Algorithmic Collusion and Third-Party Pricing Providers: Insights for Executives and Attorneys

Ai Deng draws on recent research to discuss the effects of third-party pricing algorithms on competition

Part one of this series explored lessons for executives and attorneys from the latest academic research on antitrust implications of using first-party pricing algorithms. Part two turns to the closely related topic of third-party pricing algorithms.

Recent antitrust litigation involving multifamily rental companies and hotel chains has centered around a key issue: these organizations outsourced their pricing decisions to third-party algorithm providers. While such algorithms could improve efficiency, outsourcing pricing decisions introduces a complex dynamic—with significant implications for competition.

Well before these recent court cases, antitrust authorities had raised concerns about firms using the same third-party algorithms. The UK’s Competition and Markets Authority warned that “[i]f a sufficiently large proportion of an industry uses a single algorithm to set prices, this could result in… the ability and incentive to increase prices.” The US Federal Trade Commission and Department of Justice issued similar warnings via their joint statement on recent litigation.

Fortunately, emerging academic research sheds new light on the effects of third-party algorithms on pricing and competition. At the core of this research is the recognition that:

1. Third-party algorithm providers and the firms adopting these algorithms often have distinct economic incentives; and
2. These incentives interact in ways that necessitate a nuanced approach to analyzing the potential for algorithmic collusion.

Independent Adoption Decisions

To begin, consider a scenario where firms independently decide to adopt the same third-party pricing algorithm. It might seem intuitive to assume that the third-party provider would have an incentive to “soften” competition among its adopters (e.g., raise prices) as it directly influences, through the collective use of its algorithm, how these firms compete.

However, research reveals that incentives of third-party providers are more complex. Raising prices might appear to benefit adopters but also could discourage adoption. For example, if a third-party algorithm raises prices, competitors that do not use the algorithm could undercut those prices and increase their profits, reducing the incentive for firms to adopt the algorithm in the first place.

Research shows that third-party providers instead are incentivized to design algorithms that make prices highly responsive to demand fluctuations. By quickly detecting or even anticipating increases in consumer demand, these algorithms can adjust prices in real time, enabling adopters to capture incremental profits. This responsiveness represents a form of economic efficiency that allows firms to allocate resources more effectively and increase profits. Importantly, this approach avoids the disincentive for adoption that arises when prices are simply raised across the board. In fact, firms operating in markets with highly variable and uncertain demand—particularly smaller firms lacking the resources to improve their responsiveness to demand on their own—are especially motivated to adopt such third-party solutions.

Therefore, while widespread adoption might raise red flags, it also could simply reflect that third-party developers can offer more effective tools than firms could develop on their own, driven by greater expertise, more access to data, and stronger incentives to invest in development. (See Harrington (2024a) for other assumptions and further discussion.) For example, in the discussion above, we assume that firms would accept the prices recommended by the third parties. To the extent that firms override pricing recommendations, the analysis would entail additional considerations.

The Role of Coordination and Collusion

The dynamics change significantly when firms coordinate their adoption of the same third-party algorithm. In such cases, the third-party provider may act as a de facto “cartel manager,” facilitating collusion among adopters. This so-called “hub-and-spoke” arrangement incentivizes the provider to raise prices for all adopters, with the resulting price increases becoming more pronounced as more firms adopt the same algorithm (all else equal).

Drawing on these insights, Harrington (2024b) proposed a method to test for coordinated adoption by examining whether the average price of adopters increases as the adoption rate rises. Evidence of such a trend could indicate collusion.

Balancing Efficiency and Supracompetitive Pricing

The discussion so far highlights a fundamental trade-off faced by independent third-party algorithm providers. On the one hand, setting supracompetitive prices on behalf of the adopters to increase their profits translates into a greater willingness to pay for the providers’ algorithmic pricing services. On the other hand, simply raising prices across the board could deter adoption, as potential adopters may fear being undercut by competitors. At the same time, improving efficiency—by making prices more responsive to demand—not only enhances adopters’ profitability but also incentivizes adoption.

Facing such a trade-off—and understanding that efficiency drives adoption—third-party providers may aim to strike a balance: introducing just enough supracompetitive pricing to maximize the adopters’ (and consequently their own) profits without discouraging adoption. In a market with highly variable and uncertain demand, efficiencies gained from adopting a third-party algorithm could outweigh costs of supracompetitive pricing, making adoption more attractive. However, this dynamic also blurs the line between efficiency and collusion. Greater algorithmic efficiency may, paradoxically, be accompanied by higher supracompetitive markups.

Additional Considerations

The design and impact of third-party algorithms depends on other market-specific factors, adding further complexity to this issue.

For example, while much of the research focuses on a single dominant third-party provider, competition among independent providers could potentially mitigate collusive effects. If one provider attempts to impose supracompetitive prices, a competing developer could undercut it, reducing the effectiveness of—and hence the incentive for—collusion. This is one reason why fostering innovation and competition in the development of third-party algorithms can be important.

Moreover, third-party algorithms may allow for significant customization by adopters. Firms may tailor their third-party algorithms to meet specific business objectives, such as increasing market share, reducing inventory, or boosting transaction volume. These objectives can also vary over time, even for the same firm. This means that firms using the same third-party algorithm may not be engaging in identical pricing strategies. As a result, competition among firms may not be limited to price alone, complicating the analyses of the effects of third parties on competition. (For other considerations, such as the role of information sharing, see Harrington (2024c).)

Conclusion

The intricate dynamics of third-party pricing algorithms call for a careful and nuanced approach to their analysis, particularly in the context of potential hub-and-spoke conspiracies. It is important to weigh the potential harms of anticompetitive behavior against the procompetitive efficiencies that algorithmic pricing can generate.

With the increasing adoption of third-party pricing algorithms in more markets, executives and attorneys should stay informed about the latest economic insights and legal developments to navigate this evolving landscape.