The ECJ’s judgment in Google Android does not signal a retreat from the effects-based approach. Instead, it clarifies that effects-based enforcement does not require adherence to any particular methodological tool. Viewed through the lens of decision theory, the judgment is ultimately about how courts assess theories of harm under conditions of uncertainty.
Introduction
On 2 July 2026, the Court of Justice delivered its long-awaited judgment in Google Android. To briefly recall the facts, Google implemented a set of interrelated contractual arrangements – namely MADAs, AFAs and RSAs. Through these arrangements, Google tied access to the Play Store to the pre-installation of Google Search and Chrome, while tying the licensing of the Play Store and Google Search to compliance obligations restricting the development and commercialization of Android forks. In addition, Google offered revenue-sharing payments conditional on the non-pre-installation of competing search services supporting its infringements. For background, see the previous posts discussing the General Court’s judgment and the Advocate General’s Opinion.
The significance of Google Android extends far beyond the Android ecosystem. The judgment is less remarkable for what it changes than for what it confirms. The Court once again makes clear that neither the as-efficient competitor (AEC) test nor a formal counterfactual analysis forms a mandatory component of every Article 102 assessment. At the same time, it reaffirms the effects-based approach and the importance of contextual analysis.
Why Economics Entered Article 102 TFEU
At first sight, the more economic approach appears to be a story about the growing influence of economics in competition law. In reality, it is equally a story about uncertainty and error costs.
Competition authorities and courts rarely know with certainty whether a particular practice restricts competition or merely disadvantages competitors. As in many areas of law, decisions under Article 102 TFEU must therefore be taken under conditions of uncertainty. Every intervention creates a risk of error: competition authorities may either fail to prohibit genuinely anticompetitive conduct (false negative) or condemn conduct that is in fact efficient and pro-competitive (false positive).
Against this background, the European Commission gradually moved away from more form-based approaches towards a stronger focus on the actual effects of conduct on competition (see the Commission’s 2008 Guidance Paper). The underlying idea was straightforward: grounding legal assessments in economic reality reduces the risk of erroneous intervention.
The EU Courts subsequently embraced this logic. Increasingly, the central question under Article 102 TFEU became not whether a practice belonged to a particular category of conduct, but whether it was capable of producing an actual or potential effect to restrict competition. As Advocate General Wahl observed in his Opinion in Intel (para 43), the objective of Article 102 TFEU is ultimately to capture behavior that has anticompetitive effects.
The more economic approach therefore sought to improve decision-making by relying more heavily on economic evidence and effects analysis. Google Android is revealing because it shows how far this idea extends.
Effects-based Enforcement and the AEC Test
If the more economic approach sought to reduce error costs through economic analysis, the as-efficient competitor (AEC) test became one of its most visible manifestations.
The idea underlying the test is closely linked to the fundamental objectives of Article 102 TFEU. EU competition law does not protect competitors as such. Rather, it seeks to protect competition on the merits and, ultimately, consumer welfare. Accordingly, the central challenge is to distinguish market displacement resulting from superior performance from exclusionary effects resulting from conduct that departs from competition on the merits.
In practice, however, this distinction is often difficult to draw. The likelihood of exclusionary effects depends on numerous factors, including the degree of market power, barriers to entry, network effects or consumer behavior. Consequently, competition authorities and courts must often assess effects whose occurrence is uncertain and merely probable to varying degrees.
The AEC test emerged as a tool to make this assessment more objective. At its core, it asks whether a competitor that is as efficient as the dominant undertaking could remain viable despite the conduct under investigation. Where an equally efficient competitor could not match the conduct without incurring losses, exclusionary effects may appear more likely.
The attraction of the AEC test lies not in its economic sophistication. More importantly, it promises to reduce the risk of false positives. By focusing on the position of an equally efficient competitor, the test seeks to avoid condemning aggressive but legitimate conduct. It therefore became closely associated with the broader ambition of the more economic approach to ground intervention in evidence-based theories of harm rather than abstract presumptions.
Google Android and the Limits of Methodological Formalism
Google’s appeal can be understood as a defense of a more methodological conception of effects-based enforcement. While framed as legal challenges to the General Court’s reasoning, several of Google’s arguments ultimately reflected a broader evidentiary concern. At their core, they questioned whether the Commission had demonstrated the exclusionary effects with a sufficient degree of economic certainty.
