The wave of populist antitrust that has been embraced by regulators and legislators in the United States, United Kingdom, European Union, and other jurisdictions rests on the assumption that currently dominant platforms occupy entrenched positions that only government intervention can dislodge. Following this view, Facebook will forever dominate social networking, Amazon will forever dominate cloud computing, Uber and Lyft will forever dominate ridesharing, and Amazon and Netflix will forever dominate streaming. This assumption of platform invincibility is so well-established that some policymakers advocate significant interventions without making any meaningful inquiry into whether a seemingly dominant platform actually exercises market power.
Yet this assumption is not supported by historical patterns in platform markets. It is true that network effects drive platform markets toward “winner-take-most” outcomes. But the winner is often toppled quickly and without much warning. There is no shortage of examples.
In 2007, a columnist in The Guardian observed that “it may already be too late for competitors to dislodge MySpace” and quoted an economist as authority for the proposition that “MySpace is well on the way to becoming … a natural monopoly.” About one year later, Facebook had overtaken MySpace “monopoly” in the social-networking market. Similarly, it was once thought that Blackberry would forever dominate the mobile-communications device market, eBay would always dominate the online e-commerce market, and AOL would always dominate the internet-service-portal market (a market that no longer even exists). The list of digital dinosaurs could go on.
All those tech leaders were challenged by entrants and descended into irrelevance (or reduced relevance, in eBay’s case). This occurred through the force of competition, not government intervention.
Why This Time is Probably Not Different
Given this long line of market precedents, current legislative and regulatory efforts to “restore” competition through extensive intervention in digital-platform markets require that we assume that “this time is different.” Just as that slogan has been repeatedly rebutted in the financial markets, so too is it likely to be rebutted in platform markets.
There is already supporting evidence.
In the cloud market, Amazon’s AWS now faces vigorous competition from Microsoft Azure and Google Cloud. In the streaming market, Amazon and Netflix face stiff competition from Disney+ and Apple TV+, just to name a few well-resourced rivals. In the social-networking market, Facebook now competes head-to-head with TikTok and seems to be losing. The market power once commonly attributed to leading food-delivery platforms such as Grubhub, UberEats, and DoorDash is implausible after persistent losses in most cases, and the continuous entry of new services into a rich variety of local and product-market niches.
Those who have advocated antitrust intervention on a fast-track schedule may remain unconvinced by these inconvenient facts. But the market is not.
Investors have already recognized Netflix’s vulnerability to competition, as reflected by a 35% fall in its stock price on April 20 and a decline of more than 60% over the past 12 months. Meta, Facebook’s parent, also experienced a reappraisal, falling more than 26% on Feb. 3 and more than 35% in the past 12 months. Uber, the pioneer of the ridesharing market, has declined by almost 50% over the past 12 months, while Lyft, its principal rival, has lost more than 60% of its value. These price freefalls suggest that antitrust populists may be pursuing solutions to a problem that market forces are already starting to address.
The Forgotten Curse of the Incumbent
For some commentators, the sharp downturn in the fortunes of the so-called “Big Tech” firms would not come as a surprise.
It has long been observed by some scholars and courts that a dominant firm “carries the seeds of its own destruction”—a phrase used by then-professor and later-Judge Richard Posner, writing in the University of Chicago Law Review in 1971. The reason: a dominant firm is liable to exhibit high prices, mediocre quality, or lackluster innovation, which then invites entry by more adept challengers. However, this view has been dismissed as outdated in digital-platform markets, where incumbents are purportedly protected by network effects and switching costs that make it difficult for entrants to attract users. Depending on the set of assumptions selected by an economic modeler, each contingency is equally plausible in theory.
The plunging values of leading platforms supplies real-world evidence that favors the self-correction hypothesis. It is often overlooked that network effects can work in both directions, resulting in a precipitous fall from market leader to laggard. Once users start abandoning a dominant platform for a new competitor, network effects operating in reverse can cause a “run for the exits” that leaves the leader with little time to recover. Just ask Nokia, the world’s leading (and seemingly unbeatable) smartphone brand until the Apple iPhone came along.
Market self-correction inherently outperforms regulatory correction: it operates far more rapidly and relies on consumer preferences to reallocate market leadership—a result perfectly consistent with antitrust’s mission to preserve “competition on the merits.” In contrast, policymakers can misdiagnose the competitive effects of business practices; are susceptible to the influence of private interests (especially those that are unable to compete on the merits); and often mispredict the market’s future trajectory. For Exhibit A, see the protracted antitrust litigation by the U.S. Department against IBM, which started in 1975 and ended in withdrawal of the suit in 1982. Given the launch of the Apple II in 1977, the IBM PC in 1981, and the entry of multiple “PC clones,” the forces of creative destruction swiftly displaced IBM from market leadership in the computing industry.
Regulators and legislators around the world have emphasized the urgency of taking dramatic action to correct claimed market failures in digital environments, casting aside prudential concerns over the consequences if any such failure proves to be illusory or temporary.
But the costs of regulatory failure can be significant and long-lasting. Markets must operate under unnecessary compliance burdens that are difficult to modify. Regulators’ enforcement resources are diverted, and businesses are barred from adopting practices that would benefit consumers. In particular, proposed breakup remedies advocated by some policymakers would undermine the scale economies that have enabled platforms to push down prices, an important consideration in a time of accelerating inflation.
Conclusion
The high concentration levels and certain business practices in digital-platform markets certainly raise important concerns as a matter of antitrust (as well as privacy, intellectual property, and other bodies of) law. These concerns merit scrutiny and may necessitate appropriately targeted interventions. Yet, any policy steps should be anchored in the factually grounded analysis that has characterized decades of regulatory and judicial action to implement the antitrust laws with appropriate care. Abandoning this nuanced framework for a blunt approach based on reflexive assumptions of market power is likely to undermine, rather than promote, the public interest in competitive markets.
The Senate Judiciary Committee is set to debate S. 2992, the American Innovation and Choice Online Act (or AICOA) during a markup session Thursday. If passed into law, the bill would force online platforms to treat rivals’ services as they would their own, while ensuring their platforms interoperate seamlessly.
The bill marks the culmination of misguided efforts to bring Big Tech to heel, regardless of the negative costs imposed upon consumers in the process. ICLE scholars have written about these developments in detail since the bill was introduced in October.
Below are 10 significant misconceptions that underpin the legislation.
1. There Is No Evidence that Self-Preferencing Is Generally Harmful
Self-preferencing is a normal part of how platforms operate, both to improve the value of their core products and to earn returns so that they have reason to continue investing in their development.
Platforms’ incentives are to maximize the value of their entire product ecosystem, which includes both the core platform and the services attached to it. Platforms that preference their own products frequently end up increasing the total market’s value by growing the share of users of a particular product. Those that preference inferior products end up hurting their attractiveness to users of their “core” product, exposing themselves to competition from rivals.
As Geoff Manne concludes, the notion that it is harmful (notably to innovation) when platforms enter into competition with edge providers is entirely speculative. Indeed, a range of studies show that the opposite is likely true. Platform competition is more complicated than simple theories of vertical discrimination would have it, and there is certainly no basis for a presumption of harm.
Consider a few examples from the empirical literature:
Li and Agarwal (2017) find that Facebook’s integration of Instagram led to a significant increase in user demand both for Instagram itself and for the entire category of photography apps. Instagram’s integration with Facebook increased consumer awareness of photography apps, which benefited independent developers, as well as Facebook.
Foerderer, et al. (2018) find that Google’s 2015 entry into the market for photography apps on Android created additional user attention and demand for such apps generally.
Cennamo, et al. (2018) find that video games offered by console firms often become blockbusters and expand the consoles’ installed base. As a result, these games increase the potential for all independent game developers to profit from their games, even in the face of competition from first-party games.
Finally, while Zhu and Liu (2018) is often held up as demonstrating harm from Amazon’s competition with third-party sellers on its platform, its findings are actually far from clear-cut. As co-author Feng Zhu noted in the Journal of Economics & Management Strategy: “[I]f Amazon’s entries attract more consumers, the expanded customer base could incentivize more third‐ party sellers to join the platform. As a result, the long-term effects for consumers of Amazon’s entry are not clear.”
2. Interoperability Is Not Costless
There are many things that could be interoperable, but aren’t. The reason not everything is interoperable is because interoperability comes with costs, as well as benefits. It may be worth letting different earbuds have different designs because, while it means we sacrifice easy interoperability, we gain the ability for better designs to be brought to market and for consumers to have choice among different kinds.
As Sam Bowman has observed, there are often costs that prevent interoperability from being worth the tradeoff, such as that:
It might be too costly to implement and/or maintain.
It might prescribe a certain product design and prevent experimentation and innovation.
It might add too much complexity and/or confusion for users, who may prefer not to have certain choices.
It might increase the risk of something not working, or of security breaches.
It might prevent certain pricing models that increase output.
It might compromise some element of the product or service that benefits specifically from not being interoperable.
In a market that is functioning reasonably well, we should be able to assume that competition and consumer choice will discover the desirable degree of interoperability among different products. If there are benefits to making your product interoperable that outweigh the costs of doing so, that should give you an advantage over competitors and allow you to compete them away. If the costs outweigh the benefits, the opposite will happen: consumers will choose products that are not interoperable.
In short, we cannot infer from the mere absence of interoperability that something is wrong, since we frequently observe that the costs of interoperability outweigh the benefits.
3. Consumers Often Prefer Closed Ecosystems
Digital markets could have taken a vast number of shapes. So why have they gravitated toward the very characteristics that authorities condemn? For instance, if market tipping and consumer lock-in are so problematic, why is it that new corners of the digital economy continue to emerge via closed platforms, as opposed to collaborative ones?
Indeed, if recent commentary is to be believed, it is the latter that should succeed, because they purportedly produce greater gains from trade. And if consumers and platforms cannot realize these gains by themselves, then we should see intermediaries step into that breach. But this does not seem to be happening in the digital economy.
The naïve answer is to say that the absence of “open” systems is precisely the problem. What’s harder is to try to actually understand why. As I have written, there are many reasons that consumers might prefer “closed” systems, even when they have to pay a premium for them.
Take the example of app stores. Maintaining some control over the apps that can access the store notably enables platforms to easily weed out bad players. Similarly, controlling the hardware resources that each app can use may greatly improve device performance. In other words, centralized platforms can eliminate negative externalities that “bad” apps impose on rival apps and on consumers. This is especially true when consumers struggle to attribute dips in performance to an individual app, rather than the overall platform.
It is also conceivable that consumers prefer to make many of their decisions at the inter-platform level, rather than within each platform. In simple terms, users arguably make their most important decision when they choose between an Apple or Android smartphone (or a Mac and a PC, etc.). In doing so, they can select their preferred app suite with one simple decision.
They might thus purchase an iPhone because they like the secure App Store, or an Android smartphone because they like the Chrome Browser and Google Search. Forcing too many “within-platform” choices upon users may undermine a product’s attractiveness. Indeed, it is difficult to create a high-quality reputation if each user’s experience is fundamentally different. In short, contrary to what antitrust authorities seem to believe, closed platforms might be giving most users exactly what they desire.
Too often, it is simply assumed that consumers benefit from more openness, and that shared/open platforms are the natural order of things. What some refer to as “market failures” may in fact be features that explain the rapid emergence of the digital economy. Ronald Coase said it best when he quipped that economists always find a monopoly explanation for things that they simply fail to understand.
4. Data Portability Can Undermine Security and Privacy
As explained above, platforms that are more tightly controlled can be regulated by the platform owner to avoid some of the risks present in more open platforms. Apple’s App Store, for example, is a relatively closed and curated platform, which gives users assurance that apps will meet a certain standard of security and trustworthiness.
Along similar lines, there are privacy issues that arise from data portability. Even a relatively simple requirement to make photos available for download can implicate third-party interests. Making a user’s photos more broadly available may tread upon the privacy interests of friends whose faces appear in those photos. Importing those photos to a new service potentially subjects those individuals to increased and un-bargained-for security risks.
As Sam Bowman and Geoff Manne observe, this is exactly what happened with Facebook and its Social Graph API v1.0, ultimately culminating in the Cambridge Analytica scandal. Because v1.0 of Facebook’s Social Graph API permitted developers to access information about a user’s friends without consent, it enabled third-party access to data about exponentially more users. It appears that some 270,000 users granted data access to Cambridge Analytica, from which the company was able to obtain information on 50 million Facebook users.
In short, there is often no simple solution to implement interoperability and data portability. Any such program—whether legally mandated or voluntarily adopted—will need to grapple with these and other tradeoffs.
