Algorithmic collusion appears to be spreading to more and more industries. And existing laws may not

RealPage, the rent price fixing software isn't the only algorithm out there providing collusion-like benefits to companies. Price competition is a vital component of free markets. This collusion-like behavior is bad for consumers and will destroy markets.

Because of how the laws around collusion are written, it is very difficult to prove. And the party that claims to be for free markets is so deep in the pockets of corporations, it is unlikely we will get new legislation making it easier to prevent this collusion-like behavior.

https://www.theatlantic.com/ideas/a..._campaign=the-atlantic&utm_content=edit-promo

The challenge is this: Under existing antitrust law, showing that companies A and B used algorithm C to raise prices isn’t enough; you need to show that there was some kind of agreement between companies A and B, and you need to allege some specific factual basis that the agreement existed before you can formally request evidence of it. This dynamic can place plaintiffs in a catch-22: Plausibly alleging the existence of a price-fixing agreement is hard to do without access to evidence like private emails, internal documents, or the algorithm itself. But they typically can’t uncover those kinds of materials until they are given the legal power to request evidence in discovery. “It’s like trying to fit a square peg in a round hole,” Richard Powers, a former deputy assistant attorney general in the DOJ antitrust division, told me. “It makes the job really hard.”

In the case of RealPage, the plaintiffs were able to make the peg fit. But in May, a Nevada judge dismissed a similar case against a group of Las Vegas hotels who used Rainmaker, concluding that there wasn’t enough evidence of a price-fixing agreement, because the hotels involved hadn’t shared confidential information with one another and weren’t required to accept Rainmaker’s recommendations, even if they allegedly did so about 90 percent of the time. “The rulings so far have set the bar very high,” Kenneth Racowski, a litigation attorney at Holland & Knight, told me. The RealPage case “was able to clear that bar, but it might prove to be the exception.”

And cases like RealPage and Rainmaker may be the easy ones. In a series of papers, Stucke and his fellow antitrust scholar Ariel Ezrachi have outlined ways in which algorithms could fix prices that would be even more difficult to prevent or prosecute—including situations in which an algorithm learns to fix prices withouts its creators or users intending it to. Something similar could occur even if companies used different third-party algorithms to set prices. They point to a recent study of German gas stations, which found that when one major player adopted a pricing algorithm, its margins didn’t budge, but when two major players adopted different pricing algorithms, the margins for both increased by 38 percent. “In situations like these, the algorithms themselves actually learn to collude with each other,” Stucke told me. “That could make it possible to fix prices at a scale that we’ve never seen.”

None of the situations Stucke and Ezrachi describe involve an explicit agreement, making them almost impossible to prosecute under existing antitrust laws. Price-fixing, in other words, has entered the algorithmic age, but the laws designed to prevent it have not kept up. Powers said he believes existing antitrust laws cover algorithmic collusion—but he worried that he might be wrong. “That's the thing that kept me up at night,” he said about his tenure at the Department of Justice. “The worry that all 100-plus years of case law on price-fixing could be circumvented by technology.”