CA-2025-000517 - [2025] EWCA Civ 1061

Analysis

Overall conclusion

Whilst the points raised by the CR are undoubtedly of interest and importance they do not reflect the way in which the issue was actually addressed by the CAT. On the facts: (i) the CAT was entitled to reject the submission that it draw an adverse inference; (ii) the adverse inference proposed by the CR was not an appropriate one to be drawn; and (iii), in any event, the CAT considered that whatever the rights and wrongs of BT’s approach, it, the CAT, could cope and arrive at conclusions on the available evidence that were justifiable. The CR’s submission raises two broad questions. First, when is it proper to draw an adverse inference? Secondly, if it is proper what sorts of inferences can be drawn? There are a series of interrelated reasons why this argument has no real prospect of success on appeal.

Burden and standard of proof in the drawing of adverse inferences / the Sainsbury’s principle

The applicant starts by arguing that the CAT failed to apply the principle that where a claimant raises a prima facie case that a defendant’s prices are excessive relative to a costs benchmark, the defendant is under a heavy evidential burden to provide evidence as to its costs because that evidence will be exclusively in its hands and it must therefore produce that evidence to forestall the drawing of adverse inferences. The applicant relies upon the Supreme Court in Sainsbury’s. There the merchants sought damages for unlawful merchant service charges ("MSC"). The defendant credit card schemes argued that the merchants had mitigated their loss by passing on the MSC to their customers. The Court observed:

The legal burden lies on the operators of the schemes to establish that the merchants have recovered the costs incurred in the MSC. But once the defendants have raised the issue of mitigation, in the form of pass-on, there is a heavy evidential burden on the merchants to provide evidence as to how they have dealt with the recovery of their costs in their business. Most of the relevant information about what a merchant actually has done to cover its costs, including the cost of the MSC, will be exclusively in the hands of the merchant itself. The merchant must therefore produce that evidence in order to forestall adverse inferences being taken against it by the court which seeks to apply the compensatory principle.”

It is argued that the same principle applies beyond issues of alleged mitigation in the form of pass-on citing Royal Mail Group Ltd v DAF Trucks Ltd [2024] EWCA Civ 181 at paragraph [174] (“Royal Mail”) which concerned the failure by a defendant to provide disclosure and factual evidence as to the operation of a cartel. The same principle has also been applied by the CAT in relation to a failure to adduce relevant evidence: PSA Automobiles SA v Autoliv AB [2025] CAT 9 at paragraphs [144] – [146] where the CAT drew adverse inferences from the defendant’s failure to call relevant witnesses to provide evidence of the scope of cartel activity. In CMA v Flynn Pharma Limited and Others [2020] EWCA Civ [339] at paragraphs [114] – [116] the Court of Appeal observed that in an excessive pricing case a defendant bears “an important evidential burden” and the extent of a competition authority’s duty to evaluate evidence adduced by the defendant “… will be affected by the nature, extent and quality of the evidence adduced by the defendant undertaking which has an evidential burden”. A similar series of observations is found in Cinven (ibid)at paragraph [118].

The law relating to the drawing of adverse inference was recently considered by the Court of Appeal in Phones 4U Limited v EE Limited and others [2025] EWCA Civ 869. The High Court rejected a claim for damages based upon an allegation of collusion. During the trial the Judge declined to draw an adverse inference from a failure on the part of one defendant, Telefonica: “… to adopt appropriate document preservation measures in 2015, when the allegations were first made, or for some time thereafter”, in "arrogant disregard", as the Judge put it, for the seriousness of the allegations made.