5. Network Effects Are Rarely Insurmountable
Several scholars in recent years have called for more muscular antitrust intervention in networked industries on grounds that network externalities, switching costs, and data-related increasing returns to scale lead to inefficient consumer lock-in and raise entry barriers for potential rivals (see here, here, and here). But there are countless counterexamples where firms have easily overcome potential barriers to entry and network externalities, ultimately disrupting incumbents.
Zoom is one of the most salient instances. As I wrote in April 2019 (a year before the COVID-19 pandemic):
To get to where it is today, Zoom had to compete against long-established firms with vast client bases and far deeper pockets. These include the likes of Microsoft, Cisco, and Google. Further complicating matters, the video communications market exhibits some prima facie traits that are typically associated with the existence of network effects.
Geoff Manne and Alec Stapp have put forward a multitude of other examples, including: the demise of Yahoo; the disruption of early instant-messaging applications and websites; and MySpace’s rapid decline. In all of these cases, outcomes did not match the predictions of theoretical models.
More recently, TikTok’s rapid rise offers perhaps the greatest example of a potentially superior social-networking platform taking significant market share away from incumbents. According to the Financial Times, TikTok’s video-sharing capabilities and powerful algorithm are the most likely explanations for its success.
While these developments certainly do not disprove network-effects theory, they eviscerate the belief, common in antitrust circles, that superior rivals are unable to overthrow incumbents in digital markets. Of course, this will not always be the case. The question is ultimately one of comparing institutions—i.e., do markets lead to more or fewer error costs than government intervention? Yet, this question is systematically omitted from most policy discussions.
6. Profits Facilitate New and Exciting Platforms
As I wrote in August 2020, the relatively closed model employed by several successful platforms (notably Apple’s App Store, Google’s Play Store, and the Amazon Retail Platform) allows previously unknown developers/retailers to rapidly expand because (i) users do not have to fear their apps contain some form of malware and (ii) they greatly reduce payments frictions, most notably security-related ones.
While these are, indeed, tremendous benefits, another important upside seems to have gone relatively unnoticed. The “closed” business model also gives firms significant incentives to develop new distribution mediums (smart TVs spring to mind) and to improve existing ones. In turn, this greatly expands the audience that software developers can reach. In short, developers get a smaller share of a much larger pie.
The economics of two-sided markets are enlightening here. For example, Apple and Google’s app stores are what Armstrong and Wright (here and here) refer to as “competitive bottlenecks.” That is, they compete aggressively (among themselves, and with other gaming platforms) to attract exclusive users. They can then charge developers a premium to access those users.
This dynamic gives firms significant incentive to continue to attract and retain new users. For instance, if Steve Jobs is to be believed, giving consumers better access to media such as eBooks, video, and games was one of the driving forces behind the launch of the iPad.
This model of innovation would be seriously undermined if developers and consumers could easily bypass platforms, as would likely be the case under the American Innovation and Choice Online Act.
7. Large Market Share Does Not Mean Anticompetitive Outcomes
Scholars routinely cite the putatively strong concentration of digital markets to argue that Big Tech firms do not face strong competition. But this is a non sequitur. Indeed, as economists like Joseph Bertrand and William Baumol have shown, what matters is not whether markets are concentrated, but whether they are contestable. If a superior rival could rapidly gain user traction, that alone will discipline incumbents’ behavior.
Markets where incumbents do not face significant entry from competitors are just as consistent with vigorous competition as they are with barriers to entry. Rivals could decline to enter either because incumbents have aggressively improved their product offerings or because they are shielded by barriers to entry (as critics suppose). The former is consistent with competition, the latter with monopoly slack.
Similarly, it would be wrong to presume, as many do, that concentration in online markets is necessarily driven by network effects and other scale-related economies. As ICLE scholars have argued elsewhere (here, here and here), these forces are not nearly as decisive as critics assume (and it is debatable that they constitute barriers to entry).
Finally, and perhaps most importantly, many factors could explain the relatively concentrated market structures that we see in digital industries. The absence of switching costs and capacity constraints are two such examples. These explanations, overlooked by many observers, suggest digital markets are more contestable than is commonly perceived.
Unfortunately, critics’ failure to meaningfully grapple with these issues serves to shape the “conventional wisdom” in tech-policy debates.
8. Vertical Integration Generally Benefits Consumers
Vertical behavior of digital firms—whether through mergers or through contract and unilateral action—frequently arouses the ire of critics of the current antitrust regime. Many such critics point to a few recent studies that cast doubt on the ubiquity of benefits from vertical integration. But the findings of these few studies are regularly overstated and, even if taken at face value, represent a just minuscule fraction of the collected evidence, which overwhelmingly supports vertical integration.
There is strong and longstanding empirical evidence that vertical integration is competitively benign. This includes widely acclaimed work by economists Francine Lafontaine (former director of the Federal Trade Commission’s Bureau of Economics under President Barack Obama) and Margaret Slade, whose meta-analysis led them to conclude:
[U]nder most circumstances, profit-maximizing vertical integration decisions are efficient, not just from the firms’ but also from the consumers’ points of view. Although there are isolated studies that contradict this claim, the vast majority support it. Moreover, even in industries that are highly concentrated so that horizontal considerations assume substantial importance, the net effect of vertical integration appears to be positive in many instances. We therefore conclude that, faced with a vertical arrangement, the burden of evidence should be placed on competition authorities to demonstrate that that arrangement is harmful before the practice is attacked.
In short, there is a substantial body of both empirical and theoretical research showing that vertical integration (and the potential vertical discrimination and exclusion to which it might give rise) is generally beneficial to consumers. While it is possible that vertical mergers or discrimination could sometimes cause harm, the onus is on the critics to demonstrate empirically where this occurs. No legitimate interpretation of the available literature would offer a basis for imposing a presumption against such behavior.
9. There Is No Such Thing as Data Network Effects
Although data does not have the self-reinforcing characteristics of network effects, there is a sense that acquiring a certain amount of data and expertise is necessary to compete in data-heavy industries. It is (or should be) equally apparent, however, that this “learning by doing” advantage rapidly reaches a point of diminishing returns.
This is supported by significant empirical evidence. As was shown by the survey pf the empirical literature that Geoff Manne and I performed (published in the George Mason Law Review), data generally entails diminishing marginal returns:
Critics who argue that firms such as Amazon, Google, and Facebook are successful because of their superior access to data might, in fact, have the causality in reverse. Arguably, it is because these firms have come up with successful industry-defining paradigms that they have amassed so much data, and not the other way around. Indeed, Facebook managed to build a highly successful platform despite a large data disadvantage when compared to rivals like MySpace.
Companies need to innovate to attract consumer data or else consumers will switch to competitors, including both new entrants and established incumbents. As a result, the desire to make use of more and better data drives competitive innovation, with manifestly impressive results. The continued explosion of new products, services, and apps is evidence that data is not a bottleneck to competition, but a spur to drive it.
10. Antitrust Enforcement Has Not Been Lax
The popular narrative has it that lax antitrust enforcement has led to substantially increased concentration, strangling the economy, harming workers, and expanding dominant firms’ profit margins at the expense of consumers. Much of the contemporary dissatisfaction with antitrust arises from a suspicion that overly lax enforcement of existing laws has led to record levels of concentration and a concomitant decline in competition. But both beliefs—lax enforcement and increased anticompetitive concentration—wither under more than cursory scrutiny.
The number of Sherman Act cases brought by the federal antitrust agencies, meanwhile, has been relatively stable in recent years, but several recent blockbuster cases have been brought by the agencies and private litigants, and there has been no shortage of federal and state investigations. The vast majority of Section 2 cases dismissed on the basis of the plaintiff’s failure to show anticompetitive effect were brought by private plaintiffs pursuing treble damages; given the incentives to bring weak cases, it cannot be inferred from such outcomes that antitrust law is ineffective. But, in any case, it is highly misleading to count the number of antitrust cases and, using that number alone, to make conclusions about how effective antitrust law is. Firms act in the shadow of the law, and deploy significant legal resources to make sure they avoid activity that would lead to enforcement actions. Thus, any given number of cases brought could be just as consistent with a well-functioning enforcement regime as with an ill-functioning one.
The upshot is that naïvely counting antitrust cases (or the purported lack thereof), with little regard for the behavior that is deterred or the merits of the cases that are dismissed does not tell us whether or not antitrust enforcement levels are optimal.
Why do digital industries routinely lead to one company having a very large share of the market (at least if one defines markets narrowly)? To anyone familiar with competition policy discussions, the answer might seem obvious: network effects, scale-related economies, and other barriers to entry lead to winner-take-all dynamics in platform industries. Accordingly, it is that believed the first platform to successfully unlock a given online market enjoys a determining first-mover advantage.
This narrative has become ubiquitous in policymaking circles. Thinking of this sort notably underpins high-profile reports on competition in digital markets (here, here, and here), as well ensuing attempts to regulate digital platforms, such as the draft American Innovation and Choice Online Act and the EU’s Digital Markets Act.
But are network effects and the like the only ways to explain why these markets look like this? While there is no definitive answer, scholars routinely overlook an alternative explanation that tends to undercut the narrative that tech markets have become non-contestable.
The alternative model is simple: faced with zero prices and the almost complete absence of switching costs, users have every reason to join their preferred platform. If user preferences are relatively uniform and one platform has a meaningful quality advantage, then there is every reason to expect that most consumers will all join the same one—even though the market remains highly contestable. On the other side of the equation, because platforms face very few capacity constraints, there are few limits to a given platform’s growth. As will be explained throughout this piece, this intuition is as old as economics itself.
The Bertrand Paradox
In 1883, French mathematician Joseph Bertrand published a powerful critique of two of the most high-profile economic thinkers of his time: the late Antoine Augustin Cournot and Léon Walras (it would be another seven years before Alfred Marshall published his famous principles of economics).
Bertrand criticized several of Cournot and Walras’ widely accepted findings. This included Cournot’s conclusion that duopoly competition would lead to prices above marginal cost—or, in other words, that duopolies were imperfectly competitive.
By reformulating the problem slightly, Bertand arrived at the opposite conclusion. He argued that each firm’s incentive to undercut its rival would ultimately lead to marginal cost pricing, and one seller potentially capturing the entire market:
There is a decisive objection [to Cournot’s model]: According to his hypothesis, no [supracompetitive] equilibrium is possible. There is no limit to price decreases; whatever the joint price being charged by firms, a competitor could always undercut this price and, with few exceptions, attract all consumers. If the competitor is allowed to get away with this [i.e. the rival does not react], it will double its profits.
This result is mainly driven by the assumption that, unlike in Cournot’s model, firms can immediately respond to their rival’s chosen price/quantity. In other words, Bertrand implicitly framed the competitive process as price competition, rather than quantity competition (under price competition, firms do not face any capacity constraints and they cannot commit to producing given quantities of a good):
If Cournot’s calculations mask this result, it is because of a remarkable oversight. Referring to them as D and D’, Cournot deals with the quantities sold by each of the two competitors and treats them as independent variables. He assumes that if one were to change by the will of one of the two sellers, the other one could remain fixed. The opposite is evidently true.
This later came to be known as the “Bertrand paradox”—the notion that duopoly-market configurations can produce the same outcome as perfect competition (i.e., P=MC).
But while Bertrand’s critique was ostensibly directed at Cournot’s model of duopoly competition, his underlying point was much broader. Above all, Bertrand seemed preoccupied with the notion that expressing economic problems mathematically merely gives them a veneer of accuracy. In that sense, he was one of the first economists (at least to my knowledge) to argue that the choice of assumptions has a tremendous influence on the predictions of economic models, potentially rendering them unreliable:
On other occasions, Cournot introduces assumptions that shield his reasoning from criticism—scholars can always present problems in a way that suits their reasoning.
All of this is not to say that Bertrand’s predictions regarding duopoly competition necessarily hold in real-world settings; evidence from experimental settings is mixed. Instead, the point is epistemological. Bertrand’s reasoning was groundbreaking because he ventured that market structures are not the sole determinants of consumer outcomes. More broadly, he argued that assumptions regarding the competitive process hold significant sway over the results that a given model may produce (and, as a result, over normative judgements concerning the desirability of given market configurations).
The Theory of Contestable Markets
Bertrand is certainly not the only economist to have suggested market structures alone do not determine competitive outcomes. In the early 1980s, William Baumol (and various co-authors) went one step further. Baumol argued that, under certain conditions, even monopoly market structures could deliver perfectly competitive outcomes. This thesis thus rejected the Structure-Conduct-Performance (“SCP”) Paradigm that dominated policy discussions of the time.
