Christopher Anderson, professor of services management at the Cornell Peter and Stephanie Nolan School for Hotel Administration, studies the use of data and analytics to predict demand, adjust prices and manage inventory in a variety of industries. In a recent article, "The Perils of Algorithmic Pricing," Anderson and his co-author discussed why pricing algorithms matter.
Chris Anderson
Question: What is algorithmic pricing and why has it come under scrutiny?
Answer: I have been researching service pricing for more than two decades, and algorithmic pricing represents a critical and evolving legal risk. While revenue management and pricing systems have been around for decades in sectors like hospitality and airlines, we are entering a new era of sophistication and automation - and the practice is now the target of numerous class-action lawsuits and regulatory scrutiny.
Algorithmic pricing uses computer code or AI to automatically set prices and adjust them in real-time based on data, such as competitors' prices, inventory and time of day. The practice maximizes profit, efficiency or both. It becomes a problem if it leads to collusion and antitrust violations.
Q: Why is this issue so urgent?
A: The key urgency is the legal theory that these systems can facilitate collusion and lead to antitrust violations without explicit agreements or intent to collude, just by the way they use data and coordinate pricing. The United States Federal Trade Commission and Department of Justice have intervened in these cases, and legislation has been proposed. This isn't a theoretical risk-it's an immediate and powerful warning for businesses that rely on these increasingly common tools.
Q: How can algorithmic pricing unintentionally lead to collusion?
A: There are multiple pending class-action antitrust lawsuits in the casino, hotel and multifamily housing sectors. By examining the legal landscape, my colleague and I developed a framework to understand what's going on.
The core finding is that the risk of antitrust violation hinges on the design and application of the algorithmic system, specifically regarding data usage and centralization.
We find that algorithmic pricing can lead to collusion through multiple mechanisms. For example, hub-and-spoke conspiracies, where a centralized vendor is the "hub" and clients are the "spokes," enables businesses to indirectly share non-public data, even without explicit agreements.
Tacit collusion could happen when businesses independently use the same or similar algorithms, unintentionally setting prices above the fair market rate.
The introduction of nonpublic, proprietary competitor data into the algorithm is a primary factor that allows the model to anticipate competitors' actions and recommend prices higher than would be found in a strictly competitive market.
While initial court dismissals cited a lack of explicit agreement, the FTC and DOJ position is that delegating pricing to a shared algorithm can itself constitute concerted action, which suggests a path towards collusion under antitrust law.
Q: What are the implications?
A: The prevailing view among U.S. regulators is that the use of these algorithms can lead to illegal price-fixing even without explicit intent. If this view is confirmed by the courts, it will pave the way for more antitrust lawsuits against algorithm vendors and their clients.
Companies and their algorithm vendors must urgently adopt new compliance strategies, focused on algorithmic design. We recommend using only available inventory and publicly available competitor data, strictly limiting or eliminating the pooling of sensitive, nonpublic information. It's also important to avoid simple rule-based pricing, such as automated price-matching or undercutting, which can unintentionally stabilize prices above competitive levels. Managers should retain the ability to manually override algorithmic recommendations to keep pricing decisions under human control and reduce the risk of implicit coordination.
Q: What are the next steps for this work?
A: My subsequent research moves from identifying the risk to explaining the precise mechanism of harm in service sectors like airlines, hotels or cloud computing.
Information sharing reduces uncertainty, which is the necessary catalyst for tacit collusion. In essence, the danger is not just that the algorithm is setting the price, but that the data feed eliminates the competitive engine of market uncertainty, making non-collusive price coordination the most rational, profit-maximizing strategy for all firms involved. This reinforces the need for strict regulatory limits on the type of data that can be shared via common revenue management platforms.
Alison Fromme is a writer for the Cornell SC Johnson College of Business.
