In yet another attempt to stretch antitrust law into new territory, the Federal Trade Commission (FTC) and Department of Justice (DOJ) released a joint statement of interest related to whether algorithmic pricing is a form of collusion by hotels. However, this joint statement relies on questionable legal arguments to sidestep the fact that the collusion requires proof of communication between the accused parties. Algorithmic pricing could benefit consumers and shouldn’t be treated as inherently suspect.
In Altman et al v. Caesars Entertainment, et al., the prosecution accused several hotels in the Atlantic City area of collusion through the use of Rainmaker software, which increases hotel revenue by automating pricing. They argued that when multiple hotels began using Rainmaker, they essentially participated in collusion to keep prices high for patrons.
In a similar case in Las Vegas, the judge dismissed a lawsuit due to a lack of evidence of collusion. The ruling found that the plaintiffs had not sufficiently proven that there was any communication indicating an agreement to collude or to use the same software. Such proof is necessary to substantiate a violation of the Sherman Antitrust Act.
The FTC and DOJ weighed in arguing that collusion can still occur even if no one involved communicated with one another, claiming that using the same pricing algorithm can lead to collusion. Specifically, they cite cases of tacit collusion where there was not an agreement but rather parallel conduct (adopting the same pricing strategies) between companies that indicated collusion.
Tacit collusion, however, is not as cut and dry as regulatory agencies would have people believe. The kind of parallel conduct involved can also arise by responses to the market without any agreement or intent to collude. Whether this constitutes collusion has been debated in U.S. antitrust law for years because it is difficult to prove. Despite having different opinions on the subject, the courts have generally required more proof of collusion than simply parallel conduct.
Algorithmic collusion, in particular, does not fit easily into existing antitrust law because it deals with new technology and there isn’t much evidence to support its occurrence. Despite the theoretical possibility of algorithms learning to collude, these programs may not be sophisticated enough to achieve this yet. Experiments have shown failure to optimize the algorithms’ pricing since they only had their dataset to rely on, not the fact another algorithm was competing, leading to higher consumer prices but not necessarily learning to collude on their own.
Whether the cases in Las Vegas and Atlantic City provide real-world examples of algorithmic collusion is difficult to prove. As one University of Nevada Las Vegas professor recently noted, higher prices could be attributed to the rise in demand following the lifting of COVID-19 restrictions and the subsequent increase in travel to these areas. In addition, using algorithmic pricing has been shown to create consumer benefits through dynamic pricing and automatically responding to market changes. Assuming any algorithm price is suspect is a mistake that could ultimately cost consumers.
The FTC and DOJ are inserting themselves into these algorithmic pricing cases without regard to whether the current law is sufficient or whether there is proof of misconduct, potentially undermining technology that could have a positive impact on consumers.
