Granville Technology Group v LG Display: Court of Appeal recalibrates damages in LCD cartel follow-on claim

The Court of Appeal has allowed, to a limited extent, an appeal by three companies in liquidation against the damages assessed in their follow-on competition claim arising from the LCD panel price-fixing cartel. In Granville Technology Group Limited (in Liquidation) & Ors v LG Display Co Limited & Anor [2026] EWCA Civ 409, Males LJ (with whom Snowden LJ and Cockerill LJ agreed) revised upwards both the overcharge percentages and the downstream pass-on rate, substituting his own figures after concluding that the trial judge had expressly applied an erroneous legal principle when arriving at his.

The claim arose from the EU Commission's 2010 decision finding a worldwide price-fixing cartel in thin film transistor LCD panels between October 2001 and February 2006. The claimants — UK-based computer manufacturers and retailers — had purchased monitors, notebooks and TVs incorporating the cartelised panels. His Honour Judge Pelling KC assessed damages at £4.4 million in total, awarding £942,438 after crediting settlements with other defendants.

The central legal issue was the judge's application of a principle drawn from Asda Stores Ltd v Mastercard Inc [2017] EWHC 93 (Comm): that where a court must use a broad brush, it should err on the side of under-compensation. That principle was disapproved by the Court of Appeal in Britned Developments Ltd v ABB AB [2019] EWCA Civ 1840, which held that the court's aim must always be to award the right amount of compensation, without erring in either direction.

The judge acknowledged the error in a footnote added after the draft judgement was circulated, stating he had not consciously applied the cautionary approach. The Court of Appeal declined to accept that assurance as sufficient. Males LJ observed that the judge's paragraph 143 — setting out his final overcharge conclusion — expressly stated he was limiting an upward adjustment from the defendants' expert's figures in order to avoid over-compensation. The text of the judgement, not the footnote, had to govern.

The judge had preferred the multiple regression model of the defendants' economist, Mr Parker, over the claimants' trend-line extrapolation approach, a finding unchallenged on appeal. Parker's model produced overcharges of 5.7% for monitors and 2.4% for notebooks. The judge adjusted these modestly upward to 8% and 4% respectively (with 14% for TVs) to account for endogeneity risk in the use of semiconductor prices as a demand proxy, and for post-cartel price persistence. The Court of Appeal held that in arriving at those figures the judge had, in terms, constrained the upward adjustment so as to avoid over-compensation. Exercising its own broad-axe assessment, the court substituted figures of 10% for monitors, 6% for notebooks and 16% for TVs.

The claimants' specific criticisms of the judge's rejection of lagged-price variables and his assessment of the model's statistical robustness were each dismissed. The judge had been entitled to prefer Mr Parker's evidence that econometric soundness — getting the right variables — mattered more than statistical test scores such as R-squared and Ramsey RESET.

The judge found that 65% of the overcharge had been passed on by the claimants to their own customers, thereby reducing their recoverable loss. The Court of Appeal confirmed that the judge had correctly identified the applicable legal test — requiring a direct and proximate causative link between the overcharge and the claimant's downstream prices — and had applied it in substance, notwithstanding his failure to use that precise formulation.

The conclusion that some pass-on had occurred was upheld as properly open on the evidence. However, paragraph 246 of the judgement again stated explicitly that the judge was erring on the side of under-compensation in conditions of uncertainty. For the same reason as on the overcharge issue, the Court of Appeal treated the body of the judgement as authoritative and substituted a pass-on rate of 60%.

The judgement confirms that where a trial judge states in terms that the disapproved Asda principle has governed a conclusion, an appellate court will not be deflected by a subsequent footnote disavowing that approach. It also underlines that in the absence of sufficient guidance from the original judgement, the appellate court will wield its own broad axe — here producing figures that, frankly, involve a degree of arbitrariness that Males LJ freely acknowledged.

Copyright & permissions