn Pharma (Holdings) Limited. These companies are referred to as “Flynn”. It is a respondent to the appeal of the CMA but is also an appellant in relation to certain findings in the Judgment which it argues are either inconsistent or vitiated by a failure to give adequate reasons and it argues that in so far as the Tribunal has made findings of fact which upon remittal bind the parties then it has erred in law.

Pfizer Inc. is a research-based global pharmaceutical company. Pfizer Limited is a UK based subsidiary of Pfizer Inc. It develops, manufactures and markets pharmaceutical products globally, including in the UK. These companies are referred to as “Pfizer”. Pfizer was a respondent to the appeal of the CMA.

The Commission of the European Union (“the Commission”) is an intervener and appears in order to support the CMA in relation to issues of law and practice. It takes the position that the Tribunal erred in law in its articulation of the test for abuse of dominance and argues that if the law is as laid down by the Tribunal it would be impracticable and unworkable.

The CMA commenced an investigation into Pfizer and Flynn in May 2013. The scope of this investigation initially concerned potentially anti-competitive agreements between Pfizer and Flynn under both section 2 CA 1998 and Article 101, and an abuse of a dominant position by Pfizer under both section 18 CA 1998 and Article 102. The investigation into potential infringements of section 18 CA 1998 and Article 102 was extended to Flynn in February 2014. The relevant parts of section 18 provide:

“18 Abuse of dominant position.”

(1)

… any conduct on the part of one or more undertakings which amounts to the abuse of a dominant position in a market is prohibited if it may affect trade within the United Kingdom.

(2)

Conduct may, in particular, constitute such an abuse if it consists in—

(a)

directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions;

…”

The prohibition in section 18 is known as “the Chapter II prohibition”. Article 102 is, mutatis mutandis, identical save only that for it to apply there must be a proven effect upon trade between Member States. Article 102(a) is identical in terminology to section 18(2)(a) CA 1998. The seminal case on the test for abusive pricing is the judgment of the Court of Justice in Case C-27/76 United Brands v Commission EU:C:1978:22 (“United Brands”).

The Decision found that Pfizer and Flynn had abused their dominant positions in the UK market for phenytoin sodium capsules under both the Chapter II prohibition and Article 102. It imposed a penalty of £84,196,998 upon Pfizer and £5,164,425 upon Flynn. The abuse comprised of Pfizer and Flynn intentionally or negligently charging excessive and unfair prices for the capsules. The Decision included directions (“the Directions”) requiring Pfizer and Flynn to reduce their prices.

On 23^rd December 2016 Flynn applied unsuccessfully to the Tribunal for interim relief to suspend the Directions pending appeal. Between January and April 2017 Pfizer and Flynn reduced their prices to comply with the Directions.

On 7^th February 2017 Pfizer and Flynn lodged appeals in the Tribunal against the Decision. The hearing occurred over 13 days between 30^th October 2017 and 24^th November 2017.

The Tribunal handed down the Judgment on 7^th July 2018: [2018] CAT 11. It found that: (i) the CMA had correctly identified the relevant geographical and product markets; (ii) Pfizer and Flynn were dominant in those markets; but (iii), the finding of

The CMA sought permission from the Tribunal to appeal to the Court of Appeal, as did Flynn on points arising from the Judgment which it considered would prejudice it upon the remittal, as well as each of Pfizer and Flynn in respect of elements of the Judgment. Permission to appeal was refused to all parties on 25^th July 2018: [2018] CAT 12.

All parties applied to the Court of Appeal for permission to appeal. Leave was given to both the CMA and Flynn (on certain of its arguments) by order on 12^th December 2018. Subsequently, the CMA sought to raise additional issues of law in particular addressing arguments it had conceded before the Tribunal. The Court of Appeal granted permission to the CMA to amend its Grounds of Appeal to raise these new points upon the basis, inter alia, that the issues arising were of considerable public importance not only to the correct approach to be adopted to the pricing of pharmaceutical products, but more generally: See [2019] EWCA Civ 1631 (4^th October 2019).

Under section 49 CA 1998 appeals lie to this Court on issues of law only. The Court has received detailed written submissions on all the issues arising. In addition, the parties placed 3 volumes of economic literature before the Court and have referred to this in support of their submissions. All the parties provided extensive and helpful written and oral submissions on the matters of law arising but also upon related issues of procedure and practice and upon the economic literature.

The basic facts are set out in the Decision and in the Judgment: see Judgment paragraphs [12] – [68]. I summarise them briefly below.

Phenytoin sodium is an anti-epileptic drug (“AED”). It was first synthesised in 1908. It is available in a variety of forms, including capsules and tablets. It is used in the control of the frequency of epileptic seizures. A phenytoin sodium capsule was first commercialised and marketed in the UK in 1938 under the brand name “Epanutin”. The capsule is available in four strengths: 25mg, 50mg, 100mg and 300mg. The tablet form is available in 100mg strength only.

In 2000 Pfizer acquired the US pharmaceutical company Warner-Lambert whose portfolio included Epanutin and thereafter Pfizer held the Marketing Authorisation

Phenytoin sodium has a narrow therapeutic index (“NTI”) such that there is a small difference between the concentration of phenytoin in the patient’s blood that is required to achieve therapeutic effect and the concentration that might cause adverse effects if exceeded. The pharmacokinetic characteristics of the drug (i.e. the manner in which it moves through the body from absorption to break-down and excretion) are non-linear and the concentration of the drug in the blood does not respond proportionately to increases or decreases in dosage. It can therefore be difficult to regulate the appropriate dose for a patient. This is a reason for the decline in use of the drug in the UK. Phenytoin sodium is therefore not generally prescribed as a first-line or single treatment for epilepsy. There are nonetheless about 48,000 patients in the UK who are already stabilised on the drug and for whom it is effective.

Prescribing and dispensing decisions in relation to AEDs are informed by clinical guidance issued by specialised bodies such as the MHRA (Medicines and Healthcare Products Regulatory Agency), NICE (National Institute for Health and Care

Phenytoin sodium capsules manufactured by Pfizer are also sold in the UK by parallel importers, usually at 100mg strength. Since April 2013 100mg strength capsules have been manufactured and supplied by NRIM under the name Phenytoin Sodium NRIM Capsules. Teva UK Limited is the principal manufacturer and supplier of phenytoin sodium tablets in the United Kingdom. Between April 2005 and October 2007, the Drug Tariff price of these tablets increased from £1.70 for a 28 x 100mg pack of tablets to £113.62, as a result of Teva increasing their prices. This led to discussions between the Department of Health (“DOH”) and Teva and in October 2008 Teva reduced its price to £29.50 per 28 x 100mg pack and the Drug Tariff price fell to £30 per pack.

The mechanisms used to regulate drug prices are complex. I provide only a simplified overview below. A more detailed summary is set out in the Judgment at paragraphs [31] – [51]. Patients do not normally pay for an AED. It is paid for by the NHS which reimburses pharmacies for medicines dispensed by it. The “Drug Tariff” sets the amounts that pharmacies can seek by way of reimbursement. It reflects the voluntary and statutory price controls applying to various pharmaceutical products and takes account of any clawback discounts.

Drugs are either branded or generic (non-branded). This has implications for the regulation of the drug and its Drug Tariff price. There are three categories of products for the purposes of calculating the Drug Tariff price: A, C and M. Category C applies to drugs not readily available in generic form and the price is determined by reference to the list price for the particular product, manufacturer or supplier. Category M applies to generics and the price is calculated upon the basis of a volume-weighted average selling price derived from information submitted to the DOH by suppliers.

Drug prices are regulated in three main ways:

Pfizer and Flynn are both members of the Pharmaceutical Price Regulation Scheme (“PPRS”), a voluntary scheme agreed between the DOH and the Association of the British Pharmaceutical Industry. It controls the overall profit that scheme members can make on the sales of branded licensed medicines to the NHS, on a portfolio basis. It has no application to generic medicines.

When branded phenytoin sodium capsules were under the PPRS and subject to price control the price was £0.66 per 28 x 25mg pack, £0.67 per 28 x 50mg pack, £2.83 per 84 x 100mg pack, and, £2.83 per 28 x 100mg pack.

A PPRS scheme characteristically endures for 5 years and controls prices by setting target rates of return applied to companies on a portfolio basis (i.e. their entire branded medicine portfolio). Target rates include allowances for research and development, information and marketing and, in addition, benefit from a margin of tolerance. In 2009 and 2014 the target rates were 6% return on sales (“ROS”) and 21% return on capital (“ROC”). The price of an individual product can be increased by applying to the DOH for approval or by price modulation whereby members are authorised to offset a price increase of up to 20% by an appropriate reduction in the prices of other products.

Details of the evolution of the agreements between Pfizer and Flynn are set out in the

A disparity existed between the Drug Tariff price of Epanutin capsules and the Drug Tariff price of the generic phenytoin sodium tablets. In 2009 Pfizer was approached by Tor Generics Ltd which sought to acquire a licence for the exclusive supply of Epanutin in the UK. It proposed that the drug be genericised to facilitate a price increase. Pfizer did not ultimately pursue the proposal. Instead it approached Flynn in January 2010 to discuss tail end products including Epanutin. In July 2010 Flynn produced a proposal that Flynn should become the holder of the MA for Epanutin and

On 17^th April 2012 Pfizer and Flynn entered into an Exclusive Supply Agreement. Pfizer continued to manufacture the capsule in Germany and supplied the capsules to Flynn for an initial term of 3 years. The agreement set out the supply prices from Pfizer to Flynn and provided for an annual price review. The supply prices were £4.50 per 28 x 25mg pack, £6.50 per 28 x 50mg pack, £39 per 84 x 100mg pack, and £39 per 28 x 300mg pack.

Various regulatory hurdles had to be overcome. On 2^nd May 2012 Flynn applied to change the name of Epanutin to “Phenytoin Sodium Capsules”. On 21^st June 2012 the MHRA contacted the Department for Health “DOH” notifying it of the de-branding plans. On 21^st June 2012 the DOH contacted Pfizer seeking details. Pfizer responded on 22^nd June 2012 describing the transaction as “still commercially sensitive” and identified the Epanutin capsules as the product being divested. On 26^th June 2012 the MHRA refused approval of the proposed name change upon the basis that it had the potential to confuse patients, prescribers and other healthcare professionals. Flynn agreed to submit a communication plan to address the concerns and on 4^th July 2012 it withdrew its application to vary the name of the products, pending submission of its communication plan. On 6^th July 2012 Flynn submitted a revised communication plan.

On 18^th July 2012 Flynn met with the DOH to discuss the price of Epanutin. Flynn said it could either genericise the product or create its own brand, if it could increase the price of the drug within the PPRS. Flynn’s communication plan was approved by the MHRA on 19^th July 2012. The DOH’s PPRS pricing committee however rejected the application to increase the price and this decision was communicated by the DOH on 26^th July 2012.

On 31^st July 2012 Flynn resubmitted its application to change the name of Epanutin to “Phenytoin Sodium Flynn Hard Capsules” and this was accepted by the MHRA on 29^th August 2012. On 23^rd September 2012 Pfizer ceased marketing the product in the UK but it continued to market and supply the drug in continental Europe.

On 24^th September 2012 Flynn de-branded Epanutin and commenced its distribution as Phenytoin Sodium Flynn Hard Capsules. The Drug Tariff price increased to £15.74 per 28 x 25mg pack, £15.98 per 28 x 50mg pack, £67.50 per 84 x 100mg pack and £67.50 per 28 x 300mg pack. The Drug Tariff price remained at this level until April 2014, when Pfizer reduced its prices in response to the entry into the market of NRIM.

On 28^th September 2012 the DOH complained to the Office of Fair Trading (the predecessor to the CMA) about the price increases. On 23^rd October 2012 the DOH contacted Flynn for information relating to its costs. On 6^th November 2012 the DOH

As already observed, the CMA commenced its investigation in May 2013 and the infringement Decision was adopted in December 2016. Key findings in the Decision were as follows:

The CMA places considerable store by what it says are stark and compelling facts which flow from the substantial price increases implemented in 2012 which could not be justified by changes in underlying costs. In 2012 Phenytoin Capsules were debranded and removed from the PPRS and were no longer price regulated. Pfizer sold the MAs to Flynn for just £1. From 24^th September 2012 Phenytoin Capsules were sold by Pfizer to Flynn, and then by Flynn, under the name “Phenytoin Sodium Flynn Hard Capsules” but there was no relevant change to the underlying cost structure. The sole change was that, after 24^th September 2012, Flynn was inserted into the supply chain and placed orders with Pfizer on a fortnightly basis. Without undergoing further investment, development, or innovation, and without any consequential benefits for patients, the hitherto stable prices for the different strengths of Phenytoin Capsules increased overnight.

To the extent necessary I refer to relevant passages of the Judgment as they arise in the analysis below. In the introduction to the Judgment the Tribunal explained the importance of the issues:

“3.

Cases of pure unfair pricing are rare in competition law. Authorities find them difficult to bring and are, rightly, wary of casting themselves in the role of price regulators. Generally, price control is better left to sectoral regulators, where they exist, and operated prospectively; ex post price regulation through the medium of competition law presents many problems. However, the law prohibits unfair pricing in certain circumstances and in such cases there is no reason in principle why competition law cannot be applied, provided this is done on the correct legal basis and the analysis of evidence is sound.

4.

We understand the CMA’s concern to deter and punish instances of unfair pricing that infringe the law. However, we have found this particular decision to be wrongly based in certain respects. Whilst we find the CMA was correct that the two companies each held a dominant position, we find the CMA’s conclusions on abuse of dominance were in error. The CMA did not correctly apply the legal test for finding that prices were unfair; it did not appropriately consider what was the right economic value for the product at issue; and it did not take sufficient account of the situation of other, comparable, products, in particular of the phenytoin sodium tablet. This means that the CMA’s findings on abuse of dominance in this case cannot be upheld.

5.

The importance of this case for the public interest makes it desirable to rectify the errors we have found. In a matter as important for government, for the public as patients and as taxpayers, as well as for the pharmaceutical industry itself, the law should be clear and any decisions made should be soundly based on proper evidence and analysis. It is important that there is a good legal foundation for any future action in this area.”

The thrust of many of the criticisms levelled by the CMA and by the Commission in this appeal relate to the articulation by the Tribunal of the test to be applied and its implications for future investigations. In paragraphs [442] - [444] the Tribunal pulled various threads together and set out a schema for the test that a competition authority “should” apply. I set them out here for convenience:

“442.

We recognise the difficulties inherent in seeking to formulate a generally applicable framework or test for abuse by unfair pricing, and we are conscious that, as United Brands itself states, there may be other ways than the two-limb test set out in that case for establishing an abuse. Nonetheless, if an authority chooses to proceed to apply the two-limb test in a structured way, as the CMA has purported to do in this case, a sensible framework would, in our view, and in light of the requirements and factors we have already set out above, be as follows.

443.

In our assessment, to apply Article 102 through the twolimb test of United Brands, in circumstances where the only alleged infringement is one of excessive pricing and the dominance of an undertaking in a given market has been established, a competition authority should:

(1)

consider a range of possible analyses, reflecting market conditions and the extent and quality of the data that can be obtained, to establish a benchmark price, or range, that reflects the price that would pertain under conditions of normal and sufficiently effective competition. On the facts of a particular situation, there might be only one basis of analysis that was credible, but the authority is not entitled to select one basis of analysis and ignore others that are also credible. The criteria for selection and application must be objective, appropriate and verifiable. The analysis must also be done on a consistent basis;

(2)

compare that price (or range) with the price that has been charged in practice and determine whether that is excessive;

(3)

for that purpose, form an assessment, for the purpose of the Excessive Limb, of whether that differential is sufficiently significant and persistent to be excessive, as a matter of its own discretion, exercised fairly and reasonably, in the light of such factors as:

(i)

the absolute size and stability of that differential;

(ii)

the reasons for it, taking account of the fact that the conditions for excessive pricing will only usually occur where the market is one where regulation, or some similar feature, or other barriers to entry, protect it from competition, or where there is regulatory failure and the relevant regulator has not intervened;

(iii)

previous decisions finding other differentials excessive, weighted for the markets applicable in those cases;

(iv)

the wider market conditions, including the evolution of pricing over time.

(4)

where there is a conclusion that the differential is excessive, then proceed to consider whether it is unfair under the Unfair Limb;

(5)

be free to use either Alternative 1 (unfair in itself) or Alternative 2 (unfair compared to competing products) to determine unfairness but give due consideration to any prima facie convincing argument that the pricing is actually fair under either Alternative and take that into account in reaching a decision under either Alternative 1 or 2;

(6)

if there is a finding of unfairness under the Unfair Limb, assess what is the economic value of the product, and whether the price charged in practice bears no reasonable relation to it;

(7)

give appropriate consideration to any objective justification advanced by the dominant undertaking;

(8)

make a finding of an infringement of Article 102 if all the conditions above are fulfilled; and: (i) the price bears no reasonable relation to the economic value; (ii) the dominant undertaking is reaping trading benefits that it would not reap under conditions of normal and sufficiently effective competition.

444.

It is for the competition authority to determine, when considering comparators, either for the application of Alternative 2 or for considering whether there are prima facie issues raised under Alternative 2 that need to be considered before proceeding under Alternative 1, or indeed if they are relevant to the Excessive Limb, what weight to be applied to them in the light of market conditions and their suitability, as comparators, for the product concerned. In making that determination, it must, but need only, act in a manner which is objective, appropriate and verifiable. It has a substantial margin of appreciation but must recognise the presumption of innocence in favour of the undertaking under investigation.”

The CMA is supported in this appeal by the Commission. Both express the overarching concern that the test mandated by the Tribunal (assuming that it is read as they suggest) interferes with the margin of manoeuvre or discretion that a competition authority should have (a) in choosing the methods or tests to apply and (b) as to the evidence it wishes to evaluate, in that it compels or requires that authorities perform certain tests in all cases. The Commission submitted:

“(1)

It misinterpreted United Brands to introduce an additional requirement of “benchmarking” the hypothetical price that would have been charged under sufficiently effective (but not idealised) competitive conditions

(2)

Advocate General Wahl’s Opinion in LatvianCopyright was treated as a general restatement of the law, relevant to the application of the United Brands test, whereas: (a) it was not, (b) insofar as used to interpret United Brands, was not followed in all respects by the Court. In particular, on a proper understanding of LatvianCopyright, the judgment does not endorse the suggestion made in the Opinion that it would be necessary to combine several different tests of analysis that are mutually corroborated to reach a conclusion on whether a particular price is unfair and therefore abusive.”

