Competition Damages: No “Err on the Side of Under-Compensation” When Using the Broad Axe (and Appellate Correction Where It Is Applied)

1. Introduction

Granville Technology Group Ltd & Ors v LG Display Co Ltd & Anor ([2026] EWCA Civ 409) concerns the quantification of damages in a “follow on” competition claim arising from the EU Commission’s finding of a worldwide LCD panel cartel (5 October 2001 to 1 February 2006) in Commission Decision of 8 December 2010 relating to a proceeding under Article 101 Treaty on the Functioning of the European Union and Article 53 of the Agreement on the European Economic Area (COMP/39.309-LCD - Liquid Crystal Displays).

The claimants (Granville, VMT and OTC—English/Jersey companies, all long since in liquidation) bought monitors, notebooks and LCD panels (and later some TVs) whose prices they alleged were inflated by the cartel. The principal defendant participants at trial were the LG entities.

The appeal focused on three quantification issues:

  1. whether the trial judge wrongly applied a principle that, where precision is impossible, the court should “err on the side of under-compensation”;
  2. whether the judge made “clear and obvious” errors in assessing cartel overcharge percentages; and
  3. whether the judge applied the correct causation test for downstream pass-on (mitigation), said to require a “direct and proximate causative link”.

2. Summary of the Judgment

The Court of Appeal allowed the appeal to a limited extent.

  • It held that the trial judge did, in the key steps of his reasoning on both overcharge and downstream pass-on, apply (or at least state he was applying) an impermissible “cautionary” approach of limiting awards to avoid over-compensation—i.e. erring on the side of under-compensation.
  • It otherwise upheld the judge’s evaluative findings on the competing econometric models for overcharge, and on the factual existence of some downstream pass-on.

Because the error affected the figures, the Court of Appeal substituted its own “broad axe” assessments:

  • Overcharge: increased from 8%/4%/14% to 10% (monitors), 6% (notebooks), 16% (TVs).
  • Downstream pass-on: reduced from 65% to 60%.

The court noted consequential recalculation issues for the claimants’ loss of profits head (tied to the extent of pass-on), which the parties said would be straightforward to compute.

3. Analysis

3.1 Precedents Cited

(a) The compensatory principle and the measure of damages

The starting point remained orthodox compensatory damages: to restore the claimant to the position it would have been in absent the wrong, as stated in Livingstone v. Rawyards Coal Co (1880) 5 App Cas 25. In competition cases, the direct loss is typically the overcharge, less any proven downstream pass-on.

The trial judge, and the Court of Appeal, located this framework in Sainsbury's Supermarkets Ltd v Mastercard Inc and others [2020] UKSC 24; [2020] 4 All ER 807 (and also, for pass-on specifically, in Sainsbury's Supermarkets Ltd v Visa Europe Services LLC [2020] UKSC 24, [2020] Bus LR 1196, as cited in the judgment).

(b) “Broad axe / broad brush” and rejection of “under-compensation” as a rule

The case squarely reaffirms (and operationalises) the Court of Appeal’s earlier rejection, in Britned Developments Ltd v ABB AB [2019] EWCA Civ 1840, [2020] 4 CMLR 7, of any supposed “rule” that where estimation is necessary the court should “err on the side of under-compensation” or “give the defendant the benefit of the doubt”.

The trial judge had relied on Asda Stores Ltd v. Mastercard Inc [2017] EWHC 93 (Comm) (Popplewell J at [307]) for such an approach, but Britned Developments Ltd v ABB AB [2019] EWCA Civ 1840, [2020] 4 CMLR 7 had already disapproved it. The Court of Appeal in the present case treated the judgment text as determinative: where the judge said he was limiting uplift “to be cautious to avoid over compensation”, the court took that at face value.

The earlier first-instance discussion of estimation and econometrics in Britned Development Limited v. ABB AB [2018] EWHC 2616 (Ch) also remained relevant background, particularly on what makes a regression “well specified”.

