Antitrust – Tools for the Fourth Industrial Revolution

[ARTICLE PDF]      

In what is being called the Fourth Industrial Revolution (4IR), the world is becoming increasingly digital, with a growing reliance on everything from electronic marketplaces and GPS guidance to emerging technologies such as artificial intelligence (AI), self-driving cars, the Internet of Things (IoT), and 3-D printing, among many others. This transformation is opening the door to new markets, innovative business models, and increased collaboration. It also raises new antitrust concerns that are likely to attract the attention of antitrust enforcers and find their way into private litigation.

In a connected, data-driven world, a number of observers are questioning how antitrust regulators should deal with the new competitive dynamics that technology creates. Over the past few years, a robust debate has emerged about the efficacy of past antitrust policies, including a call to reexamine the degree to which antitrust laws should police highly concentrated markets. This suggestion has stirred up controversy by resurrecting an older concept: analyzing competition enforcement through the lens of market structure rather than consumer welfare. Although a great deal of the attention has been focused on the largest consumer-facing technology firms, this stance could also affect a wider variety of markets and firms.

This point of view has not yet gained traction with U.S. regulators or courts or resulted in any immediate change in legal doctrine. “But it is a good example of how 4IR technology innovations—and the new business models, markets, and intermediaries they create—are reshaping the antitrust discussion,” says John Gibson, a partner in Crowell & Moring’s Antitrust Group and chair of the firm’s 3-D Printing Digital Transformation Working Group.

When Software Sets Prices

One of the topics that is receiving a great deal of attention in both the EU and the U.S. is the use of computer software to adjust prices in response to consumer or competitor activity, a practice referred to as algorithmic pricing. This concern dates back to 1993, when the Department of Justice brought an antitrust lawsuit against several travel industry participants for allegedly using a shared online reservation system to signal ticket prices to one another. But the issue has become more prominent as the proliferation of e-commerce and the growing sophistication of software make it easier for competitors to coordinate prices in real time.

In 2015, the DOJ filed its first e-commerce pricing algorithm-related lawsuit, in United States v. Topkins. There, an online seller of wall posters pleaded guilty to working with others to use software to coordinate prices for their products in an online marketplace, resulting in a $20,000 criminal fine. More recently, a federal class action price-fixing suit filed in the Southern District of New York alleged that a ride-sharing company conspired with its drivers to use the company’s pricing algorithm to set the prices charged to passengers. “They were saying that it was a hub-and-spoke conspiracy, where the drivers agreed to set prices together, rather than set their prices independently,” Gibson explains. That case was ordered into arbitration in March 2018, however, so the key questions it raises will not be sorted out publicly in court.

"Regulators and the courts are saying essentially that we’ve been using a reliable set of antitrust tools for more than 100 years and those can apply to algorithmic pricing." —John Gibson

 

The Court Weighs in on Suing Online Platforms The U.S. Supreme Court could  soon  decide  in  Apple Inc. v. Pepper which purchasers can pursue private antitrust challenges to the conduct of online platforms. Plaintiffs who purchased apps for Apple devices allege that Apple has inflated prices by (1) requiring that apps for its devices be sold only in its online store and (2) charging a commission to app developers—which they allegedly recoup through higher prices.  Apple argues that plaintiffs lack standing under the 1977 U.S. Supreme Court decision in Illinois Brick Co. v. Illinois, which held that only direct purchasers  can  sue  under the federal antitrust laws. That  is,  the  plaintiffs  here are “indirect” purchasers  because  they  are  customers of the app developers, not Apple. Plaintiffs rely on their direct purchases from Apple, which allegedly monopolizes the distribution of apps. At  the  November  26, 2018, oral argument, some justices expressed that this “closed system” may distinguish Illinois Brick. “The Court appears prepared to revisit the Illinois Brick doctrine and evaluate its application to online platforms,” says Crowell & Moring’s John Gibson. “The Court could also broadly define direct purchasers, thus exposing firms to greater antitrust liability.”

