Protecting consumers from AI-powered collusion

An economic test for unlawful agreements among competing firms to adopt a pricing algorithm

Research by Joseph Harrington develops a new approach for uncovering economic evidence of an unlawful agreement between competing firms to adopt a third party’s pricing algorithm. The study, prepared for the journal Economic Policy, shows that if firms have an agreement, then the prices produced by the algorithm will be higher as more firms adopt and, on average, adopting firms price higher than non-adopting firms.

The economic evidence provided by the proposed new test could be used in conjunction with other evidence to prove a violation of Article 101 of the Treaty on the Functioning of the European Union. This test could also be used by competition authorities to screen markets where evidence that prices are increasing in the adoption rate suggests an investigation is warranted – which may include the markets for apartment rentals, gasoline and hotels.

Big Data and artificial intelligence (AI) are fuelling a growing market in the supply of algorithms by software and data analytic companies to assist firms in the pricing of their products and services. With superior programming expertise, computing power and data, the algorithms of these third parties can allow a firm to adjust its prices more rapidly to changes in demand and supply, which can increase transaction volumes and social welfare, and to tailor prices to more narrowly-defined market segments, which enhances price discrimination and can increase profits and benefit lower-income consumers who are able to buy at a lower price.

At the same time, there is the possibility of anticompetitive harm should competitors contract with the same third party who could then coordinate price increases. There is evidence, and claims of evidence by plaintiffs in private litigation, that these anticompetitive effects have occurred in the markets for apartment rentals, gasoline and hotels.

The primary policy challenge is to permit firms to adopt a pricing algorithm on efficiency grounds while preventing its adoption when it is anticompetitive. Central to meeting that challenge is determining when a third party and competitors in a market have an unlawful agreement to adopt and implement a pricing algorithm.

This study develops an approach for uncovering economic evidence of an unlawful agreement. The approach is based on understanding how a third party’s design of its pricing algorithm depends on whether or not there is a latent agreement.

The study shows that if firms have an agreement, then the prices produced by the pricing algorithm will be higher as more firms adopt and, on average, adopting firms price higher than non-adopting firms. In contrast, if there is no agreement, then adopting firms’ prices do not change with the adoption rate and, on average, adopting and non-adopting firms price the same.

This empirical test was conducted in a recent study of retail gasoline markets in Germany which found evidence that the average price of adopting firms is higher when more firms adopt (Assad et al, 2023).


An Economic Test for an Unlawful Agreement to Adopt a Third-Party’s Pricing Algorithm

Author:

Joseph E Harrington, Jr (Department of Business Economics & Public Policy, The Wharton School, University of Pennsylvania)