In relation to the analysis of the test of unfairness, the Commission submitted:

“(1)

Although the CAT ostensibly accepted that United Brands allows unfairness to be established in one of two ways, the effect of the judgment is to require Alternative 2 (unfair compared to competing products) to be applied, even if Alternative 1 (unfair in itself) is satisfied. This is unsupported by United Brands, and if followed would impose an unjustifiable burden on a competition authority or other party with the burden of establishing an infringement of Article 102(a) TFEU.

(2)

The Tribunal’s misinterpretation of the general scheme of United Brands and of the copyright cases, including Latvian Copyright, led it to impose a number of additional requirements and a further test of applying demand side considerations to the “economic value” of the product in determining whether the price was unfair. In this context, the Tribunal overstated the import of the Commission’s decision in Scandlines.”

In justice to the Tribunal I should point out that before the Tribunal the CMA, quite deliberately and after due consideration, made a key concession about the test to be applied which embraced what is now a central plank in the objections of both the CMA and the Commission to the Judgment. In particular, the CMA conceded that if a defendant undertaking adduced prima facie relevant evidence which was different in type to that relied upon by the CMA then the CMA was required fairly to evaluate that evidence. They have now withdrawn this concession.

Having read the transcript of the hearing below it is apparent that this court has received more detailed arguments on this point than occurred below. This is addressed more fully under the first Ground of Appeal below.

It is convenient to identify and summarise the Grounds of Appeal arising in the appeal. There is a degree of overlap between them. I have addressed them below in the following order.

The CMA’s first Ground of Appeal raises the following issues which concern the alternative fairness tests in paragraph [252] of United Brands:

The CMA’s second Ground of Appeal addresses whether a competition authority is required to use a hypothetical benchmark price or range of prices as part of its evaluation of whether an actual price is excessive. In particular, it considers whether non-price benchmarks such as cost or other related benchmarks such ROS and ROCE as sufficient.

The CMA’s third Ground of Appeal addresses the relevance of comparator evidence. It focuses upon the existence of a “margin of manoeuvre” or discretion for competition authorities and whether, assuming it exists, it serves to limit the jurisdiction of the Tribunal to reject findings or conclusions which amount to judgment calls of the authority. It also raises the issue of the limits of the powers of the Tribunal, including the question of materiality.

The CMA’s fourth Ground of Appeal concerns the meaning and effect of the expression “economic value” as that phrase is used in paragraph [250] of United Brands. It focuses upon whether the test is legal or economic in nature and its relationship to other aspects of the test for abuse as set out by the Court.

The fifth Ground of Appeal, advanced by Flynn, concerns the proper interpretation of the Judgment and whether any finding of the Tribunal in relation to abuse bind the parties upon the remittal.

I turn to the first Ground of Appeal which flows out of the finding by the Tribunal that if a defendant undertaking adduced prima facie relevant evidence based upon a method not used by the authority then the CMA was under a duty to fully investigate it: See Judgment paragraphs [366] and [443(5)]. There is also a suggestion in the Judgment that in any case where the authority relies upon one alternative it should also use the other alternative as a “sanity check” (Judgment paragraphs [294(14)] and [368]), adopting the “authoritative” analysis of Advocate General Wahl in Latvian Copyright. On my reading of the Judgment I was unclear as to whether the Tribunal was indicating that the authority should therefore always use a multi-test, combined, approach or whether that was required only when it was raised by a defendant undertaking. For the avoidance of doubt, I address both possibilities. The nub of this Ground flows out of the position that the CMA adopted in the Decision. There the CMA stated that if it establishes enough evidence on cost plus to prove abuse (i.e. under the “in itself” alternative in United Brands paragraph [252]) then “it was not necessary” to address any other sort of evidence (such as under the second alternative,

The CMA also stated in paragraph [5.244]: “… it is sufficient to demonstrate that one of its limbs is satisfied in order to establish an infringement”. It followed from the analysis in the Decision that if the CMA was correct then if a competition authority seeks to establish an abuse upon the basis of one type of evidence only (e.g. CostPlus) then it is not under a legal duty to evaluate alternative types of evidence (e.g. based upon competing products) regardless of whether such evidence is adduced either by an undertaking in its rebuttal evidence or as a “sanity check”.

During the hearing before the Tribunal the CMA conceded that the CMA would always examine good comparators at some stage of the analysis and would not ignore a relevant consideration. But following Judgment, in which this concession is recorded, the CMA applied to the Court of Appeal to withdraw the concession and to amend its Grounds of Appeal and therefore revert to the strict position set out in the Decision.

In their written submissions on appeal the CMA reinforced this by arguing: “There is no obligation to have recourse to these alternative methodologies in order to supplement or corroborate the two-limb United Brands test”.

In closing submissions to this Court Mr Hoskins QC for the CMA made clear that where a legal test was expressed in case law as alternatives and where one of those alternatives was satisfied then it was “…sufficient as a matter of law, to meet the test”. It followed that “…evidence relevant to the other alternative” need not be “taken into account”. It could not be “exculpatory” to the first alternative.

The starting point is United Brands. All parties describe the judgment in this case as “seminal”. Paragraphs [248] – [253] read as follows:

“248.

The imposition by an undertaking in a dominant position directly or indirectly of unfair purchase or selling prices is an abuse to which exception can be taken under Article 86 of the Treaty.

249.

It is advisable therefore to ascertain whether the dominant undertaking has made use of the opportunities arising out of its dominant position in such a way as to reap trading benefits which it would not have reaped if there had been normal and sufficiently effective competition.

250.

In this case charging a price which is excessive because it has no reasonable relation to the economic value of the product supplied would be such an abuse.

251.

This excess could, inter alia, be determined objectively if it were possible for it to be calculated by making a comparison between the selling price of the product in question and its cost of production, which would disclose the amount of the profit margin; however the Commission has not done this since it has not analysed UBC's costs structure.

252.

The questions therefore to be determined are whether the difference between the costs actually incurred and the price actually charged is excessive, and, if the answer to this question is in the affirmative, whether a price has been imposed which is either unfair in itself or when compared to competing products.

253.

Other ways may be devised - and economic theorists have not failed to think up several – of selecting the rules for determining whether the price of a product is unfair.”

The core of this Ground focuses upon paragraph [252] which describes two ways of determining whether a price is unfair, namely the “in itself” test and the “competing products” test. The CMA and Commission refer to the disjunctive “or” in the paragraph which ostensibly differentiates the two tests. They contend that their interpretation is borne out by subsequent case law of the Court and is buttressed by sensible economic logic. For the reasons set out below I conclude that the reading of the test in United Brands by the CMA is unduly rigid and literal and invests far too much significance in the disjunctive “or” in paragraph [252].

In the text below I start by considering the judgment of the Court in United Brands and the implications of subsequent case law. Then I address the legal, procedural and practical issues which flow from this analysis and I consider the economic arguments identified by the parties which they contend bear upon the issues. Finally, I set out a summary of the governing principles. It is common ground that there is no difference in the law as it applies to the Chapter II prohibition and Article 102.

I start by considering the structure of the reasoning in paragraphs [248] – [253] of United Brands. These paragraphs have three components to them. They commence by setting out the legal test for abuse, they then describe what this means in economic terms, and then they describe the methods and evidence that might be relevant to proving the abuse.

The Court starts (paragraph [248]) by setting out the basic test which of course derives from Article 102, which is fairness.

Then (in paragraphs [249] and [250]) the Court describes two central economic features of an abuse of unfairness. These are (i) that the undertaking has reaped “trading benefits” which could not have been obtained in “normal and sufficiently competitive”conditions; and (ii) that a selling price that is “excessive” in that it bears no reasonable relation to the “economic value” of the product or service in question is an example of an abuse. These paragraphs are connected: charging a price with no reasonable nexus to its economic value and which is therefore excessive (paragraph

In paragraphs [251] and [252] the Court moves to consider how in evidential and methodological terms such an abuse can be “determined objectively”. It gives an example of one way (but only one way - cf “inter alia” in paragraph [251]) to determine whether a price was unfair, namely the Cost-Plus method. The first stage or limb entails comparing the price charged with the costs of production (paragraph [251]), to see whether it is excessive, and the second stage or limb involves determining whether, if it is excessive, it is also unfair “in itself” or by reference to “competing products”. It is these two alternative tests of unfairness which are at the heart of this first Ground of Appeal.

Paragraph [253] is also important because it acknowledges that there are other economic ways of devising rules for determining whether a price is unfair. In other words the tests or methods referred to by the Court are not intended by the Court to lay down the only ways in which an unfair price might be determined. As such this militates strongly against any suggestion that the test in paragraphs [251] and [252] is to be construed as if it set down a fixed and definitive methodology.

Commentators point out that these paragraphs contain significant ambiguities. They do so to emphasise that they need to be read contextually and not over-rigidly. These ambiguities matter only if one reads the guidance in United Brands in a dispositive, literal and rigid manner. As the Court of Appeal observed in Attheraces Limited v BHB [2007] EWCA Civ 38 at paragraph [115] “…it would be wrong to read this passage too literally”. The Court of Appeal observed that the judgment had to be “read and applied with care”. I agree. I give three examples of ambiguities referred to in literature.

First, there is no definition or explanation of terms such as “reasonableness” or “economic value”. There is however no indication that the Court intended these to be precise terms of legal or economic art.

Second, the Court in paragraph [250] equates (without more) a price that is

Third, in relation to the use of a benchmark to determine whether the dominant undertaking has made use of the opportunities arising out of its dominant position in such a way as to reap trading benefits which it would not have reaped if there had been normal and sufficiently effective competition, the Court (paragraph [249]) says only that this is “advisable”, i.e. not required.

The facts of United Brands itself are illuminating and undermine the CMA’s argument that if it establishes abuse using one method or alternative then it can ignore evidence of another type adduced by a defendant undertaking. The Commission adopted a decision finding that the pricing of United Brands was “excessive” by reference to comparative evidence of (1) the difference in prices charged by United Brands in different Member States and (2) the policy adopted by United Brands of discriminatory pricing (cf judgment paragraph [258]). The Commission concluded that in some Member States United Brands was charging a price which was 50% higher than in other Member States (judgment paragraph [260]). The Court, however, criticised the Commission for failing to “take into account in its reasoning” evidence from United Brands that its pricing in Ireland “had produced a loss” in four out of the five previous years (judgment paragraphs [261], [262]). The Court was critical of the quality of the evidence adduced by United Brands (judgment paragraphs [263], [264]) but nonetheless concluded that the Commission had not “effectively refuted” the evidence (Judgment paragraph [265]). The Court observed:

“[264] However unreliable the particulars supplied by UBC may be (and in particular the document mentioned previously which works out the “losses” on the Irish market in 1974 without any supporting evidence), the fact remains that it is for the Commission to prove that the applicant charged unfair prices.

[265] UBC's retraction, which the Commission has not effectively refuted, establishes beyond doubt that the basis for the calculations adopted by the latter to prove that UBC's prices are excessive is open to criticism and on this particular point there is a doubt which must benefit the applicant, especially as for nearly 20 years banana prices, in real terms, have not risen on the relevant market.”

In consequence, the Court (paragraph [267]) concluded that the Commission had “not adduced adequate legal proof of the facts and evaluations which formed the foundation of its finding” that the company had engaged in the imposition of unfair selling prices. For this reason, the decision had to be annulled.

The paragraphs in the judgment which are the focus of attention in this appeal must be seen in this context. The Commission embarked upon an investigation premised upon comparator evidence. The defendant undertaking adduced evidence of a different (Cost-Plus) nature (viz., that it had been consistently loss making). The Court annulled the decision because the Commission had failed to take this rebuttal evidence “into account”. The Commission had not met its burden and standard of proof and the benefit of any resulting doubt had to be accorded to the undertaking.

The parties referred to many other authorities. I turn now to the principal case law. These address the construction of the United Brands test and a series of related points including as to the duty of a competition authority to address evidence and arguments of an exculpatory nature adduced by undertakings being investigated.

This was a reference from the Cour D’Appel Aix-en-Provence and, inter alia, addressed the test to be applied to determine when royalty rates imposed by a dominant copyright collecting society were abusive as unfair.

One issue concerned the practical utility of a comparison between different rates charged in different Member States. The third question posed by the national court asked whetherArticle 102 was infringed where a dominant copyright-management society fixed: “… a scale and rate of royalty which is several times greater than that applied by all copyright-management societies in the member countries of the EEC without any objectively justifiable ground and is unrelated to the sums redistributed to the authors, so that the royalty is disproportionate to the economic value of the service provided?”. The Court made clear that such comparisons could provide: “…useful indications regarding the possible abuse of a dominant position by a national copyright-management society”. At paragraph [38] it stated that when comparisons led to the conclusion that the fees charged were “appreciably higher than those charged in other Member States” then “… that difference must be regarded as indicative of an abuse of a dominant position”. Although the Court did not use the language of a switching burden of proof it went on to add: “In such a case it is for the undertaking in question to justify the difference by reference to objective dissimilarities between the situation in the Member State concerned and the situation prevailing in all the other Member States”. If an undertaking did adduce an objective justification it was then for the authority to consider it and in effect refute it. The Court did not say that an undertaking was limited as to the evidence it could adduce in its defence.

In this case the Commission rejected a complaint by operators at Athens airport that certain airport charges were abusive. The Commission did not conduct a Cost-Plus analysis. It concluded upon the basis of the evidence submitted to it that there was no Community interest in pursuing a full evaluation and closed the file. On appeal the complainant argued that the Commission had failed to carry out an adequate or sufficiently full evaluation. The Court of First Instance rejected this contention. On further appeal to the Court of Justice it was argued that the lower Court had erred in not condemning the Commission for failing to conduct a Cost-Plus investigation as required by United Brands. The Court rejected this argument. The Commission had carefully examined the rates of the impugned charges by comparison with those applied in other airports and there was no dispute as to the material accuracy of that analysis. The Court of First Instance had not erred in holding that the Commission

The CMA and the Commission argue that this supports their contention that paragraph [252] of United Brands lays down two strict alternatives and that they are not obliged to adopt a combinatorial approach. I disagree. The judgment is certainly inconsistent with the proposition that the Commission must always analyse more than one appropriate test. But that is not the same as saying that in an appropriate case, the Commission can ignore exculpatory evidence of another type if it is prima facie relevant. On the facts of this case the Commission chose to examine comparator data. The Court held that excessive prices “may” be determined by reference to whether the price is “unfair in itself” or competing product data, but it was not obliged (“must”) to perform a Cost-Plus exercise. The Court did not say that the alternatives were legally mutually exclusive and/or sufficient in their own right. The ruling is a far cry from saying that if the Commission selects one method, but a defendant undertaking serves relevant exculpatory evidence upon some other basis then the Commission can avoid fairly evaluating that evidence. On the facts the reason why the Commission did not have to adopt more than one method was because the evidence the complainants served was “unconvincing” (cf paragraph [70]). The only relevant evidence was hence the comparator evidence and absent any prima facie relevant rebuttal evidence the Commission had sufficient before it to make a finding (in this case not to pursue the case and to close the file).

This is a Commission decision and is cited by Pfizer and Flynn as indicative of an approach followed by the Commission in a case which is inconsistent with the stricter arguments now advanced by the CMA and Commission. Before this Court however the Commission urged caution. It says that this was a Commission decision only, it was limited to very particular facts, and is one of the very few cases (ever) where the Commission has taken account of demand side factors in analysing economic value. The significance of the decision is that the Commission (cf paragraph [103]) apparently construed paragraph [252] of United Brands in a broader manner than is now said to be justified. In paragraphs [100] – [103] the Commission stated:

“100.

The Court did not specifically set out how the “economic value” of a product should be determined, although it stated in paragraph 251 of its judgement that “the excess could, inter alia, be determined objectively if it were possible for it to be calculated by making a comparison between the selling price of the product in question and its cost of production, which could disclose the amount of the profit margin”.

101.

The Court further stated in paragraph 252 that “[t]he questions therefore to be determined are whether the difference between the costs actually incurred and the price actually charged is excessive, and, if the answer to this question is in the affirmative, whether a price has been imposed which is either unfair in itself or when compared to competing products”.

102.

It is important to note that the decisive test in United Brands focuses on the price charged, and its relation to the economic value of the product. While a comparison of prices and costs, which reveals the profit margin, of a particular company may serve as a first step in the analysis (if at all possible to calculate), this in itself cannot be conclusive as regards the existence of an abuse under Article 82.

103.

In this decision, the Commission will follow the methodology set out by the Court in paragraph 252 of the United Brands judgement. The Commission will therefore assess the costs actually incurred by HHAB in providing the products/services in question (the costs of production) and make a comparison with the prices actually charged (section II.B.2.1). The Commission will then assess whether the prices are unfair when compared to prices charged to other users or by other ports (section II.B.2.2), or whether the prices are unfair in themselves (section II.B.2.3).”

Pfizer and Flynn point out that here the Commission did not treat the alternatives in paragraph [252] as dispositive and that it generally supports the broader position adopted by the Tribunal, which the CMA conceded before the Tribunal was the proper course to adopt.

This case concerned a reference relating to an allegation that royalties charged by a copyright collecting society were abusive. The collecting society was fined by the competition authority for imposing unfair rates. The society appealed to the Competition Council. There was an appeal to a first instance court and then an appeal to the Supreme Court (Administrative Cases Division) therefrom. The court referred 7 questions to the Court of Justice (cf judgment paragraph [24]). This was not a case involving a Cost-Plus analysis since in cases involving intangible property, such as copyright, it is recognised that such an analysis might be artificial. The questions referred sought clarification as to how to apply the unfairness test in Article 102 in such circumstances. The questions raised concerned issues such as: the sufficiency for the purpose of Article 102 of an authority conducting a limited exercise comparing national rates with those in neighbouring states; the sufficiency of using the PPP index of gross national product; whether if comparisons must be made they had to be based upon rates for each segment or averages; when differences in rates would be considered appreciable; and, as to the nature of the evidence that was to be expected from a defendant undertaking to prove the fairness of rates alleged to be unfair. The questions did not specifically distinguish between the different limbs or stages of the

The Tribunal however placed great store upon the “authoritative” Opinion of Advocate General Wahl in this case (cf Judgment paragraphs [293] and [294]). A central theme running throughout the Opinion is the very great importance attached to competition authorities applying a combination of different tests to establish abuse. An important illustration is in paragraph [43] (and the Tribunal’s endorsement of the Opinion and that paragraph) where under the heading “Combining different approaches” the Advocate General stated:

“43.

