Follow-on cartel damages: binding force of Commission recitals (including post-leniency “play-along”), adverse inferences, and counterfactual causation without naming the missing bidder

1) Introduction

Nexans France SAS & Anor v London Array Ltd & Ors (Re Permission to Appeal) [2026] EWCA Civ 887 is a Court of Appeal (Civil Division) decision refusing permission to appeal from the Competition Appeal Tribunal (“CAT”). The underlying CAT judgment held that members of the London Array offshore windfarm joint venture (the respondents/claimants) suffered an unlawful overcharge on the tender for “Export Cables” supplied by a group company of the applicants (the applicants/defendants), in circumstances where the European Commission had found a worldwide high-voltage power cable cartel (Case AT.39610 – Power Cables (2 April 2014), “the Decision”).

The application concerned four proposed grounds, all directed at the CAT’s findings on collusion, inference, counterfactual causation and econometric proof. Lord Justice Green refused permission on all grounds, emphasising (i) the binding effect of the Commission Decision in follow-on litigation, (ii) the legitimacy of adverse inferences where cartelists withhold explanatory evidence, (iii) the evaluative nature of counterfactual causation in cartel tender cases, and (iv) appellate restraint in relation to expert/economic evidence.

Key parties (role-based)

• Applicants/defendants: Nexans France SAS and Nexans SA (together “Nexans”).
• Respondents/claimants: London Array project companies (members of the joint venture constructing the windfarm).
• Relevant non-party actors: other cartel participants referenced in the Decision (including ABB and Prysmian), and the bidding entity Nexans Norway (a group company which submitted the bid).

Key issues on the permission application

1. Whether the CAT impermissibly found collusion involving a leniency applicant after its “end date” in Article 1 of the Decision, and whether such findings were unpleaded/unfair/irrational.
2. Whether the CAT irrationally/unfairly found that another cartelist submitted a cover bid and that Nexans knew this.
3. Whether the CAT erred in law on causation by finding it “entirely likely” that some additional competition would have emerged in the counterfactual, without identifying which bidder would have acted differently.
4. Whether the CAT erred by using “group-to-group” econometric comparisons to support causation/overcharge, allegedly contrary to BritNed.

2) Summary of the Judgment

Permission to appeal was refused. The Court held, in substance, that the proposed grounds either: (a) misunderstood the binding content of the Commission Decision (including recitals and annexes treated as integral), (b) amounted to non-appealable challenges to factual evaluation and expert evidence, (c) failed to show procedural unfairness on the pleaded case, or (d) sought to impose an unrealistic and legally unnecessary requirement of “pinpointing” a single counterfactual bidder in a tender market distorted by a long-running cartel.

3) Analysis

3.1 Precedents cited and their role

(a) Pleading and procedural fairness

• Al-Medenni v Mars UK Ltd [2005] EWCA Civ 1041, Popely v Ayton [2022] EWHC 3217 (Ch) (“Popely”), Vogon International Ltd v Serious Fraud Office [2004] EWCA Civ 104 (“Vogon”): cited by Nexans to argue that a party must know the case it has to meet and that adverse findings should not be made on unpleaded points or without an opportunity to respond. The Court distinguished these authorities as fact-specific: in Vogon and Popely the impugned findings were unnecessary and outside the pleaded case; here, the claimants pleaded reliance on the specific Commission recital (paragraph [444]) and the CAT made findings against Nexans within that framework.

(b) Adverse inferences and evidential burdens

• Le Patourel v BT plc [2025] EWCA Civ 1061: referenced to show that adverse inferences are not automatic; the question is whether it is appropriate on the facts.
• Prest v Petrodel Resources Ltd [2013] UKSC 34: cited as authority for drawing adverse inferences where a party fails to produce evidence within its control.
• Royal Mail Group Ltd v DAF Trucks Ltd [2024] EWCA Civ 181: relied upon both for the permissibility of adverse inferences and for the broader competition damages context, including the recognition that cartel participation is ordinarily expected to confer some benefit, especially where defendants do not explain how the cartel operated.

(c) Appellate restraint and expert evidence in cartel damages

• BritNed Development Limited v ABB AB & Another [2019] Bus LR 718: invoked by Nexans to argue that averaging across projects is inappropriate in bespoke tender markets. The Court relied on BritNed (including the appellate caution it records at paragraph [123]) to reinforce that challenges to econometric methodology and sample selection are typically factual/evaluative matters for the trial tribunal, and that BritNed did not lay down a rigid rule against group-to-group comparisons.

3.2 Legal reasoning

(A) Binding effect of the Commission Decision: operative part and supporting reasoning (Ground 1a/1c)

A central theme is the Court’s treatment of what in the Commission Decision binds the CAT in a follow-on claim. Nexans argued that because Article 1 recorded ABB’s infringement ending on 17 October 2008 (its leniency marker), the CAT could not rely on evidence of ABB communications in early November 2008 to find tender rigging affecting the first-round bids.

