EU and US competition law are anthropocentric in terms of the concept of agreements and concerted practices as they require the concurrence of wills, mental consensus, awareness, or intent – concepts that are inherent to humans. A key aspect is distinguishing these from conscious parallelism, which is not prohibited. Algorithms, increasingly entrusted by companies with the task of setting prices, blur the line between conscious parallelism and concerted practices and significantly impact the traditional factors of collusion and cartel stability. The forms of use of pricing algorithms relevant for competition law are categorized into four scenarios. Under current competition law, the use of pricing algorithms as tools for prior agreements between humans does not pose problems. When using pricing algorithms from the same provider, the liability of horizontal competitors for hub-and-spoke collusion requires at least awareness of possible anti-competitive behaviour, and the provider of these algorithms can also be liable. The scenarios of the predictable agent and fully autonomous algorithmic coordination represent conscious parallelism, except in certain cases. Due to the numerous efficiencies and advantages of using algorithms, the agreement or concerted practice can be exempted from prohibition under EU competition law, while in US antitrust, this is only possible if the rule of reason applies. Pricing algorithms also present a significant challenge for detecting and proving agreements or concerted practices and the issue of culpability for criminal liability and liability for minor offences. Conscious parallelism can currently be addressed via two competition law alternatives – merger control and abuse of collective dominance.
|