Thus, in the absence of an ubiquitous test and given the limitations inherent in all existing methods, it is in my view crucial that in order to avoid (or, more correctly, to minimise) the risk of errors, competition authorities should strive to examine a case by combining several methods among those which are accepted by standard economic thinking and which appear suitable and available in the specific situation. It seems to me that those which can be found in the Court’s case-law (and that have been illustrated in points 18 and 19 above) may serve that purpose.”

Elsewhere the Advocate General cites with approval the judgment of the Tribunal in Napp (see below) as an illustration (paragraph [44]) of the approach at work and says that this “…is also consistent with suggestions made in international discussion fora of those authorities as well as in contemporary economic literature”. He accepts however that the approach has been criticised upon the basis that the “combined application of several imprecise methodologies, even where producing mutually consistent results, may not lead to a more reliable conclusion.” Weaknesses in the application of one method are not necessarily remedied by applying another equally weak method (cf paragraph [45]).

The Advocate General was not entirely consistent in his choice of language and he does not always distinguish, when referring to the use of different tests, whether he is referring to the first limb of United Brands (excessiveness) or the second limb

The Tribunal seems to have construed the Opinion and judgment as endorsing the mandatory use of a combinatorial approach at the least in relation to the first limb of the United Brands test (excessiveness): Judgment paragraph [443(1)]. In relation to the fairness limb of United Brands the Tribunal also relies upon the Advocate General as indicating that when an authority uses one of the alternative fairness tests it should also use the other as a “sanity check”: Judgment paragraph [293(14)].

The CMA and Commission interpret the Judgment as compelling a combinatorial approach i.e. the mandatory use of more than one of tests or methods to establish abuse. The CMA and the Commission disagree with this analysis and conclude that if correct it would make enforcement unworkable.

The judgment of the Court however supports the conclusion that a combinatorial approach is not mandatory in all cases at either limb or stage of the United Brands test. In paragraph [37] the Court endorsed paragraph [36] of the Opinion which reads:

“(a)

No single method or test

36.

It can be safely stated that, at the current stage of legal and economic thinking, there is no single method, test or set of criteria which is generally accepted in economic writings or across jurisdictions for that purpose. Different authorities as well as lawyers and economists have suggested a number of methods of analysis (as well as a variety of criteria, tests or

‘screens’) to that end. However, in point of fact, each of those methods reveals some inherent weaknesses.”

Nothing in paragraph [36] however suggests that a competition authority can use only one test and ignore evidence adduced by an undertaking relating to another. Indeed, the opposite conclusion would seem to apply. At paragraph [49] the Court was explicitly linking the absence of a single test with the authority’s margin of manoeuvre and discretion: “It falls to the competition authority concerned to make the comparison and to define its framework, although it should be borne in mind that the authority has a certain margin of manoeuvre and that there is no single adequate method”. The Advocate General also emphasised the existence of a margin of manoeuvre or discretion for competition authorities:

That conclusion is supported by reference to the cases the Court endorses. It cited

The judgment of the Court is not authority for the proposition that in law competition authorities must use more than one method for determining abuse (under either limb or stage of the United Brands test). The use of a combinatorial approach might be good practice or might be requisite on the facts of a particular case, but that is not the same as saying that it is a universal rule of law. But neither is this case authority for the proposition that an authority can ignore prima facie relevant evidence adduced by a defendant undertaking. The case supports the existence of a margin of manoeuvre for competition authorities.

The Commission issued a Statement of Objections against Intel for concluding rebate agreements with customers alleged to have the capability of foreclosing competition to rival chip manufacturers. Intel adduced rebuttal economic evidence (by means of an AEC test) designed to show that the rebates in fact exerted no foreclosing effects. The Commission decision addressed this evidence at length. However, on appeal before the General Court the Commission argued that there was no need for the Court to consider Intel’s AEC defence because the abuse was established if the rebate agreements in dispute had the capability to harm competition, irrespective of whether in fact they exerted such an effect. The Court agreed and did not address Intel’s defence that the Commission had erred in its analysis of the AEC test. The General Court held “…that the Commission must only show that a practice is capable of restricting competition”. The fact that the EU Commission had conducted an analysis of effects in the decision did not alter the conclusion that it was under no obligation in law to do so.

On appeal Intel argued that where an undertaking subject to an investigation adduced exculpatory evidence to rebut the inculpatory evidence of the Commission, it was under a duty to consider it and (it followed) the General Court had erred in failing to evaluate the analysis of the Commission. A Grand Chamber of the Court agreed. In paragraphs [137] – [147] the Court held that prima facie if a dominant undertaking entered into rebate agreements designed to tie customers to the dominant supplier that was an abuse. However, where the undertaking submitted by way of defence supporting evidence (of a different type) that its conduct was not capable of restricting competition and, in particular, of producing the alleged foreclosure effects, then the position was different:

“139.

In that case, the Commission is not only required to analyse, first, the extent of the undertaking’s dominant position on the relevant market and, secondly, the share of the market covered by the challenged practice, as well as the conditions and arrangements for granting the rebates in question, their duration and their amount; it is also required to assess the possible existence of a strategy aiming to exclude competitors that are at least as efficient as the dominant undertaking from the market (see, by analogy, judgment of 27 March 2012, Post Danmark, C-209/10, EU:C:2012:172, paragraph 29).

140.

The analysis of the capacity to foreclose is also relevant in assessing whether a system of rebates which, in principle, falls within the scope of the prohibition laid down in Article 102 TFEU, may be objectively justified. In addition, the exclusionary effect arising from such a system, which is disadvantageous for competition, may be counterbalanced, or outweighed, by advantages in terms of efficiency which also benefit the consumer (judgment of 15 March 2007, British Airways v Commission, C-95/04 P, EU:C:2007:166, paragraph 86). That balancing of the favourable and unfavourable effects of the practice in question on competition can be carried out in the Commission’s decision only after an analysis of the intrinsic capacity of that practice to foreclose competitors which are at least as efficient as the dominant undertaking.”

Since the Commission was under a duty to consider Intel’s AEC defence the General Court had erred in finding otherwise. This judgment is important. First, it makes clear that if an undertaking adduces evidence of a type unlike that which the competition authority relies upon to establish an abuse then the authority is under a duty to consider that evidence (cf “required” in paragraphs [139] and in [144]). Second, it follows that had Intel not adduced exculpatory economic evidence it would have been open to the Commission to rely upon the more limited category of evidence related to capacity in order to establish abuse i.e. where there may be more than one category of evidence relevant to the establishment of an abuse there is no duty on the competition authority initially to consider all possible categories of evidence, provided of course that the evidence relied upon is sufficient.

In Napp the Tribunal endorsed the use by the competition authority of a broad combinatorial approach which deployed a variety of different methods based upon both costs and comparables. The case was concerned with abusive pricing by a dominant pharmaceutical company. The judgment was cited with approval by the Advocate General in Latvian Copyright (cf Opinion paragraph [44]).

In the decision (paragraph [203]) the Director stated that, in principle, a price was excessive “…if it is above that which would exist in a competitive market and where it is clear that high profits will not stimulate successful new entry within a reasonable period. Therefore, to show that prices are excessive, it must be demonstrated that (i) prices are higher than would be expected in a competitive market, and (ii) there is no effective competitive pressure to bring them down to competitive levels, nor is there likely to be.” The Tribunal commented:

“391.

While there may well be other ways of approaching the issue of unfair prices under section 18(2)(a) of the Act, the Director’s starting point, as stated in paragraph 203 of the Decision, seems to us to be soundly based in the circumstances of the present case.”

The combinatorial approach included: (i) Napp’s community prices relative to those of competitors; (ii) Napp’s price in the community segment over time notwithstanding the expiry of its formulation patent in 1992; (iii) a comparison of Napp’s price in the community segment of the market (where it did not face competition) with its price in the hospital segment of the market, where it faced competition; (iv) a comparison of Napp’s prices in the community segment relative to its prices for export on a contract manufacture basis where it faced competition; (v) a comparison of Napp’s gross profit margin on sales to the community segment compared with the margin on other products sold to the NHS; and (vi) Napp’s gross profit margin on sales to the community segment compared with gross profit margins of its next most profitable competitor.

The Tribunal held:

“397.

In our view those comparisons, taken together, amply support the Director’s conclusions that Napp’s prices in the community segment were, during the period of the infringement, well above what would have been expected in competitive conditions. Thus we agree with the Director’s finding, at paragraph 211 of the Decision, that it is only in the community segment, where buyers are less price sensitive, and where there is an absence of effective competition, that Napp can sustain a premium of 40 per cent over competitors. With the exceptions mentioned in paragraph 360 above, Napp’s prices have remained unchanged for 20 years, including nearly 10 years since the expiry of Napp’s formulation patent. At the same time, Napp has retained a market share of the community segment of some 96 per cent. Those facts also support the proposition that Napp’s community prices, unlike its hospital prices, have not been subject to competitive pressure, as the Director found in paragraph 213 of the Decision”

The Tribunal recorded the detailed rebuttal evidence Napp adduced during the administrative procedure. The judgment is an exemplar of a combinatorial approach.

This case concerned the lawfulness of the price at which a party in sole possession of valuable information (pre-race data about British horse races) was willing to supply it to another party. The claim posed the question: “… when is the price charged by the person controlling access to...information so high as to be excessive or unfair?” The Court of Appeal (paragraph [6]) identified the many imponderables arising:

“Is there a pricing principle which can be applied to such a case? If so, what is a non-abusive "right price" and how is it to be ascertained by the court? Is it, as was held in this case, the cost of production of the information plus a reasonable profit (called "cost +")? If the possessor of the information may only lawfully charge a price calculated in this way, how does the court set about ascertaining the cost + price? In comparing the price charged and the cost incurred, what should be included in the allowable costs incurred? Is it only the costs directly involved in the secondary activity of creating, collating and compiling the information, or does it include, or reflect, all, or only some, and, if so, which, of the costs incurred in conducting the primary activity to which the information relates?”

The parties disagreed about the approach to be adopted. The claimants argued that the test was Cost-Plus. The defendant (BHB, who owned the data) argued that the test was, upon the basis of United Brands, the economic value of the product. At paragraphs [213] and [218] the Court held that there was no single test:

“213.

As already noted, the Commission's decision in Scandlines supports the view that the exercise under Article 82, while it starts from a comparison of the cost of production with the price charged, is not determined by the comparison. This in itself is sufficient to exclude a cost + test as definitive of abuse. Mr Roth accepts that there is no single methodology or litmus test of abuse: the court has a choice of methods, but not an unlimited one. His contention is that the judge has gone outside the admissible limits of method in coming to his conclusion. Mr Hollander, also contending that the choice of methodology is for the court, defends both the choice made by the judge and the way he has implemented it.” And:

“218.

For all the above reasons we conclude that, in holding that the economic value of the pre-race data was the cost of compilation plus a reasonable return, the judge took too narrow a view of economic value in Article 82. In particular he was wrong to reject BHB's contention on the relevance of the value of the pre-race data to ATR in determining the economic value of the pre-race data and whether the charges specified by BHB were excessive and unfair.”

I would draw the following general conclusions from the case law about the test to be applied:

The above is not an exhaustive list of the points arising from case law. There are other points of importance relating to the burden and standard of proof on competition authorities and the nature and extent of the evidential burden which is placed upon undertakings. I deal with these elsewhere in the judgment at paragraphs [110] – [117] below.

The parties rely upon economic principle and literature in aid of their arguments. Three volumes of economic literature were before the Court which focused upon how competition authorities (in the EU and elsewhere) and economists viewed the tests to be applied. My conclusion from this is that the economic literature supports and is consistent with the conclusions of law that I have drawn from the case law (above). I would summarise the gist of the arguments of the parties as follows.

The CMA relies upon the literature to support its arguments on: the need for competition authorities to have a wide margin of appreciation; the need to avoid imposing upon competition authorities overly rigid rules (such as it says the Tribunal has); and, as supporting its position that the facts of the present case are just the sort of case where ex post intervention is justified. The pharmaceutical companies rely upon this material to point out that proving abuse is an essentially economic exercise and that accordingly there can be no artificial evidential rules applied (such as it says the CMA and Commission are arguing for). The companies also argue that in economic terms the logic for ex post intervention in pricing is weak since markets tend to be self-rectifying and regulatory intervention risks creating false and perverse incentives. Over and above these broad considerations the literature is referred to in relation to the following specific issues: (i) the economic logic of drawing a clear legal distinction between the “in itself” and the “competing products” tests in United Brands paragraph [252]; (ii) the practical workability of the test described by the Tribunal; and (iii), the economic logic of requiring a hypothetical price comparator to be used in all cases.

Mr Hoskins QC for the CMA cautioned that some of the literature was indicative of what the law “should be” as opposed to “what it was”. I am conscious of this risk. The literature does not lay down the law but is indicative of the economic policy which underlies it. All parties acknowledged that the Court of Justice has explained that the test of unfairness is in large measure economic in nature. In United Brands (ibid) the Court (paragraph [253]) expressly observed that “economic theorists” had identified various ways of determining excessive prices over and above the methods which the Court itself identified. In Latvian Copyright Advocate General Wahl (Opinion paragraph [36]) also observed that there was no single test that economists had identified for determining whether prices were fair. In paragraph [37] of its judgment in that case the Court endorsed paragraph [36] of the Opinion and pointed out that the Court itself (in paragraph [253] of United Brands) had made the same point. Elsewhere the Advocate General referred to economic literature as of importance, for instance in paragraphs [43] and [107] and the Court approved paragraph [107] in its Judgment at paragraph [56]. In short, it is proper and helpful in a case such as this to look to economic literature for insight.

In October 2018 the OECD published a Background Paper entitled “Excessive Prices in the Pharmaceutical Markets” (“the OECD Paper”) which, in a balanced and comprehensive manner, summarises and lists the principal literature on the issue. All parties cited extensively from this document. It places economic thinking in the context of EU case law. It refers to the recent cases worldwide where competition authorities have intervened ex post in cases of alleged exploitative pharmaceutical pricing and the OECD Paper includes commentary on the judgment of the Tribunal in the present case. It was treated by the parties as a fair summary of current economic thinking.

Mr Hoskins QC drew our attention to various features of pharmaceutical markets recognised in the literature, to support the submissions of the CMA favouring a wide margin of appreciation for competition authorities and the importance of ex post intervention. The OECD Paper explains why such ex post intervention is relatively unusual but also how it can be very important in certain types of case. Medicines are subject to a “dense and comprehensive” regulatory framework that recognises the limited ability of competition enforcement agencies to lower prices. The framework is less comprehensive in relation to off-patent drugs, where inter-brand competition is relied upon to contain prices. But competition concerns can arise even in the offpatent sector where there can remain an absence of therapeutic and inter-brand competition even upon expiry of patents. This can lead to a “lack of price elasticity of demand, particularly as regards ‘essential’ drugs”. These developments appear, observationally, to have occurred in tandem with the emergence of business strategies that identified market segments where prices could be successfully increased. Companies identify niche essential drugs that are not under patent but whose market is so small that no competitors will enter the market, or where supply is limited for regulatory or contractual reasons. In its submissions the CMA identified analogous factors specific to the drug in question in these proceedings which it contended made the present case apt for ex post intervention. These included that phenytoin was not in patent. It is an “old” product first marketed in 1938 which was used for a declining patient population but where the suppliers benefited from regulatory clinical guidance which substantially precluded switching even between clinically identical, molecular, substitutes. This served to maintain barriers to switching and inter-brand competition and was an important factor in establishing the dominant positions of Pfizer and Flynn in their respective markets and their power over price.

These features served to distinguish the present case from other markets where patent expiry removed the principal obstacle to market entry. Where there are no material barriers to entry high prices can act as a magnet to entry which, in due course, drives prices down. Many markets are thus self-correcting. In the absence of entry barriers regulatory intervention can risk prolonging a monopoly situation by blocking efficient signals which would otherwise promote market entry. A belief in market forces “…is often bolstered by the (perceived high) likelihood of regulatory failure, a risk which is compounded in the case of price regulation”. The investigation of ex post cases of alleged excessive pricing faces significant difficulties in terms of data availability and analysis, identifying appropriate assessment standards, and of designing and implementing suitable remedies: “This has led some to consider that the identification of excessive prices is a ‘daunting, if not, impossible task’… The issues are still more extreme when trying to set clear rules that allow for ex ante compliance with excessive pricing rules. The key problem here is that it is not clear what the appropriate benchmark should be.”. In written submissions the CMA argued that the economic difficulties in applying competition law to excessive pricing must not be allowed to render “…the law a dead letter”. It was the task of the competition authority to exercise its judgment as to when ex post intervention was apt, and courts should avoid articulating rules which made this inherently difficult task unworkable or excessively difficult.

The pharmaceutical companies cited the OECD Paper because it described the many different approaches which authorities use to evaluate whether prices are exploitative/unfair. The OECD Paper cites United Brands as laying down the seminal test in the EU and observes that European competition authorities and courts have made use of a variety of different methods, all said to be consistent with the case law, to determine whether a price is excessive and unfair. In some cases, a comparison between production costs and prices is used but price/cost analysis is not feasible in all cases due to lack of data or because the disputed price relates to an intangible good such as an IP right. Other methods are also used such as benchmarking “of some sort”. Price-based benchmarks are used by comparing the investigated price with prices charged by the dominant firm in different markets or over time or by comparing the prices charged by the dominant firm and those charged by other firms, either in the same market or in other markets. Another benchmark focuses upon the profitability of the dominant firm by comparing such profits either with a normal competitive profit or the profits of other firms. Other methods are also identified. The guiding factor in each case is the availability and suitability of the evidence and data. Competition authorities often adopt a pick and mix or combinatorial approach to the evidence to be relied upon. There are no fixed rules, assumptions or presumptions. Everything depends upon the facts of the case.

The pharmaceutical companies pointed out that the literature recognised that there were limitations inherent in all the tests which might be used: “… all the methods to determine whether a price is excessive under competition law have weaknesses”. Accordingly, analysis “…should be carried out according to as many of the methods indicated above as possible and should look for robust evidence that prices are indeed excessive”. Economists would view a combinatorial approach as good or best practice. But the companies argued the literature does not support the contention of the CMA that it could treat its choice of one method only as sufficient and ignore relevant exculpatory evidence from investigated undertakings.