The Court rejected this as a category error. The claim was against Nexans, not ABB, and the Commission’s reasoning explicitly treated the November 2008 exchange as part of the “cartel activities” supporting the infringement findings against Nexans:

• Decision paragraph [444] recorded that Nexans contacted ABB to discuss price levels for London Array.
• Annex I (integral to the Decision) contained a corresponding entry (Item 422) describing an exchange of price information for “London Array”.
• Decision paragraphs [493(b)] and [493(d)] characterised such exchanges as cartel activity including floor prices/coordination of price levels.

In other words, the CAT did not “invent” liability: it applied binding findings against Nexans, and it was entitled to treat the exchange as evidence of cartelised conditions affecting Nexans’ bidding conduct (including through its group bidding vehicle).

(B) “No inconsistency” where a leniency applicant continues limited contact to protect inspections (Ground 1a)

Nexans sought to frame an internal inconsistency: how could the Decision recite conduct by ABB after 17 October 2008 while Article 1 ended ABB’s infringement on that date? The Court agreed with the CAT that there was no true inconsistency. The Decision itself contained a caveat (via a footnote) that ABB’s post-marker conduct could occur insofar as “reasonably necessary to preserve the integrity of the inspections”. The Court accepted the CAT’s explanation that a leniency applicant may be required to “play along” temporarily so as not to compromise the Commission’s dawn raids.

Crucially, that explanation did not exculpate Nexans: it meant the November 2008 exchange could remain part of the factual matrix establishing Nexans’ cartelised behaviour, even if ABB did not incur regulatory liability for that limited post-marker engagement.

(C) Group liability and “taint” of the bidding entity (Nexans Norway)

Nexans also faced the CAT’s finding that the bidding conduct of the group entity that submitted the tender (Nexans Norway) was “tainted” by the cartel conduct of the addressee entities. The Court pointed to Decision paragraphs [716]–[717], where the Commission rejected the group’s “subsidiary independence” narrative, citing reporting lines and implementation structure. This substantially undercut any appealable contention that the CAT lacked a basis to link group cartel conduct to the bid actually submitted.

(D) Pleading sufficiency where the Commission recital is redacted publicly (Ground 1b)

Nexans argued unfairness because the claimants did not plead ABB’s specific involvement. The Court held the pleading was adequate: the claim form expressly relied on Decision paragraph [444]. The public version redacted the counterparty’s identity, but Nexans would have had access to unredacted material and knew who was contacted. The pleaded reliance on [444] therefore put the issue “squarely on the table”.

As to alleged unfairness to ABB, the Court stressed that the CAT did not determine ABB’s liability; it made findings against Nexans by applying the Decision.

(E) Cover bids and adverse inferences in the face of limited explanatory evidence (Ground 2)

Nexans challenged the CAT’s inference that another cartelist’s first-round bid was a non-genuine cover bid, known within Nexans, and affecting the tender dynamics. The Court treated this as a quintessential evidential evaluation: the CAT construed opaque communications in the context of the Decision (including the Commission’s characterisation of those communications as cartel activity) and considered the significant price disparity in bids.

An important practical point is the Court’s endorsement of the CAT’s approach to evidential silence. Where defendants do not adduce evidence explaining how the cartel operated internally or how bid teams were insulated from cartel knowledge, the tribunal may (where justified) draw adverse inferences. The Court supported this with Prest and Royal Mail Group Ltd v DAF Trucks Ltd.

(F) Counterfactual causation: no need to identify the “missing” bidder by name (Ground 3)

Nexans argued the CAT erred by not finding, on the balance of probabilities, which specific manufacturer would have bid differently in a cartel-free world. The Court rejected this as imposing an unrealistic and legally unnecessary level of precision on a hypothetical inquiry.

The Court accepted that the legal test was balance of probabilities, but held the CAT was entitled to find actionable harm where it was “entirely likely” that some manufacturer would have provided incremental competitive pressure absent a decade-long cartel—especially given:

• the Decision’s findings about credible non-European competitors (e.g., Japanese/Korean manufacturers) being excluded/managed by the cartel;
• the specific cartelised features of the tender (including the Nexans-ABB exchange and the cover-bid dynamics);
• econometric evidence consistent with cartel-era margins being higher than post-cartel margins; and
• the defendants’ limited evidence explaining cartel operations and bid formation.

The Court also observed the “heads I win, tails you lose” nature of Nexans’ complaint: had the CAT selected a particular counterfactual bidder on thin evidence, the complaint would likely have been that the finding was speculative.

