Opinion
Architecture and Hardware Law and technology

What Role for Antitrust in Regulating Platforms?

Using regulation to protect competition and innovation.
Posted
  1. Introduction
  2. Exclusion
  3. Mergers and Acquisitions
  4. Copying
  5. Conclusion
  6. References
  7. Author
  8. Footnotes
Amazon, Google, and Facebook symbols, illustration

Google, Facebook, and Amazon stand accused of various tactics to thwart competition and protect their market position. The Justice Department, Federal Trade Commission (FTC), and others are investigating the platforms for possible antitrust violations. These developments call to mind the browser wars and Microsoft's legal battles in the 1990s and 2000s. They also raise an important question: What role does antitrust have to play this time around?

As I explain, antitrust law is highly relevant to some—but not all—of the critics' complaints. If Google uses customer contracts to weaken Bing, antitrust law can and should step in. Likewise if Facebook bought Instagram, back in 2012, to neutralize it as a competitive threat. More challenging are complaints about product design, such as a platform's arrangement of search results. Even further afield are concerns the platforms are copying their rivals' best ideas. If Amazon copies another seller's decision to market a particular product, no antitrust issue is raised, but copying combined with deception would raise serious concern.

Back to Top

Exclusion

Platforms enhance competition in several respects. One way is by competing in adjacent businesses beyond their "home" market.1 For example, Google's Android mobile operating system provides important competition for Apple. Platforms also level the playing field for others. Amazon's cloud computing services enable startups to scale up without investing in infrastructure, while Google's search ads help them connect with customers at low cost.

Moreover, competing hard is not an antitrust violation. For example, the leading platforms have made enormous investments in machine learning and related hardware. The result is a better algorithm, with more relevant ads and product recommendations, and an improved user interface such as digital assistants. These product improvements are competition in action, even if they make life difficult for competitors.


Contracts that weaken a rival's offering by limiting its access to customers are a major focus of platform antitrust enforcement.


An important antitrust problem arises when a platform, rather than competing on the merits, excludes a rival through its customer contracts. For example, during the browser wars, Microsoft's contracts with PC manufacturers restricted their promotion of the Netscape browser. Such contracts risked depriving Netscape of a critical mass of customers, thereby inhibiting its development into a Windows rival. Contracts that weaken a rival's offering by limiting its access to customers are a major focus of platform antitrust enforcement.

Critics see a replay of the browser wars in Google's contracts with Android handset manufacturers, which the European Commission has condemned as unduly restricting user access to competing search engines and browsers. One challenged provision had required the manufacturers to pre-install the Google search and Chrome browser apps as a condition of access to the Google Play Store, a must-have utility for Android users seeking third-party apps. The Commission was not persuaded by the fact that downloading the Bing app today is a lot easier than browser distribution in the 1990s. A second set of contracts pertains to Google's strong position as an intermediary between advertisers and ad inventory providers. Here, the concern is that Google might have used various contracts to weaken the competitive prospects of rival intermediaries.

Platforms can also exclude rivals without using contracts, by making design choices that favor their own related businesses over rivals. For example, starting in 2007, Google introduced specialized product search results alongside the traditional "10 blue links" on the search engine results page. These so-called "product universals" pointed to third-party merchants offering the product for sale. Critics allege that by promoting these product universals, Google undermined competing comparison-shopping search engines by pushing them lower in its results. At the same time, Google arguably improved its user experience by sending users more directly to product listings rather than indirectly to another search engine.

The overall competitive effect of a design change can be difficult to assess, particularly where a change constitutes a genuine product improvement yet also has a tendency to weaken rivals. One influential approach has been to ask: Does the profitability of the change depend on its tendency to weaken rivals? A change having both positive and negative effects is unlikely to fail this test. This lax attitude toward product design choices reflects a deep-seated reluctance to condemn product improvements.

Moreover, even where an adverse effect is likely, remedies can pose a further barrier to action. Antitrust enforcers routinely block enforcement of a problematic contract. They are much more reluctant to wade into the complex details of product design, particularly where doing so requires ongoing supervision of the design.

Back to Top

Mergers and Acquisitions

A second antitrust problem arises when a platform acquires its rival. For example, imagine Microsoft had bought Netscape instead of acting to exclude it by contract. (In fact, Microsoft approached Netscape about buying or licensing Netscape's browser code.) The harm would be similar, but without the conflict or complaining victim.

Platform acquisitions that thwart competition are illegal. If Facebook tried to buy Snap, the FTC would likely move to block the deal on concerns ranging from reduced privacy to higher ad prices and slower innovation. But some platform acquisitions are less clear cut. If Microsoft had bought Netscape, would the Justice Department have perceived a sufficiently clear competitive threat to challenge the deal?