Pulling the strands together I conclude that the economic literature supports the conclusions of law that Iderive from the case law summarised above (cf paragraph [97]): There are many different tests which might be used to determine whether a price is excessive and unfair; there are or may be difficulties with all tests and much will depend upon the availability of evidence and data; all cases are highly fact and context specific; there is a need for competition authorities to be able to intervene ex post in pharmaceutical cases; and, it is economically rational that competition authorities should have a margin of appreciation as to the choice of method and evidence that they seek to rely upon.

I next address a particular argument of the CMA that its narrow analysis of paragraph [252] of United Brands as laying down strict alternatives is economically logical. The CMA cited the dichotomy drawn in Article 101 TEU between “object” and “effect” in relation to the prohibition of restrictive agreements. It was argued by analogy that the Treaty itself recognised the economic logic of a strict demarcation between object and effect and this was analogous to the equally strict distinction drawn by the Court in paragraph [252] between an “in itself” and a “competing products” fairness test. I disagree and the economic literature provides no support for this narrow argument:

I finally address the argument that any test other than that advanced by the CMA and the Commission would leave the law as unworkable. Insofar as the concerns of the CMA and Commission rest upon their interpretation of the Judgment as compelling a combinatorial approach the statement of relevant principles set out in this judgment (and the consistent position set out in the judgment of the Chancellor) should address that concern. The OECD Paper is generally sanguine about fears of over-complexity: “… even if it were true that the assessment of excessive pricing can be challenging, it would be wrong to overstate the difficulties of pursuing such an assessment. While it can be hard to set out simple guidance that draws a clear line between excessive and lawful pricing, there may nevertheless be cases where pricing is so extreme that it becomes relatively easy to demonstrate that it is excessive by reference to a variety of different measures. Acknowledging the difficulties in identifying a single standard which is able to determine whether prices are excessive in all situations, many competition authorities have instead applied several methodologies in parallel in order to minimise the possibility of erroneous intervention”. Very much the same argument was advanced by the Commission in Intel before the General Court in relation to the AEC defence advanced by Intel (see paragraphs [87] – [89] above). As is clear from the Commission decision in that case the AEC analysis was hugely complex and detailed, yet this was irrelevant to the judgment of the Court which held that since an AEC test was in principle relevant then it was the duty of the Commission (and the court on appeal) fairly to evaluate it.

I turn next to an issue of broader practical importance concerning the extent of the duty on competition authorities to investigate. The Tribunal described the nature of the investigation that the CMA must conduct into competing evidence adduced by an undertaking using different language. In several places it referred to a “sufficient investigation” but on one occasion appeared to equate this with a “full investigation”

I consider however that the fears of the CMA are largely unjustified. The Tribunal did use different language (as the Chancellor points out in his judgment) and on one view the terminology used (in particular the word “full”) does imply a general burden upon the CMA which goes beyond that which is consistent with the existence of a margin

Elsewhere it emphasised that on the remittal the CMA would be able to exercise its

“margin of discretion” as to the evidence to be examined (e.g. Judgment paragraph [392]). I see the force in the submission of the CMA as to the lack of clarity in the description of the test which the CMA must apply both generally and upon remittal. To remove any doubt, I address the CMA’s concerns upon the basis that the Tribunal has defined a general duty of evaluation which does circumscribe the margin of manoeuvre of the CMA.

I resolve this issue by reference to first principles. I have already set out my conclusions on the general test following United Brands (cf paragraphs [60] – [63] and [97] above). By the very nature of the legal test for abuse a competition authority has a margin of manoeuvre or discretion as to the method(s) it uses and the evidence it relies upon. How it goes about evaluating the evidence will be fact and context specific and will take into account the evidence adduced by a defendant undertaking. I have set out the law on burden and standard of proof at paragraphs [114] and [115] below.

At base the CMA has a duty to conduct a fair evaluation of all the evidence before it. What this means in a given case is impossible to say in advance and will depend upon the facts of the case. A degree of proactivity might be needed, in some cases, but not in others. For instance, at both the UK and EU level the CMA and Commission have the statutory power to seek information from third parties by the use of statutory demands for information, if they consider this necessary to establish an infringement, and such powers are routinely considered and used even in investigations which might be described as relatively light touch (and certainly not “full”). What this ultimately boils down to is the existence of a margin of manoeuvre or discretion that the authority has in relation to the evaluation of evidence before it. The law does not predetermine in advance how intensive any particular evaluation by the authority will be. The notion of a duty to evaluate evidence fairly encapsulates the contextual nature of the duty. If the CMA fails in this duty the Tribunal exists to remedy any such failing. In short there is a duty of fair evaluation upon the CMA in relation to the evidence before it. But it has a discretion as to how it performs that duty which includes as to the depth and intensity of the inquiry and there is no general duty to perform a “full” investigation in all cases. In coming to this conclusion I am not precluding the possibility that in a specific case the Tribunal might, having found an error in the analysis and findings of the CMA, remit the matter with a very specific direction in the judgment (under paragraph 3 of Schedule 8 CA 1998) as to the steps that must be taken during a fresh evaluation to correct the error. But the present case is not such a case; the remittal was in general terms. To the extent that there is an issue on the facts of this case about the scope of the remittal, this is dealt with under the fifth (Flynn) Ground of Appeal (paragraphs [174] – [182] below).

I turn now to a different point. The CMA expressed the concern that if the duty imposed upon it was to investigate “fully”, that undertakings could “game the system” and adopt a strategic approach to proceedings designed to heighten the burden on the authority and thereby deter enforcement. I do not consider that the law permits this to occur. Case law makes clear that the extent of the duty on an authority to evaluate evidence adduced by an undertaking (taking account of its due margin) is affected by the quality of that evidence. There is an important evidential burden upon an undertaking being investigated. Case law exemplifies this. In United Brands the Court condemned the Commission for failing to investigate evidence on costs and profits adduced by the undertaking even though it criticised that evidence as in some respects “unreliable”. This was because, limitations apart, the evidence refuted key elements of the Commission’s analysis and cost data was readily available in the form of a recent UNCTAD report. The undertaking had met its evidential burden and the burden of fair evaluation lay upon to the Commission. But the evidential burden can be quite exacting. A different result arose in Athens Airport (ibid) where the Court found that the complainant had not adduced evidence of sufficient quality to shift the burden back to the Commission (see paragraphs [74] – [75] above) and it therefore had no obligation to proceed to evaluate the category of evidence that the complainant wished the Commission to assess. In Case T-216/13 Telefonica v Commission (28^th June 2016) the General Court explained that where the Commission had demonstrated the existence of an infringement it was not enough for a defendant undertaking simply to raise the “…possibility that a circumstance arose which might affect the probative value” of the Commission evidence - speculation will not do. It “…was not capable of affecting the probative value” of the Commission evidence. In Case C-105/04P NFVV v Commission in her Opinion of 8^th December 2005 Advocate General Kokott, in the context of a discussion about the relative burdens of the Commission and investigated undertakings, observed that where the Commission has drawn a conclusion based upon “…objectively verifiable evidence from stated sources…” then such findings could not be undermined simply by the undertaking “unsubstantiatedly disputing them” (ibid page I-8748 paragraph [74]) - assertion will not do.

I deal finally with one argument raised by the CMA which if correct could have wide ramifications. Mr Hoskins QC sought to argue, as a matter of principle, that the CMA was justified in not conducting a full investigation of a defendant undertaking’s evidence by virtue of the fact that the Tribunal had a merits jurisdiction and could, if needs be, correct failings on the part of the CMA. This meant that there was always a safeguard against poor administrative decision making: no harm would come of it. I disagree. I can see no basis upon which the fuller, merits, jurisdiction, of the Tribunal governs or affects the scope of the duty of the competition authority at first instance. As Ms Bacon QC for Flynn pointed out by reference to case law of the Court of Justice and other considerations there were many reasons why competition authorities must conduct a sufficient, self-contained, evaluation: (i) the jurisdiction of competition authorities is treated as quasi-criminal because it can lead to the imposition of potentially vast fines and no undertaking should be subjected to a criminal penalty without have been subjected to a fully adequate evaluation of the evidence; (ii) case law establishes that the burden of proof remains on the competition authority to prove its case with evidence that is “objective, appropriate and verifiable” (e.g. Latvian Copyright paragraph [41]) and the presumption of innocence lies with the investigated undertaking throughout; (iii) competition authorities are subject to duties of good administration which do not contemplate that sub-optimal evaluations are acceptable merely because there is an appeal route which might cure that failing (see e.g. Case C-272/09P KME Germany v Commission (8^th December 2011) at paragraphs [107] and [108] - the duty of good administration means that the

In short, the authority has a duty to conduct a fair evaluation of the evidence. It has a margin of manoeuvre or discretion in how it goes about meeting this obligation. This might, depending upon the facts, involve the taking of proactive steps, such as the issuance of requests for information to third parties, but it will not inevitably do so. The extent of the duty will be affected by the nature, extent and quality of the evidence adduced by the defendant undertaking which has an evidential burden. The fact that upon an appeal the Tribunal might review the evaluation is not a factor which affects the nature and extent of the prior duty imposed upon the competition authority.

This ground of appeal has raised a number of issues where both I and the Chancellor have concluded that the position adopted by the Tribunal needs to be clarified. However the core question arising concerned the correctness of the position adopted by the CMA to the effect that the “in itself” and “competing products” tests were “true alternatives” in the sense that if the CMA relied upon one alternative to find abuse then it had no obligation in law to evaluate other prima facie evidence that prices were fair adduced by a defendant undertaking. On this basis, I accept that the Tribunal was right to say that the “in itself” and “competing products” tests were not strict alternatives. I therefore disagree with the CMA on this central issue. Subject, therefore, to the clarification that both I and the Chancellor have made, the appeal on this ground fails.

The second Ground of Appeal concerns the interpretation of paragraph [249] of United Brands (see paragraph [56] above). The CMA argues that the Tribunal erred in that it mandated that a competition authority “should", as part of its analysis, construct a hypothetical benchmark price or range of prices against which to measure the actual prices charged. The Tribunal held (paragraph [443(1)] – see paragraph [40] above) that the CMA should: “…consider a range of possible analyses, reflecting market conditions and the extent and quality of the data that can be obtained, to establish a benchmark price, or range, that reflects the price that would pertain under conditions of normal and sufficiently effective competition.”. Elsewhere, the Tribunal referred with approval to the need to compare the actual price with a “hypothetical” price (cf paragraphs [294(11)], [312], [313] and [316]). The Tribunal cites the Opinion of the Advocate General in Latvian Copyright as authority for this proposition: See paragraph [79] above.

The CMA and the Commission argue that there is no basis in law for this requirement, which they say is not to be found in United Brands or later case law and to the extent

The answer to this ground lies in the summary at paragraph [97] above: The authority has a margin of manoeuvre or discretion as to how it goes about proving its case, subject always to the appellate jurisdiction of the Tribunal. To the extent therefore that the Tribunal compelled the use of a particular test then in my view it has misconstrued the case law. It is not entirely clear what the Tribunal was referring to when it used the expression “hypothetical”price. If this was intended to refer to an artificially constructed price, then I agree with the CMA and the Commission. But it might well be that the Tribunal was referring simply to the exercise of calculating a benchmark ROS or ROCE and/or the exercise of looking to external comparators. Nonetheless, given the uncertainty which has arisen in respect of the phrase I consider it necessary to consider what sorts of evidence should be used in the analysis.

First, as to the expression “hypothetical” nothing suggests that in every case there is a need for the creation of a hypothetical benchmark, in the sense of an artificial construct. Indeed, the thrust of the OECD Paper and the literature it cites suggests that the counterfactuals of greatest practical value are often those drawn from real life, as opposed to some hypothetical model. The case law supports this conclusion. I note that in some cases the Commission does use hypothetical modelling (part of the AEC test used in the Intel case was based upon just such hypothetical modelling) but the point is that it is not a necessary part of the analysis. Any suggestion by the Advocate General in Latvian Copyright that the use of hypothetical price benchmarking was mandatory is not a proposition that was endorsed by the Court which, as already observed, emphasised the flexibility of the margin of manoeuvre of competition authorities.

Second, as to whether that benchmark must relate to price, I agree with the CMA and the Commission. I also agree with the submissions of Ms Bacon QC for Flynn (who ultimately did not support the reasoning of the Tribunal, if the Judgment was to be construed as requiring a hypothetical benchmark price in every case) that in both the law and in economics all that is required is that there be “a” benchmark or standard against which to measure excess or fairness. The need for a comparator is economically logical since the concepts of fairness, excessiveness and reasonableness are all relative concepts. They must be compared with their counterfactual e.g. unfairness, normality or unreasonableness. But case law and literature make clear that there are numerous counterfactuals which might be used, and importantly this includes the costs of the dominant undertaking as well as benchmarks set by reference to ROS or ROCE or some other similar measure. As was pointed out in argument the overarching description of an abuse in United Brands at paragraph [249] is by reference to a comparison with “trading benefits” realised in conditions of normal and sufficiently effective (i.e. workable) competition. This necessarily comparative exercise does not exclude a benchmark premised upon the undertaking’s own cost

Third, I note that in paragraph [249] the Court says only that it is “advisable” to ascertain whether the undertaking had exploited its dominance in a way which it could not have “… if there had been normal and sufficiently effective competition”, these being the words said to create the requirement for a hypothetical benchmark price. There is no specific reference to price in the paragraph and in any event the expression “advisable” is inconsistent with the Court intending to provide anything more than guidance as to best practice. It would have used more directive language had it intended to lay down a fixed rule.

Fourth, the Court in Latvian Copyright approved of the statement in the Opinion that there was no single method or test or set of criteria which is inconsistent with the Court having approved of any statement to the effect that use of a hypothetical benchmark price was mandatory: see paragraph [84] above.

In my view by the nature of the abuse in issue there needs to be “a” benchmark. But, in the first instance at least, the choice of benchmark is for the competition authority to choose and can be based upon the costs of the undertaking being investigated or it can be based upon comparables such as the prices charged by the same or different undertakings in the same or different geographical markets or indeed any other benchmark or combinations thereof capable of providing a “sufficient” indication that the prices charged are excessive and unfair. It follows from the above that assuming the Tribunal was mandating the use in all cases of a hypothetical benchmark price which did not include the costs of the undertaking or some other benchmark related to the undertaking, then I respectfully disagree with the Tribunal. I would allow this Ground of Appeal.

The CMA argues, in line with the position set out in the Decision, that: (i) because it applied an “in-itself” test there was no obligation upon it also to apply any alternative (“competing products”) test; (ii) that in any event it did conduct an analysis of competing product comparators and found that they were not appropriate; and (iii), the findings of the CMA in the Decision lay within its margin of manoeuvreor appreciation and the Tribunal erred in law in not respecting that discretion.

I have addressed the first argument under the First Ground of Appeal i.e. on the facts of this case there was an obligation upon the CMA properly and fairly to evaluate the comparator evidence because it was adduced by the undertakings as part of their defences. It was not therefore open to the CMA to ignore that evidence simply because it had, in its judgment, conducted a sufficient analysis. In the text below I therefore concentrate upon issues (ii) and (iii). I start with the issue of the margin of appreciation.

This Ground raises an important point about the existence of a margin of “manoeuvre” or “appreciation” on the part of competition authorities as to the assessment that they make of complex factual matters and the extent, if any, to which supervisory courts and tribunals should respect that exercise of judgment. The CMA, especially in written submissions, contends that the Tribunal erred in finding that the prices were not abusive in themselves. It argues, in essence, that it conducted a detailed (and sufficient) analysis of the evidence, that the Tribunal was unable to find serious fault with the preponderant part of it, and that the only flaw identified by the Tribunal was one of degree – namely that the CMA should have carried out a more intense evaluation of the comparables evidence adduced by the defendant undertakings. This was not therefore a finding that the CMA ignored a relevant consideration but went only to the CMA’s judgment call about the evidence.

The evidence relied upon by the CMA and which it holds up as “sufficient” is set out in detail in the Decision. I summarise it briefly in the following way.

In relation to excessiveness Pfizer and Flynn charged prices that materially exceeded their costs attributable to Phenytoin Capsules plus a ROS calculated at 6%. Pfizer’s average selling price (“ASP”) exceeded Cost Plus by an average excess of 443% across all capsule strengths (the excesses ranged between 29% and 705%), accruing approximately £53.9m excess profit. The ASP of Flynn exceeded Cost-Plus by an average of 41% (the range was between 31% and 133%) though these percentages understate the extent of the actual excess, because Flynn paid high supply prices to Pfizer thereby artificially inflating its costs. In absolute terms, Flynn earned £29.8m in excess profit which was more than half the £53.9m excess earned by Pfizer, despite Flynn’s limited activities and risks (as noted in the Judgment, paragraph [346]). The differential between the prices charged pre-September 2012 and those charged by each of Pfizer and Flynn post September 2012, which did not reflect any material change in costs, risk or innovation were “dramatic”. The price of Phenytoin Capsules charged by Pfizer in the UK was many multiples of its price for the same product in other Member States. These were findings of fact set out in the Decision and were not put in issue by the Judgment.

In relation to unfairness the CMA refers to the following: (i) the disparity between the prices charged and the economic value of the Phenytoin Capsules was substantial; (ii) the differential between prices charged pre and post September 2012, with no justification in terms of costs, innovation, or additional commercial risks, was dramatic; (iii) the high absolute returns generated for the parties were high, particularly for Flynn given that it undertook minimal risk and added no significant value to the supply chain; (iv) there was limited competitive pressure on the parties and the relevant markets did not function in a way likely to produce a reasonable relationship between Pfizer and Flynn’s price and economic value; (v) the prices charged by Pfizer and Flynn were significantly higher than those charged by Pfizer in other Member States; (vi) the prices charged exerted a material and detrimental effect upon the NHS, costing approximately £50m pa with no improvement in patient care and leading to disinvestment in other medical services; and (vii), both Pfizer and Flynn were aware that the prices charged were unacceptably high and, the reason for inserting Flynn as a “middleman” into the supply chain was to manage the “reputational risk” attaching to the significant price increases, rather than to benefit patients or add value to the NHS.

