Those concerned over the size, apparent market dominance, and influence of widely-used internet platforms often focus on a “Big Four” featuring Apple, Google, Facebook, and Amazon. These and other tech companies have drawn politicized attacks from the left and the right, and regulatory action from the Federal Trade Commission (FTC) and Department of Justice (DOJ). Those leading the charge on this front within the Biden administration are part of a new intellectual movement in antitrust economics: New Brandeisians.
Legal scholar Lina Khan, among the movement’s thought leaders, now chairs the Federal Trade Commission. Her 2017 article “Amazon’s Antitrust Paradox” outlines the basic economic rationale cited by those concerned that the size of today’s leading internet platforms stifles market competition:
For the purpose of competition policy, one of the most relevant factors of online platform markets is that they are winner-take-all. This is due largely to network effects and control over data, both of which mean that early advantages become self-reinforcing. The result is that technology platform markets will yield to dominance by a small number of firms…Network effects arise when a user’s utility from a product increases as others use the product. Since popularity compounds and is reinforcing, markets with network effects often tip towards oligopoly or monopoly.
Published the same year as Khan (2017), antitrust economists David Evans and Richard Schmalensee distill both technical academic work and historical experience into a concise and convincing rejection of Khan’s basic argument:
Unfortunately, the simple network effects story leads to naïve armchair theories that industries with network effects are destined to be monopolies protected by insurmountable barriers to entry, and media-friendly slogans like “winner-take-all.”
The authors conclude that New Brandeisians have not caught up with mainstream economists’ more sophisticated understanding of network effects. They are correct, but this critique does not go far enough. The realities of internet platforms and associated technology have radically altered the competitive landscape and point toward an even stronger rejection of New Brandeisian thinking.
Telephones and VCRs
Evans and Schmalensee argue that economists’ “view of network effects evolved from a seminal economic contribution to a set of slogans that don’t comport with the facts.” Two of the first industries where economists identified and studied network effects, landline telephone service and VCRs, remain canonical examples of the phenomenon:
“A telephone was useless if nobody else had one. A telephone was more valuable if a user could reach more people. Economists called this phenomenon a direct network effect; the more people connected to a network, the more valuable that network is to each person who is part of it.”
VCRs illustrate the phenomenon of indirect network effects. Two incompatible technical standards (VHS and Betamax), “roughly comparable in cost and performance,” competed for consumers in the early market for VCRs. More consumers adopting a given standard incentivized sellers of video tapes to provide more offerings using that standard. The early industry is widely believed to have reached a tipping point in favor of VHS, which dominated the home movie market until the introduction of DVDs.
Antitrust concerns arise in cases where “winning” firms or technical standards reach a critical mass and become locked in. Potential new entrants must build large consumer bases to become competitive, a highly risky proposition for entrepreneurs and investors alike. The result is significant market power, where the incumbent can set high prices and leverage its power in markets for complimentary products. Both direct and indirect network effects provide opportunities for anticompetitive behavior by a dominant incumbent that further hinders entrants from becoming big.
Khan and fellow New Brandeisians such as Columbia University law professor Timothy Wu draw heavily from these basic early examples of direct and indirect network effects when they argue that online platforms are “winner take all” and market competition is an insufficient check on the power of winning firms.
Old Models and New Reality
Evans and Schmalensee survey later work by economists on network effects that call these basic stories into question. For example, users of a given online platform interact in many different ways, blurring the line between direct and indirect network effects and casting doubt on the idea that sheer size is a ticket to unstoppable market dominance. Facebook began as a platform specifically targeting college students. The restaurant reservation platform OpenTable succeeded when it began focusing on connecting diners and restaurants in specific cities. Evidently, there are many paths for new entrants to build large user bases, and much scope for dominant incumbents to fail to innovate and make strategic errors.
But one can go further than Evans and Schmalensee in criticizing the New Brandeisians’ applications of early network-effect theory to today’s online platforms. A fundamental change took place when “high tech” industries went from telephones and VCRs to e-commerce, social networking, and online search. In the former cases, market entry required large investments in physical capital, such as laying telephone lines and building factories. Similarly, consumers often faced large upfront hardware costs to “join a new network,” such as buying a new VCR or telephone.