Among Google’s various arguments, four closely related issues are particularly relevant here. First, it challenged the General Court’s decision to take the RSAs into account when assessing the effects of the MADAs, despite the RSAs themselves not ultimately forming a finding of abuse. Second, it disputed the Commission’s reliance on a status quo bias, arguing that the low number of downloads of competing search applications could equally be explained by user preferences and the superior quality of Google’s services. Third, Google contended that the Commission had failed to assess the effects of the MADAs against a realistic counterfactual scenario. Fourth, Google argued that the Commission had failed to establish that the MADAs were capable of foreclosing as-efficient competitors.
Viewed collectively, these arguments reveal a common theme. They reflect a vision of effects-based enforcement that seeks to eliminate evidentiary uncertainty through increasingly refined economic analysis. The underlying premise is that the more uncertain the causal relationship between conduct and market outcomes, the more demanding the economic inquiry must become.
The Court was unwilling to go that far.
A first illustration can be found in its treatment of the RSAs. Exclusionary effects must be examined in their legal and economic context. The Court confirmed that conduct does not have to constitute an independent abuse in order to remain relevant as part of that context (para 191).
A second illustration concerns the debate over user behavior. Google argued that users remained free to download competing search applications and that their choices could therefore be attributed to the quality of Google’s services. The Court accepted that superior performance may explain consumer behavior (para 206). However, it equally recognized that where the choice architecture itself is distorted by the conduct of a dominant undertaking, observed market outcomes cannot automatically be attributed to competition on the merits (para 203 and 211). The practical absence of downloads therefore became more relevant than the theoretical possibility of downloading competing applications (para 207).
Finally, consistent with its established case law, the Court confirmed once again that neither an AEC test nor a formal counterfactual analysis constitutes a mandatory component of every Article 102 assessment (para 255 and 272). Both may provide valuable evidence. Neither enjoys the status of a universal legal requirement.
Taken together, these findings point in the same direction. The Court’s concern is ultimately not whether a sufficient number of economic checklists have been ticked off, but whether the available evidence shows that the conduct is capable of producing actual or potential anticompetitive effects.
Decision-Making Under Uncertainty: More Economics, Less Methodological Formalism?
From a decision-theoretic perspective, the Court’s approach is unsurprising.
As mentioned above, judicial decision-making does not occur under conditions of certainty. The courts must infer anticompetitive effects from imperfect and incomplete evidence. The Court itself acknowledged in Google Android that there are situations in which it is impossible to determine whether users’ behavior is attributable to the qualitative characteristics of a product or service or to the conduct of the dominant undertaking (para 203). Any assessment of anticompetitive effects therefore necessarily involves probabilistic reasoning.
In such an environment, insisting on a predefined methodological toolkit risks replacing legal formalism with methodological formalism. The relevant question is not whether a particular model, test or econometric technique has been applied. Rather, the question is whether the body of evidence, taken as a whole, is capable of pushing the alleged theory of harm beyond the threshold implied by the applicable standard of proof.
How much economics does Article 102 TFEU still require? There is no universal answer. The required quantity and quality of evidence – and therefore also the depth of economic analysis – depend on the factual circumstances of the case and, ultimately, on how courts balance the risks of false positives and false negatives.
Arguably, this is the real lesson of Google Android. The judgment does not reject economics, nor does it signal a retreat from effects-based enforcement. Instead, it recognizes the inherent uncertainty of competition law adjudication. Far from signaling less economics under Article 102 TFEU, the judgment endorses a more pragmatic understanding of economic analysis. Economics remains indispensable, but courts do not decide whether a theory of harm is true; they decide whether the evidence makes it sufficiently probable.
Are the Forthcoming Article 102 Guidelines Consistent with Google Android?
At first glance, the Draft Guidelines seem difficult to reconcile with the Court’s judgment in Google Android. While the judgment requires exclusionary effects to be assessed in their legal and economic context and in light of all relevant factual circumstances, the Draft Guidelines introduce presumptions for certain categories of conduct.
However, from a decision-theoretic perspective, these presumptions reflect the Commission’s inference that certain forms of conduct are associated with a sufficiently high prior probability of anticompetitive effects. This does not imply that every presumption contained in the Draft Guidelines is necessarily well founded. Rather, Google Android suggests that the legality of presumptions with Article 102 TFEU depends less on their existence as such than on the Commission’s willingness to engage with evidence capable of rebutting them.
The judgment therefore reinforces two important points for the future application of Article 102 TFEU: first, rebutting exclusionary effects should be clearly distinguished from objective justification; second, presumptions cannot relieve the Commission of its duty to evaluate the evidence put forward by the undertaking concerned.