In the Decision the CMA held that having reached the conclusion that each impugned price was unfair “in itself” (Decision paragraph [5.476] – set out at paragraph [51] above) there was no need to proceed to reach a conclusion on whether those prices were also unfair when compared to competing products. Nonetheless, “for completeness” the CMA proceeded to consider whether “such a comparison could be conducted” (Decision paragraph [5.478]). Between paragraphs [5.479] and [5.526] the CMA considers the comparators with the brunt of the analysis explaining why, at a relatively broad level, the CMA did not consider the posited alternatives to be true or adequate comparators. I would observe that the companies adduced a wider range of evidence on this matter before the Tribunal than they did before the CMA. Mr Brealey QC, for Pfizer, drew our attention to the evidence that was before the Tribunal and as to its potential significance.

On appeal the Tribunal, upon the evidence before it which it deployed to determine whether the Decision was well-founded, concluded that the analysis of the CMA was insufficiently deep. The Tribunal accepted that “… there is no intrinsic reason why [the CMA] could not find the test of ‘unfair in itself’ met in the light of a consideration of such issues” (Judgment paragraph [369]). The error found by the Tribunal was the approach taken to the two “Alternatives” in United Brands at paragraph [252]. It was “… not so much the factors that were considered under Alternative 1 but, rather, the decision to select only Alternative 1…”. An authority cannot “simply ignore a prima facie valid argument that a price is fair under one Alternative” (Judgment paragraph [367]). Rather, an authority must “consider whether a prima facie case of fairness under one Alternative undermines the basis for the finding of unfairness under the other Alternative and produce a reasoned basis for determining that the Unfair Limb is satisfied” (Judgment paragraphs [367] and [392]). The CMA had conducted some analysis, but it had not done so “…in sufficient depth” (Judgment paragraphs [362], [379] and [392]).

The CMA argues that having correctly found that the CMA had to be accorded a “substantial margin of appreciation” (Judgment paragraph [444]) it then wrongly interfered in a legitimate exercise of judgment when finding that the investigation of the CMA of comparables was of insufficient depth. The margin of appreciation applies at each stage of the analysis, including: (a) the choice of methodology; (b) the assessment of whether a price is excessive and unfair, including its consideration of comparators when applying the unfair limb of the United Brands test and the weight attached thereto;and (c) the assessment of economic value and the overall evaluation of whether a price bears no reasonable relation to the economic value of a product.

To determine this Ground of Appeal it is necessary to be clear, from the outset, as to the difference between the judgment call that competition authorities must make under the Chapter II prohibition and Article 102, and the powers of courts and tribunals called upon supervise the decisions of such authorities. The CMA wrongly elides two quite different principles. I accept that the CMA has a “margin of manoeuvre” (the terms used by the Court in Latvian Copyright) or “appreciation” or

“discretion” (there is no magic in the different terms). This flows from the fact that the legal test under Section 18(2)(a) CA 1998 and Article 102(a) is broad brush and necessarily confers a significant latitude upon a competition authority as to the methods and evidence bases that it resorts to in order to prove an abuse of unfair pricing. This much is well established in case law.

But this is quite different in principle to the question whether the Tribunal, as a supervisory judicial body, must pay deference to that exercise of judgment. Under the CA 1998 the Tribunal has a merits jurisdiction as to both law and fact and upon the basis of established case law it is not bound to defer to the judgment call of a competition authority. It is empowered under the legislation to come to its own conclusions on issues of disputed fact and law and can hear fresh evidence, not placed before the CMA, to enable it to do so. The conferral of a merits jurisdiction upon the Tribunal flows from important legal considerations relating to the rights of defence and access to a court, under fundamental rights such as Article 6 of the Convention. The starting point is that competition law is treated as a species of criminal law. There is a wealth of case law establishing this. The Tribunal recognised this in Napp (ibid) at paragraphs [99] – [100], applying the jurisprudence of the Court in Case C-235/92 P Montecatini v EC Commission [1999] ECR I-4575 at paragraphs [175] and [176]. This conclusion was subsequently confirmed by the European Court of Human Rights in Menarini Diagnostica SRl v Italie (27^th September 2011) (“Menarini”) in relation to Italian competition law and Article 6. In Argos/Littlewoods v Office of Fair Trading [2006] EWCA Civ 1318 at paragraph [18], on an appeal in relation to alleged price fixing, the Court of Appeal recognised the necessary connection between a full merits hearing and Article 6: “The appeals to the Tribunal in the present cases were, in effect, full hearings with such relevant evidence as any party wished to adduce, witnesses being cross-examined if appropriate. That is necessary so as to ensure that Article 6 of the European Convention on Human Rights is satisfied."

The consequences of this are significant. There are several cases on point, but one suffices to illustrate the implications. In Case C-501/11P Schindler Holdings Ltd and ors v Commission (18^th July 2013) the Court addressed a ground of appeal objecting that a penal sanction had been imposed by the Commission, an administrative body, for breach of EU competition law and not therefore a court. Accordingly, since only a court could impose criminal sanctions the finding of infringement and the imposition of a penalty by the Commission was unlawful under the Convention and equivalent instruments of EU law. The Court rejected this argument. It referred with approval to the classification of competition law as criminal for the purpose of Article 6 in Menarini (ibid). However, the imposition of a penalty by a (non-judicial) competition authority was consistent with Article 6 if the undertaking concerned had the chance to challenge the decision before a court “… that offers the guarantees provided for in

Such a judicial body had to be competent to substitute its own appraisal for that of the decision maker (ibid paragraph [36]). This had implications for the “margin of discretion” enjoyed by the Commission:

“37 As regards the review of legality, the Court has pointed out that the European Union judicature must carry it out on the basis of the evidence adduced by the applicant in support of the pleas in law put forward and that it cannot use the Commission’s margin of discretion – either as regards the choice of factors taken into account in the application of the criteria mentioned in the 1998 Guidelines or as regards the assessment of those factors – as a basis for dispensing with the conduct of an in-depth review of the law and of the facts….”

The Court came to this conclusion in the context of a jurisdiction that is avowedly one of judicial review; it is not a merits jurisdiction such as that conferred under the CA 1998. It would follow that the above analysis applies a fortiori in the context of the domestic merits appeal.

From case law it is possible to draw various conclusions about the role of judicial bodies in relation to the margin of appreciation of a competition authority: (i) for a (non-judicial) administrative body lawfully to be able to impose quasi-criminal sanctions there must be a right of challenge; (ii) that right must offer guarantees of a type required by Article 6; (iii) the subsequent review must be by a judicial body with “full jurisdiction”; (iv) the judicial body must have the power to quash the decision “in all respects on questions of fact and law”; (v) the judicial body must have the power to substitute its own appraisal for that of the decision maker; (vi) the judicial body must conduct its evaluation of the legality of the decision “on the basis of the evidence adduced” by the appellant; and (vii), the existence of a margin of discretion accorded to a competition authority does not dispense with the requirement for an “in depth review of the law and of the facts” by the supervising judicial body.

Notwithstanding the above the jurisdiction of the Tribunal is not unfettered. This flows primarily from the fact that the appeal is not a de novo hearing but takes the decision as its starting, middle and end point. Under section 46 CA 1998 the appeal is

“against, or with respect to,” the decision and includes “whether” there has been an infringement. That focus upon the impugned decision is reflected in the procedural rules of the Tribunal. The appellant must identify the decision under appeal and set out why it is in error. The Grounds must set out the “extent” to which the decision “is based on an error of fact or was wrong in law”: see Rule 9(4)(d) Tribunal Rules (SI 2015/1648) (“the Rules”).

The Tribunal can hear evidence, including fresh evidence not before the CMA, and make findings of both fact and law. The Tribunal is empowered to ensure that the right to adduce new evidence is not abused: See Rule 21(2). The Notice of Appeal must include a statement identifying any evidence (whether witness statement or documentary) which was “not before the maker of the disputed decision”. The Tribunal can ensure that appeals are conducted in a focused manner including taking steps to ensure that appellants identify the “main” issues arising out of the decision with clarity and precision: see the combined effects of Rules 4 (Governing Principles) and 19 (Directions).

In T-Mobile v Ofcom [2008] EWCA Civ 1373 it was observed that the task of the

First, materiality is not an exact science. The Tribunal might be able to do no more than conclude that an error might make a difference to the final outcome or to some significant component thereof; certainty might not be possible. An error of fact or law might not be material to the ultimate question (breach or no breach) but could be material to some significant aspect of the Decision such as duration of the breach, or geographical spread, or the number of customers or consumers affected etc. These might be relevant to penalty or remedial directions.

Second, there is no fixed list of errors that the Tribunal might consider material. Case law indicates that the following might be relevant: failing to take account of relevant evidence; taking into account irrelevant evidence; failing properly to construe significant documents or evidence; drawing inferences of fact from evidence about relevant matters which are illogical or unjustified; failing adequately or sufficientlyto investigate an issue that the Tribunal considers to be relevant or potentially relevant to the analysis. Ms Bacon QC, for Flynn, cited Case C-272/09P KME Germany v Commission EU:C:2011:810 (8^th December 2011) at paragraph [94] as illustrative and analogous. The Court, in the context of a judicial review, explained that because the Commission enjoyed a “margin of discretion with regard to economic matters” that did not mean that the Court would refrain from reviewing the Commission’s interpretation of the evidence, its factual accuracy, its reliability, its consistency and also “…whether that evidence contains all of the information which must be taken into account in order to assess a complex situation and whether it is capable of substantiating the conclusion drawn from it…”.

Third, but importantly, it is consistent with a merits appeal for the Tribunal, even having heard the evidence, to conclude that the approach taken by the CMA and its resultant findings are reasonable in all the circumstances and to refrain from interfering upon that basis. If the Tribunal considers that the findings of the CMA are reasonable it might be difficult to say that any findings that it arrives at which differ from those of the CMA are material. The Tribunal in the present case indicated as much at various points in the Judgment. This point was also made by the Tribunal in Albion Waterv Water Services Regulation Authority [2008] CAT 31 at paragraph [72]. Because the Tribunal has a full merits jurisdiction and can hear fresh evidence there could of course arise circumstances where the Tribunal finds that on the evidence before the CMA it arrived at a reasonable conclusion but on the basis of the new evidence before the Tribunal the CMA’s conclusions were nonetheless wrong. Such cases may be rare, but the possibility necessarily arises because of the power of the Tribunal to receive and assess fresh evidence.

Fourth, I would expect that in a judgment the Tribunal would set out its reasoning on the materiality of errors so found. If the Tribunal annulled a decision upon the basis of an error that was very slight or de minimis and/or gave no reasoning to justify the annulment that might be considered an error of law, subject to an appeal.

With these considerations in mind I turn to consider the CMA’s argument that the Tribunal wrongly interfered in the CMA’s margin of appreciation.

I start by summarising the evidence identified by the Tribunal as indicating that the CMA had conducted an insufficient examination. Phenytoin sodium may be dispensed in either capsule or tablet form. Pfizer manufactures capsules and rivals, including Teva, produce tablet versions for the UK market. During the administrative procedure before the CMA Pfizer adduced evidence on the relationship between phenytoin sodium capsules and tablets (Judgment paragraphs [209] – [213] and [374] – [393]). The capsules and tablets are clinically identical. They are sold in the UK to the same ultimate customer, the NHS, who paid a price for the tablet double that charged by Pfizer for its capsules. It was not in dispute that Pfizer benchmarked its capsule prices against the tablet price. During the administrative procedure Pfizer encouraged the CMA to evaluate the tablet as a comparator as relevant to the economic value of phenytoin sodium. The CMA decided that in law it could ignore the tablet as a comparator and only considered the tablet for sake of “completeness”(Decision paragraph [5.478]). For this reason, the nub of the CMA’s submissions to the Tribunal focused upon why it excluded tablets, rather than upon the intrinsic merits or demerits of the comparators. The CMA adduced no evidence concerning any comparators, or possible comparator products. The expert medical evidence of Professor Walker for Pfizer stood unchallenged. The evidence of an economist for the

During the hearing below the CMA raised a new point about the relative value of Average Selling Prices (between the Pfizer and Teva products) instead of a comparison of the respective Drug Tariffs Prices. The Tribunal observed that this argument had been first raised by the CMA late on, at the end of the hearing. It was a potentially important issue but there was very limited evidence. This was not an issue properly explored in the Decision or in evidence or in pleadings (Judgment paragraph [386]). The Tribunal accepted that although the point was raised late it was nonetheless based upon material set out in the Decision (Judgment paragraph [388]). The Tribunal decided that it was not possible in these circumstances to give proper consideration to the evidence or the arguments which flowed therefrom.

In paragraphs [389] - [392] the Tribunal explained why and how this evidence might be relevant:

“389.

It is not possible, within the scope of the present proceedings, for us to give full and adequate consideration to the competitive situation in relation to tablets during the Relevant Period. We have been given at best some isolated pieces of information, and certainly not enough to form a conclusion.

390.

However, if it is indeed the case that new entrants have entered the tablet sector and that as a result price competition has reduced the tablet ASP, a matter on which we can make no finding on the evidence before us, this would suggest that one of the material reasons given in the Decision by the CMA for disregarding the tablet as a meaningful comparator, namely that it was subject to the same restrictions on competition as the capsule, would be wrong. However, that process would also be highly germane to seeking to establish the benchmark price in conditions of sufficient competition, as well as being informative on the question of unfairness. Assessing whether or not that remains the case, however, is clearly a matter for the CMA.

391.

The Decision states (para 3.448) that the CMA had considered making a formal investigation into tablets following the DH’s complaint in 2012 but decided against it on grounds of administrative priorities. This is entirely understandable. Nevertheless, in this case, the CMA should, in our view, have done a sufficient investigation into the competitive conditions surrounding the most obvious comparator product properly to inform its decision on Pfizer-Flynn Capsules. It is not an answer to state there was no obligation to conduct a full investigation. That is so in relation to the CMA's discretion in relation to the price of tablets; but it is not right in terms of obtaining sufficient evidence properly to apply Article 102 to the price of Pfizer-Flynn Capsules. As Mr Hoskins conceded in a different context (see paragraph 365 above), no authority should leave a relevant factor unclear.

392.

All this suggests to us that the phenytoin tablet as a meaningful comparator should not have been wholly rejected on the grounds relied on by the CMA. There was enough material to make it pause to consider, at the very least, whether there was a prima facie case of fairness under Alternative 2. We accept that there is an element of circularity in this. The authority must investigate possible comparator candidates to see if they are likely to be meaningful on objective, verifiable and appropriate criteria. On the other hand, the authority has a margin of discretion as to the possible comparators that it needs to examine. How is it to know whether the comparators are likely to be meaningful unless it examines them? That difficulty does not avoid the need, in our view, for the authority, at least, to examine any prima facie good comparator, as the CMA accepted.”

In my judgment it follows from the above that the Tribunal’s findings were made within its jurisdiction. It specified the areas where it found the evaluation lacking. It was not bound by the CMA’s margin of “manoeuvre” or discretion. It has explained why in its view the error could be material. I can detect no error in the approach adopted by the Tribunal. At base this is an objection to a finding of fact and I therefore reject this Ground of Appeal.

This Ground of Appeal addresses the concept of “economic value” as that phrase is used in paragraph [250] of United Brands. It focuses upon: whether the test is a legal or economic test; whether it is the same as or different to other components of the test; whether it is capable of taking account of demand side factors; and, whether a competition authority has a margin of manoeuvreor appreciation in relation to the evaluation of “economic value” which the Tribunal should respect.

The concept of economic value is not defined. In broad terms the economic value of a good or service is what a consumer is willing to pay for it. But this cannot serve as an adequate definition in an abuse case since otherwise true value would be defined as anything that an exploitative and abusive dominant undertaking could get away with. It would equate proper value with an unfair price. This is a well-known conundrum in international competition law. The same point was made by the Court of Appeal in Attheraces (ibid) at paragraph [205]. The issue was first identified in US antitrust and arose from criticisms of the judgment of the Supreme Court in US. v Du Pont 351 US 377 (1956) when it attracted the soubriquet "the cellophane fallacy". To overcome this in United Brands in paragraph [250] the Court held that there must be a "reasonable" relationship between price and economic value.

The simple fact that a consumer will or must pay the price that a dominant undertaking demands is not therefore an indication it reflects a reasonable relationship with economic value. But a proxy might be what consumers are prepared to pay for the good or service in an effectively competitive market, hence the relationship between the two descriptions of abuse in paragraphs [249] and [250] and the fact that the economic value description is said to be an example of the broader description of an abuse in paragraph [249].

The issue is relevant because the CMA advanced an argument to the Tribunal that due to clinical guidance which prevented switching (see paragraphs [21] and [22] above) patients were in effect tied to the manufacturer’s brand and the payer (the NHS) had no option but to pay the price demanded. In such circumstances it was not possible to say that the therapeutic advantages patients derived from the drug amounted to an indication of genuine economic value.

The particular “economic value” said to be in issue is “patient benefit”, i.e. the benefit that epilepsy patients derive from their use of the capsule and its ability to keep their condition under control. The specific complaint of the CMA is that the Tribunal erred in finding that the CMA had attributed a nil value to patient benefit (see analysis in Judgment paragraph [412] set out below).

The CMA however explains that it did take account of economic value generally, and

“patient benefit” specifically. This can be seen from detailed analysis of those issues in the Decision in the context of the “Plus” component of the Cost-Plus test.

Having established that there was no additional economic value in the Pfizer-Flynn Capsule beyond Cost Plus, the CMA went on to find that Pfizer’s and Flynn’s prices were unfair in themselves, as they bore no reasonable relation to the economic value of the product. The CMA had regard, in particular, to: (i) the substantial disparity between Pfizer’s and Flynn’s prices and the economic value of their products; (ii) the fact that competitive conditions prevailing on both relevant markets demonstrated that the relevant markets did not function in a manner that was likely to produce a reasonable relation between price and economic value; (iii) the fact that Pfizer’s and Flynn’s prices had an adverse effect on the end customer (in this case the NHS in the form of CCGs) and that Pfizer and Flynn were aware of this; (iv) the age of the drug;

The CMA infers from the Tribunal’s conclusion about nil value that it has (i) misunderstood the nature of the concept of “economic value” as used in United Brands and subsequent jurisprudence of the Court; (ii) failed to appreciate that the CMA did not assume a zero or nil value for “patient benefit” since it was taken into account in the Cost-Plus analysis; and (iii), adopted an inconsistent approach because elsewhere in the Judgment (paragraph [272]) the Tribunal acknowledged that the CMA had considered “economic value” as part of the Cost-Plus analysis.

To determine this Ground of Appeal it is necessary to summarise the conclusions of the Tribunal. In paragraph [272] the Tribunal records the matters taken into account by the CMA which I have identified above.