Regulators Stay the Course— For Now

Algorithmic pricing is on the radar of both the FTC and the DOJ. But so far, U.S. regulators have not seen a need to adjust their approach to antitrust enforcement. Their view appears to be that while algorithms can make it easier to collude, using them does not in itself constitute collusion. As former FTC Commissioner Maureen K. Ohlhausen explained in 2017, “Some of the concerns about algorithms are a bit alarmist. From an antitrust perspective, the expanding use of algorithms raises familiar issues that are well within the existing canon. An algorithm is a tool, and like any other tool, it can be put to either useful purposes or nefarious ends.”

“Regulators and the courts are saying essentially that we’ve been using a reliable set of antitrust tools for more than 100 years and those can apply to algorithmic pricing,” says Gibson. “That means that unless there are at least two people or two companies getting together and agreeing to do something anticompetitive—like set prices—there is no antitrust violation.” In short, unless there is evidence suggesting that the design and adoption of algorithms by rivals was the means used in a conscious effort to coordinate pricing, it is unlikely that mere reliance on algorithms to track market trends and inform unilateral pricing decisions will rise to the level of an antitrust violation. If the rule were otherwise, antitrust enforcement could inhibit the development of innovative, technology-driven ways of improving market efficiency.

With that point of view in mind, says Gibson, there will probably be minimal antitrust enforcement action from regulators around algorithms in the near future. But he adds one caveat: Congress has recently expressed concern about the size and power of big tech companies, data aggregators, and platforms, which could translate into growing scrutiny of algorithmic pricing. Meanwhile, he says, “it seems likely that whatever litigation we see in this area is going to come from private-sector plaintiffs and intermittent government intervention, in extreme cases.”

As so often happens, rapidly evolving technology may eventually prompt regulators to adopt new approaches. Well-established law differentiates between unlawful pricing decisions that reflect a conscious choice by rivals to coordinate and lawful decisions that reflect unilateral choices, even when they lead to parallel pricing. In time, AI software, rather than humans, may well write pricing algorithms—and in the pursuit of greater market efficiency, that software could conceivably design systems that fix prices among competitors without human intervention, and perhaps without humans even knowing it. “When that happens, it will be a watershed moment,” says Gibson. And it will raise numerous questions: Can machines conspire with each other for antitrust purposes? Who should be held responsible if machines are writing such algorithms? In that world, he says, “the old tools may no longer work anymore to identify collusion. New tools may have to be crafted, and the fire of litigation will probably help forge them.”

AmEx and the Online Platform In June 2018, the U.S. Supreme Court ruled that American Express’s anti-steering rules, which prevent merchants from promoting other payment cards to consumers at the point-of-sale, did not violate antitrust laws. In Ohio v. American Express, the Court held that AmEx was a “two-sided transaction platform” where a sale to one side of the platform cannot be made without a simultaneous sale to the other, and that antitrust claims in two-sided platforms must take into account how restraints on competition affect parties on both sides of the platform—here, merchants and cardholders. “You now have to look at the net competitive harm,” says Crowell & Moring’s John Gibson. “The Court said that when examined that way, AmEx’s relatively higher merchant fees may actually benefit merchants because these fees are used to fund AmEx’s rewards program, which in turn brings in affluent customers.” AmEx may have a significant impact on online platforms. “For example, ride-sharing platforms would likely be considered two-sided for antitrust purposes,” says Gibson. “They connect two groups that depend on the platform to process transactions, which occur simultaneously, and both sides benefit.” Other online platforms, which might connect advertisers and users, for example, would probably be considered one-sided, unless a simultaneous transaction occurs. These points are important, says Gibson, “because whether a court defines the relevant market as one-sided or two-sided will have significant ramifications for the plaintiff’s burden of proof on its theory of competitive harm and the burden on the defendants of coming forward with a pro-competitive justification.”