The Court of Appeal endorsed the principles set out below and upheld the Judge who, whilst obviously critical, did not, on the basis of the evidence as a whole, consider it necessary to draw any form of adverse inference. In Efobi v Royal Mail Group Ltd [2021] UKSC 33 ("Efobi") at paragraph [41] Lord Leggatt emphasised that the position was not legal or technical but was a matter of rationality and common sense. In Volpi v Volpi [2022] EWCA Civ 464 Lewison LJ observed:

Tribunals are free to draw, or to decline to draw, inferences from the facts of the case before them using their common sense. Whether any positive significance should be attached to the fact that a person has not given evidence, or to the lack of contemporaneous documentation, depends entirely on the context and particular circumstances…”

On an appeal, where it is argued that an adverse inference should, or should not, have been drawn an appellant will need to demonstrate that the court or tribunal below acted outside the realm of its legitimate discretion, or to put it another way, irrationally. Important points were made by Morritt LJ in Malhotra v Dhawan [1997] EWCA 1096; [1997] 8 Med LR 319 at page [322] about the need to balance the need for an adverse inference with other evidence and factual findings. The Judge said:

"First, if it is found that the destruction of the evidence was carried out deliberately so to as hinder the proof of the plaintiff's claim, then such finding will obviously reflect on the credibility of the destroyer. In such circumstances it would enable the Court to disregard the evidence of the destroyer in the application of the principle... Second, if the Court has difficulty in deciding which party's evidence to accept, then it would be legitimate to resolve that doubt by the application of the presumption. But, thirdly, if the judge forms a clear view, having borne in mind all the difficulties which may arise from the unavailability of material documents, as to which side is telling the truth, I do not accept that the application of the presumption can require the judge to accept evidence he does not believe or to reject evidence he finds to be truthful."

What does this mean in the present case? Here, the CAT was undoubtedly critical of the failure of BT to generate the data in issue. However, that is far from being the end of the story because the CAT was content for the issue, at least as a matter of case management, to be addressed through expert evidence and it rejected the submission that an adverse inference be drawn, and did not make a finding that BT had strategically gamed the system.

The point of departure is that in law a breach can prima facie be established where the decision maker establishes that the price exceeds Cost-Plus: see case law summarised at Cinven (ibid) paragraphs [24] and [77]. If there is no such justification for an excess, then Cost-Plus can mark the divide between lawful and unlawful. How does evidence of justification for the differential come into play? If in a regulatory investigation there was a feature of a market which was transparently capable of justifying an excess of price over Cost-Plus then the decision maker could not ignore it. However, in most cases evidence relating to justification lies in the hands of the defendant undertaking and if it is not raised by that undertaking the decision maker cannot be criticised for not taking it into account. In Cinven the issue was considered in relation to alleged patient benefits said to flow from the emergence of competitive drug offerings: see (ibid) paragraphs [118] and [182] citing from Case C-307/18 Generics (UK) Ltd and Others v CMA ECLI:EU:C:2020:52 at paragraphs [165] and [166]. There it was made clear that because the evidence will usually be in the possession of the defendant there is an evidential burden upon it to produce that evidence and explain how it correlates with the differential. Where that occurs, it is then for the decision maker to address that evidence fully and fairly (Footnote: ¹).

The argument that an adverse inference should have been drawn in this case rests upon the false premise that the criticisms made of the expert evidence (due to the failure of BT to generate the relevant data) were such that the CAT effectively dismissed that evidence altogether; but that, by contrast, the criticisms made by the CAT of the expert evidence of the CR were such that it was still entitled to material weight. This is not an accurate description of the approach of the CAT. The position it adopted towards the evidence was far more nuanced. When the CAT came to assess the Common Costs Starting Point, it rejected the opinion of the CR’s expert that the common costs were very low (and therefore increased the differential between Cost-Plus and price):

“We should add that in this context, there is no Common Costs Starting Point figure from Mr Duckworth which can feed into the analysis here. This is because the RFS contains no information on the split between incremental and common costs, and neither Mr Duckworth nor Mr Parker provided any estimation of BT Consumer common costs except to argue that they were very low. For the reasons given above, we do not accept that they were very low.”

With this in mind the CAT rejected the argument that it should draw an adverse inference from the failure on BT to generate the relevant data:

As set out above, and save on the question of direct costs, the methods employed by Mr Duckworth on the one hand and Dr Jenkins on the other, to reach a competitive benchmark are very different. Each model has been formulated and explained in considerable detail.