The Tribunal noted that counsel for Pfizer and for the CMA both accepted that economic value was relevant and had to be taken into account at some stage, but neither was prepared to be more specific. The Tribunal concluded that it was “clear” that “economic value” was a “legal rather than an economic term” (Judgment paragraph [407]). The evaluation could include supply and demand side factors and was highly fact specific and “very much a matter of judgment” (ibid) and that it needed to be considered separately from other components of the test. The nub of the Tribunal’s ruling on the facts is found at paragraph [412]:

“412.

The CMA was criticised by the parties for not considering patient benefit although it did indeed describe, in broad outline in the Decision, the nature of epilepsy and phenytoin's role in its treatment. The CMA has not, however, contested the evidence of Professor Walker and has, in effect, conceded that phenytoin remains a useful and effective treatment for a significant number of patients. That being so, we find the outright rejection of any value at all to patients surprising. The CMA seems to have placed some reliance on the age of the drug, which is irrelevant in therapeutic terms. We think there is clearly some economic value to be derived from the therapeutic benefit to patients of phenytoin capsules.”

The Tribunal then addressed an argument raised by the CMA about dependency. The basic idea is that because patients are unable to switch away from the Pfizer drug, because of clinical guidance (see paragraph [21] above), they (or the state as the payer) are bound to pay whatever price is demanded. The simple fact that patients value the drug highly and must remain wedded to it is not therefore to be treated as a discrete patient benefit; it is a feature of the regulatory regime surrounding the drug, but not the drug itself. The CMA relied upon the Opinion of Advocate General Jacobs in Tournier (ibid). That Opinion arose in the context of requests for preliminary rulings in a series of cases concerning the conduct of the French copyright management society SACEM towards discothèque owners in relation to charges for music in France. The Court was requested to consider whether the royalty rates charged were discriminatory and excessive, particularly by reference to the royalties charged in other EU Member States. The Advocate General considered the relevance of the importance of music to French discothèques, i.e. the benefit to users. He said (ibid paragraph [65]) that the idea that those who need music more should pay more was “superficially attractive” but the “usefulness of such criterion breaks down” when users were “completely dependent” on the supply of the music in question and there was no other possible source of supply. The CMA concluded, by parity of reasoning, that because patients stabilised upon capsules were dependent, zero value should be ascribed to patient benefit when determining the economic value of capsules.

The Tribunal did not disagree “generally” with the Advocate General’s assessment in Tournier but observed “… the facts here are a little different” (Judgment paragraph [416]). It was true that a competitive market was difficult to apply where there was only one supplier and the buyer was medically dependent upon the supply but it did “… not follow that the value to be attributed to the demand side is zero”. On the facts it was common ground that for patients already stabilised on a particular formulation

I do not accept this Ground of Appeal. The Tribunal made findings of fact which it was entitled to make and which are not capable of being challenged. The CMA’s criticisms are in my view unwarranted. There are, though, certain aspects of the reasoning of the Tribunal with which I would respectfully disagree. I start by explaining why this Ground of Appeal cannot succeed.

The CMA objects that the Tribunal found that the CMA attributed a nil value to patient benefit as economic value whereas, in fact, it was analysed as part of the CostPlus test carried out by the CMA. I agree that demand side factors may be capable of generating economic value but on a fair reading of the Judgment the Tribunal was not saying that the CMA failed to address its mind to the issue and for that reason ignored economic value. It was saying that having addressed itself to the issue (as part of Cost-Plus) it had failed adequately to take account of evidence that there might be “some” (albeit unspecified) value to be attributed to patient benefit, and that the reasons given by the CMA for rejecting patient benefit as relevant (namely dependency) was itself an issue of fact and degree (and not principle) and did not mean that the CMA could ignore relevant evidence. The Tribunal articulated the issue in the following manner: “The question is whether the CMA was correct, on the facts of this case, to exclude from its calculation of Pfizer’s and Flynn’s economic value all factors other than those that formed part of the Cost-Plus calculation” (Judgment paragraph 411). To the extent that it is argued that the findings made by the CMA fell within its margin of appreciation and that the Tribunal should not have interfered I do not repeat the conclusions at paragraphs [128] – [134] above which apply here. At base this was a finding by the Tribunal on the evidence and that specific finding is not challenged in this appeal.

Insofar as it is argued that the Advocate General in Tournier was laying down an absolute and immutable rule that whenever there is dependency there was no residual scope for any economic value to arise, I agree with the Tribunal that this is not what

I turn to materiality. In this connection the CMA pointed out that the sheer scale of the prices charged by Pfizer and Flynn over Cost-Plus (i.e. above a 6% ROS) meant that any errors or omissions made by it were innately unlikely to exert any effect upon the ultimate conclusion, of breach. For instance even if the CMA should have used a higher figure for ROS than 6% (because as the Tribunal found the PPRS might not be the ideal benchmark for ROS) the fact that at times the percentage actually being applied by the undertakings was so vastly in excess of ROS (see the table at paragraphs [38] above) meant that even if the CMA’s calculation of ROS was out by a factor of (say) 100% or 200% that would have made no real difference to the end conclusion. The CMA says that neither Pfizer nor Flynn set out to attribute precise quantitative monetary values to their arguments about patient benefit. Even if the CMA should have attributed “some” quantitative value to patient benefit (as part of the assessment of economic value) the likelihood of that altering the conclusion based upon Cost-Plus was remote.

There is a brief analysis in the Judgment of materiality. In an ideal world it might have been fuller. The nub of the reasoning was that: the undertakings had raised a series of challenges to the methodology adopted by the CMA; the Tribunal identified a series of individual errors in that methodology; the Tribunal itself was not in a position to conduct a review of these matters; they went to relevant parts of the test for abuse; they could individually and/or cumulatively be material but the Tribunal was not in a position to form any conclusions about this. The CMA does not however argue that read fairly the Judgment discloses that the Tribunal ignored materiality altogether or applied illogical or irrational reasoning or that the reasoning was so lacking in detail as to amount to an error of law. Nor does it argue that the undertakings themselves failed altogether to adduce evidence on materiality as part of their evidential burden and that in such circumstances the Tribunal should have found that the errors were immaterial. There are no grounds of appeal which raise such matters. At base therefore, the criticism of the CMA is one of fact and, as such, outside the scope of the statutory right of appeal. Such criticisms must necessarily therefore fail.

As observed, there are certain paragraphs in the Judgment relating to economic value that I do have a concern about and in respect of which we received submissions which went beyond anything advanced to the Tribunal. Since this matter is being remitted, I set out my views on these below.

First, the Tribunal observed that this was clearly a legal test. The categorisation of this as a “legal” concept seemingly led the Tribunal to treat economic value as a discrete component of the test in law to be applied. It is “legal” in the strictly limited sense that it has been ascribed a meaning in a court judgment but, at base, it is an economic concept which describes what it is that users and customers value and will reasonably pay for and it arose in the United Brands judgment as an economic description of the abuse of unfair pricing: see the analysis at paragraph [61] above.

Second, the Tribunal did not agree with the submissions of all parties that economic value was simply a matter to be taken into account as part of other components of the test. The Tribunal held that it was not part of the “in itself” test but was part of “a more general assessment” (Judgment paragraphs [427] and [443(6)]). I agree with the parties on this. It is evident from the judgment in United Brands that the reference to “economic value” is as part of the overall descriptor of the abuse; it is not the test. The test should therefore, when properly applied, be capable of evaluating economic value. So, for instance, as the CMA argues, when evaluating patient benefit it would be possible to measure its economic value in the Plus element of Cost-Plus, or even in the fairness element. Equally, if there is evidence of the prices being charged in relevant, comparator, markets which were effectively competitive then those prices could be capable of acting as proxy evidence of the economic value of patient benefit.

In conclusion, this Ground of Appeal fails. The Tribunal made findings about the evidence which properly are categorised as findings of fact. These were to the effect that the CMA had failed to accord at least “some” weight to economic value. These criticisms would apply to the CMA analysis at whatever stage of the test economic value was taken into account, including in the Plus element of a Cost-Plus test. The CMA has advanced what seem to me to be plausible submissions that given the very high disparity existing between cost, ROS and ultimate price the possibility of any

“economic value” attributable to patient benefit exerting any effect on the outcome is remote. The Tribunal did not suggest otherwise. Whether this ultimately turns out to be so, will be for the CMA to consider on the remittal.

Flynn takes issue with two findings by the Tribunal. First, its “apparent” conclusion that comparisons put forward by Flynn with the profitability of other products in

It became apparent during oral submissions that the real concern of Flynn was that, upon remittal, the CMA would treat various ostensible findings by the Tribunal as definitive and binding and Flynn would be precluded from re-arguing these points, even if it adduced new evidence. Flynn was particularly concerned with the implications to be drawn from paragraph [345] of the Judgment:

“345.

However, suitable and meaningful comparators do not have exactly to mimic the features of phenytoin, and if there were prima facie evidence of a meaningful comparator which helped establish a benchmark price for Pfizer-Flynn Capsules in conditions of normal and sufficiently effective competition, it should have been examined carefully. On the various suggested combinations of other companies put forward by Flynn, whether or not their activities included manufacture, or their portfolios included or they were subject to a sufficient degree of competition, it is not apparent from the Decision (paras 5.164 and 5.193-4) that the CMA examined these fully and may have been too ready to dismiss them entirely (i.e. in a binary fashion) because of other factors such as Flynn’s low level of risk and high supply price. Nevertheless, with the exception of the tablets issue discussed below and subject to our findings on the CMA’s overall approach, we do not find that any of the comparators suggested to us, in themselves, presented such a clear evidential picture (given the difficulties, on the material before us, of understanding how relevant any given comparator was) that they undermined the conclusions reached by the CMA in deciding on a reasonable rate of return. We do not, however, regard the CMA’s overall approach as valid (see paragraph 310 above).”

Flynn launched a detailed attack upon the logic of the reasoning in this paragraph. By way of illustration it was argued that the Tribunal appears to suggest, on the one hand, that the CMA failed to analyse properly the comparator evidence advanced by Flynn but, on the other hand, that none of the comparators “in themselves” presented such a “clear evidential picture” that they undermined the conclusions of the CMA in relation to a reasonable rate of return. In the course of the oral hearing Ms Bacon QC, for Flynn, explained Flynn’s concern was “…precisely that upon remittal the CMA

Mr Hoskins QC, for the CMA, helpfully clarified that upon remittal the CMA would revisit the question of the reasonable rate of return and would take into account any evidence contained in submissions advanced by Flynn said to be relevant to that consideration. The CMA would reconsider the position that it had formally adopted. There would be a “new basket” of evidence for the CMA to consider and he accepted that it would be “artificial” for the CMA to take the position that certain findings by the Tribunal were “somehow binding”. The CMA would reconsider the matter with an “open mind” and would do this carefully and impartially and would then come to a

In my judgment the criticisms advanced by Flynn as to alleged omissions and/or inconsistencies in the language used in the Judgment do not ultimately matter. This is because, properly analysed, the Tribunal remitted the entire issue of abuse to the CMA and did not intend any of its findings, howsoever expressed, to be binding upon that remittal. The pragmatic and sensible position adopted by the CMA, as confirmed by counsel during the hearing, represents the correct position. Indeed this was, in effect, the stance also adopted by Mr Brealey QC, for Pfizer, who described the outcome of the appeal to the Tribunal as a “classic merits based appeal” where the Tribunal found that the CMA had not properly considered various matters and remitted the issues in their entirety to be further considered as the CMA “saw fit”.

In paragraph [358] the Tribunal observed that whilst it had upheld certain aspects of the approach adopted by the CMA nonetheless it should have placed less weight on various matters (such as the PPRS, in identifying an appropriate ROS) and it should have examined “… more closely the various comparators put forward by Flynn, amongst other factors, appropriately weighted, to establish the right benchmark price…”. In paragraphs [467] and [468] the Tribunal stated as follows:

“467.

In the present case, however, although our essential finding is that the CMA misapplied the test for unfair pricing, the correct application of that test as we have described it would involve detailed consideration of further information, some of which may need to be obtained and properly tested, and the careful assessment of what normal competitive conditions might have been. A particular example is a better understanding of the evolution of the tablet market and tablet pricing. These are not things that the Tribunal is, in practice, in this case, in a position properly to do.

468.

We therefore confirm the Decision save for that part of it that relates to abuse (and any consequential findings, including penalties). That part we set aside. Our provisional view is that we will remit the matter, insofar as it deals with abuse, to the CMA for further consideration as it sees fit. However, before making an order to that effect, we will invite written submissions from the parties on whether to remit the matter to the CMA and the scope of any such remittal.”

Subsequently, the Tribunal endorsed this provisional position in its judgment on remittal: [2018] CAT 12. It is apparent from this judgment that the entire issue of abuse was remitted. There is no suggestion that, in directing that the CMA was to reconsider the matter in accordance with the Judgment, it was directing that certain specific findings of the Tribunal were to be treated as res judicata, i.e. bound the

Given that the purpose behind the Flynn Grounds of Appeal was to ensure that there was no fetter created by findings in the Judgment, upon either the ability of Flynn to adduce new evidence or the CMA to re-investigate, there is no need for this Court to express a view upon the particular arguments advanced by Flynn. The correctness or otherwise of Flynn’s Grounds of Appeal would in any event have required this Court to construe a number of allegedly ambiguous paragraphs in the Judgment which would have been a difficult and possibly impossible exercise without the Court having engaged in a detailed analysis of the background facts, and none of the parties have made reference to that underlying evidence. The appeal proceeded as an exercise in interpretation of the Judgment only.

This Ground of Appeal therefore fails upon the basis that it is advanced upon an inaccurate interpretation of the Judgment and order for remittal.

I turn to consider the disposal of these appeals.

On the CMA’s first Ground of Appeal I find against the CMA upon its main point

On the CMA’s second Ground of Appeal (hypothetical price benchmarks) (paragraphs [118] – [125] above) I would allow the appeal: see below at paragraph [189] for the implications of this.

On the CMA’s third Ground of Appeal (comparators) (paragraphs [126] – [152] above) I reject this challenge. The finding of the Tribunal that the investigation of the CMA was insufficiently deep or intense is a finding of fact which the Tribunal was entitled to make, and against which there can be no appeal.

On the CMA’s fourth Ground of Appeal (patient benefit and economic value) (paragraphs [153] – [173] above) I reject this challenge. On a fair reading of the

On Flynn’s Ground of Appeal (scope of remittal) (paragraphs [174] – [182] above) on a true construction of the Judgment nothing therein on abuse is intended to bind the parties on the remittal. Such findings as the Tribunal made as recorded in the Judgment do not amount to res judicata. The Judgment does not fetter the right of any party to adduce new evidence or argument on the issues remitted. Equally the parties can, if they so wish, rely upon any findings in the Judgment on these matters as indicative evidence. It will then be for the CMA to consider the probative value or weight that can properly be attached to any such finding. I therefore reject the appeal.

My conclusions in relation to the CMA’s first and second Grounds of Appeal are primarily of importance upon the remittal. They do not affect the correctness of the Tribunal’s judgment on the facts. The Judgment of the Tribunal is therefore upheld, including the order made for a remittal of the issues of abuse and penalties and any other consequential matter. Where there is a difference as between the views this Court expresses and those of the Tribunal then the judgments of this Court will govern the principles to be applied on the remittal.

I agree with the judgments of Green LJ and Sir Geoffrey Vos. Although some of their reasons are differently expressed, I detect no difference of substance between them. Sir Geoffrey Vos, Chancellor of the High Court

I agree with Green LJ’s judgment. I will nonetheless explain my reasons for the essential conclusions we have both reached in my own words.

It is important to start by noting two fundamentals of the judgment of the Competition Appeals Tribunal (the “CAT”), which are not in dispute on this appeal, concerning market definition and dominant position. First, at [198], the CAT upheld the CMA’s findings on market definition at [4.188] of its decision where it said that the relevant markets in this case were: (a) the manufacture of Pfizer-manufactured phenytoin sodium capsules that are distributed in the UK (which include Parallel Imports as they are distributed in the UK), and (b) the distribution of Pfizer-manufactured phenytoin sodium capsules in the UK. That decision was not challenged before us. Secondly, at [251], the CAT concluded that the CMA was correct to find that Flynn and Pfizer had each held dominant positions in their respective relevant markets. This decision was also not challenged before us.

Article 102 of the Treaty on the Functioning of the European Union (“Article 102”) provides that:-

“Any abuse by one or more undertakings of a dominant position within the internal market or in a substantial part of it shall be prohibited as incompatible with the internal market in so far as it may affect trade between Member States.

Such abuse may, in particular, consist in:

(a)

directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions; …”.

It must also be borne in mind that what the CMA and the CAT were deciding was whether Pfizer had imposed unfair selling prices on Flynn, and whether Flynn had imposed unfair purchase prices (mostly) on the Department of Health (“DH”).

In footnote 22 of its judgment, the CAT recorded that the Competition and Markets Authority (the “CMA”) had found that the conduct in question (of Pfizer and Flynn) may have affected (and indeed did affect) trade in the buying and selling of drugs within the whole or part of the UK and was capable of affecting trade between EU Member States. That finding was not challenged.

It is first worth recording the main elements of the CAT’s legal conclusions. This short summary is far from complete, but should suffice to identify the parts of the judgment that are challenged as being wrong in law.

In relation to the excessive limb of the test in United Brands v. Commission (Case C27/76) EU:C:1978:22 (“United Brands”), the CAT stated its conclusions at [310] (Footnote: ¹) that the CMA had been wrong in law “(a) … to restrict its Excessive Limb assessment to a Cost Plus approach, and to exclude other methodologies, rather than seeking to establish a benchmark price (or range) that would have pertained in circumstances of normal and sufficiently effective competition using the evidence more widely available; (b) … to adopt a Cost Plus methodology that produced a result that would have pertained in circumstances of perfect or, more accurately (for the purpose of the present case), idealised competition, rather than the ‘real world’”. It also held that the CMA “made an error of assessment by relying only on the Cost Plus approach that it selected”.

At [339], the CAT said that it did “not think the CMA was right to place such reliance on the Pharmaceutical Price Regulation Scheme’s (the “PPRS”) 6% rate of return … as in itself confirming, far less determining, what was a reasonable rate of return for

At [357], the CAT expressed the view that, in Pfizer’s case, “the CMA’s theoretical approach may understate what the appropriate benchmark price for Pfizer would notionally have been under conditions of normal and sufficiently effective competition”, but that, without further investigation, it was not in a position to say whether that was the case.

