What courts really examine in cartel and abuse of dominance cases
In competition litigation, the distance between allegation and proof is rarely covered by legal labels alone. Calling a practice exclusionary, exploitative or collusive does not prove the case. What matters is something more demanding: explaining how the market worked, which competitive mechanism was disrupted, what effects followed, and whether the evidence supports that conclusion.
This is where economic analysis becomes decisive.
Not as a display of technical complexity. Not as an exercise in equations for their own sake. But as the discipline that organises facts, assumptions and market data into a coherent account of reality. Courts do not reward the most sophisticated model. They reward the model that answers the right question and explains the market as it actually operated.
Competition enforcement has moved decisively from form to effects. This is visible across the full spectrum of cases: cartels, abuse of dominance, loyalty rebates, predatory pricing, margin squeeze and damages actions. The central question is no longer simply what form the conduct took. It is whether, how and to what extent the conduct harmed the competitive process.
The tools reflect this shift.
In dominance cases, the As-Efficient-Competitor test asks whether an equally efficient rival could have survived under the same pricing conditions. In cartel damages actions, econometric models seek to estimate the price that would have prevailed absent the infringement. In digital markets, economic modelling must go further still: it must capture how ranking, selection, display and algorithmic mechanisms actually shape competition.
Yet courts are not impressed by mathematics as such. They examine four things above all. First, whether the model answers the legal and economic question before the court. Second, whether its assumptions reflect commercial reality. Third, whether the data are suitable and reliable. Fourth, whether the conclusions remain robust when reasonable assumptions change.
Technical impression counts for little. Evidential credibility counts for everything.
The case law makes this point clearly.
Intel transformed the treatment of loyalty rebates. Conduct that was once close to being condemned by category now requires a closer assessment of market coverage, duration, rebate conditions and the ability of an as-efficient competitor to respond. Even familiar theories of harm must pass through the filter of economic evidence.
BritNed v ABB, the first cartel damages judgment of the English High Court, turned heavily on econometrics. The decisive issue was not whether the model appeared sophisticated. It was whether it could isolate the effect of the infringement from all the other factors that affected price. The message was clear: regression analysis persuades only when the variables, sample and counterfactual are demonstrably connected to the facts of the market.
Cabo v MGA pushed the point further. The court was not prepared to accept theoretical projections of how the market might have developed. It asked whether the damages model and sales forecasts corresponded to what could realistically have happened in the toy market. The lesson is uncomfortable but essential: numerical precision without commercial plausibility creates the appearance of evidence, not evidence itself.
The Amazon Buy Box proceedings add the digital dimension. In platform markets, economic analysis cannot stop at prices and market shares. It must engage with the actual mechanics of the market: algorithms, rankings, selection criteria and user behaviour. The Competition Appeal Tribunal scrutinised whether the proposed methodology aligned with the proper counterfactual and the actual operation of the Buy Box. In digital markets, simplifying assumptions rarely survive judicial scrutiny. The model must approximate the machinery of the market itself.
The practical conclusion is straightforward. An economic model persuades when it is intelligible, empirically grounded and anchored in the real operation of the market. It fails when it projects precision without foundation, relies on abstractions or substitutes complexity for facts.
In modern competition litigation, cases are increasingly won or lost on the quality of the economic analysis. The economist’s task is not to impress the court. It is to explain how the market actually works and to convert an allegation into evidence.
Seen in this light, economic analysis is not a supporting detail of the case.
It is the case.
This article was prepared by Marina Kouloumbri during her summer 2026 internship at Trojan Economics.
