Court Dismisses Algorithmic Price-Fixing Case, but Opens Door to Amended Complaint

In an early test of antitrust claims based on alleged algorithmic price fixing, a federal judge dismissed a class-action complaint alleging that hotel operators conspired to unlawfully fix the prices of hotel rooms on the Las Vegas Strip using pricing software. Last week, a judge in the District of Nevada dismissed allegations that Las Vegas Strip hotel operators colluded to use pricing software to fix room rates, finding that Plaintiffs failed to plausibly allege that there was an agreement among the hotels to use the same pricing algorithm or even the same software product; which hotel operators were involved in the purported agreement; or that there was any confidential information exchanged via the software to support a “hub-and-spoke” conspiracy. The court did, however, grant Plaintiffs leave to submit an amended complaint. Thus, the case (Gibson v. MGM Resorts) provides initial guideposts for how companies using pricing software might reduce potential antitrust risks.

Background 

In a class-action complaint, Cendyn Group and a group of hotel operators were accused of artificially inflating the price of hotel rooms on the Las Vegas Strip by colluding to use algorithmic pricing software offered by Rainmaker, a subsidiary of Cendyn. According to the complaint, Cendyn collected pricing and occupancy information from competing hotel operators, and its algorithmic pricing software then used this information to recommend prices to the hotel Defendants. Plaintiffs alleged that these pricing algorithms allowed the hotels to maximize profits by charging supracompetitive rates because the hotels knew that the other operators would follow the rates recommended by the pricing algorithm. 

In their complaint, Plaintiffs referred to a speech by former FTC Acting Chair Maureen K. Ohlhausen who explained that, while computer algorithms themselves may not be inherently unlawful, these systems may allow competitors to “reach an agreement on prices or output levels” and to “essentially to fly under the radar, so their unlawful agreements can escape detection by the enforcement agencies.”

Ultimately, however, the court concluded that Plaintiffs had not plausibly alleged a conspiracy to restrain trade and dismissed the complaint.

Court Requires the Who, What, To Whom, and When

The court dismissed the complaint because Plaintiffs failed to plausibly allege a conspiracy by answering the questions: “who, did what, to whom (or with whom) . . . and when.”

1. What Was Agreed To? The court found that Plaintiffs failed to properly allege Defendants had entered into an agreement—a predicate for a Section 1 Sherman Act violation. While Plaintiffs generally alleged that Defendants used Rainmaker software, the court found that Plaintiffs failed to state which one of three Rainmaker products they used, that they all used the same software product, that they used the same pricing algorithm, or that they used that software in their Las Vegas Strip hotels (rather than in other locations). According to the Court, these pleading deficiencies made it impossible to infer that all hotel operators agreed to use the same software pricing algorithms in their Las Vegas Strip hotels.
   
   The court also said that Plaintiffs failed to allege that the hotels were required to accept the prices that were recommended by the software. Although Plaintiffs had pointed to evidence that Rainmaker’s GuestRev pricing recommendations were generally accepted 90% of the time, the court said this did not show how many Las Vegas Strip hotels used Rainmaker’s GuestRev software or that they had agreed to accept the GuestRev pricing recommendations at those hotels. This was a “fatal deficiency in the Complaint.”
   
   
2. Who Agreed? The court likewise held that Plaintiffs failed to sufficiently allege which hotel operators entered into the purported agreement. For example, Plaintiffs alleged that Rainmaker hosted annual conferences where Defendant “Hotel Operators” had the opportunity to meet and network. But there was no allegation that Defendants attended the conferences. The court concluded it was not enough to generically allege that unspecified “Hotel Operators” entered into the agreement; rather, Plaintiffs were required to identify which specific Defendant did what.
   
   
3. When Was it Agreed? Plaintiffs conceded that they were unable to identify when Defendants entered into the purported agreement. The court said it was unable to “plausibly infer” that the Defendants began to use the software “around the same time,” meaning it also could not infer that they had entered into any agreement. Moreover, without plausible allegations of parallel conduct, the court could not consider any alleged “plus factors” Plaintiffs put forth.

No Proof of Hub-And-Spoke Conspiracy

Finally, the court found that Plaintiffs failed to adequately allege a hub-and-spoke conspiracy via the use of Cendyn’s software. The court said that Plaintiffs did not clearly allege that the pricing information Defendants purportedly obtained from the pricing algorithm was nonpublic, as opposed to publicly available pricing information. While Plaintiffs successfully alleged that confidential pricing information is placed into the software, they failed to plead that confidential information from other competitors came out.

Despite the “numerous pleading deficiencies,” the court noted that if Plaintiffs were to identify which pricing algorithms particular hotel operators were using on the Las Vegas Strip, it might support a plausible inference that there was an agreement between Defendants to use the same pricing algorithms, at least under a rule of reason theory of harm.

Takeaways

 The court’s decision suggests multiple ways to help reduce antitrust risk when using pricing software:

1. A company using its own propriety pricing software, or at least software that differs from its competitors, is likely to face lower risk than if competitors are using the same software. Of course, competitors should never agree or coordinate to use the same pricing software or particular pricing algorithms.
   
   
2. When entering into an agreement with vendors of algorithmic pricing software, make it clear in the agreement and with the company’s internal pricing team that any algorithmic pricing recommendation is only a suggestion and not a requirement. Companies should continue to make independent pricing decisions, including whether or not to follow the pricing recommended by an algorithm.
   
   
3. If there is a chance a pricing algorithm the company uses is also being used by a competitor, it is safest to ensure that the pricing software uses only publicly available information as inputs into the pricing algorithm. If a company receives pricing recommendations from an algorithm that includes nonpublic pricing information, there will be less risk if the nonpublic pricing information is the company’s own information and not that of its competitors.

The district court left the door open to an amended complaint and suggested that certain evidence could suffice to support a complaint that overcomes a motion to dismiss. Meanwhile, class actions in other industries have also been filed based on the alleged common use of algorithmic pricing software. Therefore, companies using algorithmic pricing software in their business should consider taking steps to mitigate potential risks while the antitrust case law on algorithmic pricing evolves.