Autonomous algorithmic coordination

Full abstract

Read the full abstract

[Purpose] The use of autonomous pricing algorithms in oligopolistic markets makes it increasingly difficult to distinguish lawful unilateral adaptation from concerted practices prohibited under Article 101(1) of the Treaty on the Functioning of the European Union. The problem becomes more acute when evidentiary presumptions developed for human interactions attribute collusive significance to market convergence resulting from the independent operation of algorithmic systems. This article examines the conditions under which technical evidence generated through compliance-by-design can support an exculpatory evidentiary mechanism capable of rebutting a presumption of autonomous algorithmic coordination while reducing the risk of Type I errors. [Methodology/design/approach] This study adopts a conceptual and doctrinal legal approach. It examines the European Union framework governing concerted practices, the allocation of the burden of proof, judicially developed evidentiary presumptions, and standards for evaluating circumstantial evidence. The analysis integrates the case law of the Court of Justice of the European Union with insights from legal epistemology, error-cost theory, and the documentation, traceability, human oversight, and auditability requirements established by Regulation (EU) 2024/1689, which serves as the methodological reference framework. [Findings] The analysis demonstrates that compliance-by-design evidence may perform an independent evidentiary function when it contemporaneously, verifiably, and auditable establishes the functional independence of a pricing algorithm. The article proposes a sequential test consisting of three cumulative stages: verification of a decoupled algorithmic architecture, a counterfactual assessment of independent rationality, and third-party auditing. Satisfaction of all three stages provides a plausible alternative explanation for parallel pricing, weakens the inference of concerted action, and raises the level of evidentiary sufficiency required of the competition authority without shifting the burden of proof. Ex ante documentation, continuous logging, human oversight, control of feedback loops, and independent external verification reduce the risks of manipulation and Type II errors. [Practical implications] Technical evidence generated through compliance-by-design provides businesses, competition authorities, and courts with a structured framework for assessing alleged cases of autonomous algorithmic coordination. Its use should be confined to demonstrating the absence of concerted action or establishing a plausible alternative explanation, without introducing efficiency considerations governed by Article 101(3) TFEU. The proposed mechanism supports a more accurate evidentiary assessment consistent with the presumption of innocence, legal certainty, and the effective enforcement of competition law.