Baumol’s main point was that industry structure is not the main driver of market “contestability,” which is the key determinant of consumer outcomes. In his words:
In the limit, when entry and exit are completely free, efficient incumbent monopolists and oligopolists may in fact be able to prevent entry. But they can do so only by behaving virtuously, that is, by offering to consumers the benefits which competition would otherwise bring. For every deviation from good behavior instantly makes them vulnerable to hit-and-run entry.
For instance, it is widely accepted that “perfect competition” leads to low prices because firms are price-takers; if one does not sell at marginal cost, it will be undercut by rivals. Observers often assume this is due to the number of independent firms on the market. Baumol suggests this is wrong. Instead, the result is driven by the sanction that firms face for deviating from competitive pricing.
In other words, numerous competitors are a sufficient, but not necessary condition for competitive pricing. Monopolies can produce the same outcome when there is a present threat of entry and an incumbent’s deviation from competitive pricing would be sanctioned. This is notably the case when there are extremely low barriers to entry.
Take this hypothetical example from the world of cryptocurrencies. It is largely irrelevant to a user whether there are few or many crypto exchanges on which to trade coins, nonfungible tokens (NFTs), etc. What does matter is that there is at least one exchange that meets one’s needs in terms of both price and quality of service. This could happen because there are many competing exchanges, or because a failure to meet my needs by the few (or even one) exchange that does exist would attract the entry of others to which I could readily switch—thus keeping the behavior of the existing exchanges in check.
This has far-reaching implications for antitrust policy, as Baumol was quick to point out:
This immediately offers what may be a new insight on antitrust policy. It tells us that a history of absence of entry in an industry and a high concentration index may be signs of virtue, not of vice. This will be true when entry costs in our sense are negligible.
Given what precedes, Baumol surmised that industry structure must be driven by endogenous factors—such as firms’ cost structures—rather than the intensity of competition that they face. For instance, scale economies might make monopoly (or another structure) the most efficient configuration in some industries. But so long as rivals can sanction incumbents for failing to compete, the market remains contestable. Accordingly, at least in some industries, both the most efficient and the most contestable market configuration may entail some level of concentration.
To put this last point in even more concrete terms, online platform markets may have features that make scale (and large market shares) efficient. If so, there is every reason to believe that competition could lead to more, not less, concentration.
How Contestable Are Digital Markets?
The insights of Bertrand and Baumol have important ramifications for contemporary antitrust debates surrounding digital platforms. Indeed, it is critical to ascertain whether the (relatively) concentrated market structures we see in these industries are a sign of superior efficiency (and are consistent with potentially intense competition), or whether they are merely caused by barriers to entry.
The barrier-to-entry explanation has been repeated ad nauseam in recent scholarly reports, competition decisions, and pronouncements by legislators. There is thus little need to restate that thesis here. On the other hand, the contestability argument is almost systematically ignored.
Several factors suggest that online platform markets are far more contestable than critics routinely make them out to be.
First and foremost, consumer switching costs are extremely low for most online platforms. To cite but a few examples: Changing your default search engine requires at most a couple of clicks; joining a new social network can be done by downloading an app and importing your contacts to the app; and buying from an alternative online retailer is almost entirely frictionless, thanks to intermediaries such as PayPal.
These zero or near-zero switching costs are compounded by consumers’ ability to “multi-home.” In simple terms, joining TikTok does not require users to close their Facebook account. And the same applies to other online services. As a result, there is almost no opportunity cost to join a new platform. This further reduces the already tiny cost of switching.
Decades of app development have greatly improved the quality of applications’ graphical user interfaces (GUIs), to such an extent that costs to learn how to use a new app are mostly insignificant. Nowhere is this more apparent than for social media and sharing-economy apps (it may be less true for productivity suites that enable more complex operations). For instance, remembering a couple of intuitive swipe motions is almost all that is required to use TikTok. Likewise, ridesharing and food-delivery apps merely require users to be familiar with the general features of other map-based applications. It is almost unheard of for users to complain about usability—something that would have seemed impossible in the early 21st century, when complicated interfaces still plagued most software.
A second important argument in favor of contestability is that, by and large, online platforms face only limited capacity constraints. In other words, platforms can expand output rapidly (though not necessarily costlessly).
Perhaps the clearest example of this is the sudden rise of the Zoom service in early 2020. As a result of the COVID pandemic, Zoom went from around 10 million daily active users in early 2020 to more than 300 million by late April 2020. Despite being a relatively data-intensive service, Zoom did not struggle to meet this new demand from a more than 30-fold increase in its user base. The service never had to turn down users, reduce call quality, or significantly increase its price. In short, capacity largely followed demand for its service. Online industries thus seem closer to the Bertrand model of competition, where the best platform can almost immediately serve any consumers that demand its services.
Conclusion
Of course, none of this should be construed to declare that online markets are perfectly contestable. The central point is, instead, that critics are too quick to assume they are not. Take the following examples.
Scholars routinely cite the putatively strong concentration of digital markets to argue that big tech firms do not face strong competition, but this is a non sequitur. As Bertrand and Baumol (and others) show, what matters is not whether digital markets are concentrated, but whether they are contestable. If a superior rival could rapidly gain user traction, this alone will discipline the behavior of incumbents.
Markets where incumbents do not face significant entry from competitors are just as consistent with vigorous competition as they are with barriers to entry. Rivals could decline to enter either because incumbents have aggressively improved their product offerings or because they are shielded by barriers to entry (as critics suppose). The former is consistent with competition, the latter with monopoly slack.
Similarly, it would be wrong to presume, as many do, that concentration in online markets is necessarily driven by network effects and other scale-related economies. As ICLE scholars have argued elsewhere (here, here and here), these forces are not nearly as decisive as critics assume (and it is debatable that they constitute barriers to entry).
Finally, and perhaps most importantly, this piece has argued that many factors could explain the relatively concentrated market structures that we see in digital industries. The absence of switching costs and capacity constraints are but two such examples. These explanations, overlooked by many observers, suggest digital markets are more contestable than is commonly perceived.
In short, critics’ failure to meaningfully grapple with these issues serves to shape the prevailing zeitgeist in tech-policy debates. Cournot and Bertrand’s intuitions about oligopoly competition may be more than a century old, but they continue to be tested empirically. It is about time those same standards were applied to tech-policy debates.
In recent years, a growing chorus of voices has argued that existing merger rules fail to apprehend competitively significant mergers, either because they fall below existing merger-filing thresholds or because they affect innovation in ways that are purportedly ignored.
These fears are particularly acute in the pharmaceutical and tech industries, where several high-profile academic articles and reports claim to have identified important gaps in current merger-enforcement rules, particularly with respect to acquisitions involving nascent and potential competitors (here, here, and here, among many others).
Such fears have led activists, lawmakers, and enforcers to call for tougher rules, including the introduction of more stringent merger-filing thresholds and other substantive changes, such as the inversion of the burden of proof when authorities review mergers and acquisitions involving digital platforms.
However, as we discuss in a recent working paper—forthcoming in the Missouri Law Review and available on SSRN—these proposals tend to overlook the important tradeoffs that would ensue from attempts to decrease the number of false positives under existing merger rules and thresholds.
The paper draws from two key strands of economic literature that are routinely overlooked (or summarily dismissed) by critics of the status quo.
For a start, antitrust enforcement is not costless. In the case of merger enforcement, not only is it expensive for agencies to detect anticompetitive deals but, more importantly, overbearing rules may deter beneficial merger activity that creates value for consumers.
Second, critics tend to overlook the possibility that incumbents’ superior managerial or other capabilities (i.e., what made them successful in the first place) makes them the ideal acquisition partners for entrepreneurs and startup investors looking to sell.
The result is a body of economic literature that focuses almost entirely on hypothetical social costs, while ignoring the redeeming benefits of corporate acquisitions, as well as the social cost of enforcement.
Kill Zones
One of the most significant allegations leveled against large tech firms is that their very presence in a market may hinder investments, entry, and innovation, creating what some have called a “kill zone.” The strongest expression in the economic literature of this idea of a kill zone stems from a working paper by Sai Krishna Kamepalli, Raghuram Rajan, and Luigi Zingales.
The paper makes two important claims, one theoretical and one empirical. From a theoretical standpoint, the authors argue that the prospect of an acquisition by a dominant platform deters consumers from joining rival platforms, and that this, in turn, hampers the growth of these rivals. The authors then test a similar hypothesis empirically. They find that acquisitions by a dominant platform—such as Google or Facebook—decrease investment levels and venture capital deals in markets that are “similar” to that of the target firm.
But both findings are problematic. For a start, Zingales and his co-authors’ theoretical model is premised on questionable assumptions about the way in which competition develops in the digital space. The first is that early adopters of new platforms—called “techies” in the authors’ parlance—face high switching costs because of their desire to learn these platforms in detail. As an initial matter, it would appear facially contradictory that “techies” both are the group with the highest switching costs and that they switch the most. The authors further assume that “techies” would incur lower adoption costs if they remained on the incumbent platform and waited for the rival platform to be acquired.
Unfortunately, while these key behavioral assumptions drive the results of the theoretical model, the paper presents no evidence to support their presence in real-world settings. In that sense, the authors commit the same error as previous theoretical work concerning externalities, which have tended to overestimate their frequency.
Second, the empirical analysis put forward in the paper is unreliable for policymaking purposes. The authors notably find that:
[N]ormalized VC investments in start-ups in the same space as the company acquired by Google and Facebook drop by over 40% and the number of deals falls by over 20% in the three years following an acquisition.
However, the results of this study are derived from the analysis of only nine transactions. The study also fails to clearly show that firms in the treatment and controls are qualitatively similar. In a nutshell, the study compares industry acquisitions exceeding $500 million to Facebook and Google’s acquisitions that exceed that amount. This does not tell us whether the mergers in both groups involved target companies with similar valuations or similar levels of maturity. This does not necessarily invalidate the results, but it does suggest that policymakers should be circumspect in interpreting those results.
Finally, the paper fails to demonstrate evidence that existing antitrust regimes fail to achieve an optimal error-cost balance. The central problem is that the paper has indeterminate welfare implications. For instance, as the authors note, the declines in investment in spaces adjacent to the incumbent platforms occurred during a time of rapidly rising venture capital investment, both in terms of the number of deals and dollars invested. It is entirely plausible that venture capital merely shifted to other sectors.
Put differently, on its own terms, the evidence merely suggests that acquisitions by Google and Facebook affected the direction of innovation, not its overall rate. And there is little to suggest that this shift was suboptimal, from a welfare standpoint.
In short, as the authors themselves conclude: “[i]t would be premature to draw any policy conclusion on antitrust enforcement based solely on our model and our limited evidence.”
Mergers and Potential Competition
Scholars have also posited more direct effects from acquisitions of startups or nascent companies by incumbent technology market firms.
Acquisitions of potential or nascent competitors by a dominant firm raise inherent anticompetitive concerns. By eliminating the procompetitive impact of the entry, an acquisition can allow the dominant firm to continue to exercise monopoly power and earn monopoly profits. The dominant firm also can neutralize the potential innovation competition that the entrant would provide.
However, these antitrust theories of harm suffer from several important flaws. They rest upon several restrictive assumptions that are not certain to occur in real-world settings. Most are premised on the notion that, in a given market, monopoly profits generally exceed joint duopoly profits. This allegedly makes it profitable, and mutually advantageous, for an incumbent to protect its monopoly position by preemptively acquiring potential rivals.
Accordingly, under these theories, anticompetitive mergers are only possible when the acquired rival could effectively challenge the incumbent. But these are, of course, only potential challengers; there is no guarantee that any one of them could or would mount a viable competitive threat.
Less obviously, it must be the case that the rival can hope to share only duopoly profits, as opposed to completely overthrowing the incumbent or surpassing them with a significantly larger share of the market. Where competition is “for the market” itself, monopoly maintenance would fail to explain a rival’s decision to sell. Because there would be no asymmetry between the expected profits of the incumbent and the rival, monopoly maintenance alone would not give rise to mutually advantageous deals.
Second, potential competition does not always increase consumer welfare. Indeed, while the presence of potential competitors might increase price competition, it can also have supply-side effects that cut in the opposite direction.
For example, as Nobel laureate Joseph Stiglitz observed, a monopolist threatened by potential competition may invest in socially wasteful R&D efforts or entry-deterrence mechanisms, and it may operate at below-optimal scale in anticipation of future competitive entry.
There are also pragmatic objections. Analyzing a merger’s effect on potential competition would compel antitrust authorities and courts to make increasingly speculative assessments concerning the counterfactual setting of proposed acquisitions.
