Algorithmic collusion: the next generation of price fixing?

July 2024 | SPOTLIGHT | COMPETITION & ANTITRUST

Financier Worldwide Magazine

July 2024 Issue

Companies today rely on algorithms to inform their commercial decisions. Effective algorithms can make recommendations that more quickly and accurately react to changes in supply and demand, costs, consumer preferences and competitor pricing. As the benefits of algorithms have become more apparent, so too have their potential risks, including with regard to antitrust and competition. Recent lawsuits alleging price fixing via algorithm suggest that the use of algorithms can raise serious antitrust concerns. Antitrust authorities are therefore increasingly focused on the use of pricing algorithms and enforcement against their potential anticompetitive effects.

The US Department of Justice’s (DOJ’s) Antitrust Division and the Federal Trade Commission (FTC) filed statements of interest in three recent algorithmic price fixing cases to detail their position that an agreement between competitors simply to delegate – to a pricing algorithm – key aspects of pricing decision making can be a per se violation of US antitrust law under section 1 of the Sherman Act. The European Commission (EC) also recently revised its guidelines on horizontal cooperation agreements to describe scenarios where the use of pricing algorithms may present competition concerns. This article outlines key aspects of the current guidance from competition law enforcement agencies on potential liability for algorithmic collusion and provides practical considerations for minimising the antitrust risks associated with the use of algorithmic pricing tools.

Algorithmic collusion under US antitrust law

Section 1 of the Sherman Act prohibits “[e]very contract, combination in the form of trust or otherwise, or conspiracy” that unreasonably restrains trade. Section 1 liability therefore turns on the satisfaction of two separate elements: a concerted action between two or more entities, and an unreasonable restraint of trade. The DOJ analysed the application of these elements to the use of pricing algorithms in a statement of interest submitted in In re RealPage, Rental Software Antitrust Litigation, a case involving allegations of price fixing by landlords using RealPage’s software to inform decision making on rents. There the DOJ underscored two important points. First, the joint delegation of pricing decisions to an algorithm constitutes a concerted action for section 1 purposes. And second, the joint action to “raise, depress, fix, peg, or stabilize prices” is a per se unreasonable restraint on trade regardless of whether it is accomplished through a human or a pricing algorithm. The agencies also clarified that “not every use of an algorithm to set price qualifies as a per se violation of Section 1 of the Sherman Act” – rather, such use is only unlawful where “competitors knowingly combine their sensitive, nonpublic pricing and supply information in an algorithm that they rely upon in making pricing decisions, with the knowledge and expectation that other competitors will do the same”.

The DOJ emphasised two additional points through its statement of interest in RealPage and was joined by the FTC in filing statements of interest in two subsequent cases – Duffy v. Yardi, another case alleging algorithmic price fixing by landlords, and Cornish-Adebiyi v. Caesars Entertainment, a case alleging price fixing by competing casino hotels using the same pricing algorithm. First, the agencies made clear that it is illegal for competitors to fix not only their final prices but also their advertised list prices or sticker prices. The agencies argued that an agreement to use an algorithm to inform the starting point of prices is similarly a Sherman Act violation, even if participants ultimately deviate from the recommended prices.

Second, the agencies emphasised that liability for algorithmic collusion under section 1 does not require an explicit agreement or even direct communication between participating competitors. Section 1 encompasses not only explicit agreements but also “tacit agreements” to fix prices that are evidenced by action alone, such as discussions about future plans combined with parallel conduct among competitors, and as the agencies explained in the statement of interest submitted in Cornish-Adebiyi, “courts have repeatedly inferred horizontal agreements among competitors based on communications to (or passed through) an organizer or other intermediary”. The DOJ argued in the RealPage statement of interest that to establish a concerted action between competitors, “it suffices to show that RealPage proposed the price-fixing scheme to competing landlords, who were each aware that its competitors were also being invited to participate in the scheme, and the competitors adhered to it – generating a common understanding among the competitors that they would increase prices collectively by using RealPage”.

These statements of interest were submitted in civil antitrust actions brought by private plaintiffs, but the agencies’ increased engagement with this issue indicates they may bring direct enforcement action, either criminal or civil, against allegedly illegal algorithmic collusion. Recent press releases and blog posts from the FTC emphasise the agencies’ “strong interest” in protecting consumers from algorithmic collusion and highlight related past and present enforcement actions involving the use of pricing algorithms and price-related information sharing among competitors. DOJ officials also warned that the agencies are scrutinising scenarios where competitors take an “I’ll show you mine, if you show me yours” approach that is common in potentially anticompetitive information exchanges.

