Competition law/algorithms has been a hot topic of debate amongst practitioners and enforcers for a while, and now two of Europe’s top antitrust enforcers, the Autorité de la Concurrence and the Bundeskartellamt, have joined forces to analyse and publish a report on the potential competition law risks relating to pricing algorithms and their potential contribution to horizontal collusion.
The joint study, “Algorithms and Competition“, initially focuses on the concept of algorithms and specifically on dynamic price-setting or self-learning algorithms. The study explores the various ways in which they can establish interactions between companies, which may lead to horizontal collusion and distort the competitive functioning of the markets.
It notes that algorithms may be used as “facilitators” in “traditional” horizontal collusion, based on previous human communication, an agreement or concerted practice. In such scenarios, the involvement of an algorithm by itself does not raise competition concerns, but may be useful to investigate to identify potential counteracting efficiencies or the negative effects of the practice.
The report considers that in such scenarios companies are responsible for the algorithms they use, even though algorithms may be provided by third parties.
The authorities appear to be particularly concerned about the potential collusion among “self-learning algorithms”, which unilaterally observe, analyse and adapt to competitors’ public behaviour, and may further decide to collude on prices. The report flags the significant question of whether the conduct of self-learning algorithms may be attributed to a company.
The report further discusses practical challenges for authorities investigating algorithms under competition law. It considers the types of evidence required, including on the role of the algorithm, its context, functioning, and decision-making process. The report highlights that the evolving and quickly-changing digital markets mean there is very low predictability about future concerns and practical difficulties. Nevertheless, the report concludes that existing competition law allows authorities to address competition risks associated with algorithms.
The Franco-German study was presented shortly after the draft 10th revision of the German Act Against Restraints of Competition was published by the Ministry of Economy and Energy. The draft follows the “Commission Competition Law 4.0” report which sets out “A New Competition Framework for the Digital Economy” and which was published in September 2019 (see our e-bulletin here). Amongst other things the draft law, which if passed may enter into force in the second half of 2020, widens the net of companies that may infringe competition law by engaging in abusive conduct to companies with “intermediate” market power, or “significant presence across markets”. The draft also seeks to focus merger control on cases of greater economic significance by increasing the domestic turnover threshold from EUR 5 to 10 million, and the de minimis threshold from EUR 15 to 20 million. Other amendments seek to accelerate proceedings and enhance cooperation between the authorities and businesses.
With the upcoming German Council Presidency of the European Council in the second half of 2020, and Commissioner Vestager having a wider portfolio in the new Commission as “Commissioner for Competition and Executive Vice-President for a Europe fit for the Digital Age”, the Franco-German report and draft German law are likely to be influential in shaping the evolution of national and EU competition law and enforcement in relation to algorithms.