In simple terms, it is far easier to determine whether a merger between McDonald’s and Burger King would lead to increased hamburger prices in the short run than it is to determine whether a gaming platform like Steam or the Epic Games Store might someday compete with video-streaming or music-subscription platforms like Netflix or Spotify. It is not that the above models are necessarily wrong, but rather that applying them to practical cases would require antitrust enforcers to estimate mostly unknowable factors.
Finally, the real test for regulators is not just whether they can identify possibly anticompetitive mergers, but whether they can do so in a cost-effective manner. Whether it is desirable to implement a given legal test is not simply a function of its accuracy, the cost to administer it, and the respective costs of false positives and false negatives. It also critically depends on how prevalent the conduct is that adjudicators would be seeking to foreclose.
Consider two hypothetical settings. Imagine there are 10,000 tech mergers in a given year, of which either 1,000 or 2,500 are anticompetitive (the remainder are procompetitive or competitively neutral). Suppose that authorities can either attempt to identify anticompetitive mergers with 75% accuracy, or perform no test at all—i.e., letting all mergers go through unchallenged.
If there are 1,000 anticompetitive mergers, applying the test would result in 7,500 correct decisions and 2,500 incorrect ones (2,250 false positives and 250 false negatives). Doing nothing would lead to 9,000 correct decisions and 1,000 false negatives. If the number of anticompetitive deals were 2,500, applying the test would lead to the same number of incorrect decisions as not applying it (1,875 false positives and 625 false negatives, versus 2,500 false negatives). The advantage would tilt toward applying the test if anticompetitive mergers were even more widespread.
This hypothetical example holds a simple lesson for policymakers: the rarer the conduct that they are attempting to identify, the more accurate their identification method must be, and the more costly false negatives must be relative to false positives.
As discussed below, current empirical evidence does not suggest that anticompetitive mergers of this sort are particularly widespread, nor does it offer accurate heuristics to detect the ones that are. Finally, there is little sense that the cost of false negatives significantly outweighs that of false positives. In short, there is currently little evidence to suggest that tougher enforcement would benefit consumers.
Killer Acquisitions
Killer acquisitions are, effectively, a subset of the “potential competitor” mergers discussed in the previous section. As defined by Colleen Cunningham, Florian Ederer, and Song Ma, they are those deals where “an incumbent firm may acquire an innovative target and terminate the development of the target’s innovations to preempt future competition.”
Cunningham, Ederer, and Ma’s highly influential paper on killer acquisitions has been responsible for much of the recent renewed interest in the effect that mergers exert on innovation. The authors studied thousands of pharmaceutical mergers and concluded that between 5.3% and 7.4% of them were killer acquisitions. As they write:
[W]e empirically compare development probabilities of overlapping acquisitions, which are, in our theory, motivated by a mix of killer and development intentions, and non-overlapping acquisitions, which are motivated only by development intentions. We find an increase in acquisition probability and a decrease in post-acquisition development for overlapping acquisitions and interpret that as evidence for killer acquisitions. […]
[W]e find that projects acquired by an incumbent with an overlapping drug are 23.4% less likely to have continued development activity compared to drugs acquired by non-overlapping incumbents.
From a policy standpoint, the question is what weight antitrust authorities, courts, and legislators should give to these findings. Stated differently, does the paper provide sufficient evidence to warrant reform of existing merger-filing thresholds and review standards? There are several factors counseling that policymakers should proceed with caution.
To start, the study’s industry-specific methodology means that it may not be a useful guide to understand acquisitions in other industries, like the tech sector, for example.
Second, even if one assumes that the findings of Cunningham, et al., are correct and apply with equal force in the tech sector (as some official reports have), it remains unclear whether the 5.3–7.4% of mergers they describe warrant a departure from the status quo.
Antitrust enforcers operate under uncertainty. The critical policy question is thus whether this subset of anticompetitive deals can be identified ex-ante. If not, is there a heuristic that would enable enforcers to identify more of these anticompetitive deals without producing excessive false positives?
The authors focus on the effect that overlapping R&D pipelines have on project discontinuations. In the case of non-overlapping mergers, acquired projects continue 17.5% of the time, while this number is 13.4% when there are overlapping pipelines. The authors argue that this gap is evidence of killer acquisitions. But it misses the bigger picture: under the authors’ own numbers and definition of a “killer acquisition,” a vast majority of overlapping acquisitions are perfectly benign; prohibiting them would thus have important social costs.
Third, there are several problems with describing this kind of behavior as harmful. Indeed, Cunningham, et al., acknowledge that this kind of behavior could increase innovation by boosting the returns to innovation.
And even if one ignores incentives to innovate, product discontinuations can improve consumer welfare. This question ultimately boils down to identifying the counterfactual to a merger. As John Yun writes:
For instance, an acquisition that results in a discontinued product is not per se evidence of either consumer harm or benefit. The answer involves comparing the counterfactual world without the acquisition with the world with the acquisition. The comparison includes potential efficiencies that were gained from the acquisition, including integration of intellectual property, the reduction of transaction costs, economies of scope, and better allocation of skilled labor.
One of the reasons R&D project discontinuation may be beneficial is simply cost savings. R&D is expensive. Pharmaceutical firms spend up to 27.8% of their annual revenue on R&D. Developing a new drug has an estimated median cost of $985.3 million. Cost-cutting—notably as it concerns R&D—is thus a critical part of pharmaceutical (as well as tech) companies’ businesses. As a report by McKinsey concludes:
The recent boom in M&A in the pharma industry is partly the result of attempts to address short-term productivity challenges. An acquiring or merging company typically designs organization-wide integration programs to capture synergies, especially in costs. Such programs usually take up to three years to complete and deliver results.
Maximizing the efficiency of production labor and equipment is one important way top-quartile drugmakers break out of the pack. Their rates of operational-equipment effectiveness are more than twice those of bottom-quartile companies (Exhibit 1), and when we looked closely we found that processes account for two-thirds of the difference.
In short, pharmaceutical companies do not just compete along innovation-related parameters, though these are obviously important, but also on more traditional grounds such as cost-rationalization. Accordingly, as the above reports suggest, pharmaceutical mergers are often about applying an incumbent’s superior managerial efficiency to the acquired firm’s assets through operation of the market for corporate control.
This cost-cutting (and superior project selection) ultimately enables companies to offer lower prices, thereby benefiting consumers and increasing their incentives to invest in R&D in the first place by making successfully developed drugs more profitable.
In that sense, Henry Manne’s seminal work relating to mergers and the market for corporate control sheds at least as much light on pharmaceutical (and tech) mergers as the killer acquisitions literature. And yet, it is hardly ever mentioned in modern economic literature on this topic.
While Colleen Cunningham and her co-authors do not entirely ignore these considerations, as we discuss in our paper, their arguments for dismissing them are far from watertight.
A natural extension of the killer acquisitions work is to question whether mergers of this sort also take place in the tech industry. Interest in this question is notably driven by the central role that digital markets currently occupy in competition-policy discussion, but also by the significant number of startup acquisitions that take place in the tech industry. However, existing studies provide scant evidence that killer acquisitions are a common occurrence in these markets.
This is not surprising. Unlike in the pharmaceutical industry—where drugs need to go through a lengthy and visible regulatory pipeline before they can be sold—incumbents in digital industries will likely struggle to identify their closest rivals and prevent firms from rapidly pivoting to seize new commercial opportunities. As a result, the basic conditions for killer acquisitions to take place (i.e., firms knowing they are in a position to share monopoly profits) are less likely to be present; it also would be harder to design research methods to detect these mergers.
The empirical literature on killer acquisitions in the tech sector is still in its infancy. But, as things stand, no study directly examines whether killer acquisitions actually take place in digital industries (i.e., whether post-merger project discontinuations are more common in overlapping than non-overlapping tech mergers). This is notably the case for studies by Axel Gautier & Joe Lamesch, and Elena Argentesi and her co-authors. Instead, these studies merely show that product discontinuations are common after an acquisition by a big tech company.
To summarize, while studies of this sort might suggest that the clearance of certain mergers might not have been optimal, it is hardly a sufficient basis on which to argue that enforcement should be tightened.
The reason for this is simple. The fact that some anticompetitive mergers may have escaped scrutiny and/or condemnation is never a sufficient basis to tighten rules. For that, it is also necessary to factor in the administrative costs of increased enforcement, as well as potential false convictions to which it might give rise. As things stand, economic research on killer acquisitions in the tech sector does not warrant tougher antitrust enforcement, though it does show the need for further empirical research on the topic.
Conclusion
Many proposed merger-enforcement reforms risk throwing the baby out with the bathwater. Mergers are largely beneficial to society (here, here and here); anticompetitive ones are rare; and there is little way, at the margin, to tell good from bad. To put it mildly, there is a precious baby that needs to be preserved and relatively little bathwater to throw out.
Take the fulcrum of policy debates that is the pharmaceutical industry. It is not hard to point to pharmaceutical mergers (or long-term agreements) that have revolutionized patient outcomes. Most recently, Pfizer and BioNTech’s efforts to successfully market an mRNA vaccine against COVID-19 offers a case in point.
The deal struck by both firms could naïvely be construed as bearing hallmarks of a killer acquisition or an anticompetitive agreement (long-term agreements can easily fall into either of these categories). Pfizer was a powerful incumbent in the vaccine industry; BioNTech threatened to disrupt the industry with new technology; and the deal likely caused Pfizer to forgo some independent R&D efforts. And yet, it also led to the first approved COVID-19 vaccine and groundbreaking advances in vaccine technology.
Of course, the counterfactual is unclear, and the market might be more competitive absent the deal, just as there might be only one approved mRNA vaccine today instead of two—we simply do not know. More importantly, this counterfactual was even less knowable at the time of the deal. And much the same could be said about countless other pharmaceutical mergers.
The key policy question is how authorities should handle this uncertainty. Critics of the status quo argue that current rules and thresholds leave certain anticompetitive deals unchallenged. But these calls for tougher enforcement fail to satisfy the requirements of the error-cost framework. Critics have so far failed to show that, on balance, mergers harm social welfare—even overlapping ones or mergers between potential competitors—just as they are yet to suggest alternative institutional arrangements that would improve social welfare.
In other words, they mistakenly analyze purported false negatives of merger-enforcement regimes in isolation. In doing so, they ignore how measures that aim to reduce such judicial errors may lead to other errors, as well as higher enforcement costs. In short, they paint a world where policy decisions involve facile tradeoffs, and this undermines their policy recommendations.
Given these significant limitations, this body of academic research should be met with an appropriate degree of caution. For all the criticism it has faced, the current merger-review system is mostly a resounding success. It is administrable, predictable, and timely. Yet it also eliminates a vast majority of judicial errors: even its critics concede that false negatives make up only a tiny fraction of decisions. Policymakers must decide whether the benefits from catching the very few arguably anticompetitive mergers that currently escape prosecution outweigh the significant costs that are required to achieve this goal. There is currently little evidence to suggest that this is, indeed, the case.
This post is authored by Nicolas Petit himself, the Joint Chair in Competition Law at the Department of Law at European University Institute in Fiesole, Italy, and at EUI’s Robert Schuman Centre for Advanced Studies. He is also invited professor at the College of Europe in Bruges.]
A lot of water has gone under the bridge since my book was published last year. To close this symposium, I thought I would discuss the new phase of antirust statutorification taking place before our eyes. In the United States, Congress is working on five antitrust bills that propose to subject platforms to stringent obligations, including a ban on mergers and acquisitions, required data portability and interoperability, and line-of-business restrictions. In the European Union (EU), lawmakers are examining the proposed Digital Markets Act (“DMA”) that sets out a complicated regulatory system for digital “gatekeepers,” with per se behavioral limitations of their freedom over contractual terms, technological design, monetization, and ecosystem leadership.
Proponents of legislative reform on both sides of the Atlantic appear to share the common view that ongoing antitrust adjudication efforts are both instrumental and irrelevant. They are instrumental because government (or plaintiff) losses build the evidence needed to support the view that antitrust doctrine is exceedingly conservative, and that legal reform is needed. Two weeks ago, antitrust reform activists ran to Twitter to point out that the U.S. District Court dismissal of the Federal Trade Commission’s (FTC) complaint against Facebook was one more piece of evidence supporting the view that the antitrust pendulum needed to swing. They are instrumental because, again, government (or plaintiffs) wins will support scaling antitrust enforcement in the marginal case by adoption of governmental regulation. In the EU, antitrust cases follow each other almost like night the day, lending credence to the view that regulation will bring much needed coordination and economies of scale.
But both instrumentalities are, at the end of the line, irrelevant, because they lead to the same conclusion: legislative reform is long overdue. With this in mind, the logic of lawmakers is that they need not await the courts, and they can advance with haste and confidence toward the promulgation of new antitrust statutes.
