Before the development of formal price theory in the early twentieth century, which included the invention of the theory of “perfect” competition, economists held an understanding of competition as active striving, leading to economic growth rather than to “optimality” in any static sense. In the twentieth century, Joseph Schumpeter would be a lonely voice reiterating and amplifying the conception of competition as a dynamic process. If competition is active and dynamic, the important form of competition is potential competition, often from sources we cannot easily foresee. If competition is active and dynamic, the market structure we observe today, or even the market structure we imagine will prevail in the future, is a poor guide to antitrust policy. We should focus instead on impediments, especially legal impediments, to entry, experiment, and contract. In this light, recent proposals to impose on Internet platforms “structural separations” defined in terms of existing technological boundaries are likely to restrict not enhance potential competition – just as such separations did in the twentieth century. Moreover, proposals to restrict platforms from diversifying, including through acquisition, are likely to mute the powerful potential competition that the large platforms represent to each other.
By Richard N. Langlois
It has long been conventional for economists to be taught the following story about the nature and history of competition. Once upon a time, economists like Adam Smith & Alfred Marshall held an oversimplified account in which many anonymous small sellers interacted with each other under conditions of perfect information. It was not until the early twentieth century that thinkers like Edward Chamberlin & Joan Robinson came along to correct this unrealistic view by creating the theory of “imperfect” competition. Like most fables, this one does not survive close scrutiny. In reality, Smith & Marshall offered no such account, but in fact saw competition in a much richer and more nuanced way than would twentieth-century price theory, including its “imperfect competition” variant. Moreover, far from having “corrected” the theory of perfect competition, it was twentieth-century price theory that, drawing on French precursors like Augustin Cournot, would actually invent the theory of perfect competition.
Using the lens of the formal price theory that developed over the course of the twentieth century, modern-day economists look back anachronistically at the origins of the Sherman Act and at the early-century debates that led to the Clayton and FTC Acts. Although the roots of later-century economic analysis were already visible, most contemporaries understood competition quite differently from today’s standard accounts. Thinkers later derided as laissez-faire economists, and often even their opponents, did not reason in terms of an efficient equilibrium of prices, quantities, and number of firms; rather, following Adam Smith, they had a dynamic view of competition as active striving. Alfred Marshall equated competition with “economic freedom and enterprise.”
Active competition could take place along many margins, including both innovation and entry. The result of active competition would be not an efficient allocation of resources at any moment but rather a dynamic process of economic growth. The key to a healthy economy was thus freedom of contract, which meant both freedom from legal restraint, especially restraint on entry, and the freedom to engage in innovative economic arrangements. In the formulation of the day, this could be assured if there was potential competition – if the door were left open for others to enter markets and to innovate along technological, organizational, and even contractual dimensions. Since competition consisted in actions not in the structure of the market, competition policy in this understanding ought to focus on whatever might stop competitive behavior.
As static formal price theory infiltrated antitrust reasoning in the years after World War II, beginning with the so-called Structure-Conduct-Performance paradigm, the importance of potential competition fell into eclipse. Among the few dissenters was Joseph Schumpeter, who upheld the older dynamic view of competition and expressed its importance with distinctive vigor. Competition, he insisted, “acts not only when in being but also when it is merely an ever-present threat. It disciplines before it attacks. The businessman feels himself to be in a competitive situation even if he is alone in his field or if, though not alone, he holds a position such that investigating government experts fail to see any effective competition between him and any other firms in the same or a neighboring field and in consequence conclude that his talk, under examination, about his competitive sorrows is all make-believe.”
Consider the famous turn-of-the century case against Microsoft. The litigants approached the case with two very different strategies. The government stuck to a focused script. It narrowly defined the relevant market as that for operating systems for Intel-compatible personal computers. Because Microsoft held some 95 percent of that market, it was a monopoly. The government then pressed its case that Microsoft’s contracting practices constituted anticompetitive exclusion that maintained its monopoly and thus violated Section 2 of the Sherman Act.
By contrast, Microsoft waged what could only be called a Schumpeterian defense. The company denied all charges, and it portrayed its position as that of a dynamic competitor in an ever-changing market, perennially besieged by threats ranging from the dimly perceptible to the radically unknown. “In the future,” one Microsoft executive was paraphrased as testifying, “users may simply plug their computers into cable outlets and get whatever programs cable providers offer. Small, handheld computing devices could wipe out the PC, just as the PC wiped out the mainframe.” A graphical exhibit depicted these threats, many of them in the form of question marks, impinging as arrows upon the company. This elicited titters from the courtroom, and the argument was widely mocked in the press. Needless to say, within a few years the twin general-purpose technologies of cloud computing and the smartphone had arisen to make Microsoft’s competitive sorrows of 1999 seem a lot less like alligator tears.
