Applying Computer Science Principles to Police Modern Cartels: A Conversation with Giovanna Massarotto

Ruled by Reason

On this episode of Ruled by Reason, AAI Senior Counsel David O. Fisher chats with legal scholar Giovanna Massarotto about what antitrust law can learn from computer science, and particularly how understanding agreement algorithms can help courts and enforcers police algorithmic price-fixing and other illegal agreements under Section 1 of the Sherman Act. The conversation centers on Massarotto’s recent paper, Detecting Algorithmic Collusion, which examines the characteristics of agreement algorithms and how they can inform the “plus factor” analysis courts use to determine the likelihood of an illegal agreement. It begins with an introduction to the concept of a "distributed system," which is any network of computers that works together to perform a common task, the Bitcoin blockchain being one notable example (5:25). It then examines the Byzantine Generals Problem, a classic story illustrating how the nodes in a distributed network can reach an agreement despite the existence of one or more unreliable nodes, which has parallels to the formation of stable cartel agreements (9:09). Massarotto explains how agreement algorithms create stability, and what they can teach courts and enforcers about how algorithmic cartels function. Specifically, she describes how agreement algorithms use digital signatures, cryptography, broadcasting, leader election, and private channels to allow stable decision-making in distributed systems (21:50). Massarotto concludes that, while broadcasting and leader election are accounted for in the existing plus-factor analysis, courts and enforcers should add the use of digital signatures, cryptography, and private channels to the list of plus factors which may indicate the existence of an illegal agreement (30:40).