That said, in oral closing arguments, Mr Beard KC somewhat downplayed the significance of Dr Jenkins’ SAC Combi test by characterising it as simply a sense-check to show that the RFS method was clearly wrong.

While we must consider these competing methods in detail, we do accept the proposition that if the RFS methodology is fundamentally or inherently defective, then the CR’s case on Limb 1 is not going to be “saved”, as it were, because the SAC Combi methodology is also defective. This observation recognises therefore that the CR has the burden of proof.

On the other hand, we also recognise that in relation to a number of arguments deployed by BT as to the inadequacy of the RFS methodology, it is important to ascertain whether such arguments are essentially speculative in circumstances where BT, had it so wished, could have adduced evidence of its own on the point but did not do so. We refer to this in paragraphs 676, 693, 703-704 and 715 below. Indeed, the CR invites us to draw appropriate adverse inferences due to the failure of BT to adduce relevant evidence - see paragraph 371 of its Closing. We do not think it necessary or appropriate to draw adverse inferences as such from the absence of such evidence. The simple point is that BT could have adduced relevant evidence and it did not, so such evidence is simply not there.

For the above reasons, it makes sense, first to consider the RFS methodology and the criticisms made of it and then, in the light of that, the SAC Combi methodology.”

The CAT did identify weaknesses in the evidence of both parties but nonetheless felt able to rely upon parts of that evidence to a modest degree. In the lengthy analysis in the Judgment up to paragraph [897], the CAT critiqued the methodologies adopted by the competing experts. In paragraph [897] the CAT then observed that it did not reject either expert’s methodologies entirely and nor did it accept the submission of both parties that no reliance could be placed upon the expert methodology of the other party. In paragraph [898] the CAT made clear that it had to “strike a balance” based upon the “weight” that it considered “should be given” to the expert evidence on the central issues:

Instead, and perhaps unsurprisingly, we considered that, because each side’s methodology contained a number of problems, it is necessary to strike a balance in terms of the outcome that reflects the weight which we consider should be given to each methodology in respect of:

the starting point for common costs, being the proportion of BT Consumer’s total indirect costs that should be represented by common costs (“the Common Costs Starting Point”);

the proportion of such common costs that should be borne by SFV Services (“the SFV Services Common Costs Contribution”); and

the application of the 13.5% margin on the total costs of sale (“Application of Margin”).”

Even if the evidence of Dr Jenkins was, relative to data that could have been generated by BT, sub-optimal, this did not prevent the CAT coming to a judgment it felt was justified and it made clear that it had taken into account BT’s failure: see for example paragraphs [816] and [825]. At the end of the day the CAT applied critical analysis to the expert evidence, concluded that it was deficient in parts, and then accorded it such residual weight as it considered fit to arrive at a final conclusion. This approach to the evidence, including those parts it considered deficient, fell four-square within its ordinary discretion as the fact finder and as the evaluator of evidence. There was no breach of any principle relating to the drawing of adverse inferences.

Case management

Secondly, during case management no application was made to the CAT for an order that BT generate, in advance of the trial, the data relating to common costs which, it was clear, BT would have been able to perform at proportionate cost and effort. BT's intention to address the costs in issue through expert evidence, and not by the generation of new data, was evident from an early stage. The CR, for instance, could have applied for an order that the data be generated by BT and/or that that any expert instructed by BT should certify what assistance had been sought from BT and whether it was provided, etc. There are many different applications that might have been advanced. It is plain that the pre-trial preparation exercise was substantially an expert led exercise.

