Last week, on the first day of the trial in Epic Games’ antitrust lawsuit against Apple, Apple’s attorney made an argument that amounted to “we shouldn’t even be here.” The dispute between the two companies began last year when Epic intentionally broke Apple’s rules by setting up its own in-app payment system in the iOS version of its game Fortnite, getting banned from the App Store in the process. Thanks to that stunt—which included a very pointed parody of Apple’s famous “1984” ad that cast the once-rebellious device-maker as the modern internet’s Big Brother—the two companies are now embroiled in a dispute that could impact the future of the App Store, one of the most important, and lucrative, gateways in tech. If Epic prevails in the suit, it would likely be free to create its own app store on iOS devices, where it could distribute its games and apps from other developers. That would break Apple’s “walled garden,” siphon off millions of dollars of Apple’s profits, and change the iPhone as we know it.
But before that can happen, the court will have to think through an important but technical question. In her opening, Apple’s attorney, Karen Dunn, insisted that “every one of Epic’s claims depends on its ability to define a relevant market.” That, in a nutshell, is Apple’s first and most important line of defense against Epic. Over the course of the trial, which ends later this month, Apple will try to persuade the judge that there is no market applicable to this case in which it has a monopoly, let alone an illegal monopoly.
Epic alleges that Apple—through its control of the App Store—has a monopoly over both the distribution of apps on iOS devices and the processing of payments for in-app purchases. As a result of its monopoly power, according to Epic, Apple can and does act anticompetitively. Epic pointed out, for instance, that Apple required all app developers to pay it a 30 percent cut of the sale of any apps or in-app products. (After Epic filed its suit, Apple decided to reduce its commissions to 15 percent for app developers making less than $1 million a year on the App Store.)
Epic’s claims primarily fall under Section 2 of the Sherman Act, which makes it illegal to “monopolize.” The Supreme Court has read this statute to apply when a firm has “monopoly power in the relevant market” and “willfully” acquired or maintained that power “as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.” In other words, a firm violates Section 2 if it has a monopoly obtained though anticompetitive means rather than superior business skills.
The first question in Section 2 cases is often the definition of the “relevant market” for the defendant’s product. Once the market is defined, the court can calculate whether the defendant’s share of that market is monopolistic (which usually translates into controlling more than two-thirds of the market). Plaintiffs usually describe the relevant market narrowly, making it more likely that the defendant has a monopoly. Defendants usually take the opposite approach.
That’s exactly what’s happening in this case. Epic argues that the relevant market is limited to the market for apps on Apple devices. By definition, Apple has a monopoly in that market because it requires that all apps on iOS devices be distributed through the App Store. (Epic makes a similar argument about the relevant market for in-app purchases on Apple devices, which Apple also exclusively controls.)
By contrast, Apple sees the relevant market as anywhere Fortnite can be downloaded or played. That includes not only the App Store, but also Android phones, PCs, and video-gaming consoles. Apple describes this as the market for digital gaming transactions, and it doesn’t have close to monopoly power there. According to its experts’ pretrial submissions, Apple controls between a quarter and a third of all digital gaming transactions. Apple’s share is even smaller when it comes to Fortnite, which earns most of its revenues on PlayStation and Xbox, with Apple accounting for less than 10 percent.
Under the law, it’s not clear whose market definition is correct. In a pretrial ruling, Judge Yvonne Gonzalez Rogers noted that this case is at “the frontier edges of antitrust law in the United States,” and she was particularly puzzled by how to define the market. “There are at least three possible perspectives on the relevant market: (1) the customer who purchases the apps or games, (2) the developer who makes the apps or games, and (3) the competing app store or digital marketplace that distributes the apps or games.” How, she asked, is she supposed to choose between these perspectives, each of which might lead to a different market definition?
Apple’s theory “is a little bit closer to the way we think the world works,” said Herbert Hovenkamp, a law professor at the University of Pennsylvania and the author of the leading treatise on antitrust law. “When you talk about buying tires, you don’t think of tire sales at the Goodyear store or tire sales at Walmart. You think of the full range of buying opportunities that are out there. If Costco, Walmart, Goodyear, Firestone, and Sears all sell tires, that’s the range of competitive choices and we would say that the market consists of all of those places.”
By narrowing its market definition to apps on iOS devices, Epic is trying to define a “single-brand market.” Courts are generally reluctant to limit the relevant market to one company’s products, but it’s not unprecedented. For instance, in 1992 the Supreme Court found a single-brand market for the service and parts of Kodak photocopiers. Although Kodak had a small share of the photocopier market, “the plaintiffs won by saying, no, you’re right that there’s an overall market for photocopiers that’s much bigger but once you buy a photocopier, then you’re stuck with replacement parts that fit in your particular photocopier,” said Hovenkamp. Similarly, in Epic’s view, once consumers buy an iPhone, they are “locked in” to the iOS ecosystem and can access apps only through the App Store. Naturally, Apple disagrees, contending that there is no “lock-in” because consumers who purchase iPhones know very well that they are entering Apple’s “walled garden” and that they can switch to Android phones if they don’t like it.
