On 10 March 2020 the Court of Appeal upheld the Competition Appeal Tribunal’s (CAT) quashing of the Competition and Markets Authority’s (CMA) decision that Pfizer and Flynn Pharma (Flynn) had abused their dominant positions in the market by pricing their epilepsy drug unfairly. Among other aspects, the Court of Appeal broadly upheld the CAT’s findings that the CMA
- misapplied the relevant legal test for unfair pricing; and
- failed adequately to consider alternative, countervailing evidence adduced by Pfizer and Flynn.
Our full review of the first instance judgment, handed down by the CAT on 7 June 2018, can be found here. In short, in December 2016, the CMA found that both Pfizer and Flynn had abused their dominant positions in the narrowly defined markets for manufacture and distribution, respectively, of Pfizer-manufactured phenytoin sodium capsules (a treatment for epilepsy). The CMA found that Pfizer and Flynn did so by increasing their prices for the capsules to an excessive level – a rare, pure “excessive pricing” decision unconnected to any other abusive practice.
The CMA’s decision essentially relied on a comparison between cost and price to determine whether the prices were excessive. The decision was based on an abstract analysis, which compared the price with a theoretical benchmark of “cost plus 6%.” Using this approach, the CMA concluded that the new price was first excessive and then unfair “in itself” because it exceeded the cost-plus benchmark (on the basis that economic value of the product was not higher than cost plus 6%). However, the CAT criticised the CMA for failing to evaluate properly the economic value of phenytoin sodium capsules and wrongly relying on only one part of the United Brands test (“price unfair in itself”), without properly assessing the prices of meaningful comparators. The most obvious comparators in this case were phenytoin sodium tablets (Pfizer and Flynn sold capsules), which were sold to the NHS at considerably higher prices (25% higher than capsules) – a price set by the Department of Health.
So the key to this assessment was the proper application of the seminal test espoused by the Court of Justice in United Brands v Commission, in particular the extent to which a competition authority should have regard to prices of comparable products after having decided that the price of the product in question is unfair “in itself”. The CAT held that that CMA gave too little consideration of the evidence adduced by Pfizer and Flynn regarding the prices of comparator drugs and quashed the CMA’s finding of abuse, but remitted the case back the CMA for further consideration in light of the judgment.
The CMA appealed the CAT’s judgment to the Court of Appeal on a series of grounds, the most significant being whether the CAT was wrong to suggest that a competition authority could not determine unfair prices by reference to the product in question alone, i.e. without regard to evidence of the prices of comparator products. The Court of Appeal endorsed the CAT on this issue. The key judgments of substance were given by Lord Justices Green and Vos, although the Court was unanimous in its decision.
The judgment also addresses important questions about competition authorities’ duty of good administration, the presumption of innocence of those investigated and the level of latitude authorities ought to be afforded when investigating and making decisions based on evidence of abusive conduct. We highlight some of the key findings of the Court of Appeal on these issues below.
“Excessive pricing”: the test in United Brands
Before assessing the Court of Appeal’s findings, it is worth reviewing the leading case on how excessive pricing can amount to an abuse of dominant position. The Court of Justice held in United Brands that a price can be unlawfully excessive where it bears “no reasonable relation to the economic value of the product supplied”, with that question being determined (among other possible methods) according to the following test:
- whether the difference between the costs incurred and the price charged is excessive (the “excessiveness limb”); and, if so,
- whether the price was unfair either (a) in itself or (b) when compared to the price of competing products (the “unfairness limb”).
Much of the debate before the Court of Appeal concerned the unfairness limb, particularly the need to consider comparator prices if a competition authority has determined the price is unfair “in itself”.
Unfairness: a disjunctive test?
Unfairness was the central issue the Court of Appeal was asked to resolve. The CMA argued that it was sufficient for it to show that the price of Pfizer/Flynn’s phenytoin capsules was unfair “in itself” and that (upon doing so) it did not need to consider the price of comparator products. In the CMA’s view, the United Brands test gave the authority the option of proving that a price was either unfair in itself or when compared to other products. Thus, on its case, it did not need to consider the alternative evidence adduced by Pfizer and Flynn supporting the proposition that the price was in fact fair when measured against the price of appropriate comparator drugs, the phenytoin sodium tablets.
Both Green and Vos LJJ were clear in stating that the assessment of unfairness (either “in itself” or by comparison) advanced by the Court of Justice in United Brands was not to be followed slavishly and was neither a purely disjunctive (i.e. “one or the other”) nor a combinatorial test. The overarching issue was whether the price charged bears no relation to the economic value of the product – and there are many ways to measure economic value. Notably, Green LJ found that
“In analysing whether the end price is unfair a competition authority may look at a range of relevant factors including, but not limited to, evidence and data relating to the defendant undertaking itself and/or evidence of comparables drawn from competing products and/or any other relevant comparable, or all of these. There is no fixed list of categories of evidence relevant to unfairness.”
The Court of Appeal therefore held that a competition authority is free to choose whatever means it sees fit to show unfair pricing, whether “in itself”, by comparison or otherwise. However, and this is one of the key findings by the Court of Appeal, the CMA must have due regard for alternative, exculpatory evidence put forward by the firm investigated (as Pfizer and Flynn did in this case – see below). Thus, with respect to the proper assessment of unfairness, the CAT’s finding that the “in itself” and “by comparison” options under the second limb of United Brands were not strict alternatives was upheld.
Having established that the unfairness test in not disjunctive, to what extent did the CMA need to consider the evidence put forward by Pfizer and Flynn showing the price was fair by comparison to other drugs? Here too the Court of Appeal upheld the CAT’s criticisms of the CMA. It found that the CMA had failed in its duty of good administration by failing adequately to consider countervailing evidence. It therefore failed to discharge its burden of proof, notwithstanding that it had a margin for manoeuvre in determining whether there had been an infringement of competition law.