This does not without more imply a failure of case management on the part of the CAT, or that a lacuna in the evidence necessarily arose, because the default position was that the issue was covered by expert opinion evidence and then subjected, in the ordinary way, to the detailed, intensive and iterative pre-trial process whereby the experts met and sought to identify and thrash out areas of agreement and disagreement. That expert opinion was then, again in the ordinary way, subjected to the adversarial trial process including the now fairly common hot tubbing process for evaluating expert evidence which enabled the Bench to probe the evidence as well as permitting of cross-examination. Further, the CAT proactively required the experts to re-run their calculations upon the basis of alternatives put forward by the CAT itself: see Judgment paragraph [37] cited below.

In assessing the position of the CAT (as opposed to that of BT), in circumstances where no one seemingly considered it necessary that BT should, pre-trial, generate the data in question, it is more difficult for the CR to allege a failure on the part of the CAT. It is no answer to say that having heard the evidence during the trial, the CAT was critical of the evidence (in the event of both sides) because that assumes that what might have been a perfectly rational pre-trial case management decision, becomes a thoroughly bad one only at or after the trial, once the default evidence has been tested and deficiencies emerge. That is not how case management works. The applicant has not alleged or shown that the position adopted by the CAT pre-trial, whereby it was content that BT’s indirect costs be addressed by expert opinion alone, was irrational as outside the necessarily wide margin of discretion that must be accorded the CAT in relation to case management decisions.

Irrationality / inadequacy of reasoning

The next complaint under this heading is about an alleged paucity of reasoning. Instead of rejecting Dr Jenkins’ evidence, the CAT relied upon it in determining the “Common Costs Starting Point” (Judgment paragraph [905]):

“In our judgment, the appropriate Common Costs Starting Point is £250m. This recognises the problems with Dr Jenkins’ approach which mean that in our view, even her low scenario common costs figure is too high.”

However, it is complained that the CAT cited no other evidence to explain how it reached this conclusion. It adopted Dr Jenkins’ low scenario estimate (of £250m) subject to a reduction of £12m (4.6%). The CAT does not explain why the problems with Dr Jenkins’ approach were cured by such a modest 4.6% reduction to her low scenario estimate.

I disagree. This is not an objection that the CAT failed altogether to adjust downwards the conclusions of Dr Jenkins but, rather, that it did not do so enough. It is an objection as to the weight the CAT attached to evidence. The CAT made its overarching approach to the expert opinions clear in paragraphs [37] and [38] where it acknowledged the differences and difficulties inherent in the analysis of the experts but also the contribution they made to the CAT’s own evaluative exercise:

We did face some difficulties with the economic expert evidence. This is because, in the case of many areas of dispute, the methodologies employed by each of the opposing experts were quite different. Perhaps because of this, at various points each party submitted that the other party’s line of analysis should be rejected entirely. As will be seen, for the most part, we did not rule out a line of reasoning as simply illegitimate per se but rather took account of possible defects so that our approach then blended the outcomes of the various models so to reach the appropriate outcome. Sometimes, for example with Dr Jenkins, she would produce a set of alternative outcomes or “sensitivities” which represented what, for her, was a more “conservative” outcome than her primary outcome. This, at least, gave us some alternative views to consider. In addition, on occasion, each side’s experts would “run” the other side’s expert’s analysis to see if it altered their outcome or not. Finally, and at our request, both sides produced alternative analyses using different figures for the purpose of the Limb 1 exercise. Here, that produced by the CR contained somewhat more variables than that produced by BT. All of this was useful, but it did not remove entirely the need, as we saw it, to consider outcomes that were not directly the product of either side’s approach. This is perhaps inevitable in cases of this kind where sets of figures, percentages and comparisons are so important and which are therefore highly fact and expert opinion sensitive.

As for the experts themselves, as with BT’s lay witnesses, this is not a case where one side’s expert can be criticised as generally lacking credibility or relevant experience. All of the experts, but particularly Mr Parker and Mr Duckworth for the CR, and Dr Jenkins for BT, are highly experienced, both in their fields and in giving evidence to tribunals such as this. On occasion, all of them could be said to have strayed into arguing their case and losing objectivity, or speculating. In general however, the experts did their best to assist us. The economic experts had to deal with our numerous questions for each of the hot tub sessions (which we notified in advance) as well as cross examination, and the ground they covered was very extensive, as will be made clear below. Their professional and constructive approach to these exchanges during the hearing, even on matters where they disagreed with one other, greatly assisted us in reaching our conclusions on the core economic issues. We should add that we found the JES a particularly useful document.”