(G) Econometric proof and appellate restraint (Ground 4)

Nexans attacked the CAT’s use of “group-to-group” (During Cartel Projects vs After Cartel Projects) comparisons and sample choices (including post-2018 projects and turnkey projects). The Court refused permission because the CAT:

• did not treat group-to-group averages as conclusive on their own, acknowledging project-specific limits;
• used both experts’ approaches as relevant evidence, not a binary choice;
• found cartel “infection” of London Array’s tender so comparators had probative value; and
• made detailed, nuanced, fact-sensitive choices on methodology and sample—matters outside the Court of Appeal’s permission jurisdiction absent an error of law.

The Court also clarified that BritNed does not establish a rule of law that bespoke tender markets forbid averaging. Rather, it illustrates fact-specific preference and, importantly, reiterates that appellate courts should be slow to interfere with evaluative findings and expert assessment.

3.3 Impact

(1) Follow-on claims can rely on Commission recitals and annexes to establish actionable cartel “infection” of a specific tender

The decision underscores that, in follow-on litigation, binding content is not confined to a headline “end date” in Article 1 read in isolation. Recitals and annexes that the Decision treats as integral may bind and can ground specific factual findings against an addressee—even where the conduct also involves a leniency applicant post-marker.

(2) Leniency “play-along” does not immunise the remaining cartelists from civil consequences

By accepting the CAT’s explanation that a leniency applicant may continue limited cartel-facing conduct to protect inspections, the Court removes a potential defence tactic: pointing to the leniency applicant’s formal end date to deny the cartelised nature of post-marker interactions. The civil relevance is the addressee defendant’s conduct and its effects, not whether the cooperating undertaking incurred further regulatory liability.

(3) Strengthened legitimacy of adverse inferences where cartel defendants do not explain their own internal processes

The Court’s reasoning aligns with a broader trend in cartel damages litigation: where defendants hold (and decline to provide) the best evidence about cartel mechanics and internal dissemination/controls, tribunals may draw adverse inferences, within principled limits, when evaluating causation and “taint”.

(4) Counterfactual causation in cartel tenders: actionable harm can be proved without naming the counterfactual competitor

The refusal of permission confirms that a claimant need not prove a single identified “but for” bidder would have submitted a lower bid, provided the tribunal can properly conclude on the balance of probabilities that the cartel materially reduced competitive constraint and that, absent the cartel, competitive conditions would likely have produced a lower price. This is practically significant for markets where cartel conduct suppresses entry and distorts who participates.

(5) Limited scope for appeals dressed as “errors of law” on econometrics and sample selection

The judgment reinforces that methodological disputes about comparators, sample windows, and model architecture are usually matters of fact/evaluation. Parties seeking appellate intervention will need to show a genuine legal misdirection, not merely a different preferred expert approach.

4) Complex concepts simplified

• Follow-on claim: a civil damages claim that relies on a competition authority’s infringement decision as a binding starting point (as opposed to proving the infringement from scratch).
• Infringement “by object”: conduct (such as bid-rigging) treated as inherently anti-competitive, so the authority need not prove actual market effects to establish liability—though civil damages still require proof of loss/causation.
• Leniency marker / immunity: a cartel participant may obtain reduced fines or immunity by reporting and cooperating. Operationally, the authority may require confidentiality and may permit limited continued contact to avoid tipping off other cartelists before dawn raids.
• Cover bid / non-genuine bid: a deliberately uncompetitive bid submitted to create an appearance of competition while ensuring a chosen firm wins.
• Counterfactual: the “but for the cartel” world used to ask what price would probably have been paid absent the infringement.
• Balance of probabilities: more likely than not. The Court accepted this standard while recognising that counterfactual reasoning may still be expressed in probabilistic, non-specific terms where appropriate.
• Adverse inference: where a party fails to produce evidence it would be expected to have, the tribunal may infer that the evidence would not have assisted that party.
• During-after econometrics; G2G vs I2G: comparing profit margins during the cartel vs after it ended. “Group-to-group” compares average margins across project groups; “individual-to-group” compares the specific project margin to a reference group.

5) Conclusion

[2026] EWCA Civ 887 is an important procedural and methodological signpost in cartel damages litigation. It confirms that tribunals may (i) treat Commission recitals and integral annexes as binding in follow-on claims, (ii) reconcile apparent “end date” tensions in leniency contexts through the accepted reality of post-marker “play-along” conduct, (iii) draw justified adverse inferences where cartel defendants decline to explain cartel operation and bid formation, and (iv) find counterfactual causation without identifying a single “missing” bidder, where the tribunal can properly conclude that competitive conditions would likely have been materially different absent a long-running cartel.

For litigants, the message is clear: attempts to recast factual and expert-evaluative disputes as errors of law will face a high barrier on permission, particularly where the tribunal has anchored its findings in the Commission Decision and has given careful, nuanced reasons.