The antitrust reform process that is unfolding is a cause for questioning. The issue is not legal reform in itself. There is no suggestion here that statutory reform is necessarily inferior, and no correlative reification of the judge-made-law method. Legislative intervention can occur for good reason, like when it breaks judicial inertia caused by ideological logjam.
The issue is rather one of precipitation. There is a lot of learning in the cases. The point, simply put, is that a supplementary court-legislative dialogue would yield additional information—or what Guido Calabresi has called “starting points” for regulation—that premature legislative intervention is sweeping under the rug. This issue is important because specification errors (see Doug Melamed’s symposium piece on this) in statutory legislation are not uncommon. Feedback from court cases create a factual record that will often be missing when lawmakers act too precipitously.
Moreover, a court-legislative iteration is useful when the issues in discussion are cross-cutting. The digital economy brings an abundance of them. As tech analysist Ben Evans has observed, data-sharing obligations raise tradeoffs between contestability and privacy. Chapter VI of my book shows that breakups of social networks or search engines might promote rivalry and, at the same time, increase the leverage of advertisers to extract more user data and conduct more targeted advertising. In such cases, Calabresi said, judges who know the legal topography are well-placed to elicit the preferences of society. He added that they are better placed than government agencies’ officials or delegated experts, who often attend to the immediate problem without the big picture in mind (all the more when officials are denied opportunities to engage with civil society and the press, as per the policy announced by the new FTC leadership).
Of course, there are three objections to this. The first consists of arguing that statutes are needed now because courts are too slow to deal with problems. The argument is not dissimilar to Frank Easterbrook’s concerns about irreversible harms to the economy, though with a tweak. Where Easterbook’s concern was one of ossification of Type I errors due to stare decisis, the concern here is one of entrenchment of durable monopoly power in the digital sector due to Type II errors. The concern, however, fails the test of evidence. The available data in both the United States and Europe shows unprecedented vitality in the digital sector. Venture capital funding cruises at historical heights, fueling new firm entry, business creation, and economic dynamism in the U.S. and EU digital sectors, topping all other industries. Unless we require higher levels of entry from digital markets than from other industries—or discount the social value of entry in the digital sector—this should give us reason to push pause on lawmaking efforts.
The second objection is that following an incremental process of updating the law through the courts creates intolerable uncertainty. But this objection, too, is unconvincing, at best. One may ask which of an abrupt legislative change of the law after decades of legal stability or of an experimental process of judicial renovation brings more uncertainty.
Besides, ad hoc statutes, such as the ones in discussion, are likely to pose quickly and dramatically the problem of their own legal obsolescence. Detailed and technical statutes specify rights, requirements, and procedures that often do not stand the test of time. For example, the DMA likely captures Windows as a core platform service subject to gatekeeping. But is the market power of Microsoft over Windows still relevant today, and isn’t it constrained in effect by existing antitrust rules? In antitrust, vagueness in critical statutory terms allows room for change.[1] The best way to give meaning to buzzwords like “smart” or “future-proof” regulation consists of building in first principles, not in creating discretionary opportunities for permanent adaptation of the law. In reality, it is hard to see how the methods of future-proof regulation currently discussed in the EU creates less uncertainty than a court process.
The third objection is that we do not need more information, because we now benefit from economic knowledge showing that existing antitrust laws are too permissive of anticompetitive business conduct. But is the economic literature actually supportive of stricter rules against defendants than the rule-of-reason framework that applies in many unilateral conduct cases and in merger law? The answer is surely no. The theoretical economic literature has travelled a lot in the past 50 years. Of particular interest are works on network externalities, switching costs, and multi-sided markets. But the progress achieved in the economic understanding of markets is more descriptive than normative.
Take the celebrated multi-sided market theory. The main contribution of the theory is its advice to decision-makers to take the periscope out, so as to consider all possible welfare tradeoffs, not to be more or less defendant friendly. Payment cards provide a good example. Economic research suggests that any antitrust or regulatory intervention on prices affect tradeoffs between, and payoffs to, cardholders and merchants, cardholders and cash users, cardholders and banks, and banks and card systems. Equally numerous tradeoffs arise in many sectors of the digital economy, like ridesharing, targeted advertisement, or social networks. Multi-sided market theory renders these tradeoffs visible. But it does not come with a clear recipe for how to solve them. For that, one needs to follow first principles. A system of measurement that is flexible and welfare-based helps, as Kelly Fayne observed in her critical symposium piece on the book.
Another example might be worth considering. The theory of increasing returns suggests that markets subject to network effects tend to converge around the selection of a single technology standard, and it is not a given that the selected technology is the best one. One policy implication is that social planners might be justified in keeping a second option on the table. As I discuss in Chapter V of my book, the theory may support an M&A ban against platforms in tipped markets, on the conjecture that the assets of fringe firms might be efficiently repositioned to offer product differentiation to consumers. But the theory of increasing returns does not say under what conditions we can know that the selected technology is suboptimal. Moreover, if the selected technology is the optimal one, or if the suboptimal technology quickly obsolesces, are policy efforts at all needed?
Last, as Bo Heiden’s thought provoking symposium piece argues, it is not a given that antitrust enforcement of rivalry in markets is the best way to maintain an alternative technology alive, let alone to supply the innovation needed to deliver economic prosperity. Government procurement, science and technology policy, and intellectual-property policy might be equally effective (note that the fathers of the theory, like Brian Arthur or Paul David, have been very silent on antitrust reform).
There are, of course, exceptions to the limited normative content of modern economic theory. In some areas, economic theory is more predictive of consumer harms, like in relation to algorithmic collusion, interlocking directorates, or “killer” acquisitions. But the applications are discrete and industry-specific. All are insufficient to declare that the antitrust apparatus is dated and that it requires a full overhaul. When modern economic research turns normative, it is often way more subtle in its implications than some wild policy claims derived from it. For example, the emerging studies that claim to identify broad patterns of rising market power in the economy in no way lead to an implication that there are no pro-competitive mergers.
Similarly, the empirical picture of digital markets is incomplete. The past few years have seen a proliferation of qualitative research reports on industry structure in the digital sectors. Most suggest that industry concentration has risen, particularly in the digital sector. As with any research exercise, these reports’ findings deserve to be subject to critical examination before they can be deemed supportive of a claim of “sufficient experience.” Moreover, there is no reason to subject these reports to a lower standard of accountability on grounds that they have often been drafted by experts upon demand from antitrust agencies. After all, we academics are ethically obliged to be at least equally exacting with policy-based research as we are with science-based research.
Now, with healthy skepticism at the back of one’s mind, one can see immediately that the findings of expert reports to date have tended to downplay behavioral observations that counterbalance findings of monopoly power—such as intense business anxiety, technological innovation, and demand-expansion investments in digital markets. This was, I believe, the main takeaway from Chapter IV of my book. And less than six months ago, The Economist ran its leading story on the new marketplace reality of “Tech’s Big Dust-Up.”
Similarly, the expert reports did not really question the real possibility of competition for the purchase of regulation. As in the classic George Stigler paper, where the railroad industry fought motor-trucking competition with state regulation, the businesses that stand to lose most from the digital transformation might be rationally jockeying to convince lawmakers that not all business models are equal, and to steer regulation toward specific business models. Again, though we do not know how to consider this issue, there are signs that a coalition of large news corporations and the publishing oligopoly are behind many antitrust initiatives against digital firms.
Now, as is now clear from these few lines, my cautionary note against antitrust statutorification might be more relevant to the U.S. market. In the EU, sunk investments have been made, expectations have been created, and regulation has now become inevitable. The United States, however, has a chance to get this right. Court cases are the way to go. And unlike what the popular coverage suggests, the recent District Court dismissal of the FTC case far from ruled out the applicability of U.S. antitrust laws to Facebook’s alleged killer acquisitions. On the contrary, the ruling actually contains an invitation to rework a rushed complaint. Perhaps, as Shane Greenstein observed in his retrospective analysis of the U.S. Microsoft case, we would all benefit if we studied more carefully the learning that lies in the cases, rather than haste to produce instant antitrust analysis on Twitter that fits within 280 characters.
[1] But some threshold conditions like agreement or dominance might also become dated.
Interrogations concerning the role that economic theory should play in policy decisions are nothing new. Milton Friedman famously drew a distinction between “positive” and “normative” economics, notably arguing that theoretical models were valuable, despite their unrealistic assumptions. Kenneth Arrow and Gerard Debreu’s highly theoretical work on General Equilibrium Theory is widely acknowledged as one of the most important achievements of modern economics.
But for all their intellectual value and academic merit, the use of models to inform policy decisions is not uncontroversial. There is indeed a long and unfortunate history of influential economic models turning out to be poor depictions (and predictors) of real-world outcomes.
This raises a key question: should policymakers use economic models to inform their decisions and, if so, how? This post uses the economics of externalities to illustrate both the virtues and pitfalls of economic modeling. Throughout economic history, externalities have routinely been cited to support claims of market failure and calls for government intervention. However, as explained below, these fears have frequently failed to withstand empirical scrutiny.
Today, similar models are touted to support government intervention in digital industries. Externalities are notably said to prevent consumers from switching between platforms, allegedly leading to unassailable barriers to entry and deficient venture-capital investment. Unfortunately, as explained below, the models that underpin these fears are highly abstracted and far removed from underlying market realities.
Ultimately, this post argues that, while models provide a powerful way of thinking about the world, naïvely transposing them to real-world settings is misguided. This is not to say that models are useless—quite the contrary. Indeed, “falsified” models can shed powerful light on economic behavior that would otherwise prove hard to understand.
Bees
Fears surrounding economic externalities are as old as modern economics. For example, in the 1950s, economists routinely cited bee pollination as a source of externalities and, ultimately, market failure.
The basic argument was straightforward: Bees and orchards provide each other with positive externalities. Bees cross-pollinate flowers and orchards contain vast amounts of nectar upon which bees feed, thus improving honey yields. Accordingly, several famous economists argued that there was a market failure; bees fly where they please and farmers cannot prevent bees from feeding on their blossoming flowers—allegedly causing underinvestment in both. This led James Meade to conclude:
[T]he apple-farmer provides to the beekeeper some of his factors free of charge. The apple-farmer is paid less than the value of his marginal social net product, and the beekeeper receives more than the value of his marginal social net product.
If, then, apple producers are unable to protect their equity in apple-nectar and markets do not impute to apple blossoms their correct shadow value, profit-maximizing decisions will fail correctly to allocate resources at the margin. There will be failure “by enforcement.” This is what I would call an ownership externality. It is essentially Meade’s “unpaid factor” case.
It took more than 20 years and painstaking research by Steven Cheung to conclusively debunk these assertions. So how did economic agents overcome this “insurmountable” market failure?
The answer, it turns out, was extremely simple. While bees do fly where they please, the relative placement of beehives and orchards has a tremendous impact on both fruit and honey yields. This is partly because bees have a very limited mean foraging range (roughly 2-3km). This left economic agents with ample scope to prevent free-riding.
Using these natural sources of excludability, they built a web of complex agreements that internalize the symbiotic virtues of beehives and fruit orchards. To cite Steven Cheung’s research:
Pollination contracts usually include stipulations regarding the number and strength of the colonies, the rental fee per hive, the time of delivery and removal of hives, the protection of bees from pesticide sprays, and the strategic placing of hives. Apiary lease contracts differ from pollination contracts in two essential aspects. One is, predictably, that the amount of apiary rent seldom depends on the number of colonies, since the farmer is interested only in obtaining the rent per apiary offered by the highest bidder. Second, the amount of apiary rent is not necessarily fixed. Paid mostly in honey, it may vary according to either the current honey yield or the honey yield of the preceding year.
But what of neighboring orchards? Wouldn’t these entail a more complex externality (i.e., could one orchard free-ride on agreements concluded between other orchards and neighboring apiaries)? Apparently not:
Acknowledging the complication, beekeepers and farmers are quick to point out that a social rule, or custom of the orchards, takes the place of explicit contracting: during the pollination period the owner of an orchard either keeps bees himself or hires as many hives per area as are employed in neighboring orchards of the same type. One failing to comply would be rated as a “bad neighbor,” it is said, and could expect a number of inconveniences imposed on him by other orchard owners. This customary matching of hive densities involves the exchange of gifts of the same kind, which apparently entails lower transaction costs than would be incurred under explicit contracting, where farmers would have to negotiate and make money payments to one another for the bee spillover.