In relation to common costs the CAT explained its reasoning. It recorded that the baseline scenario used by Dr Jenkins as a starting point for BT Consumer common costs was £390m out of a total indirect cost figure of £983m (i.e. the former represented approximately 40% of the latter). The CAT observed that on Dr Jenkins’ low scenario such costs became £262m or 27% of the total indirect costs. There was no common costs starting point figure in the CR’s expert report nor even an estimation of BT Consumer common costs. The CR had limited his argument to contending that the common costs were “very low”, a proposition the CAT rejected. The judgment of the CAT was that the appropriate common costs starting point was £250 million which was below the low scenario postulated by BT of £262m and this conclusion took into account its analysis of Dr Jenkins’ opinion and the fact that her opinion was adversely affected by BT’s approach to evidence generation.

It is true that there is no more detailed explanation as to how the CAT arrived at the £250m figure. I do not accept, however, the criticism of the CAT which flows from this. There are two points to make:

First, this raises a question as to how much detail the CAT can be expected to set out in a judgment to support a conclusion of this sort. This Court must be careful not to demand a spurious level of accuracy or detail. Here, the CAT used the lowest, baseline, figure in Dr Jenkins model from which to form its own (downward) view. Her low-end conclusion was no more than a marker along the route to the wielding of the broad axe and the making of an educated but necessarily imprecise estimate of where the balance lay between competing and incomplete evidence. The level of reasoning set out was adequate. It would place an unreasonable burden on the CAT to require more. The CAT is entitled to say, as it did albeit using different language, that having set out its analysis of such evidence as was available, it had arrived at the point where, to move to a conclusion, it had to rely upon its own experience and judgment, this being as much a skilled art as a science.

Secondly, this challenge raises the question as to what is needed to persuade the CAT to draw an adverse inference. In all such cases it will be important to identify the particular fact or matter the CAT is invited to accept, by way of adverse inference, and how this is connected to the omission alleged. There may be no difficulty where, for example, a defendant found by a prior decision to have engaged in price fixing declines to provide disclosure or tender witnesses about the effect of the price fixing cartel in circumstances where the defendant contends that the cartel has a zero impact upon the marketplace. Such a failure to tender relevant evidence might justify the CAT concluding that there was some material impact upon the market. It might, however, be difficult to go further and draw the adverse inference that the impact on the market was of the particular magnitude contended for by the CR. On the other hand, where the CR has, using it best endeavours, generated an estimation of the likely quantum of effect then, absent contradictory evidence from the defendant, the CAT might, still exercising its own judgment, justifiably incline towards a figure in the region of the CR’s. In the present case the CR adopted the extreme position namely that the CAT should attribute no value to common costs or that they should bear an immaterial or greatly reduced weighting (see paragraph [18] above). The CR did not however provide his own reasoned calculation identifying where the CAT should draw the line. If the CAT, succumbing to the CR’s invitation, had treated the allocation as zero or close to zero the outcome risked being punitive and unfair given that the CAT plainly thought that there was material value (albeit below that advanced by BT) to be allocated. In circumstances where the CAT is seeking to be as fair and accurate as it can be, such a result would not be satisfactory.