Although present-day antitrust policy continues to be informed principally by price theory, and thus to be focused on existing competitors in existing markets, there has arisen a new – or perhaps retro – challenger: the so-called Neo-Brandeisian view. Like the Schumpeterian approach, the Neo-Brandeisian approach rejects price theory as the sole guide – or perhaps as any guide – to assessing competition. Lina Khan, one of the principal exponents of this view, has argued for a return to “process and structure” in understanding markets.
Yet what she has in mind is quite different from the understanding of process and structure in Adam Smith, Alfred Marshall, or Joseph Schumpeter. Khan calls for a return to “the tradition of structural separations.” Like the Progressives of old, Khan laments the invention of general incorporation statutes, fondly recalling the restrictive state charters that “generally limited the size, scope, and duration of operations and steered business activity toward serving community purposes.” She recalls the country’s history with separation regimes in railroads, bank holding companies, television networks, and telecommunication carriers. And she suggests imposing a similar regime of separation on today’s Internet platforms, restricting them only to those activities that are truly technically integrated – that are “unique infrastructural assets” – and forbidding any that represent mere “commerce.”
A careful assessment of the Neo-Brandeisian position is beyond the scope of the present effort. But it is clear that a regime of separations is if anything the opposite of a Schumpeterian approach to antitrust. An essential feature of Schumpeterian competition is the tearing down of existing categories of business. The inevitable mutation of technologies and markets would transform “line-of-business restrictions from apparently sensible protections against unfair competition to archaic rules that impede competition and innovation.” As happened in the twentieth century, market segmentation would impede firms from creating and deploying capabilities across segment boundaries. As happened in the twentieth century, market segmentation would ultimately end up protecting incumbents from the genuinely new. Schumpeterian competition is not about maintaining the structure of competition. It is about destroying and replacing it.
Some Neo-Brandeisians announce themselves to be protective of innovation – even, in the case of Tim Wu, of putatively Schumpeterian innovation. But they generally mean this in an exceedingly narrow sense: antitrust should be concerned with preventing existing large firms from unilaterally excluding or buying up (small, new) competitors. In Kahn’s separation scheme, this would mean flatly outlawing self-preferencing: platforms themselves would be barred from competing with any firms that use the platform. The European Commission recently brought charges against Apple after complaints from Spotify that Apple was preferencing its own music-streaming apps over those of Spotify. In a very generous reading, one could see behind this a dynamic theory of leveraging. Maybe modern-day platforms are somehow trying to leverage themselves into large new markets. Perhaps by preferencing its music-streaming apps over those of Spotify, Apple is trying to monopolize the market for music streaming.
The problem with this view is that it misses the nature and sources of genuine competition in the modern Internet economy – and, indeed, in any economy. It misses the importance – and the true sources – of potential competition.
Market segmentation is an attempt to prevent the large platform firms from diversifying. But, especially in an industry underpinned by a general-purpose technology like cloud computing, the ability to diversify is what makes the Internet giants powerful competitors against one another. In the highly unlikely event Apple were to destroy Spotify, it would still have to deal with the likes Amazon and Google, which both have music-streaming services. The Big Tech firms already compete in a wide variety of overlapping niches. Much more significantly, however, these firms have developed underlying technological and market capabilities that, although by no means identical, are fundamentally similar. Computer engineers and data centers are flexible factors of production that can be easily redeployed when a new opportunity emerges. If one firm stumbles in any market niche, the others stand ready to pounce. The most disruptive form of new entry has often come not in the form of a small startup but of a large firm in a related area that can bring to bear the necessary complementary capabilities.
Is it not also true that Schumpeterian competition often comes from small firms with new ideas and capabilities? Many fear that today’s Big Tech firms are buying up startups that might otherwise have blossomed into potential competitors – so-called killer acquisitions. (An example frequently given is Facebook’s acquisition of Instagram and WhatsApp, even if possession of those properties has not stanched the flow of valuable younger users to platforms like Snapchat and TikTok.) Merger policy may be the area where antitrust policy has the greatest hope of militating in favor of dynamic competition. Yet even here caution is in order. Buying a startup with capabilities complementary to its own may allow an acquiring firm to compete more effectively against its other large rivals; and society may profit more from the synergies of acquisition than if the startup had been left to flounder on its own. Until antitrust policy made the purchase of outside technology difficult in the middle of the twentieth century, there had long existed a healthy market for independently developed ideas. Technology purchased on the market has featured importantly in the development of major systemic innovations like the iPhone and the Alexa voice-recognition system. Moreover, empirical evidence suggests that a well-functioning market for mergers and acquisitions creates an incentive for startups to innovate.
