The Application of Current Antitrust Law to Explicit Collusion by Autonomously Acting Pricing Algorithms

Abstract

This paper aims to answer the question whether the current (EU) antitrust law is well-equipped to deal with explicit collusion by pricing algorithms that act independently from the companies using them. The paper unfolds in three main parts: first, it outlines the exact circumstances of the later discussed scenario and the differentiation to other ways in which algorithms can be related to infringements of antitrust law. Then it investigates whether the use of algorithms specified in the first part falls under the scope of antitrust law in general, that is whether the term 'agreement' according to Art. 101 (1) TFEU is applicable to independently acting algorithms or whether any other tool of current antitrust law (here: concerted practices) enables for the application of Art. 101 (1) TFEU to algorithmic price-setting. Finally, it assesses if the company using the algorithm can be held responsible for any breach of antitrust law that the algorithm is engaged in. It does so by applying the concept of attribution of liability for acts of employees to the employing firm. My hypothesis is that there is no fundamental difference between the conduct of deep learning algorithms and human employees and that as a result the same principles of attribution of liability to firms can be applied.

Keywords: algorithmic collusion, deep learning machines, price-setting algorithms, cartel law and algorithms, explicit collusion