In oral argument, Ms Kreisberger KC, for the CR, laid emphasis upon the fact that whilst BT failed to generate relevant data covering the alleged infringement period it had generated such data historically and the CR had relied upon this data but, inconsistently, the CAT then paid no real regard to the CR’s analysis of this material. I do not accept this argument. The CR relied upon regulatory financial statements (“RFS”) provided by BT to Ofcom for a period up to 2009. BT had been required to provide this data as part of its regulatory obligations as the possessor of statutory SMP (significant market power) which was the equivalence of dominance: see Judgment paragraph [145], [146], [513]. The CR argues that absent BT generated data, far greater weight should have been attached to the historical RFS data from which the CR extrapolated that the value of BT’s common costs to be attributed to SFV services was so small as to be irrelevant. In fact, the CAT examined the strengths and the weaknesses of RFS data in detail between paragraphs [663] – [780]. It concluded that the evidence did not establish that the common costs of BT Consumer were so insignificant as to be irrelevant. Equally, in paragraph [779] it rejected as “quite impossible” the argument of BT that the RFS model relied upon by the CR was “not fit for purpose”. The truth lay in between. The reliance that could be placed upon the data had to be “tempered”, in particular by the fact that: the outcomes advanced by the CR were built upon a limited data set from 2009; the up-rating by use of a CPI index, likely, under-estimated cost increases; and, the RFS methodology had no flexibility so far as recovery of common costs was concerned. The argument before this Court that the CAT erred in not attaching greater weight to this evidence is untenable. The detailed and thorough analysis of the CAT is plainly justifiable and, in fact, was not really challenged save in broad overview terms by the CR. But more pertinently this objection is one about findings of fact over which the Court has no jurisdiction.

The failure to cross examine

The CAT observed that the CR did not cross-examine Dr Jenkins on her scoring exercise (Judgment paragraph [793]). The applicant says that this is indicates that the CAT justified its reliance, at least in part, upon the flawed evidence of Dr Jenkins because parts had not been subjected to cross examination. This is said to be an error for three reasons. First, the matter was thoroughly traversed in the hot-tub where Dr Jenkins conceded under questioning from Mr Ridyard, an economist panel member, that the exercise was based on her own judgment. Secondly, the PTR Order expressly provided for concurrent expert evidence and stated that the parties “will not be required to put their full case to the experts in cross-examination” and the CAT’s reliance upon the CR not cross-examining Dr Jenkins as a basis for giving weight to her evidence was therefore procedurally unfair. Thirdly, Dr Jenkins’ approach was thoroughly critiqued by the CR’s expert, Mr Duckworth, in his Second Expert Report and in the Joint Expert Statement.

I do not accept the point. It is unclear whether the CAT was merely recording the fact that there had been no cross-examination on the issue or whether it was treating the failure to cross-examine as supportive of the conclusion that Dr Jenkins’ opinion was entitled to some weight. In context, it is difficult to see that cross examination on the point in issue would have led the CAT to change its position on the evidence which had been tested during the hot-tub by Mr Ridyard, the experienced economist member of the panel, and subjected to detailed criticism by the CR’s expert. The applicants do not say that they were prohibited from cross examining the experts. At the end of the day this is a small point, and I cannot see that it could be material to the outcome.

The need for deprecation.

Finally, the applicant argues that if this Court condones the approach of the CAT, it will set a bad precedent; there is a need for deprecation. I disagree. The discretion of a court to draw an adverse inference is not intended to be punitive. An important consideration is whether the failure obstructs the ability of the CAT to perform its role. Here, there is no indication that the CAT considered BT’s failure left it unable to cope. Of course, the CAT needs to be astute to the risk of litigants gaming the evidence preparation process. Here the CAT was aware of the risk that BT had declined, strategically, to generate potentially awkward data but instead instructed an expert and then deprived her of necessary help and assistance. That much was put forcibly by the CR in argument and addressed by the CAT. However, the CAT factored its criticisms of BT into its conclusions about the reliability of the conclusion of Dr Jenkins (see for instance Judgment paragraphs [816] and [885]). If the CAT had considered that it was materially hampered it could have gone further and drawn additional inferences erring even further on the side of the CR’s case, where there was a benefit of the doubt to be given. Alternatively, it could have accepted any estimates (if provided) provided by the CR upon the basis that they had not been responded to with directly relevant evidence from BT. In the event the CAT did not consider that it needed to go to these extremes.