In short, not only did the bee/orchard externality model fail, but it failed to account for extremely obvious counter-evidence. Even a rapid flip through the Yellow Pages (or, today, a search on Google) would have revealed a vibrant market for bee pollination. In short, the bee externalities, at least as presented in economic textbooks, were merely an economic “fable.” Unfortunately, they would not be the last.
The Lighthouse
Lighthouses provide another cautionary tale. Indeed, Henry Sidgwick, A.C. Pigou, John Stuart Mill, and Paul Samuelson all cited the externalities involved in the provision of lighthouse services as a source of market failure.
Here, too, the problem was allegedly straightforward. A lighthouse cannot prevent ships from free-riding on its services when they sail by it (i.e., it is mostly impossible to determine whether a ship has paid fees and to turn off the lighthouse if that is not the case). Hence there can be no efficient market for light dues (lighthouses were seen as a “public good”). As Paul Samuelson famously put it:
Take our earlier case of a lighthouse to warn against rocks. Its beam helps everyone in sight. A businessman could not build it for a profit, since he cannot claim a price from each user. This certainly is the kind of activity that governments would naturally undertake.
He added that:
[E]ven if the operators were able—say, by radar reconnaissance—to claim a toll from every nearby user, that fact would not necessarily make it socially optimal for this service to be provided like a private good at a market-determined individual price. Why not? Because it costs society zero extra cost to let one extra ship use the service; hence any ships discouraged from those waters by the requirement to pay a positive price will represent a social economic loss—even if the price charged to all is no more than enough to pay the long-run expenses of the lighthouse.
More than a century after it was first mentioned in economics textbooks, Ronald Coase finally laid the lighthouse myth to rest—rebutting Samuelson’s second claim in the process.
What piece of evidence had eluded economists for all those years? As Coase observed, contemporary economists had somehow overlooked the fact that large parts of the British lighthouse system were privately operated, and had been for centuries:
[T]he right to operate a lighthouse and to levy tolls was granted to individuals by Acts of Parliament. The tolls were collected at the ports by agents (who might act for several lighthouses), who might be private individuals but were commonly customs officials. The toll varied with the lighthouse and ships paid a toll, varying with the size of the vessel, for each lighthouse passed. It was normally a rate per ton (say 1/4d or 1/2d) for each voyage. Later, books were published setting out the lighthouses passed on different voyages and the charges that would be made.
In other words, lighthouses used a simple physical feature to create “excludability” and prevent free-riding. The main reason ships require lighthouses is to avoid hitting rocks when they make their way to a port. By tying port fees and light dues, lighthouse owners—aided by mild government-enforced property rights—could easily earn a return on their investments, thus disproving the lighthouse free-riding myth.
Ultimately, this meant that a large share of the British lighthouse system was privately operated throughout the 19th century, and this share would presumably have been more pronounced if government-run “Trinity House” lighthouses had not crowded out private investment:
The position in 1820 was that there were 24 lighthouses operated by Trinity House and 22 by private individuals or organizations. But many of the Trinity House lighthouses had not been built originally by them but had been acquired by purchase or as the result of the expiration of a lease.
Of course, this system was not perfect. Some ships (notably foreign ones that did not dock in the United Kingdom) might free-ride on this arrangement. It also entailed some level of market power. The ability to charge light dues meant that prices were higher than the “socially optimal” baseline of zero (the marginal cost of providing light is close to zero). Though it is worth noting that tying port fees and light dues might also have decreased double marginalization, to the benefit of sailors.
Samuelson was particularly weary of this market power that went hand in hand with the private provision of public goods, including lighthouses:
Being able to limit a public good’s consumption does not make it a true-blue private good. For what, after all, are the true marginal costs of having one extra family tune in on the program? They are literally zero. Why then prevent any family which would receive positive pleasure from tuning in on the program from doing so?
However, as Coase explained, light fees represented only a tiny fraction of a ship’s costs. In practice, they were thus unlikely to affect market output meaningfully:
[W]hat is the gain which Samuelson sees as coming from this change in the way in which the lighthouse service is financed? It is that some ships which are now discouraged from making a voyage to Britain because of the light dues would in future do so. As it happens, the form of the toll and the exemptions mean that for most ships the number of voyages will not be affected by the fact that light dues are paid. There may be some ships somewhere which are laid up or broken up because of the light dues, but the number cannot be great, if indeed there are any ships in this category.
Samuelson’s critique also falls prey to the Nirvana Fallacy pointed out by Harold Demsetz: markets might not be perfect, but neither is government intervention. Market power and imperfect appropriability are the two (paradoxical) pitfalls of the first; “white elephants,” underinvestment, and lack of competition (and the information it generates) tend to stem from the latter.
Which of these solutions is superior, in each case, is an empirical question that early economists had simply failed to consider—assuming instead that market failure was systematic in markets that present prima facie externalities. In other words, models were taken as gospel without any circumspection about their relevance to real-world settings.
The Tragedy of the Commons
Externalities were also said to undermine the efficient use of “common pool resources,” such grazing lands, common irrigation systems, and fisheries—resources where one agent’s use diminishes that of others, and where exclusion is either difficult or impossible.
The most famous formulation of this problem is Garret Hardin’s highly influential (over 47,000 cites) “tragedy of the commons.” Hardin cited the example of multiple herdsmen occupying the same grazing ground:
The rational herdsman concludes that the only sensible course for him to pursue is to add another animal to his herd. And another; and another … But this is the conclusion reached by each and every rational herdsman sharing a commons. Therein is the tragedy. Each man is locked into a system that compels him to increase his herd without limit—in a world that is limited. Ruin is the destination toward which all men rush, each pursuing his own best interest in a society that believes in the freedom of the commons.
In more technical terms, each economic agent purportedly exerts an unpriced negative externality on the others, thus leading to the premature depletion of common pool resources. Hardin extended this reasoning to other problems, such as pollution and allegations of global overpopulation.
Although Hardin hardly documented any real-world occurrences of this so-called tragedy, his policy prescriptions were unequivocal:
The most important aspect of necessity that we must now recognize, is the necessity of abandoning the commons in breeding. No technical solution can rescue us from the misery of overpopulation. Freedom to breed will bring ruin to all.
As with many other theoretical externalities, empirical scrutiny revealed that these fears were greatly overblown. In her Nobel-winning work, Elinor Ostrom showed that economic agents often found ways to mitigate these potential externalities markedly. For example, mountain villages often implement rules and norms that limit the use of grazing grounds and wooded areas. Likewise, landowners across the world often set up “irrigation communities” that prevent agents from overusing water.
Along similar lines, Julian Morris and I conjecture that informal arrangements and reputational effects might mitigate opportunistic behavior in the standard essential patent industry.
These bottom-up solutions are certainly not perfect. Many common institutions fail—for example, Elinor Ostrom documents several problematic fisheries, groundwater basins and forests, although it is worth noting that government intervention was sometimes behind these failures. To cite but one example:
Several scholars have documented what occurred when the Government of Nepal passed the “Private Forest Nationalization Act” […]. Whereas the law was officially proclaimed to “protect, manage and conserve the forest for the benefit of the entire country”, it actually disrupted previously established communal control over the local forests. Messerschmidt (1986, p.458) reports what happened immediately after the law came into effect:
Nepalese villagers began freeriding — systematically overexploiting their forest resources on a large scale.
In any case, the question is not so much whether private institutions fail, but whether they do so more often than government intervention. be it regulation or property rights. In short, the “tragedy of the commons” is ultimately an empirical question: what works better in each case, government intervention, propertization, or emergent rules and norms?
More broadly, the key lesson is that it is wrong to blindly apply models while ignoring real-world outcomes. As Elinor Ostrom herself put it:
The intellectual trap in relying entirely on models to provide the foundation for policy analysis is that scholars then presume that they are omniscient observers able to comprehend the essentials of how complex, dynamic systems work by creating stylized descriptions of some aspects of those systems.
Dvorak Keyboards
In 1985, Paul David published an influential paper arguing that market failures undermined competition between the QWERTY and Dvorak keyboard layouts. This version of history then became a dominant narrative in the field of network economics, including works by Joseph Farrell & Garth Saloner, and Jean Tirole.
The basic claim was that QWERTY users’ reluctance to switch toward the putatively superior Dvorak layout exerted a negative externality on the rest of the ecosystem (and a positive externality on other QWERTY users), thus preventing the adoption of a more efficient standard. As Paul David put it:
Although the initial lead acquired by QWERTY through its association with the Remington was quantitatively very slender, when magnified by expectations it may well have been quite sufficient to guarantee that the industry eventually would lock in to a de facto QWERTY standard. […]
Competition in the absence of perfect futures markets drove the industry prematurely into standardization on the wrong system — where decentralized decision making subsequently has sufficed to hold it.
Unfortunately, many of the above papers paid little to no attention to actual market conditions in the typewriter and keyboard layout industries. Years later, Stan Liebowitz and Stephen Margolis undertook a detailed analysis of the keyboard layout market. They almost entirely rejected any notion that QWERTY prevailed despite it being the inferior standard:
Yet there are many aspects of the QWERTY-versus-Dvorak fable that do not survive scrutiny. First, the claim that Dvorak is a better keyboard is supported only by evidence that is both scant and suspect. Second, studies in the ergonomics literature find no significant advantage for Dvorak that can be deemed scientifically reliable. Third, the competition among producers of typewriters, out of which the standard emerged, was far more vigorous than is commonly reported. Fourth, there were far more typing contests than just the single Cincinnati contest. These contests provided ample opportunity to demonstrate the superiority of alternative keyboard arrangements. That QWERTY survived significant challenges early in the history of typewriting demonstrates that it is at least among the reasonably fit, even if not the fittest that can be imagined.
In short, there was little to no evidence supporting the view that QWERTY inefficiently prevailed because of network effects. The falsification of this narrative also weakens broader claims that network effects systematically lead to either excess momentum or excess inertia in standardization. Indeed, it is tempting to characterize all network industries with heavily skewed market shares as resulting from market failure. Yet the QWERTY/Dvorak story suggests that such a conclusion would be premature.
Killzones, Zoom, and TikTok
If you are still reading at this point, you might think that contemporary scholars would know better than to base calls for policy intervention on theoretical externalities. Alas, nothing could be further from the truth.
For instance, a recent paper by Sai Kamepalli, Raghuram Rajan and Luigi Zingales conjectures that the interplay between mergers and network externalities discourages the adoption of superior independent platforms:
If techies expect two platforms to merge, they will be reluctant to pay the switching costs and adopt the new platform early on, unless the new platform significantly outperforms the incumbent one. After all, they know that if the entering platform’s technology is a net improvement over the existing technology, it will be adopted by the incumbent after merger, with new features melded with old features so that the techies’ adjustment costs are minimized. Thus, the prospect of a merger will dissuade many techies from trying the new technology.
Although this key behavioral assumption drives the results of the theoretical model, the paper presents no evidence to support the contention that it occurs in real-world settings. Admittedly, the paper does present evidence of reduced venture capital investments after mergers involving large tech firms. But even on their own terms, this data simply does not support the authors’ behavioral assumption.
And this is no isolated example. Over the past couple of years, several scholars have called for more muscular antitrust intervention in networked industries. A common theme is that network externalities, switching costs, and data-related increasing returns to scale lead to inefficient consumer lock-in, thus raising barriers to entry for potential rivals (here, here, here).
But there are also countless counterexamples, where firms have easily overcome potential barriers to entry and network externalities, ultimately disrupting incumbents.
Zoom is one of the most salient instances. As I have written previously:
To get to where it is today, Zoom had to compete against long-established firms with vast client bases and far deeper pockets. These include the likes of Microsoft, Cisco, and Google. Further complicating matters, the video communications market exhibits some prima facie traits that are typically associated with the existence of network effects.
Along similar lines, Geoffrey Manne and Alec Stapp have put forward a multitude of other examples. These include: The demise of Yahoo; the disruption of early instant-messaging applications and websites; MySpace’s rapid decline; etc. In all these cases, outcomes do not match the predictions of theoretical models.
More recently, TikTok’s rapid rise offers perhaps the greatest example of a potentially superior social-networking platform taking significant market share away from incumbents. According to the Financial Times, TikTok’s video-sharing capabilities and its powerful algorithm are the most likely explanations for its success.
While these developments certainly do not disprove network effects theory, they eviscerate the common belief in antitrust circles that superior rivals are unable to overthrow incumbents in digital markets. Of course, this will not always be the case. As in the previous examples, the question is ultimately one of comparing institutions—i.e., do markets lead to more or fewer error costs than government intervention? Yet this question is systematically omitted from most policy discussions.