(c) Appellate restraint on factual/evaluative findings (especially expert evidence)

The Court of Appeal emphasised the narrow grounds on which it can interfere with a trial judge’s findings—especially where conclusions rest on dense expert evidence: Fage UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5, [2014] FSR 29, Henderson v Foxworth Investments Ltd [2014] UKSC 41, [2014] 1 WLR 2600, Volpi v Volpi [2022] EWCA Civ 464, [2022] 4 WLR 48, and Thomson v Christie Manson & Woods Ltd [2005] EWCA Civ 555, [2005] PNLR 38. This framing mattered: most technical attacks on model choice and robustness failed because the judge’s preference for Parker’s approach was within his evaluative range.

(d) Pass-on causation: “direct and proximate causative link” as factual causation

The court treated the “direct and proximate causative link” requirement (for downstream pass-on as mitigation) as part of factual causation, not a free-standing policy bar. The key authorities were:

  • Royal Mail Group Ltd v. Daf Trucks Ltd and others [2023] CAT 6; [2023] 5 C.M.L.R. 6 (CAT articulation of the test; non-exhaustive “indicators”).
  • Royal Mail Group Ltd v DAF Trucks Ltd [2024] EWCA Civ 181 (Court of Appeal endorsement of CAT’s approach; distinction between factual and legal causation; “indicators” not determinative).
  • Stellantis Auto SAS v Autoliv AB [2025] CAT 9 (confirmation that Royal Mail factors are not exhaustive or determinative).

Importantly, the court rejected an over-formalistic reading of the “four factors” in Royal Mail: they are evidential signposts, not legal prerequisites. Even if absent, pass-on can still be proven by other evidence—but the evidential burden becomes correspondingly heavier.

3.2 Legal Reasoning

(a) Overcharge: accepting the model choice, correcting the quantification step where the wrong “caution” was applied

The trial judge accepted Mr Parker’s multiple regression framework (including the semiconductor proxy for demand), rejecting Dr Van der Laan’s trend extrapolation and most challenges to model specification (including price lags within the cartel period). The Court of Appeal held these were fact-laden evaluative choices within the permissible range and not “plainly wrong”.

The successful attack was narrower: in adjusting Parker’s 5.7% (monitors) and 2.4% (notebooks) to account for (i) endogeneity risk and (ii) post-cartel price persistence, the judge said he made a “modest” uplift but limited it to avoid over-compensation. That invoked the disapproved approach. The Court of Appeal therefore substituted its own broad-axe uplift to 10% and 6%, and fixed TVs at 16%.

(b) Downstream pass-on: correct test applied in substance, but wrong “caution” used to select the percentage

On the legal test, the claimants argued the judge failed to apply the “direct and proximate” link. The Court of Appeal held that, although the judge did not use that phrase, his inquiry was in substance the right one: whether the claimants’ prices were higher than they would otherwise have been because of the overcharge. Given the close relationship between LCD panel costs and the claimant products (computers/notebooks incorporating those panels), the case lay nearer the “trucks” end of the spectrum than the “postage stamps” end.

Factually, the judge was entitled to reject both extremes (0% and 100%) and to infer some pass-on from market features, variable cost significance, competitive dynamics, and the limited surviving documents. However, when selecting 65% within a 50–100% range, the judge again expressly invoked erring on the side of under-compensation. The Court of Appeal therefore substituted 60% as its own broad-axe assessment.

(c) Remedial technique: appellate substitution rather than remission

Having found the legal error affected figures, the court considered remittal disproportionate and, given the trial judge’s footnote stance that no change was needed, impractical. It therefore performed the “least bad” course: substituting broad-axe figures itself, openly acknowledging the arbitrariness that can attend selection within a range where the evidence cannot discriminate more finely.

3.3 Impact

(a) Practical rule for competition damages: “broad axe” is neutral, not defendant-favouring

The judgment reinforces that, in cartel damages, judicial estimation is permissible and often unavoidable, but it must be directionally neutral: the court aims for the right figure “without erring in either direction”. Trial judges should avoid language suggesting an asymmetrical “benefit of doubt” for defendants.

