Algorithms and Antitrust Compliance
Algorithms have become ubiquitous in everyday life in the digital economy, from choosing the content we see on social media to recommending the next song on our playlists. The ABA Section of Antitrust Law’s Compliance and Ethics Committee in its Fall 2018 special report published Algorithms: Challenges and Opportunities for Antitrust Compliance that explores the impact of this increasingly transformative new technology. This report, co-authored by Anita Banicevic (Davies Ward Phillips & Vineberg), Gabrielle Z.A. Kohlmeier (Verizon), Dajena Pechersky (Davies Ward Phillips & Vineberg), and Ashley Howlett (Hogan Lovells, seconded to Verizon), is the second in a series of special reports by the Tech Compliance Working Group on the impact of new technologies on antitrust compliance.
Varying in their complexity, algorithms — programmed code containing a series of rules and formulas that are performed in a specific order to accomplish a specified task — have transformed markets through their ability to instantaneously aggregate and process data. Algorithms are deployed in many of the systems that consumers rely on every day, from popular ride-sharing algorithms that use dynamic pricing, and comparison shopping platforms that allow consumers to scan and compare market prices, to more sophisticated algorithms that learn individual behaviors to provide customized services. Businesses predominantly use algorithms for predictive analytics and business process optimization.
As the report highlights, there are many procompetitive benefits that algorithms provide, including faster price adjustments, lower consumer search and transaction costs, lower prices, higher quality products and services, and more efficient resource allocation through better matching between supply and demand. As this emerging technology continues to be deployed across various sectors of the economy, competition authorities are taking note. This report provides an interesting dive into the novel challenges that algorithms pose for competition authorities, sharing perspectives about antitrust risks and compliance challenges from around the globe with particular attention to the increased potential for collusion, both tacit and overt.
“Collusion” occurs when competing firms coordinate to set prices with the objective of increasing profits. “Overt collusion” refers to anticompetitive conduct that is facilitated through explicit written or oral agreements, whereas “tacit collusion,” also referred to as “conscious parallelism,” is anticompetitive conduct achieved absent explicit coordination. Because there is no underlying illegal behavior, conscious parallelism is not illegal. The report asserts that existing antitrust tools are effective to address overt algorithmic collusion; however, the challenge may arise with tacit algorithmic collusion. While algorithms allow real-time market assessments and adjustments, which could facilitate collusion, scholars debate whether there is real concern at this stage; tacit collusion is mere speculation at this point, as AI technology is far from developing algorithms with those capabilities.
US antitrust enforcers recognize that algorithm-enabled anticompetitive behavior may make external detection more difficult. Nonetheless, they have staked the position that algorithm-enabled behavior is to be analyzed under the same legal standard, with the technology involved having no bearing on the analysis. The report also reviews the global competition landscape, analyzing the strategies of Canada, Australia, the UK, the EU, Japan, Singapore, and Indonesia. Many nations appear to be taking a measured approach to concerns of collusion resulting from evolving technology. Confident that existing analytical principles and enforcement tools are appropriate, international counterparts have adopted measured approaches focused on monitoring developments and applying targeted, fact-specific action as appropriate.
Concluding with eight practical considerations for practitioners, the report notes that competition authorities are already using their own algorithms to detect anticompetitive behavior and enhance their own economic modeling. As sophistication and facility with evaluating algorithms evolves, algorithms may be used as evidence in evaluating mergers. As a result, compliance efforts should ensure that these algorithms are not engaged in anticompetitive behavior. Antitrust compliance by design, ardently endorsed by the European Commission, should be a consideration for all compliance personnel. Considerations for developing a compliance program could include training and communication, AI-enhanced compliance screening, reporting channels, and evaluation and governance procedures, with periodic review and updates. Compliance personnel may also consider expanding algorithmic monitoring to third-party users. On the whole, the report emphasizes the centrality of algorithms in digital economy, concomitant with the opportunity to spur innovation and the potential for facilitating anticompetitive conduct. Practitioners should stay abreast of developments in this area and employ compliance tools to mitigate their clients’ exposure to antitrust liability.
The report, Algorithms: Challenges and Opportunities for Antitrust Compliance, was published in the American Bar Association Section of Antitrust Law Compliance and Ethics Committee Spotlight (Fall 2018) and shortlisted for an Antitrust Writing Award.
Tawanna Lee is a 2L at The George Washington University Law School and Legal Intern at CCIA. Follow her on Twitter at @TawannaLee