Another case that could be instructive is United States v. Microsoft, which defined the relevant market as “Intel-compatible PC operating systems,” which only Microsoft produced at the time. “Microsoft made arguments similar to Apple’s—the market is any operating system on hand-held devices and other PCs—but the court bought this single-product market definition,” said Rebecca Haw Allensworth, a law professor at Vanderbilt who teaches a course on Big Tech and antitrust. “Once Microsoft had lost on this market definition issue, they had basically lost the case.”
How the court ultimately defines the relevant market will depend on arcane factual distinctions. Much of the trial so far, for example, has focused on the comparison of iPhones and gaming consoles. One of Epic’s witnesses—Lori Wright, a gaming executive at Microsoft—made the case last week that consoles are “special-purpose” devices dedicated only to games, unlike iPhones, which are “general-purpose” devices that are useful for all kinds of nongaming activities. “We certainly don’t view the iPhone as a competing device” to Microsoft’s Xbox, she said. This week, Epic’s expert, David Evans, is on the witness stand. He crafted Epic’s theory that the App Store is the relevant market, and he was even more adamant that iPhones don’t compete with gaming consoles, even with respect to games. He said that the “key distinction” is that iPhones, unlike consoles, can be used “anywhere, anytime.” He also pointed out that of the top 50 downloaded games on iOS, only four are also available on consoles. Apple’s opportunity to define its own market will come later this week, when it begins its case.
But the experts I spoke with said that focusing too much on the market definition obscures the underlying issue in this case: determining Apple’s market power and anticompetitive practices.
“I think the real question should be whether Apple has substantial market power, and it appears to have substantial power,” said Eleanor M. Fox, a law professor at New York University who specializes in antitrust and who has done some consulting work for Epic. “Litigating whether it should be called a monopolist should be recognized as a distraction from the real issue and a legal construction.”
“Defining a market isn’t what competition law is about. What’s required is showing market power.” Allensworth said. “I think it’s very clear that Apple has market power, and probably enough to say they have monopoly power.”
The custom of determining a firm’s market power by calculating its market share dates back many decades. “Historically, going to before World War II, we measured market power by defining a relevant market and then computing the share, usually as a percentage of output or sales. That is the classic, traditional way of how market power has been assessed,” said Hovenkamp. Judges continue to follow this approach, but it can be an imprecise barometer for market power. “Most of the judges who’ve done it, they’re not trained as economists or scientists at all. They kind of eyeball markets and say this looks like a market and they compute a share of it. It does tend to understate power to the extent that we tend to define markets too broadly.”
But more recently, economists have been able to apply sophisticated statistical methods to measure firms’ market power directly, bypassing the need for a strict market definition. Some antitrust scholars believe that the requirement to define a market should be dispensed with entirely.
“There’s a significant group of people out there that think that the market power issue in this case can be addressed more accurately without even defining a market but only by looking at price responses to cost changes,” said Hovenkamp. “The legal requirement points toward requiring a market, but the economics really suggests that a market definition is not necessary.”
“Market definition is only a construct of lawyers, not economists,” said Fox. “We want to find out if a firm has market power. To do that we need to know whether the firm acts under constraints that tend to push its prices down to a competitive price—cost plus a reasonable profit—or not. There are direct ways of doing that: What are the forces that tend to push prices down? What firms exert the pressure? Are they close or far away?” Fox added that the market-definition approach can be “totally misleading.” For instance, “having a large share in a market that is very easy to enter doesn’t even mean market power.”
The requirement that the relevant market be defined could end up being a huge boost for Apple’s case. If it persuades the judge of its market definition, Apple might be able to prevail even if Epic would otherwise be able to prove Apple’s market power and anticompetitive practices more directly. “[The market definition] is a first screen that must be passed before you go on to the next really important step, which is whether the conduct is anticompetitive and should be prohibited,” said Fox. “Thus at the first screen stage, we can end the analysis without even being able to consider the true merits of the case—whether the firm has engaged in conduct that we want to condemn.”
As the Epic case shows, digital markets aren’t easy to define, and fights over the relevant market will likely be a recurring theme in antitrust cases against Big Tech companies. Facebook is already challenging the government’s market definition in its own antitrust case against the social-media giant.
This is, obviously, how the large tech companies like it. “If courts insist that plaintiffs must prove a market in traditional technical neoclassical microeconomic ways,” said Fox, “none of the Big Tech firms would be caught by the Sherman Act.”
