As previously announced, I am incredibly happy and honored to publish guest articles written by several of the world’s most renowned antitrust scholars every month of the year 2020. The one for June is authored by Scott Hemphill, Moses H. Grossman Professor of Law at NYU School of Law and co-director of the Engelberg Center on Innovation Law and Policy. In it, Scott discusses how to deal with uncertainty in the context of (killer) merger control. I am confident that you will enjoy reading it as much as I did. Scott, thank you very much!
All the best, Thibault Schrepel
Uncertain Harms: The Case of Nascent Competitors
Economics plays a central role in modern antitrust. When parties seek clearance for a merger, they offer sophisticated predictions about price, quality, and innovation. When decision-makers fashion a rule to govern a set of cases, they try to minimize costly errors—so-called false positives and false negatives—in selecting the best rule.
Yet economic thinking in antitrust has a blind spot. It struggles when confronted with uncertain or low probability competitive harms. It is intuitive, even obvious that, say, a 20 percent chance of a billion-dollar consumer harm is something we ought to avoid if we can. Avoiding a harm with a large expected value (here, $200 million) is a valid and even pressing goal for antitrust enforcement. If we ignore such harms, effectively discounting low probability harms to zero, we permit serious anticompetitive conduct to slip through the cracks.
Lately, I’ve been thinking about this problem in the context of acquisitions of nascent competitors—firms whose prospective innovation represents a serious future threat to an incumbent. For example, Instagram and WhatsApp were nascent competitors for Facebook at the time Facebook acquired them. A promising but unproven cure for a disease represents nascent competition for a therapy that is the current standard of care. Upstart nascent competitors are an important source of innovation, both in their own right and because they spur the incumbent to innovate in response. What to do about the elimination of nascent competition is therefore an urgent concern for antitrust authorities and commentators all over the world.
Tim Wu and I have a new paper, available here, that examines nascent competition as a distinct analytical category and considers what antitrust law can do about it. Here I want to focus on one piece of the puzzle, the uncertain nature of the harm. Nascent competitors often pose a uniquely potent threat to an entrenched incumbent. But the firm’s potency as a competitor is as yet not fully developed and hence unproven. The firm’s eventual significance is uncertain. This is particularly true in the environments of rapid technological change in which nascent threats tend to arise. That uncertainty, along with a lack of present, direct competition, may make enforcers and courts hesitant or unwilling to prevent an incumbent from acquiring or excluding a nascent threat. A hesitant enforcer might insist on strong proof that the competitor, if left alone, probably would have grown into a full-fledged rival, yet in so doing, neglect an important category of anticompetitive behavior.
To understand a deal that eliminates a nascent threat, the right starting point is the size—that is, the expected value—of the resulting harm. The expected value is a simple and familiar tool in economics and cost-benefit analysis. It has two components: the size of the benefits lost had the nascent threat grown into a viable competitor, multiplied by the likelihood of that outcome. No exact calculation is needed to recognize that the expected harm from the acquisition of Instagram might be very large, even if the probability of Instagram’s emergence as a successful competitor was modest. Moreover, the net harm from eliminating a nascent rival may remain high even after taking account of any merger-specific benefits of the deal.
There are of course costs of stopping these mergers, but it is important not to overstate them. Critics worry that the incumbent may uniquely incubate the startup or enjoy synergies that other buyers can’t match. To some degree, this point is correct; if these benefits are big enough, then we should allow the deal. But the claim that the firm most threatened by a startup is also a uniquely valuable patron is something that must be proved, not presumed. Critics also worry that preventing these deals will dry up an important source of funding for startups, but in fact most deals would be unaffected and for those that are, typically there are other buyers.
Other approaches to the problem are flawed. For example, one strand of U.S. merger law (unfortunately called “actual potential competition”) requires that the target “probably” would have successfully entered and competed with the incumbent. Similarly, the U.K.’s “balance of probabilities” approach requires that successful competition by the nascent rival was more probable than not. These narrow approaches allow anticompetitive conduct to go unchecked. They also fail to minimize error costs, by setting a rule in which the frequency of false negatives (harmful clearances) may be low but their size is large. A rule that ignores uncertain harms has a further problem—that it creates a perverse incentive for the incumbent to accelerate its anticompetitive conduct. The incumbent gets a free pass if it acts quickly enough.
Merger control can learn here from other areas of antitrust practice. One useful source of wisdom comes from horizontal agreements to settle patent litigation, in which the patentee pays the alleged infringer millions of dollars not to produce the patented product until the patent expires. This issue has arisen repeatedly in the pharmaceutical industry: brand name drug makers pay generics to abandon or delay their efforts to enter, thereby eliminating the threat of competition. Such settlements are anticompetitive because they reduce the expected amount of competition had the litigation proceeded to completion.
The harm from eliminating this competitive threat has been recognized by antitrust authorities and courts in many jurisdictions. Notably, there is a large consumer harm even if the probability of patent validity and infringement is high. If an early generic entry would bring $1 billion in consumer benefits, and was 20 percent likely to succeed, thwarting that prospective entry results in a large consumer harm. The U.S. Supreme Court has recognized this point, rejecting the argument that low probability harms should simply be ignored. That this claim was repeatedly pressed and came close to succeeding, is a further illustration of the blind spot.
Protecting nascent competition, even where the likelihood of successful entry is modest, is not a new idea. Indeed, it was central to one of the most important enforcement actions in the last twenty years, challenging Microsoft’s efforts to exclude competitive threats to Windows. The most important threat, the Netscape browser, had not developed into an operating system and might never have done so. Yet enforcers refused to grant the incumbent—in the memorable words of a U.S. appellate court—“free rei[n] to squash nascent, albeit unproven, competitors at will,” and proceeded despite the uncertain nature of the harm. A similarly ambitious approach to mergers is a worthy use of scarce enforcement resources, given the lack of alternative challengers to powerful incumbents and the importance of the innovation at stake.
Citation: C. Scott Hemphill, Uncertain Harms: The Case of Nascent Competitors, CONCURRENTIALISTE (June 16, 2020)
Read the other guest articles over here: link