Algorithmic Tacit Collusion: With a Focus on the CPH-LHR Route

Abstract

This thesis investigates the growing risk of algorithmic tacit collusion in the commercial airline industry, specifically the route CPH-LHR and examines how current EU competition law may prove insufficient in addressing this phenomenon. Specifically, it explores pricing algorithms such as “Predictable Agent” and “Digital Eye”, which, while operating autonomously and without direct communication between competitors, can lead to stable, supra-competitive prices resembling traditional collusion. Since such behavior often falls outside the current interpretation of Article 101(1) TFEU, it creates a regulatory gap where coordination may occur without legal consequences. The thesis applies a threefold analytical framework. First, the economic analysis examines the competitive dynamics on the CPH–LHR route, where the market structure is characterised as a duopoly. By applying game theory, particularly Bertrand competition and repeated games, the analysis shows how pricing algorithms can modify firms’ strategic behaviour. This fosters conditions under which tacit coordination becomes economically rational. Finally, the theory of harm is used to assess how such algorithm-driven interactions may undermine the intensity of competition, entrench market stability, and reduce the disciplining effect of price wars. Second, the legal analysis provides a doctrinal assessment of whether such algorithmic behavior can be captured under existing EU competition law. It analyses the criteria of agreements, concerted practices, and information exchange under Article 101(1), in light of recent case law and the 2023 Horizontal Guidelines. The findings suggest that current legal concepts often fall short in capturing coordination that arises through autonomous algorithmic responses, rather than explicit intent or contact. Third, the integrated analysis proposes a normative solution, a design-based compliance model inspired by developments in adjacent EU legal frameworks such as the GDPR (privacy by design), the Corporate Sustainability Due Diligence Directive (CSDDD), the AI Act, and the Digital Markets Act (DMA). This model shifts regulatory focus from ex-post sanctioning to ex-ante prevention. Such an approach redistributes regulatory responsibility from authorities to firms and is particularly suited for addressing the structural and behavioral risks posed by autonomous systems. The thesis concludes that a targeted, risk-sensitive compliance model may serve as a necessary supplement to traditional competition law, provided it adheres to legal certainty, proportionality, and is practically enforceable in technologically complex environments.