In Conclusion
My argument is not that models are without value. To the contrary, framing problems in economic terms—and simplifying them in ways that make them cognizable—enables scholars and policymakers to better understand where market failures might arise, and how these problems can be anticipated and solved by private actors. In other words, models alone cannot tell us that markets will fail, but they can direct inquiries and help us to understand why firms behave the way they do, and why markets (including digital ones) are organized in a given way.
In that respect, both the theoretical and empirical research cited throughout this post offer valuable insights for today’s policymakers.
For a start, as Ronald Coase famously argued in what is perhaps his most famous work, externalities (and market failure more generally) are a function of transaction costs. When these are low (relative to the value of a good), market failures are unlikely. This is perhaps clearest in the “Fable of the Bees” example. Given bees’ short foraging range, there were ultimately few real-world obstacles to writing contracts that internalized the mutual benefits of bees and orchards.
Perhaps more importantly, economic research sheds light on behavior that might otherwise be seen as anticompetitive. The rules and norms that bind farming/beekeeping communities, as well as users of common pool resources, could easily be analyzed as a cartel by naïve antitrust authorities. Yet externality theory suggests they play a key role in preventing market failure.
Along similar lines, mergers and acquisitions (as well as vertical integration, more generally) can reduce opportunism and other externalities that might otherwise undermine collaboration between firms (here, here and here). And much of the same is true for certain types of unilateral behavior. Tying video games to consoles (and pricing the console below cost) can help entrants overcome network externalities that might otherwise shield incumbents. Likewise, Google tying its proprietary apps to the open source Android operating system arguably enabled it to earn a return on its investments, thus overcoming the externality problem that plagues open source software.
All of this raises a tantalizing prospect that deserves far more attention than it is currently given in policy circles: authorities around the world are seeking to regulate the tech space. Draft legislation has notably been tabled in the United States, European Union and the United Kingdom. These draft bills would all make it harder for large tech firms to implement various economic hierarchies, including mergers and certain contractual arrangements.
This is highly paradoxical. If digital markets are indeed plagued by network externalities and high transaction costs, as critics allege, then preventing firms from adopting complex hierarchies—which have traditionally been seen as a way to solve externalities—is just as likely to exacerbate problems. In other words, like the economists of old cited above, today’s policymakers appear to be focusing too heavily on simple models that predict market failure, and far too little on the mechanisms that firms have put in place to thrive within this complex environment.
The bigger picture is that far more circumspection is required when using theoretical models in real-world policy settings. Indeed, as Harold Demsetz famously put it, the purpose of normative economics is not so much to identify market failures, but to help policymakers determine which of several alternative institutions will deliver the best outcomes for consumers:
This nirvana approach differs considerably from a comparative institution approach in which the relevant choice is between alternative real institutional arrangements. In practice, those who adopt the nirvana viewpoint seek to discover discrepancies between the ideal and the real and if discrepancies are found, they deduce that the real is inefficient. Users of the comparative institution approach attempt to assess which alternative real institutional arrangement seems best able to cope with the economic problem […].
In current discussions of technology markets, few words are heard more often than “platform.” Initial public offering (IPO) prospectuses use “platform” to describe a service that is bound to dominate a digital market. Antitrust regulators use “platform” to describe a service that dominates a digital market or threatens to do so. In either case, “platform” denotes power over price. For investors, that implies exceptional profits; for regulators, that implies competitive harm.
Conventional wisdom holds that platforms enjoy high market shares, protected by high barriers to entry, which yield high returns. This simple logic drives the market’s attribution of dramatically high valuations to dramatically unprofitable businesses and regulators’ eagerness to intervene in digital platform markets characterized by declining prices, increased convenience, and expanded variety, often at zero out-of-pocket cost. In both cases, “burning cash” today is understood as the path to market dominance and the ability to extract a premium from consumers in the future.
This logic is usually wrong.
The Overlooked Basics of Platform Economics
To appreciate this perhaps surprising point, it is necessary to go back to the increasingly overlooked basics of platform economics. A platform can refer to any service that matches two complementary populations. A search engine matches advertisers with consumers, an online music service matches performers and labels with listeners, and a food-delivery service matches restaurants with home diners. A platform benefits everyone by facilitating transactions that otherwise might never have occurred.
A platform’s economic value derives from its ability to lower transaction costs by funneling a multitude of individual transactions into a single convenient hub. In pursuit of minimum costs and maximum gains, users on one side of the platform will tend to favor the most popular platforms that offer the largest number of users on the other side of the platform. (There are partial exceptions to this rule when users value being matched with certain typesof other users, rather than just with more users.) These “network effects” mean that any successful platform market will always converge toward a handful of winners. This positive feedback effect drives investors’ exuberance and regulators’ concerns.
There is a critical point, however, that often seems to be overlooked.
Market share only translates into market power to the extent the incumbent is protected against entry within some reasonable time horizon. If Warren Buffett’s moat requirement is not met, market share is immaterial. If XYZ.com owns 100% of the online pet food delivery market but entry costs are asymptotic, then market power is negligible. There is another important limiting principle. In platform markets, the depth of the moat depends not only on competitors’ costs to enter the market, but users’ costs in switching from one platform to another or alternating between multiple platforms. If users can easily hop across platforms, then market share cannot confer market power given the continuous threat of user defection. Put differently: churn limits power over price.
Contrary to natural intuitions, this is why a platform market consisting of only a few leaders can still be intensely competitive, keeping prices low (down to and including $0) even if the number of competitors is low. It is often asserted, however, that users are typically locked into the dominant platform and therefore face high switching costs, which therefore implicitly satisfies the moat requirement. If that is true, then the “high churn” scenario is a theoretical curiosity and a leading platform’s high market share would be a reliable signal of market power. In fact, this common assumption likely describes the atypical case.
AWS and the Cloud Data-Storage Market
This point can be illustrated by considering the cloud data-storage market. This would appear to be an easy case where high switching costs (due to the difficulty in shifting data among storage providers) insulate the market leader against entry threats. Yet the real world does not conform to these expectations.
While Amazon Web Services pioneered the $100 billion-plus market and is still the clear market leader, it now faces vigorous competition from Microsoft Azure, Google Cloud, and other data-storage or other cloud-related services. This may reflect the fact that the data storage market is far from saturated, so new users are up for grabs and existing customers can mitigate lock-in by diversifying across multiple storage providers. Or it may reflect the fact that the market’s structure is fluid as a function of technological changes, enabling entry at formerly bundled portions of the cloud data-services package. While it is not always technologically feasible, the cloud storage market suggests that users’ resistance to platform capture can represent a competitive opportunity for entrants to challenge dominant vendors on price, quality, and innovation parameters.
The Surprising Instability of Platform Dominance
The instability of leadership positions in the cloud storage market is not exceptional.
Consider a handful of once-powerful platforms that were rapidly dethroned once challenged by a more efficient or innovative rival: Yahoo and Alta Vista in the search-engine market (displaced by Google); Netscape in the browser market (displaced by Microsoft’s Internet Explorer, then displaced by Google Chrome); Nokia and then BlackBerry in the mobile wireless-device market (displaced by Apple and Samsung); and Friendster in the social-networking market (displaced by Myspace, then displaced by Facebook). AOL was once thought to be indomitable; now it is mostly referenced as a vintage email address. The list could go on.
Overestimating platform dominance—or more precisely, assuming platform dominance without close factual inquiry—matters because it promotes overestimates of market power. That, in turn, cultivates both market and regulatory bubbles: investors inflate stock valuations while regulators inflate the risk of competitive harm.
DoorDash and the Food-Delivery Services Market
Consider the DoorDash IPO that launched in early December 2020. The market’s current approximately $50 billion valuation of a business that has been almost consistently unprofitable implicitly assumes that DoorDash will maintain and expand its position as the largest U.S. food-delivery platform, which will then yield power over price and exceptional returns for investors.
There are reasons to be skeptical. Even where DoorDash captures and holds a dominant market share in certain metropolitan areas, it still faces actual and potential competition from other food-delivery services, in-house delivery services (especially by well-resourced national chains), and grocery and other delivery services already offered by regional and national providers. There is already evidence of these expected responses to DoorDash’s perceived high delivery fees, a classic illustration of the disciplinary effect of competitive forces on the pricing choices of an apparently dominant market leader. These “supply-side” constraints imposed by competitors are compounded by “demand-side” constraints imposed by customers. Home diners incur no more than minimal costs when swiping across food-delivery icons on a smartphone interface, casting doubt that high market share is likely to translate in this context into market power.
Deliveroo and the Costs of Regulatory Autopilot
Just as the stock market can suffer from delusions of platform grandeur, so too some competition regulators appear to have fallen prey to the same malady.
A vivid illustration is provided by the 2019 decision by the Competition Markets Authority (CMA), the British competition regulator, to challenge Amazon’s purchase of a 16% stake in Deliveroo, one of three major competitors in the British food-delivery services market. This intervention provides perhaps the clearest illustration of policy action based on a reflexive assumption of market power, even in the face of little to no indication that the predicate conditions for that assumption could plausibly be satisfied.
Far from being a dominant platform, Deliveroo was (and is) a money-losing venture lagging behind money-losing Just Eat (now Just Eat Takeaway) and Uber Eats in the U.K. food-delivery services market. Even Amazon had previously closed its own food-delivery service in the U.K. due to lack of profitability. Despite Deliveroo’s distressed economic circumstances and the implausibility of any market power arising from Amazon’s investment, the CMA nonetheless elected to pursue the fullest level of investigation. While the transaction was ultimately approved in August 2020, this intervention imposed a 15-month delay and associated costs in connection with an investment that almost certainly bolstered competition in a concentrated market by funding a firm reportedly at risk of insolvency. This is the equivalent of a competition regulator driving in reverse.
Concluding Thoughts
There seems to be an increasingly common assumption in commentary by the press, policymakers, and even some scholars that apparently dominant platforms usually face little competition and can set, at will, the terms of exchange. For investors, this is a reason to buy; for regulators, this is a reason to intervene. This assumption is sometimes realized, and, in that case, antitrust intervention is appropriate whenever there is reasonable evidence that market power is being secured through something other than “competition on the merits.” However, several conditions must be met to support the market power assumption without which any such inquiry would be imprudent. Contrary to conventional wisdom, the economics and history of platform markets suggest that those conditions are infrequently satisfied.
Without closer scrutiny, reflexively equating market share with market power is prone to lead both investors and regulators astray.
[TOTM: The following is part of a digital symposium by TOTM guests and authors on the law, economics, and policy of the antitrust lawsuits against Google. The entire series of posts is available here.]
As one of the few economic theorists in this symposium, I believe my comparative advantage is in that: economic theory. In this post, I want to remind people of the basic economic theories that we have at our disposal, “off the shelf,” to make sense of the U.S. Department of Justice’s lawsuit against Google. I do not mean this to be a proclamation of “what economics has to say about X,” but merely just to help us frame the issue.
In particular, I’m going to focus on the economic concerns of Google paying phone manufacturers (Apple, in particular) to be the default search engine installed on phones. While there is not a large literature on the economic effects of default contracts, there is a large literature on something that I will argue is similar: trade promotions, such as slotting contracts, where a manufacturer pays a retailer for shelf space. Despite all the bells and whistles of the Google case, I will argue that, from an economic point of view, the contracts that Google signed are just trade promotions. No more, no less. And trade promotions are well-established as part of a competitive process that ultimately helps consumers.
However, it is theoretically possible that such trade promotions hurt customers, so it is theoretically possible that Google’s contracts hurt consumers. Ultimately, the theoretical possibility of anticompetitive behavior that harms consumers does not seem plausible to me in this case.
Default Status
There are two reasons that Google paying Apple to be its default search engine is similar to a trade promotion. First, the deal brings awareness to the product, which nudges certain consumers/users to choose the product when they would not otherwise do so. Second, the deal does not prevent consumers from choosing the other product.
In the case of retail trade promotions, a promotional space given to Coca-Cola makes it marginally easier for consumers to pick Coke, and therefore some consumers will switch from Pepsi to Coke. But it does not reduce any consumer’s choice. The store will still have both items.
This is the same for a default search engine. The marginal searchers, who do not have a strong preference for either search engine, will stick with the default. But anyone can still install a new search engine, install a new browser, etc. It takes a few clicks, just as it takes a few steps to walk down the aisle to get the Pepsi; it is still an available choice.
If we were to stop the analysis there, we could conclude that consumers are worse off (if just a tiny bit). Some customers will have to change the default app. We also need to remember that this contract is part of a more general competitive process. The retail stores are also competing with one another, as are smartphone manufacturers.