The economics of internet platforms and many online businesses present a different competitive reality. Utilizing already-existing physical infrastructure (broadband and wireless data transmission) and user hardware (computers and smartphones), new entrants face vastly lower startup costs. Platform users face almost no upfront costs at all. In the cases of telephones and VCRs, the upfront hardware costs for consumers were so high relative to the benefits of adoption that economists often called them “switching costs.” In contrast, those reading this article may have windows currently open to Facebook, Google, and Zoom. They may switch between applications for the same function on a regular basis, or use them simultaneously for different purposes.
Note the last name on that list. Since 2017, New Brandeisians have maintained a steady drumbeat that Facebook and Google’s user bases would prevent innovation and new entry in applications already offered on their platforms, leaving users stuck with inferior services shielded from competition. During the same period, Zoom has gone from a mostly-unknown startup to a market leader in virtual meeting platforms, eclipsing offerings from Facebook and Google.
Evans and Schmalensee recognize the significance of reduced upfront user costs when they observe that “network effects can work in reverse.” They cite the now-famous list of once dominant platforms, such as Friendster and MySpace, that went from being portrayed as nearly unstoppable in the media to digital ghost towns in only a few years’ time. This “churn” in leading online platforms is indeed among the most salient critiques of the New Brandeisian antitrust approach.
Evolution Beats Intelligent Design
Network effects only “tip markets toward monopoly or oligopoly” when competitors and consumers face high up-front costs to creating and joining new networks. Many successful online ventures, including some of today’s members of the “big tech” club, began as much smaller projects by garage-sized startups and hobbyists. The primary threat of entry faced by today’s big-tech platforms is not from well-capitalized startups with business models nearly identical to the big players. The bigger threat comes from new innovators that dominant firms cannot identify and effectively fight off, often because such innovators do not yet realize they are that competitive threat.
This dramatically different type of competition stems from the radical uncertainty of a new and still-evolving business model. This perspective is more commonly associated with Austrian economics than the mathematical models and statistical analyses forming the basis of Evans and Schmalensee’s critiques. But combining these two ideas suggests network effects in today’s digital industries may actually fuel competition over time instead of stifling it.
The large stock of potential entrepreneurs and their ability to experiment and quickly pivot their business models to learn what consumers want and how to provide it fuels a learning process that leads to new ideas that eventually overtake the best guesses of even the sharpest big-tech CEOs.
Ever wonder why the brand names of so many of today’s tech giants have become words in common usage, such as Googling a topic, “friending” someone, or more recently, “zooming” one’s colleagues? Verbs for these platform services often did not exist before today’s large firms invented the services they provide. In most cases, these inventions were borne not from a single big idea, but learned in a process of experimenting, tinkering, and ultimately competing.
Online platforms grow and succeed through evolution rather than intelligent design. End results are not fully planned but far more robust for precisely this reason. Today’s giants benefitted from similar competitive processes, and given the difference in the way network effects interact with the digital world’s radically different cost structure, one struggles to find a reason the process will cease to happen.
The basic but somewhat outdated logic of the earliest network-effects industries studied by economists forms a central pillar of New Brandeisians’ aggressive stance toward big tech. On their own, mainstream antitrust economists like Evans and Schmalensee, as well as Austrian economists focused on dynamic innovation and entrepreneurship, each offer serious challenges to those who would break up today’s giants. Combining the ideas of both critics reveals the notion of “winner take all” in online platforms as unsound economic thinking.
The Promise of Entry
The novel competitive realities of online platforms and other e-commerce markets convincingly reject New Brandeisian thinking. But those who wish to see the behavior of big tech through rose-colored glasses or reject antitrust policy out of hand will also be disappointed. The highly important and still-evolving platform industry raises many questions, but seriously considering these questions requires dispensing with antitrust thinking that amounts to little more than applying ideas about 20th century industrial giants to 21st century tech giants.
The evolutionary process that yields a steady stream of new and unexpected challengers is far from unique to internet-era competition. One sees echoes of these ideas in Clyde Christensen’s “The Innovators’ Dilemma,” which provides numerous examples of disruptive technologies (like DVDs versus VCRs) that dominant firms using the old technology could neither foresee nor effectively compete with. The unique cost structure and rapid pace of change in online markets is not a new phenomenon, but one that is sped up to the point old models of competition no longer apply.
The process of rapid innovation and learning almost inevitably gives “winners” considerable market share for at least a short period of time. This, along with the unique capabilities of online platforms to serve consumers but also influence society, deserves careful thought. Antitrust economists used to speak of the idea that monopolists could be disciplined by the “threat of entry.” The uncertainty, fast-moving innovation, and large pool of ideas that characterize online platforms make new competition over time less a threat and more a promise.