Facebook's acquisition of Instagram presents just this issue. Instagram—today a major growth engine for Facebook—was in 2012 a fast-growing mobile photo-sharing app. Had Instagram stayed independent, it might have become a full-fledged competitor to Facebook. Acquisition targets sometimes become giant-killers—consider, for example, Yahoo's missed opportunity to buy Google. Instagram in particular had ample venture capital support, and commentators at the time recognized Instagram had found and exploited Facebook's struggles in mobile photo sharing. Facebook recognized Instagram as a rising threat, and indeed, a top company official reportedly wrote colleagues that the point of the deal was to eliminate a competitive threat.


Antitrust enforcement is a powerful tool for blocking problematic contracts and mergers, less so for questions of platform design.


Where a powerful platform such as Facebook is concerned, protecting competition is particularly important.2 The public benefits of competition are unusually great, given that Facebook faces limited competition within the market and few plausible challengers in the near term. Meanwhile, Facebook has the capacity and incentive to suppress competition, even from relatively long-shot rivals.

One broad lesson of the Microsoft litigation is that antitrust law protects "nascent, albeit unproven competitors … particularly in industries marked by rapid technological advance and frequent paradigm shifts."a Netscape was a nascent competitor; so was Instagram. As applied to acquisitions, protecting platform competition requires a willingness to block a deal even when we are unsure what will happen otherwise. It also sometimes means revisiting a consummated deal after the fact. For example, suppose Facebook indeed bought Instagram in order to eliminate a competitive threat, but failed to share its true intention with the FTC at the time. An after-the-fact antitrust suit to challenge the acquisition might be appropriate, once its purpose and effects are better understood.

Back to Top

Copying

A third complaint is that platforms copy the products and marketing decisions of smaller rivals. For example, Amazon has a knack for spotting profitable products sold on amazon.com—Duracell batteries, say—and introducing its own version. Antitrust's first response is likely to be, so what? Private-label goods are a familiar sight in supermarkets. The copyist's strategy does not depend on weakening the copied firm. Often, copying supports competition and lower prices. Thus, copying is ordinarily of little interest to antitrust.

Platform copying can be more complex than the supermarket example, due to the platform's participation in multiple related businesses and its informational advantage at the center of a web of sellers and customers. For example, Amazon runs a popular Marketplace service for third-party sellers. If Amazon observes a Marketplace seller is having success selling a particular toy, it might decide to sell the the same toy itself. But here, too, copying tends to intensify competition.

Though not ordinarily an antitrust issue, copying can raise important policy questions. In principle, platform appropriation might be so rapid, systematic, and effective as to discourage certain innovations. A Marketplace seller might not invest in marketing new products, lest Amazon appropriate the benefits. If Google repackages content from the New York Times, while keeping a sizable fraction of the associated advertising revenues, that might harm the incentive to gather news. When Facebook copies Snap's Stories feature, a sort of temporary slideshow, this might give pause to the next would-be Snap. The problem is compounded if the copying is one-way, with the platform always the pirate, never the pirated.

Platform copying becomes an antitrust issue when combined with platform deception, such as manipulation of search results. Suppose an Amazon customer is led to believe the top search result represents the best deal, when actually it delivers the greatest profit for Amazon. The FTC has not faced a case exactly like this. However, it has challenged deception as a competitive harm on multiple occasions. And outside the antitrust sphere, the FTC has repeatedly reminded search engines of their obligation to clearly distinguish ads from "natural" search results. Manipulation of product discovery, if proven, would be a problematic distortion of the competitive process.

Back to Top

Conclusion

This brief discussion of platform competition issues supports several conclusions. Antitrust enforcement is a powerful tool for blocking problematic contracts and mergers, less so for questions of platform design. That does not mean that design choices, if established to harm competition, should escape scrutiny. Instead, new rules or regulators may be called for. Beyond antitrust, copying may raise important questions of innovation policy. However, the crucial underlying empirical question—to what extent is innovation really being suppressed—remains unanswered.

    1. Hemphill, C.S. Disruptive incumbents: Platform competition in an age of machine learning. Columbia Law Review 119, 1973 (2019).

    2. Hemphill, C.S. and Wu, T. Nascent competitors. University of Pennsylvania Law Review (forthcoming 2020).

    a. United States v. Microsoft Corp., 253 F.3d 34, 79 (D.C. Cir. 2001) (en banc) (per curiam).

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More