Lawmakers are also pushing for greater agency enforcement in this area. Amy Klobuchar, a US senator, sent a 29 April 2024 letter to the FTC and DOJ encouraging the agencies to investigate insurance companies’ reliance on algorithmic pricing tools offered by healthcare cost management company MultiPlan. Senator Klobuchar also recently introduced the Preventing Algorithmic Collusion Act of 2024. The Act would prohibit the use of pricing algorithms that rely on non-public competitor data, as well as establish disclosure requirements and audit tools for competition law enforcement to monitor non-prohibited algorithm use.

EC guidance on algorithmic collusion

The EC is also focused on the use of pricing algorithms by competitors, which may constitute illegal collusion whether or not there is an explicit agreement between competitors to fix prices.

The EC’s revised guidelines on horizontal cooperation agreements, adopted in June 2023, reiterate its existing position that an explicit agreement among competitors to use the same pricing algorithm is considered a “by object” infringement of article 101 of the Treaty on the Functioning of the European Union. The guidelines take the position that the mere use of a common pricing algorithm may be considered an article 101 infringement where the algorithm relies on users’ submission of commercially sensitive pricing information, and where the user is aware that competitors are using the same algorithm to inform their own pricing decisions. The revised guidelines further suggest that pricing algorithm providers themselves may be liable under article 101 if it was reasonably foreseeable that their pricing algorithm would be used to facilitate the unlawful exchange of commercially sensitive pricing information among competitors.

Independent use of an algorithmic pricing tool that relies on publicly available data is considered presumptively lawful under the revised guidelines. Without offering specific guidance, the guidelines raise the unsettling possibility that future pricing algorithms may be advanced enough to develop their own coordination strategies, and that such inter-algorithm coordination could potentially subject even independent users to liability for algorithmic collusion.

Practical tips to minimise antitrust risk from algorithm use

Companies should ensure that the use of pricing algorithms is still consistent with independent decision making in price setting. Certain considerations, like not entering into explicit agreements with competitors to use the same pricing algorithm in the same way, may seem obvious. But using pricing algorithms that are marketed as a shared resource for industry participants and offer pricing recommendations based on non-public pricing information may also increase the risk of antitrust liability, as recent guidance demonstrates.

Guidance and enforcement actions addressing potential algorithmic collusion will continue to evolve alongside algorithms themselves. But there are concrete steps companies can take to minimise antitrust risk even while taking advantage of these increasingly powerful and often beneficial pricing tools. First, companies can minimise potential antitrust risk by employing their own proprietary algorithms that rely on publicly available data. Companies may continue to use third-party algorithms but should confirm that there are appropriate guardrails in place that protect against the sharing of deanonymised and competitively sensitive information. To mitigate antitrust liability associated with the use of third party pricing tools, companies should create documentation reflecting that pricing decisions are the result of independent, unilateral decision making, including business factors independent of algorithmic pricing recommendations where applicable, particularly where a company is aware that the same pricing algorithm is used by competitors. Companies interested in third party pricing algorithms should also consider modelling their use on FTC and DOJ guidance concerning the permissible exchange of competitively sensitive information in healthcare markets. This guidance identified a “safety zone” within which the exchange of information between competitors is unlikely to raise antitrust concerns – namely where the information exchange is managed by a third party and is properly anonymised and aggregated, where it reflects data that is at least three months old, and where it includes data from at least five participants. The policy statements reflecting this guidance have since been rescinded and DOJ officials recently warned they will bring cases that fit within the framework of legal precedent, particularly for data that allows competitors to understand forward-looking behaviour. Still, absent updated guidance, these criteria remain useful guideposts for companies using or considering third party pricing algorithms because they ensure that the data received is sufficiently anonymised and not forward-looking, and therefore less likely to facilitate collusion. Companies should also conduct training with relevant teams to ensure compliance and to determine how best to avoid risk.

Nana Wilberforce is a partner, John W. O’Toole is counsel and Lydia J. Turnage is an associate at Wilmer Hale. Ms Wilberforce can be contacted on +1 (213) 443 5376 or by email: nana.wilberforce@wilmerhale.com. Mr O’Toole can be contacted on +1 (202) 663 6256 or by email: john.o’toole@wilmerhale.com. Ms Turnage can be contacted on +1 (212) 230 8855 or by email: lydia.turnage@wilmerhale.com.

Nana Wilberforcer, John W. O’Toole and Lydia J. Turnage

Wilmer Hale