At [358], the CAT said that, in Flynn’s case, the CMA should have placed less weight on the PPRS in identifying an appropriate Return on Sales (“ROS”) for this purpose and “should have examined more closely the various comparators put forward by Flynn, amongst other factors, appropriately weighted, to establish the right benchmark price …”.

At [360], the CAT said that “the CMA’s approach to aspects of its Cost Plus analysis for Flynn is affected by its failure to set a benchmark price or range in circumstances of normal and sufficiently effective competition”, and that “that may in turn have affected the way it examined possible comparator products and companies and its view of the PPRS”.

As regards the relationship between the two alternatives of the United Brands unfairness test, the CAT said at [367] that it could not be right for the CMA to “ignore a primafacie valid argument that a price is fair under one Alternative and proceed to find an infringement of Article 102 solely on the basis of the other Alternative establishing that prices are unfair”. The CMA did not need to succeed on both alternatives, but the CMA “must consider whether a prima facie case of fairness under one Alternative undermines the basis for the finding of unfairness under the other Alternative and produce a reasoned basis for determining that the Unfair Limb is satisfied”. That was necessary [at 368] “in order to accord with the burden of proof and respect the presumption of innocence” and to accord with “the approach in AG Wahl’s Opinion that Alternative 2 of Limb 2 functions as a “sanity check””. (Footnote: ²)

The CAT summarised the CMA’s conclusions on the first unfairness alternative at

The CAT then addressed from [370] onwards, the question of whether there were “meaningful comparators giving rise to a prima facie case of fairness to which the

At [377] onwards, the CAT dealt in some detail with tablets, referring first to the CMA’s reasons for rejecting tablets as a meaningful comparator at [5.496-5.526]. It concluded that “the CMA clearly gave some consideration to the suitability of tablets as a comparator”, but that it was not clear that “it did so in sufficient depth”. It said at [379] that “[g]iven the inherent difficulty in making assessments in this area of competition law” it was “all the more important to conduct a full and proper examination”. The CAT did not think at [389] that it was possible for it to give “full and adequate consideration to the competitive situation in relation to tablets”. It had “at best some isolated pieces of information, and certainly not enough to form a conclusion”.

At [390], the CAT said that if there were new entrants to the tablet sector, as the evidence suggested ([380] and [385]), reducing the ASP (average selling price) for tablets, that would suggest that “one of the material reasons given in the Decision by the CMA for disregarding the tablet as a meaningful comparator, namely that it was subject to the same restrictions on competition as the capsule, would be wrong”. The CAT said that that process would “also be highly germane to seeking to establish the benchmark price in conditions of sufficient competition, as well as being informative on the question of unfairness”.

The CAT determined at [391] that “the CMA should, in our view, have done a sufficient investigation into the competitive conditions surrounding the most obvious comparator product properly to inform its decision on Pfizer-Flynn Capsules”, and that it was not “an answer to state there was no obligation to conduct a full investigation”. That was “so in relation to the CMA’s discretion in relation to the price of tablets”, but “not right in terms of obtaining sufficient evidence properly to apply Article 102 to the price of Pfizer-Flynn Capsules”. Mr Mark Hoskins QC, leading counsel for the CMA, is recorded as having conceded for the CMA at [365] that “no authority should leave a relevant factor unclear”.

At [392], the CAT concluded that tablets “as a meaningful comparator should not have been wholly rejected on the grounds relied on by the CMA”. There was enough material to make the CMA “pause to consider, at the very least, whether there was a prima facie case of fairness under Alternative 2”. Having acknowledged an element of circularity, the CAT said that the CMA “must investigate possible comparator candidates to see if they are likely to be meaningful on objective, verifiable and appropriate criteria”, but the CMA had “a margin of discretion as to the possible comparators that it needs to examine”.

The CAT criticised at [393] the fact that the evidence of the CMA’s expert, Mr

The CAT said at [269-271] and [405] that “whilst the CMA considered economic value before unfairness and concluded that there were no non-cost related factors which would increase the economic value of the capsule product beyond Pfizer’s and Flynn’s Cost Plus”, it considered economic value after unfairness as “the right and logical place to do so”, because “one of the over-arching questions for a finding of abuse is whether the price complained of “bears no reasonable relation to the economic value of the product supplied””.

In considering therapeutic benefit in the context of economic value at [411] the CAT identified the question as being “whether the CMA was correct, on the facts of this case, to exclude from its calculation of Pfizer’s and Flynn’s economic value all factors other than those that formed part of the Cost Plus calculation”. The CAT said at [412] that the CMA was criticised for not considering patient benefit, even though it had described “the nature of epilepsy and phenytoin’s role in its treatment”. The CAT found “the outright rejection of any value at all to patients surprising”, and that there was “clearly some economic value to be derived from the therapeutic benefit to patients of phenytoin capsules”. It concluded at [419] that the CMA’s decision was defective in its treatment of the economic value that might be derived from patient benefit.

At [422], the CAT concluded that the CMA’s consideration of tablets as a meaningful comparator was also insufficient in relation to economic value.

In considering at [424ff] whether the CMA made an appropriate overall assessment of unfairness, the CAT agreed at [426-427] that the CMA had made the assessment, namely whether “Pfizer’s and Flynn’s prices were unfair by reference to the fact that they bore no reasonable relation to the economic value of the product and allowed the dominant undertakings to reap trading benefits that they would not have reaped if there had been normal and effective competition”. (Footnote: ³) The CAT held, however, that the CMA appeared to have made its assessments “mainly as part of its consideration of unfairness ‘in itself’ under the Unfair Limb, rather than as an over-arching assessment, as we think it should have done”. The CAT said that where the CMA had attempted a more general assessment, it was expressed “more in terms of the Price Comparison over Time, as an assessment of the price increase involved, rather than as a comparison of the current price with economic value”. As a consequence, the CAT said any overall assessment was heavily dependent either on the Price Comparison over Time or on the CMA’s calculation of economic value on the basis of what “it rated as no more than the value of Cost Plus”. For that reason, the CAT concluded that the CMA’s overall assessment of unfairness was “likely to be defective”.

At [428], the CAT did not think that the CMA could re-present its findings under the

The CAT expressed its conclusions on the approach to unfairness at [443] as follows:-

“In our assessment, to apply Article 102 through the two-limb test of United Brands, in circumstances where the only alleged infringement is one of excessive pricing and the dominance of an undertaking in a given market has been established, a competition authority should:

(1)

consider a range of possible analyses, reflecting market conditions and the extent and quality of the data that can be obtained, to establish a benchmark price, or range, that reflects the price that would pertain under conditions of normal and sufficiently effective competition. On the facts of a particular situation, there might be only one basis of analysis that was credible, but the authority is not entitled to select one basis of analysis and ignore others that are also credible. The criteria for selection and application must be objective, appropriate and verifiable. The analysis must also be done on a consistent basis;

(2)

compare that price (or range) with the price that has been charged in practice and determine whether that is excessive;

(3)

for that purpose, form an assessment, for the purpose of the Excessive Limb, of whether that differential is sufficiently significant and persistent to be excessive, as a matter of its own discretion, exercised fairly and reasonably, in the light of such factors as:

(i)

the absolute size and stability of that differential;

(ii)

the reasons for it, taking account of the fact that the conditions for excessive pricing will only usually occur where the market is one where regulation, or some similar feature, or other barriers to entry, protect it from competition, or where there is regulatory failure and the relevant regulator has not intervened;

(iii)

previous decisions finding other differentials excessive, weighted for the markets applicable in those cases;

(iv)

the wider market conditions, including the evolution of pricing over time.

(4)

where there is a conclusion that the differential is excessive, then proceed to consider whether it is unfair under the Unfair Limb;

(5)

be free to use either Alternative 1 (unfair in itself) or Alternative 2

(unfair compared to competing products) to determine unfairness but give

due consideration to any prima facie convincing argument that the pricing is actually fair under either Alternative and take that into account in reaching a decision under either Alternative 1 or 2;

(6)

if there is a finding of unfairness under the Unfair Limb, assess what is the economic value of the product, and whether the price charged in practice bears no reasonable relation to it;

(7)

give appropriate consideration to any objective justification advanced by the dominant undertaking;

(8)

make a finding of an infringement of Article 102 if all the conditions above are fulfilled; and:

(i)

the price bears no reasonable relation to the economic value;

(ii)

the dominant undertaking is reaping trading benefits that it would not reap under conditions of normal and sufficiently effective competition”.

The CAT said at [444] that it was for the CMA to “determine, when considering comparators, either for the application of Alternative 2 or for considering whether there are prima facie issues raised under Alternative 2 that need to be considered before proceeding under Alternative 1, or indeed if they are relevant to the Excessive Limb, what weight to be applied to them in the light of market conditions and their suitability, as comparators, for the product concerned. In making that determination, it must, but need only, act in a manner which is objective, appropriate and verifiable. It has a substantial margin of appreciation, but must recognise the presumption of innocence in favour of the undertaking under investigation”.

Finally, at [463-464], the CAT accepted that Pfizer and Flynn were essentially able to set and sustain high prices for phenytoin capsules and that they did not face sufficient competitive pressure, whether from within or from outside the relevant market, to constrain their behaviour, because they each held dominant positions. But the CAT found that the CMA’s conclusions on abuse of dominance were in error, even recognising that Latvian Copyright, which contained useful guidance, was decided after the CMA’s decision. The CAT held that the “CMA did not correctly apply the legal test for finding that prices were unfair as laid down in the United Brands case”.

The issues for decision in the CMA’s appeal shifted as its appeal progressed. They were never properly defined at the outset. Instead, each party has tended to formulate

The issues all relate, however, to whether Pfizer had imposed unfair selling prices on Flynn, and whether Flynn had imposed unfair purchase prices on DH. They have tended to coalesce around the requirements drawn from the CJEU’s decision in United Brands. Despite detailed submissions as to the facts and findings in the case, neither side ultimately sought to argue that United Brands was not good and applicable law. I will, therefore, repeat here the critical paragraphs from that judgment as follows:-

“248 The imposition by an undertaking in a dominant position directly or indirectly of unfair purchase or selling prices is an abuse to which exception can be taken under Article [102] of the Treaty.

Some light has obviously been thrown on these paragraphs by the opinion of Advocate General Wahl and the decision of the CJEU in Latvian Copyright. It is not, however, necessary to set out those passages.

I have mentioned the essential parts of the CAT’s judgment at some length because it became apparent in the course of argument that, in one important respect, it was not entirely consistent. The question that emerged as central to the appeal was as to the obligations of the CMA in relation to potential comparator products that might be relevant to each of (a) the economic value of phenytoin sodium capsules, and (b) whether the price charged for them was (i) excessive, and/or (ii) unfair.

The CAT’s judgment used different formulations in different places to describe the CMA’s obligations. It talked variously about a requirement on the CMA to (a) investigate, and (b) consider comparators. In this regard, I refer by way of example to the following four different formulations (with emphasis added):-

I have also not found the CMA’s grounds of appeal easy to follow. Accordingly, I have formulated my own list of the main issues as they appeared from the written and oral arguments of the parties. They are not identical to Green LJ’s analysis, but I identify below where they align with Green LJ’s approach:-

I agree with Green LJ’s judgment as to the issues raised by, and the disposal of, Flynn’s appeal (which he describes as the fifth ground of appeal) and will therefore say nothing more about it.

I do not intend to summarise the entirety of the lengthy arguments presented to us orally and in writing. It is, however, useful to identify the main points of the argument, as it ended up during the hearing.

In oral argument, Mr Hoskins explained his case somewhat differently from his skeletons. He submitted on the excessive limb (and generally), that there was no legal obligation on a claimant (whether a competition authority or a private complainant) to identify a checklist of each prima facie available and informative benchmark, methodology, or factor, and then to address each one. Instead, the relevant legal question was whether the methodologies and the evidence that a claimant had chosen to rely upon meet the standard of proof. A defendant is, of course, legally permitted to put forward evidence to the CMA and the CAT to seek to undermine the claimant’s case. Advocate General Kokott’s opinion at [75] in FEG v. Commission: (Case C105/04 P) C:2005:751, and the CJEU’s decision at [124] and [130] in Telefonica v. Commission (Case T-216/13) EU:T:2016:369 [2018] 4 C.M.L.R. 21 show that a defendant cannot put forward a prima facie fact, and require the claimant to investigate and see whether it is made out.

The CMA submitted, therefore, that the CAT was wrong in law to say that the CMA ought to have used and investigated other benchmarks applicable in circumstances of reasonable competition. The CMA can decide to ignore the evidence put forward by the defendant or can investigate it. Whether it does so or not, it must reach its decision on the normal burden of proof and with the normal presumption of innocence. The CJEU’s decision in Intelv. Commission (Case C 413/14 P) EU:C:2017:632 [2017] 5 C.M.L.R. 18 was, according to the CMA, consistent with these submissions. [310] of the CAT judgment was wrong in law.

The CMA submitted that the selection of the cost plus methodology was a reasonable exercise of judgment by the CMA, and that it should have been given a reasonable margin of manoeuvre to choose cost plus over other methods. The CAT was wrong to say that the adoption of a cost plus methodology produced a result that would have pertained in circumstances of perfect or idealised, rather than real world, competition ([310(b)], [318]-[321], and [324]), when the use of such a methodology was endorsed in United Brands, Attheraces (Footnote: ⁴) at [207]-[209], and Scandlines (Footnote: ⁵) at [102].

The CMA submitted that the CAT’s criticism of the CMA’s reliance on the PPRS’s 6% rate of return ([310(c)], [339], and [358]) was in the context of its misconception that it was necessary to seek to set a benchmark price rather than engage in a price/cost comparison. Even if there might be some utility in looking at alternative or complementary methodologies, that was not necessary in this case, where Pfizer’s and Flynn’s prices exceeded cost plus by such significant amounts.

Mr Hoskins submitted that the CAT had ignored the fact that the CMA’s cost plus analysis was corroborated by a number of factors, namely: (i) the very significant increases in Pfizer and Flynn’s prices in September 2012 without a material change in costs, risks or activities; (ii) the absence of any economic justification for Pfizer’s price increases, other than its desire to earn supra-competitive profits; (iii) Pfizer’s prices being significantly increased in the UK and nowhere else, such that the UK prices were much higher than in other Member States; (iv) if phenytoin capsules had generated a 6% ROS, they would have been Flynn’s most profitable products in 2013;and (v) the limited commercial activity undertaken by Flynn that could not justify the prices charged or the absolute returns earned by it.

As regards the “unfair limb” of the United Brands test, the CMA submitted that the CAT made legal errors at [366]-[388], where it said that an authority had to consider whether a prima facie case of fairness under the second alternative undermined the basis for the finding of unfairness under the first alternative. That was both inconsistent with the clear case law and internally incoherent. The CAT failed to deal with the argument that, if a price was fair under the second alternative, it could not be held to be unfair. On the CAT’s analysis, the alternatives under the unfair limb would cease to be alternatives, and become cumulative, which is contrary to every statement

Mr Hoskins submitted that the CAT was also wrong in law (at [379]) to say that the CMA had to make a full and proper examination, because that might lead to the conclusion that there is not unfairness when compared to competing products. That would emasculate the CAT’s own conclusion that they are indeed alternatives: “the snake eats its own tail”.

The CMA’s alternative submission was that, if the CMA were obliged to consider comparator markets put forward by the defendants, then the CMA’s decision did address all the factors put forward by Pfizer and Flynn.

Mr Hoskins accepted that his primary case was that the CAT was wrong to say that there had to be a full investigation. But even if the CMA were obliged to investigate comparators put forward by the undertakings, when the CAT came to evaluate whether the CMA had done a full investigation, they failed properly to apply the test they had established. Mr Hoskins said that was a question of law. He accepted, however, that the CMA had not challenged the factual heads upon which the CAT determined that the investigation had not been adequately conducted.

On remedies, Mr Hoskins conceded in the context of his submissions on Flynn’s cross appeal that, if the matter went back to the CMA for it to consider other comparators put forward by Flynn, the CMA would have to reconsider its position. It would be artificial for it to do so with a free hand in relation to some, but not other, of the elements of the basket. In that position, the CMA would have to consider all the factors put forward at that stage by it or by the undertakings with an open mind, carefully and impartially.

Mr Mark Brealey QC, counsel for Pfizer, submitted, in the context of economic value, that the CAT had been right to say at [391] that “in this case, the CMA should, in our view, have done a sufficient investigation into the competitive conditions surrounding the most obvious comparator product properly to inform its decision on Pfizer-Flynn Capsules”. But he said that the question of whether the CMA had a duty to investigate comparators as the CAT had said was a question of fact and balance. It was, he said, a duty to consider and to gather more information so as to make a proper decision. The CMA did not, in the context of economic value, have to go down blind alleys. It had to consider, and, if necessary, investigate. That was not over burdensome in the context of economic value.

Mr Brealey submitted that the substantive law was the same for the CMA and a private claimant, but that the procedural law was completely different. The CMA had a duty under public law to be fair, and to consider defences and relevant considerations raised by the undertakings (see Intel (Footnote: ⁶) at [138-139] and Schindler (Footnote: ⁷) at

Mr Brealey submitted that, if the authority is to discharge its legal burden, it has a positive duty to consider the prices that the purchaser is paying elsewhere. And in a case of unfair pricing, where the CMA is looking at economic value, before the undertaking is involved, there is a legal burden on the CMA to consider comparator evidence, that is here what the DH is paying elsewhere on the market. That would inform its decision on fairness.

Ms Kelyn Bacon QC, leading counsel for Flynn, submitted that in United Brands, the Commission had found excessive prices by just looking at the prices, but once the applicant had put forward a defence that those prices were, in part, loss making, the Commission could not reject that without carrying out a proper analysis of profitability along the lines suggested in [252]. United Brands was the reverse of this case. In that case, the finding of unfair prices was annulled because the Commission only looked at prices and not at profitability, when profitability was specifically raised by United Brands. In this case, it was the other way around. The CMA looked only at profitability, and Flynn submits that it wrongly failed to complement that by looking at competitive prices.