(b) Draft-judgment footnotes cannot cure operative reasoning if the judgment text says otherwise

A notable procedural lesson is the court’s insistence that “the text of the judgment must speak for itself”. Even where a judge says (in a footnote added after circulation in draft) that they did not “consciously” apply the wrong approach, express reasoning in the body indicating that they did will prevail—at least where it plausibly affected numerical outcomes.

(c) Pass-on remains fact-sensitive; “direct and proximate” is not a straitjacket

For downstream pass-on, the judgment consolidates the Royal Mail line: “direct and proximate” is essentially about evidential proof of factual causation, not an additional policy hurdle. Economic evidence and market features may suffice where direct tracing evidence is unavailable, particularly in long-running historic cartel claims involving dissolved or liquidated entities.

(d) Litigation strategy implications

  • Claimants should not assume that absence of documents prevents a pass-on finding; courts may accept expert-driven inference, especially where variable costs are material.
  • Defendants should be prepared to prove pass-on with a “direct and proximate” causal narrative tied to the claimant’s pricing mechanics, not merely to generic budgeting processes.
  • Both sides should expect appellate restraint on model-choice issues, but a willingness to intervene where a judge articulates a legally incorrect “tilt” in the broad-axe exercise.

4. Complex Concepts Simplified

“Follow on” claim
A damages claim that relies on an infringement already found by a competition authority (here, the EU Commission Decision), rather than re-proving the cartel.
Overcharge
The percentage by which the claimant paid more than it would have paid in the “but for” world without the cartel.
Upstream pass on vs downstream pass on

Upstream pass on: suppliers above the claimant in the supply chain pass the overcharge down to the claimant (here it was common ground that upstream suppliers passed on 100%).

Downstream pass on: the claimant raises its own prices (or otherwise recovers the cost) so the claimant avoids some loss; it is treated as mitigation and must be proven by the defendant.

“Broad axe / broad brush”
A metaphor for judicial estimation where precision is impossible. This case emphasises it is not a licence to bias outcomes toward under-compensation; it is a tool to reach the best estimate the evidence permits.
Multiple regression analysis (econometrics)
A statistical method to estimate the effect of the cartel on prices while accounting for other drivers (costs, supply, demand), often by using “proxy” variables and a “dummy” cartel-period variable.
Endogeneity
A risk that an explanatory variable (e.g. semiconductor prices as a demand proxy) is itself influenced by the dependent variable (LCD prices), potentially biasing results.
Price persistence and “lagged” variables
The idea that today’s prices may be influenced by prior months’ prices/costs/demand (a “lag”). The court distinguished (i) alleged within-cartel lags (rejected on the evidence) from (ii) post-cartel unwinding (accepted to a limited extent).
R-squared and Ramsey RESET
Statistical diagnostics sometimes used to assess model fit/specification. The judge preferred the view that economic sensibility and correct specification of key drivers mattered more than any single diagnostic.
“Direct and proximate causative link” (pass-on)
A fact-based requirement that the claimant’s downstream prices were higher because of the overcharge—not merely that all costs feed into general budgeting. The Royal Mail “four factors” are evidential indicators, not mandatory legal checkboxes.

5. Conclusion

Granville Technology Group Ltd & Ors v LG Display Co Ltd & Anor confirms that, in competition damages where quantification is necessarily approximate, courts must apply the “broad axe” neutrally: there is no principle of erring on the side of under-compensation to avoid over-compensation. Where a judgment’s operative reasoning shows that tilt, appellate intervention may follow, including substitution of revised percentages for overcharge and downstream pass-on.

The decision also consolidates the modern approach to downstream pass-on: “direct and proximate” causation is an evidential/factual inquiry, not a policy veto, and can be proven by a combination of expert evidence, market structure, and reasonable inference—especially in historic cartel litigation where granular pricing records and witnesses no longer exist.