Facebook can be a monopolist over a cluster of noncompeting products that do not fit the standard economic definition of a “market.” The key is to identify situations in which clustering non-substitute products itself creates market power.


In late 2020, the Federal Trade Commission brought an antitrust suit accusing Facebook of monopolization. Every antitrust case claiming an unlawful monopoly must identify a “market” that the defendant is monopolizing. Past defendants often produced a single, readily identifiable product such as aluminum ingot, cellophane, or Intel-based computer operating systems. Often, parties dispute the boundaries of these markets. For example, should the cellophane market be broadened to include wax paper and tin foil? Or should the operating system market in Microsoft be broadened to include Apple’s operating system?

Facebook has generated a different kind of dispute, however, which is that the grouping of products it offers is not a “market” at all. A fundamental proposition of economics since the nineteenth century is that markets are made up of close substitutes. Competition occurs inside a market because it defines the range of a customer’s choices. For example, we say that three gasoline stations in a two-block area are in the same market. Customers can choose among them, so they must compete for that customer’s business. A station fifty miles away is not in the same market if it is not a realistic option, nor is a grocer that is nearby but does not sell gasoline.

The FTC claims that Facebook monopolizes a market for “personal social networking services.” That includes services that are quite dissimilar, however. For example, Facebook offers general messaging, two-party chatting, posting of photographs and videos, discussion boards, a marketplace and digital advertising, and even a kind of dating platform. Facebook moved to dismiss the case by stating that the FTC “has not alleged a plausible relevant market.”  A similar issue is likely to arise in the Google antitrust case, as well as a potential future case against Amazon. To date, the lawsuits against Apple have focused mainly on its control of app sales through its Appstore.

So what binds Facebook’s diverse assembly of products into a “market”? Facebook’s individual services are clearly not close substitutes for one another. Further, many firms offer individual services that compete with one of Facebook’s services. For many of these, Facebook is not the biggest. For example, it is not the biggest messaging app, platform for hosting photos or videos, or even digital advertising platform. This is also true of Amazon, which has less-than-dominant market shares in most of the individual products that it sells, save ebooks.

How do you identify monopoly if a firm’s business involves a large number of non-competing services that do not satisfy the traditional economic definition of a single “market”? Courts have wrestled with this problem before by developing a theory of “cluster markets,” which I explore in a new paper. Banks, hospitals, retailers, and even patent portfolios operate in a variety of markets and have more power than they would if each of their individual products or services were treated separately. The key is to identify situations in which clustering non-substitute products itself creates market power.

“Banks, hospitals, retailers, and even patent portfolios operate in a variety of markets and have more power than they would if each of their individual products or services were treated separately.”

Market power is measured by comparing a firm’s price to its costs. A competitive firm is forced to charge a price close to its costs, but a firm with market power can profit by charging more. Two phenomena of clustering can increase market power. First, economies of scope, or joint costs, can make it cheaper for a firm to offer multiple products in combination. Second, combining individual products can increase the value that customers place on the overall product. While reducing costs or creating value are both good things, here we are not condemning a firm for that reason, but only inquiring whether clustering accounts for its power. Then we might want to pursue other harmful practices that market power enables.

As an example of joint costs, it is cheaper for Uber to add UberEats to its existing drivers and dispatch software than it is for a new firm to offer food delivery services separately. If that is true, then a firm that offered the two services together would have higher margins of price over cost than two different firms that offered the services separately. We can then speak of the cluster of Uber rides and UberEats as a “market” even though these two services do not compete with one another. That is, a customer typically wants one or the other, but not both at the same time. 

Alternatively, a firm can increase consumer value by combining complements, which are products that customers value more highly when they are used together. For example, being able to exchange messages and post photos or videos on the same platform might be more valuable than exchanging messages on one platform and posting pictures on another. The combination would attract a larger number of users, and the result would be that the firm could increase its advertising sales or other revenue producing activity.

A finding of increased market power would require not merely that the combination reduces costs or increases value, but also that attaining this status is something that others could not readily duplicate. This is where another feature of platforms, network externalities, comes in. Facebook’s value accrues from its large variety of services offered on the same platform, plus the fact that it has a very large installed base, which is something that users also independently value. That is, the value of being on Facebook increases as the number of other people on Facebook (2.7 billion active users as of April 2021) grows, and also as the variety of services that Facebook offers is larger.

These facts suggest two things: first, Facebook can be a monopolist over a cluster of noncompeting products that do not fit the standard economic definition of a “market.” Second, however, they suggest caution about remedies: antitrust law should not be used to destroy the value resulting from a profitable design. That makes breakups a perilous remedy and suggests alternatives such as compelled interoperability, injunctions against anticompetitive conduct, or more aggressive prohibitions of mergers.