European competition authorities are pursuing exploitative abuse cases under Article 102 Treaty on the Functioning of the European Union (TFEU) with renewed intensity yet are lacking the consensus on how to operationalise fairness as a legal standard.
At the CRA Annual Brussels Conference 2025, regulators and academics debated this shift alongside new experimental evidence challenging how welfare and market definitions are assessed in platform markets.
The panel featured Professor Pinar Akman (University of Leeds), Professor Leonardo Bursztyn (University of Chicago), Professor Michael Jacobides (London Business School), Linsey McCallum (DG Competition), and Martijn Snoep (Netherlands Authority for Consumers and Markets) and moderated by Mikaël Hervé (Charles River Associates).
The exploitative abuse revival without conceptual consensus
Recent Commission cases have increasingly targeted unfair trading conditions:
– Apple Music Streaming (price information restrictions)
– Facebook Marketplace (data use)
– Meta AI’s WhatsApp access and Google’s AI terms for content creators
Snoep described competition law as having both a “sword” function against exclusion and a “shield” function against unfair treatment, with the shield having been “the unloved child of competition law” since the 1970s. McCallum explained these cases address “mixed incentives” where platforms simultaneously seek intermediation from business users while competing with them.
Yet the panel’s sharpest exchange revealed fundamental disagreement.
Professor Akman identified three unresolved questions: neither the Treaty nor case law defines fairness consistently; fairness depends on perspective; and where fairness doesn’t correlate with consumer welfare, tension arises. Her position: fairness alone cannot be a theory of harm. Exploitative abuse cases should demonstrate both exploitation and exclusion.
Snoep disagreed. Fairness has been addressed in Article 102 since 1958, and authorities have a legal obligation to enforce it. Article 102 creates a distinct legal regime for dominant firms with special responsibilities, not general consumer protection.
Key takeaway: Fairness-based enforcement is proceeding without a settled framework, creating jurisdictional variation in how it’s operationalized.
Platform usage doesn’t equal welfare
Professor Bursztyn presented experimental research fundamentally challenging welfare analysis:
– 50% of surveyed Instagram users wished the platform didn’t exist, yet continued using it
– Users would like to be paid $50 to stop using TikTok individually, but would be ready to pay $30 for TikTok not to exist at all
This reveals “non-user utility”: platforms creating value not through benefits to users, but by imposing costs on non-users. Users participate to avoid social exclusion rather than derive positive benefits.
Market definition implications: Individual deactivation leads users to substitute with non-social activities; collective deactivation (the relevant counterfactual) shifts substitution toward other social platforms. Conventional choice data may systematically overestimate market breadth.
Both McCallum and Snoep indicated receptiveness to experimental evidence. The Commission’s Apple Music Streaming case used a Spotify consumer experiment; the Google Shopping case involved a Bing experiment. This opens evidentiary pathways, distinguishing value creation from exit costs in ways traditional data cannot.
Ecosystems challenge market definition
Professor Jacobides argued that market definition frameworks struggle with ecosystem power. His example: if Match.com with Tinder left Apple’s ecosystem, it would lose not just iOS users but potentially its entire market, as users won’t segregate dating by phone operating system.
Professor Jacobides stated, “we’re stretching the boundaries of market definition to define things that are not really market.” The Commission’s Booking.com/eTraveli case represented the first explicit ecosystem market definition.
Tying, AI and remedies
On tying: McCallum clarified that while draft Article 102 guidelines reference presumptions for traditional industrial goods, this does not apply to digital markets. Every Commission digital tying case included full effects assessments. The ECJ judgement in Google Android will provide further clarity.
On AI: No consensus emerged on whether it will disrupt or reinforce existing dominance. Professor Jacobides warned that capital-intensive investments and ecosystem integration may strengthen incumbent positions.
On remedies: A Berlin court characterized Google Shopping remedies as largely cosmetic eight years post-decision. McCallum noted the Commission is adapting through collaborative remedy design and questioned whether the legal hierarchy requiring behavioural remedies before structural ones remain appropriate for digital markets.
What practitioners need to know
Three developments require attention:
First: Professor Bursztyn’s experimental methodology demonstrates that conventional welfare analysis fails in network markets where users participate to avoid exclusion rather than because platforms create value.
Second: Fairness-based enforcement is expanding without consensus. The Akman-Snoep exchange revealed fundamental disagreement on whether fairness requires demonstration of competitive harm or functions as a standalone standard.
Third: McCallum’s clarification that tying presumptions don’t apply to digital markets confirms that effects-based analysis remains necessary for product integration decisions.
The Article 102 framework is adapting to network effects and ecosystem dynamics that differ fundamentally from traditional markets. How authorities resolve the fairness question and incorporate experimental evidence on non-user utility will determine enforcement boundaries going forward.
CRA will continue to provide economic analysis and strategic guidance as these policies evolve.