Green and Vos LJJ each held that competition authorities were free to choose how they demonstrate unfair pricing and that it would be wrong to say that authorities are obliged in all circumstances to consider both unfairness by comparison with other products and “in itself”. However, the Court of Appeal was firm in saying that a competition authority may not bury its head in the sand; it cannot simply ignore a prima facie valid argument that a price is fair. This is because competition authorities are not typical complainants in that they both investigate and reach decisions on infringements of competition law in a quasi-criminal manner. Indeed, it is part of a competition authority’s “duty of good administration” to give consideration to evidence advanced to the contrary.
Green LJ referenced Intel v Commission in this regard. In that case the Court of Justice found that that the General Court erred in failing to examine all of Intel’s arguments concerning the way the Commission had conducted an “as efficient competitor” test, in order to show whether the rebate schemes at issue were capable of having foreclosure effects on as efficient competitors. In those circumstances, the General Court was required to consider the arguments raised by Intel.
In Green LJ’s words, Intel “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.”
But how much consideration? And to what extent should competition authorities have to discover alternative evidence for themselves? Both Green and Vos LJJ held that this was matter of fact and degree, which will be dependent on the particular circumstances of the case. However, Vos LJ perhaps put it best when stating that
“[t]he 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 has a considerable margin of manoeuvre and it may decide how it wishes to deal with comparators put forward by an undertaking. If it rejects 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.”
The Court of Appeal therefore upheld the CAT’s quashing of the CMA’s decision on the basis that the CMA failed properly to apply the correct legal analysis and, in turn, adequately evaluate all the evidence before it. The case is now referred back to the CMA, which can assess Pfizer/Flynn’s pricing of phenytoin sodium capsules de novo. In light of the Court of Appeal’s judgment, that must include taking due account of Pfizer and Flynn’s countervailing evidence, notably on the price of phenytoin sodium tablets.
The “benchmark issue”: was the CAT wrong to mandate a higher benchmark for measuring excessiveness?
When considering the first limb of the United Brands test above, the CMA submitted that it was for it to choose whatever methodology it wishes – whether using hypothetical or real-world comparator prices – to demonstrate excessiveness. It would then be for the firms in question to demonstrate the inappropriateness of that methodology. The CAT on the other hand had held that some objective, hypothetical benchmark beyond the standard return on sales of 6% under the Pharmaceutical Price Regulation Scheme, should be considered by competition authorities when measuring excessiveness.
Both Green and Vos LJJ were agreed in holding that the CAT was wrong on this point and that the CMA has a “margin of manoeuvre” in deciding how to prove its case. The CMA may select from and use a variety of counterfactuals, including the “Costs Plus” method.
However, the Court of Appeal also found that the question of whether the particular benchmark employed by the CMA is a correct assessment of economic value is a question of fact and degree to be decided on the facts of the particular case. Here, the Court of Appeal upheld the CAT’s criticisms of the CMA’s failure to consider alternative means of assessing economic value beyond the Costs Plus 6% standard. It also upheld the CAT’s finding of fact that the CMA did not properly consider Pfizer/Flynn’s assessment of value according to patient benefit. Again, the CMA failed to give adequate assessment of the alternative evidence (undermining the Costs Plus 6% standard) put before it.
Firms investigated for excessive pricing practices can take some comfort from the Court of Appeal’s findings that competition authorities must fairly evaluate evidence and arguments put forward to show that a price is fair, even if those arguments are based on a different set of economic criteria to those being pursued by the authority. Whilst authorities have the benefit of a margin for manoeuvre in demonstrating excessive pricing, they must be flexible within that margin and give all plausible evidence and arguments due consideration. They must do so, not only as a matter of properly applied legal test, but also as a matter of good administration and respect for the presumption of innocence. A failure to do so will render decisions vulnerable to challenge.
The judgment also conveys a broader lesson to authorities: while you may have discretion in how you build your case, you must consider properly evidence of a different type that casts doubt on the existence of an infringement. In that respect, the Court of Appeal shows us that the Court of Justice’s Intel judgment is of broad application and not merely limited to the as efficient competitor cost/price test that was at issue in that case.
The opinions expressed are personal. The authors represented Pfizer in the investigation.
 6% representing the standard return on sales (“ROS”) under the Pharmaceutical Price Regulation Scheme of which both Pfizer and Flynn are members (but which did not apply to the phenytoin sodium capsules sold by Flynn).
 Case 27/76, United Brands Company and United Brands Continentaal BV v Commission, EU:C:1978:22.
 Flynn also appealed the CAT’s judgment on one ground, but which is not discussed in this alert.
 See paragraph 97(vi).
 See further the General Court’s judgment in Case T‑216/13, Telefónica SA v Commission, EU:T:2016:369. At paragraph 164, the Court explains that the duty requires competition authorities to establish the facts and relevant circumstances, and “examine carefully and impartially the relevant aspects of the case”.
 Case C‑413/14 P, Intel Corporation Inc. v Commission, EU:C:2017:632. See Green LJ at paragraphs 88 and 89.
 See paragraph 270.
 Or a return on capital employed (“ROC”) of 21%, among other possible benchmark ranges.
 Green LJ expressed some doubt about whether the CAT had indeed compelled the CMA to use a benchmark test, but found in favour of the CMA to the extent the CAT had made such a finding.
 See in particular Green LJ at paragraphs 165-167.
 Note that “mere speculation” will not suffice – Telefonica v Commission (see above).