As either Niels Bohr or Yogi Berra once said, prediction is hard, especially about the future. That means that we can’t incorporate potential competition into our analysis merely by trying to guess what kind of competition the future will bring. We will get it wrong. But we can focus on a (Schumpeterian version of) process and structure. That may mean a greater attention to intellectual property rights than to antitrust. It might also mean thinking less about firms and markets and more about economic capabilities and where they come from. And, at the very least, it will mean being careful to avoid creating new institutional barriers to potential competition.
1 Department of Economics, the University of Connecticut.
2 See for example Huw Dixon, “A Brief History of Imperfect Competition,” CPI Antitrust Chronicle, October 2021. https://www.competitionpolicyinternational.com/a-brief-history-of-imperfect-competition/.
3 Brian J. Loasby, Choice, Complexity, and Ignorance, Cambridge 1976, pp. 173-192; Scott Moss, “The History of the Theory of the Firm from Marshall to Robinson and Chamberlin: The Source of Positivism in Economics,” Economica 51(203): 307-318 (1984).
4 Paul J. McNulty, “A Note on the History of Perfect Competition,” Journal of Political Economy 75(4): 395-399 (1967); Mary S. Morgan, “Competing Notions of ‘Competition’ in Late Nineteenth-Century American Economics,” History of Political Economy 25(4): 563-604 (1993).
5 Alfred Marshall, Principles of Economics, Eighth Edition. London 1920, VI.iii.2.
6 Benjamin J. Klebaner, “Potential Competition and the American Antitrust Legislation of 1914,” The Business History Review 38(2): 163-185 (1964).
7 Joseph A. Schumpeter, Capitalism, Socialism, and Democracy, New York 1950, p. 84.
8 A. Douglas Melamed & Daniel L. Rubinfeld, “U.S. v. Microsoft: Lessons Learned and Issues Raised,” in Eleanor M. Fox & Daniel A. Crane, eds., Antitrust Stories, New York 2007, pp. 287-311.
9 William Saletan, “Microsoft Plays Dead,” Slate, January 28, 1999.
10 Farhad Manjoo, “Steve Jobs Was Right: Smartphones and Tablets Killed the P.C.,” The New York Times, November 13, 2019.
11 Lina M. Khan, “Amazon’s Antitrust Paradox,” The Yale Law Journal 126(3): 710-805 (January 2017).
12 Lina M. Khan, “The Separation of Platforms and Commerce,” Columbia Law Review 119(4): 973-1098 (2019).
13 But see for example A. Douglas Melamed & Nicolas Petit, “The Misguided Assault on the Consumer Welfare Standard in the Age of Platform Markets,” Review of Industrial Organization 54(4): 741-774 (2019).
14 Richard J. Gilbert, “Separation: A Cure for Abuse of Platform Dominance?” Information Economics and Policy 54: 100876 (2021).
15 Tim Wu, “Taking Innovation Seriously: Antitrust Enforcement If Innovation Mattered Most,” Antitrust Law Journal 78(2): 313-328 (2012).
16 Sam Schechner, “EU Charges Apple with App Store Antitrust Violations in Spotify Case,” The Wall Street Journal, April 30, 2021.
17 Nicolas Petit, Big Tech and the Digital Economy: The Moligopoly Scenario. Oxford 2020; Hal Varian, “Seven Deadly Sins of Tech?” Information Economics and Policy 54: 100893 (2021).
18 David J. Teece, “Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing, and Public Policy,” Research Policy 15(6): 285-305 (1986).
19 So say internal Facebook documents. Keach Hagey & Jeff Horwitz, “Facebook Tried to Make Its Platform a Healthier Place. It Got Angrier Instead,” The Wall Street Journal, September 15, 2021.
20 David A. Hounshell, “The Evolution of Industrial Research in the United States,” in Richard S. Rosenbloom & William J. Spencer, eds., Engines of Innovation: U.S. Industrial Research at the End of an Era, Boston 1996, p. 26; David C. Mowery, “Plus Ça Change: Industrial R&D in the ‘Third Industrial Revolution,’” Industrial and Corporate Change 18(1): 1-50 (February 2009).
21 On the many startups Amazon needed to acquire and integrate for the Alexa project, see Brad Stone, Amazon Unbound: Jeff Bezos and the Invention of a Global Empire, New York 2021, pp. 21-53.
22 Gordon M. Phillips & Alexei Zhdanov, “R&D and the Incentives from Merger and Acquisition Activity,” The Review of Financial Studies 26(1): 34-78 (2013).