Ms Bacon submitted that the argument on benchmark prices was somewhat semantic. What United Brands makes clear is that there has to be a counterfactual of some sort, because a price has to be excessive by reference to something. That cannot be an abstract question. One always has to look at the counterfactual of what is not excessive or what is reasonable or fair. The counterfactual is contained in [249] and [250] of United Brands, in particular, in the requirement that the question is whether the undertaking obtains benefits that it would not have obtained if there had been normal and sufficiently effective competition. Depending on the nature of the product, and the nature of the market, the appropriate counterfactual could be a benchmark price or a benchmark profitability level that, in effect, sets the benchmark price, or it could be both. There just has to be some standard against which the price is measured and that standard, whatever it is, has to reflect conditions of normal and sufficiently effective competition.

Flynn submitted that there does not necessarily have to be a single benchmark. In many or even most cases, that may be unrealistic. It may be spurious precision to search for a pinpointed single benchmark or single counterfactual for the nonexcessive price. More realistically, in most cases, there are likely to be several different benchmarks or comparators that inform the benchmark price which could suggest different results.

Ms Bacon submitted that if a competition authority has one piece of evidence pointing in the direction of unfairness and another piece of evidence suggesting that the price is fair, it cannot possibly, consistently with the authorities (e.g. [91]-[97] of CISAC v. Commission (Case T-442/08) EU:T:2013:188 [2013] 5 C.M.L.R. 15), close its eyes to the latter and say that it can find against the undertaking on the basis of a rigid construction of the word “or” in [252] of United Brands. If an undertaking submits to the CMA a comparator that is prima facie “suitable and available” (see AG Wahl at [43]), in the sense that it is prima facie meaningful, then the authority should look at it under the principle of good administration and having regard to all relevant considerations. The CMA accepted before the CAT that it did have to examine a prima facie good comparator. The standard was whether prima facie the comparator provided or might provide some evidence, even if it is not conclusive evidence, of price or profitability under conditions of normal and sufficiently effective competition. Then, the CMA should carry out a proportionate and reasonable analysis of it. The concept of “consider” is not qualitatively different from “investigate”; they are on a spectrum.

The CAT’s decision is in my view somewhat repetitive. It was quite easy to lose sight of a stark reality, which was that, literally overnight, Pfizer and Flynn increased their prices for phenytoin sodium capsules by factors of between approximately 7 and 27, when they were in a dominant position in each of their markets. That did not, of course, abrogate the need for a rigorous reasoned approach to the legal and factual questions before the CAT, but it was important to keep in mind.

Neither United Brands nor AG Wahl’s opinion in Latvian Copyright should be read as deeds. The CMA has to be able to do its job depending on the economic circumstances of the case. This was a case where costs and reasonable profit margin (cost plus) could reasonably be assessed, unlike, for example, the performing rights cases. It was also a case where the alleged comparators themselves had a lengthy economic history. It would be undesirable to establish an approach or a methodology that is so complex and time-consuming that the CMA has neither the time nor the resources to deal with cases of alleged unfair pricing.

The CAT’s decision is readable and approachable. It is, however, unfortunate that it used a variety of terminologies to describe the problems that it identified in the CMA’s decision. It is important in these undoubtedly complex cases for the CAT to maintain some precision, in order to avoid uncertainty as to what exactly is being decided as a matter of law.

I have been conscious, therefore, in determining the issues before the court of four factors particularly: (i) the need to allow the CMA a margin of appreciation in deciding how it seeks to establish a case of unfair pricing, (ii) the need to avoid imposing prescriptive rules and processes on the CMA save where they are clearly dictated by law, (iii) the fact that the CMA bears the legal burden of proof throughout and (iv) that, whilst the standard of proof is the civil one, the sanctions imposed for infringements are quasi-criminal ones.

Ultimately, as my description of counsel’s arguments demonstrates, there was less between the parties on this issue than originally appeared. The CMA submitted that it was for it to choose which methodology to use. The undertakings could then seek to demonstrate the inappropriateness of that methodology, but there was no requirement to seek to identify any particular benchmark, let alone a benchmark price. Flynn submitted that it was necessary to identify one or more appropriate benchmarks as the CAT had said. Ms Bacon nonetheless accepted that, in any particular case, the CMA that could properly fix upon a single methodology to identify the appropriate counterfactual with which the actual prices could be compared.

It seems to me that, as a matter of law, the CAT was wrong to suggest at [310(a)] and [443(1)]] that the CMA was required in considering the excessive limb as a matter of law to seek “to establish a benchmark price (or range) that would have pertained in circumstances of normal and sufficiently effective competition using the evidence more widely available”. Such an approach might be appropriate in some cases, but has not been specifically endorsed by the CJEU in either United Brands or Latvian Copyright, and certainly did not automatically vitiate the CMA’s methodology in this case.

It is true, as Ms Bacon submitted, that [249] in United Brands made it clear that the overarching exercise was to ascertain whether the dominant undertaking had made use of “the opportunities arising out of its dominant position in such a way as to reap trading benefits which it would not have reaped if there had been normal and sufficiently effective competition”. It is not, however, right to suggest that the only possible comparison is between “benchmark prices” in such a competitive situation and the undertaking’s actual prices. Indeed, United Brands explained at [250] that in that case charging a price which was excessive would be an abuse if it had “no reasonable relation to the economic value of the product supplied”. At [251], the CJEU expressly acknowledged that the comparison could be between the selling price and the “cost of production, which would disclose the amount of the profit margin”. Finally, in this connection, at [252], the first question to be determined is “whether the difference between the costs actually incurred and the price actually charged is excessive”.

The CJEU in Latvian Copyright did not approve AG Wahl’s suggestion at [17] of his opinion that the first step in the analysis was to determine whether there was an excess between the price actually charged by the dominant undertaking and what he called the “benchmark price”, namely “the price which that undertaking would hypothetically have charged had there been effective competition in the market”. Indeed, the CJEU at no stage even mentioned the concept of a “benchmark price” upon which AG Wahl had placed such repeated reliance.

The CJEU does, however, seem at [37] to have approved AG Wahl’s suggestion at [35-36], in the context of the authority’s margin of manoeuvre with respect to the methodology that may be followed to determine an excessive price, that there was “no single method, test or set of criteria which is generally accepted in economic writings or across jurisdictions for that purpose”, that “[d]ifferent authorities as well as lawyers and economists [had] suggested a number of methods of analysis”, and that “each of those methods [revealed] some inherent weaknesses”.

In my judgment, the first step in the analysis for the excessive limb is likely in most cases to be for the competition authority to consider whether the costs of production or the costs actually incurred in relation to the product in question, including of course a reasonable rate of return, can be ascertained. In some cases, that simply cannot be done, and in others, it may provide an inappropriate counterfactual. But, where it can be done, there is no reason, based on the applicable authorities, why the authority should not use that methodology to ascertain an appropriate counterfactual for the excessive limb of the analysis. In other cases, it may be necessary to determine the excessive limb by other methods.

It is true that the cost plus calculation must take some account in the ‘plus’ part of the calculation of the economic value of the product, but once again, I do not think that the CMA is required to adopt any particular approach to the determination.

I agree, therefore, with the CMA that the CAT fell into legal error when it held at [310(1)] and [443(1)] that it had to establish a benchmark price or a range of prices, beyond a cost plus calculation, in order to determine whether the prices charged by Pfizer and Flynn were excessive. I, therefore, agree with Green LJ’s analysis under his section on the CMA’s Second Ground of Appeal.

I will deal with the question of comparators at the unfair stage of the analysis below.

I was struck in the course of argument by the way in which this issue seemed to turn on semantic and black-letter reasoning. As I have said, CJEU decisions and AG opinions are not to be construed as deeds. As is well known, they are generally and deliberately more nuanced and open to interpretation than decisions of domestic courts.

The CAT acknowledged at [366] that it was clear from [47] of the General Court in Scippacercoloa, (Footnote: ⁸) and [101] and [103] of the Commission Decision in Scandlines that the two tests of unfairness in [252] of United Brands are “alternatives, in the sense that an authority can, as a matter of law, establish a breach of Article 102 under either Alternative 1 or 2 and does not need to succeed under both”.

The question under this issue is whether the CAT was right when it went on at [366] and [443(5)] to say that: (i) the CMA did not have an unfettered choice between the two; (ii) a breach of Article 102 could not be established by selecting only the first alternative, when a prima facie argument had been raised that under the other alternative, the pricing was fair; and (iii) the CMA could not find that there is an infringement where one alternative demonstrates unfairness and the other does not, as a matter of logic, to accord with the burden of proof, to respect the presumption of innocence, and to accord with the sanity check mentioned by AG Wahl at [124].

In my judgment, the CAT was wrong to say that the CMA was obliged to consider the second alternative of the unfairness test having decided that it was appropriate in all the circumstances to adopt the first alternative. As has been repeatedly said, the tests are alternatives. But I do agree with the CAT when it said at [367] that the CMA could not “simply ignore a primafacie valid argument that a price is fair” whichever alternative it chose to adopt.

It, therefore, seems to me that the question of whether the choice between the two limbs of the unfairness test adumbrated in United Brands is a binary one, is an academic and irrelevant one. As will appear under issue 3 below, I take the view that the competition authority will always need, at least as part of its duty of good administration, to give some consideration to prima facie valid comparators advanced evidentially by the undertakings. That is so whether or not the CMA chooses to proceed eventually under the “unfair in itself” alternative of the unfairness test. Even in that situation, the fact that comparators are expressly mentioned under the second alternative does not absolve the CMA from giving whatever proper attention is required to comparators raised by the undertakings. In these circumstances, I am in substantive agreement with Green LJ’s conclusions on this point. I turn then to the question of what proper attention is required to comparators under the next issue.

I have already pointed out the disparate language used by the CAT to explain precisely what it said the CMA needed to do when presented with comparators by undertakings. It seems to me, however, that the issue can be quite easily resolved from first principles.

The CMA emphasised that it was to be governed by the same principles as any other claimant advancing a similar claim. It is true that, insofar as the adversarial process in the CAT is concerned, the CMA will generally be governed by the same rules as any other litigant. It is not, however, true to say that the CMA exercising its public functions under the Competition Act 1998 is in the same position as an ordinary complainant.

The approach to be followed by the CMA in dealing with an alleged infringement of the Chapter II prohibition (as defined by section 18(4) of the Competition Act 1998) and an alleged infringement of the prohibition in Article 102 is clearly provided for in the Competition Act 1998 (Competition and Markets Authority’s Rules) Order 2014 (SI 2014 No. 458) (the “CMA Rules 2014”) (see Rule 2). Rules 5 and 6, for example, set out a specific procedure for the CMA to provide undertakings with statements of objections if it proposes to make an infringement decision. Rule 6 makes provision for the undertakings’ opportunity to respond, by requiring by rule 6(2)-(4) for the CMA to give them a reasonable opportunity to inspect the documents, and to attend an oral hearing in order to make oral representations. The CMA must obviously give due consideration to the submissions so made.

Moreover, the authorities make clear that the CMA has a duty of sound administration in carrying out its functions in investigating and reaching a decision on an alleged Part II and Article 102 prohibition infringement.

In Telefonica v. Commission (Case T-216/13) EU:T:2016:369 [2018] 4 C.M.L.R. 21 at [164], the CJEU said that “the principle of sound administration requires the [competition authority] to play its part, using the means available to it, in establishing the facts and the relevant circumstances”, and to “examine carefully and impartially all the relevant aspects of the individual case”. (Footnote: ⁹) In that case, the applicant had not shown that the Commission had not sufficiently examined the evidence adduced by the parties, and the CJEU said that it could “not be required to use its powers of investigation in order to prove a matter which [was] only alleged, but which [was] not supported by any element of the evidence adduced by the parties”. (Footnote: ¹⁰) In Kingdom of the Netherlands v. ING Groep NV (Cases T-29/10 and T-33/10), the General Court said at [108] that the authority was “required, in the interests of sound administration of the rules …, to conduct a diligent and impartial examination of the evidence at its disposal”. (Footnote: ¹¹)

Moreover, the CMA has the legal burden of proof and has a margin of appreciation in deciding how to meet that burden. An undertaking can present evidence or arguments in its defence which it says undermines the CMA’s case. In Albion Water II [2008] CAT 31, the CAT said at [72] that it would “whilst still carrying out an assessment of the merits of the case, give due weight to a finding which is arrived at by an appropriate and reliable methodology, even if a dissatisfied party could suggest other ways of approaching the issue which would also have been reasonable and which might have resulted in a resolution more favourable to its case”.

In my view, these various factors provide the answer to the question of how the CMA should deal with comparators that the undertakings raise and demonstrate by evidence to be prima facie valid in the sense that they tend to show that the undertakings’ prices may indeed be fair.

It is important, in my judgment, not to confuse the CMA’s own processes with the position that pertains when any infringement decision is appealed to the CAT. With respect to the CMA’s submissions, they tended to blur the distinction and assume that all cases would be appealed to the CAT. That is not the case, and in any event, the processes adopted by the CMA in reaching its decisions must be fair and appropriate and in accordance with the principle of sound administration, whether or not any particular ultimate decision is appealed to the CAT.

I can start with the approach to a determination of the unfairness limb. Plainly, as the

CAT acknowledged in this case, the CMA is at liberty to choose to engage the first

“in itself” limb of the unfairness test. Whether or not it makes that clear in its Statement of Objections, the undertakings faced with such a Statement will be at liberty to raise arguments and evidence suggesting comparators that they contend are prima facie valid in the sense that they tend to show that the undertakings’ prices may be fair.

In such a situation, in my judgment, the CMA is obliged to evaluate the arguments and evidence advanced by undertakings under the CMA Rules 2014 fairly and impartially. It may reject comparators so advanced, but should give reasons for doing so. In my judgment, however, the obligation that the CAT imposed on the CMA at [379] and [391] was not correct in law. The CMA does not have any duty actively to investigate in every case, in the sense of obtaining evidence about, any comparators put forward by the undertakings. It may do so, of course. It may even be desirable for it to do so (or even necessary in some cases), but it has a considerable margin of manoeuvre and it may decide how it wishes to deal with the comparators put forward by an undertaking. If it rejects the comparators wrongly or without giving appropriate reasons, its infringement decision will be more vulnerable, if and when the matter comes before the CAT on appeal.

I have no doubt that the same approach holds good in relation to comparators put forward by the undertakings in relation to the excessive limb of the test and in relation to the appropriate economic value to be attributed to the product in question.

I conclude, therefore, under this issue that the CAT was wrong in law to hold that the CMA had in every case to investigate comparators raised by the undertakings, which could be said to support a prima facie case that the prices were fair.

Green LJ also concludes at [113] that (i) the CMA has no duty in every case proactively to investigate all comparators put forward by an undertaking that prima facie demonstrate that the prices charged were fair, and that (ii) the CMA does, however, have a duty fairly to evaluate any such comparators. As I have said, it may well be prudent for the CMA to make its own investigations, but it is not under a legal duty to do so. If the CMA wrongly ignores evidence of comparators, and those comparators turn out to be relevant or important, their analysis will fail at the CAT. In my judgment, the suggestion of an obligation in every case to conduct any investigation is not warranted in law.

The CMA’s treatment of comparators began at [5.476], where it said that “[h]aving reached the conclusion that each of Pfizer’s Prices and Flynn’s Prices is unfair in itself” it did not need to address the second limb to establish whether those prices were also unfair “when compared to competing products”.

At [5.478], the CMA said that “for completeness, and because the [undertakings] submitted representations to the CMA on the issue of whether their respective prices [were] unfair when compared to competing products”, the CMA would consider “whether such a comparison could be conducted”. The CMA then concluded for reasons given at [5.479 ff] that there were “no products that would provide a

The CAT’s conclusions on these paragraphs were, as I have said, at [379] that it was not clear that the CMA had considered the suitability of tablets in sufficient depth; that it had not conducted a full and proper examination, and at [390], that, if there were new entrants to the tablet sector reducing the prices, one of the CMA’s reasons for rejecting tablets as a meaningful comparator, would be wrong. At [391], the CAT again said that the CMA should have done a sufficient investigation into the competitive conditions surrounding tablets, and at [392] that tablets as a meaningful comparator should not have been wholly rejected on the grounds relied on by the CMA. There was, the CAT thought, enough material to make the CMA pause to consider whether there was a prima facie case of fairness and that the CMA should have investigated possible comparator candidates to see if they were likely to be meaningful on objective, verifiable and appropriate criteria, even bearing in mind its margin of discretion as to the possible comparators that needed to be examined.

I note that these factual findings have not in themselves been appealed to this court. Instead, the CMA has confined its challenge to the questions of (i) whether the CAT was wrong to say that it had in every case to investigate comparators (which I have dealt with above) and (ii) whether, when the CAT came to evaluate whether the CMA had done a full investigation, they had failed properly to apply the test they had established.

In my judgment, since the factual findings referred to above have not been challenged, the CMA cannot avoid the matter being sent back to it for a proper evaluation of the comparators in accordance with the principles set out at issue 3 above. It is true that the CAT, at times, used the language of “investigation” to describe what it thought the CMA should have done, and also true that I would hold it was wrong to hold that an investigation had to be undertaken in every case. But that does not, in my view, affect the CAT’s substantive factual findings as to the CMA’s treatment of comparators. The CAT held that its treatment was inadequate for a series of factual reasons which have not and could not be challenged. In those circumstances, I take the view that the issue of the comparators raised by the undertakings must be reconsidered by the CMA. Mr Hoskins accepted, as I have said, that if that was our decision, the CMA would need to look afresh at the issue of comparators generally and would wish to do so. I agree that it would be artificial for the CMA to reconsider comparators (in relation to each limb of the test) without a free hand to look, as Mr Hoskins put it, at each of the “elements of the basket”. He accepted, and again I agree, that the CMA ought to consider all the factors put forward by the undertakings when the case goes back to it “with an open mind, carefully and impartially”.

I do not, therefore, think that the CAT was wrong in law to conclude that the CMA’s treatment of comparators was not sufficient. It was entitled to reach the factual conclusions it did. Those conclusions were not vitiated by the CAT’s legal error in thinking wrongly that the CMA had been obliged to undertake a full investigation into prima facie valid comparators put forward by the undertakings.

I, therefore, agree with Green LJ’s treatment of this issue under what he identifies as the third ground of appeal.

I agree with Green LJ’s treatment of the CMA’s Fourth Ground of Appeal. The question of patient benefit will need to be revisited when the matter is reconsidered by the CMA. The CMA cannot succeed on this issue which raises a factual criticism of the CAT’s judgment.

As Green LJ has explained under the Fourth Ground of Appeal, the CAT was wrong to require a free-standing assessment of economic value in addition to the assessments of excessiveness and unfairness.

In these circumstances, the disposition of this appeal will be as proposed by Green LJ.