Despite popular claims to the contrary, Apple cannot charge anything it wants for its phone. It is competing with Samsung, etc. Therefore, Apple has to pass through some of Google’s payments to customers in order to compete with Samsung. Prices are lower because of this payment. As I phrased it elsewhere, Google is effectively subsidizing the iPhone. This cross-subsidization is a part of the competitive process that ultimately benefits consumers through lower prices.
These contracts lower consumer prices, even if we assume that Apple has market power. Those who recall your Econ 101 know that a monopolist chooses a quantity where the marginal revenue equals marginal cost. With a payment from Google, the marginal cost of producing a phone is lower, therefore Apple will increase the quantity and lower price. This is shown below:
One of the surprising things about markets is that buyers’ and sellers’ incentives can be aligned, even though it seems like they must be adversarial. Companies can indirectly bargain for their consumers. Commenting on Standard Fashion Co. v. Magrane-Houston Co., where a retail store contracted to only carry Standard’s products, Robert Bork (1978, pp. 306–7) summarized this idea as follows:
The store’s decision, made entirely in its own interest, necessarily reflects the balance of competing considerations that determine consumer welfare. Put the matter another way. If no manufacturer used exclusive dealing contracts, and if a local retail monopolist decided unilaterally to carry only Standard’s patterns because the loss in product variety was more than made up in the cost saving, we would recognize that decision was in the consumer interest. We do not want a variety that costs more than it is worth … If Standard finds it worthwhile to purchase exclusivity … the reason is not the barring of entry, but some more sensible goal, such as obtaining the special selling effort of the outlet.
How trade promotions could harm customers
Since Bork’s writing, many theoretical papers have shown exceptions to Bork’s logic. There are times that the retailers’ incentives are not aligned with the customers. And we need to take those possibilities seriously.
The most common way to show the harm of these deals (or more commonly exclusivity deals) is to assume:
There are large, fixed costs so that a firm must acquire a sufficient number of customers in order to enter the market; and
An incumbent can lock in enough customers to prevent the entrant from reaching an efficient size.
Consumers can be locked-in because there is some fixed cost of changing suppliers or because of some coordination problems. If that’s true, customers can be made worse off, on net, because the Google contracts reduce consumer choice.
To understand the logic, let’s simplify the model to just search engines and searchers. Suppose there are two search engines (Google and Bing) and 10 searchers. However, to operate profitably, each search engine needs at least three searchers. If Google can entice eight searchers to use its product, Bing cannot operate profitably, even if Bing provides a better product. This holds even if everyone knows Bing would be a better product. The consumers are stuck in a coordination failure.
We should be skeptical of coordination failure models of inefficient outcomes. The problem with any story of coordination failures is that it is highly sensitive to the exact timing of the model. If Bing can preempt Google and offer customers an even better deal (the new entrant is better by assumption), then the coordination failure does not occur.
To argue that Bing could not execute a similar contract, the most common appeal is that the new entrant does not have the capital to pay upfront for these contracts, since it will only make money from its higher-quality search engine down the road. That makes sense until you remember that we are talking about Microsoft. I’m skeptical that capital is the real constraint. It seems much more likely that Google just has a more popular search engine.
The other problem with coordination failure arguments is that they are almost non-falsifiable. There is no way to tell, in the model, whether Google is used because of a coordination failure or whether it is used because it is a better product. If Google is a better product, then the outcome is efficient. The two outcomes are “observationally equivalent.” Compare this to the standard theory of monopoly, where we can (in principle) establish an inefficiency if the price is greater than marginal cost. While it is difficult to measure marginal cost, it can be done.
There is a general economic idea in these models that we need to pay attention to. If Google takes an action that prevents Bing from reaching efficient size, that may be an externality, sometimes called a network effect, and so that action may hurt consumer welfare.
I’m not sure how seriously to take these network effects. If more searchers allow Bing to make a better product, then literally any action (competitive or not) by Google is an externality. Making a better product that takes away consumers from Bing lowers Bing’s quality. That is, strictly speaking, an externality. Surely, that is not worthy of antitrust scrutiny simply because we find an externality.
And Bing also “takes away” searchers from Google, thus lowering Google’s possible quality. With network effects, bigger is better and it may be efficient to have only one firm. Surely, that’s not an argument we want to put forward as a serious antitrust analysis.
Put more generally, it is not enough to scream “NETWORK EFFECT!” and then have the antitrust authority come in, lawsuits-a-blazing. Well, it shouldn’t be enough.
For me to take the network effect argument seriously from an economic point of view, compared to a legal perspective, I would need to see a real restriction on consumer choice, not just an externality. One needs to argue that:
No competitor can cover their fixed costs to make a reasonable search engine; and
These contracts are what prevent the competing search engines from reaching size.
That’s the challenge I would like to put forward to supporters of the lawsuit. I’m skeptical.
Zoom, one of Silicon Valley’s lesser-known unicorns, has just gone public. At the time of writing, its shares are trading at about $65.70, placing the company’s value at $16.84 billion. There are good reasons for this success. According to its Form S-1, Zoom’s revenue rose from about $60 million in 2017 to a projected $330 million in 2019, and the company has already surpassed break-even . This growth was notably fueled by a thriving community of users who collectively spend approximately 5 billion minutes per month in Zoom meetings.
To get to where it is today, Zoom had to compete against long-established firms with vast client bases and far deeper pockets. These include the likes of Microsoft, Cisco, and Google. Further complicating matters, the video communications market exhibits some prima facie traits that are typically associated with the existence of network effects. For instance, the value of Skype to one user depends – at least to some extent – on the number of other people that might be willing to use the network. In these settings, it is often said that positive feedback loops may cause the market to tip in favor of a single firm that is then left with an unassailable market position. Although Zoom still faces significant competitive challenges, it has nonetheless established a strong position in a market previously dominated by powerful incumbents who could theoretically count on network effects to stymie its growth.
Further complicating matters, Zoom chose to compete head-on with these incumbents. It did not create a new market or a highly differentiated product. Zoom’s Form S-1 is quite revealing. The company cites the quality of its product as its most important competitive strength. Similarly, when listing the main benefits of its platform, Zoom emphasizes that its software is “easy to use”, “easy to deploy and manage”, “reliable”, etc. In its own words, Zoom has thus gained a foothold by offering an existing service that works better than that of its competitors.
And yet, this is precisely the type of story that a literal reading of the network effects literature would suggest is impossible, or at least highly unlikely. For instance, the foundational papers on network effects often cite the example of the DVORAK keyboard (David, 1985; and Farrell & Saloner, 1985). These early scholars argued that, despite it being the superior standard, the DVORAK layout failed to gain traction because of the network effects protecting the QWERTY standard. In other words, consumers failed to adopt the superior DVORAK layout because they were unable to coordinate on their preferred option. It must be noted, however, that the conventional telling of this story was forcefully criticized by Liebowitz & Margolis in their classic 1995 article, The Fable of the Keys.
Despite Liebowitz & Margolis’ critique, the dominance of the underlying network effects story persists in many respects. And in that respect, the emergence of Zoom is something of a cautionary tale. As influential as it may be, the network effects literature has tended to overlook a number of factors that may mitigate, or even eliminate, the likelihood of problematic outcomes. Zoom is yet another illustration that policymakers should be careful when they make normative inferences from positive economics.
A Coasian perspective
It is now widely accepted that multi-homing and the absence of switching costs can significantly curtail the potentially undesirable outcomes that are sometimes associated with network effects. But other possibilities are often overlooked. For instance, almost none of the foundational network effects papers pay any notice to the application of the Coase theorem (though it has been well-recognized in the two-sided markets literature).
Take a purported market failure that is commonly associated with network effects: an installed base of users prevents the market from switching towards a new standard, even if it is superior (this is broadly referred to as “excess inertia,” while the opposite scenario is referred to as “excess momentum”). DVORAK’s failure is often cited as an example.
Astute readers will quickly recognize that this externality problem is not fundamentally different from those discussed in Ronald Coase’s masterpiece, “The Problem of Social Cost,” or Steven Cheung’s “The Fable of the Bees” (to which Liebowitz & Margolis paid homage in their article’s title). In the case at hand, there are at least two sets of externalities at play. First, early adopters of the new technology impose a negative externality on the old network’s installed base (by reducing its network effects), and a positive externality on other early adopters (by growing the new network). Conversely, installed base users impose a negative externality on early adopters and a positive externality on other remaining users.
In terms of the Coase theorem, it is very difficult to design a contract where, say, the (potential) future users of HDTV agree to subsidize today’s buyers of television sets to stop buying NTSC sets and start buying HDTV sets, thereby stimulating the supply of HDTV programming.
And yet it is far from clear that consumers and firms can never come up with solutions that mitigate these problems. As Daniel Spulber has suggested, referral programs offer a case in point. These programs usually allow early adopters to receive rewards in exchange for bringing new users to a network. One salient feature of these programs is that they do not simply charge a lower price to early adopters; instead, in order to obtain a referral fee, there must be some agreement between the early adopter and the user who is referred to the platform. This leaves ample room for the reallocation of rewards. Users might, for instance, choose to split the referral fee. Alternatively, the early adopter might invest time to familiarize the switching user with the new platform, hoping to earn money when the user jumps ship. Both of these arrangements may reduce switching costs and mitigate externalities.
Danial Spulber also argues that users may coordinate spontaneously. For instance, social groups often decide upon the medium they will use to communicate. Families might choose to stay on the same mobile phone network. And larger groups (such as an incoming class of students) may agree upon a social network to share necessary information, etc. In these contexts, there is at least some room to pressure peers into adopting a new platform.
Finally, firms and other forms of governance may also play a significant role. For instance, employees are routinely required to use a series of networked goods. Common examples include office suites, email clients, social media platforms (such as Slack), or video communications applications (Zoom, Skype, Google Hangouts, etc.). In doing so, firms presumably act as islands of top-down decision-making and impose those products that maximize the collective preferences of employers and employees. Similarly, a single firm choosing to join a network (notably by adopting a standard) may generate enough momentum for a network to gain critical mass. Apple’s decisions to adopt USB-C connectors on its laptops and to ditch headphone jacks on its iPhones both spring to mind. Likewise, it has been suggested that distributed ledger technology and initial coin offerings may facilitate the creation of new networks. The intuition is that so-called “utility tokens” may incentivize early adopters to join a platform, despite initially weak network effects, because they expect these tokens to increase in value as the network expands.
A combination of these arrangements might explain how Zoom managed to grow so rapidly, despite the presence of powerful incumbents. In its own words:
Our rapid adoption is driven by a virtuous cycle of positive user experiences. Individuals typically begin using our platform when a colleague or associate invites them to a Zoom meeting. When attendees experience our platform and realize the benefits, they often become paying customers to unlock additional functionality.
All of this is not to say that network effects will always be internalized through private arrangements, but rather that it is equally wrong to assume that transaction costs systematically prevent efficient coordination among users.
Misguided regulatory responses
Over the past couple of months, several antitrust authorities around the globe have released reports concerning competition in digital markets (UK, EU, Australia), or held hearings on this topic (US). A recurring theme throughout their published reports is that network effects almost inevitably weaken competition in digital markets.
Because of very strong network externalities (especially in multi-sided platforms), incumbency advantage is important and strict scrutiny is appropriate. We believe that any practice aimed at protecting the investment of a dominant platform should be minimal and well targeted.
There are considerable barriers to entry and expansion for search platforms and social media platforms that reinforce and entrench Google and Facebook’s market power. These include barriers arising from same-side and cross-side network effects, branding, consumer inertia and switching costs, economies of scale and sunk costs.
Finally, a panel of experts in the United Kingdom found that:
Today, network effects and returns to scale of data appear to be even more entrenched and the market seems to have stabilised quickly compared to the much larger degree of churn in the early days of the World Wide Web.
To address these issues, these reports suggest far-reaching policy changes. These include shifting the burden of proof in competition cases from authorities to defendants, establishing specialized units to oversee digital markets, and imposing special obligations upon digital platforms.
The story of Zoom’s emergence and the important insights that can be derived from the Coase theorem both suggest that these fears may be somewhat overblown.
Rivals do indeed find ways to overthrow entrenched incumbents with some regularity, even when these incumbents are shielded by network effects. Of course, critics may retort that this is not enough, that competition may sometimes arrive too late (excess inertia, i.e., “ a socially excessive reluctance to switch to a superior new standard”) or too fast (excess momentum, i.e., “the inefficient adoption of a new technology”), and that the problem is not just one of network effects, but also one of economies of scale, information asymmetry, etc. But this comes dangerously close to the Nirvana fallacy. To begin, it assumes that regulators are able to reliably navigate markets toward these optimal outcomes — which is questionable, at best. Moreover, the regulatory cost of imposing perfect competition in every digital market (even if it were possible) may well outweigh the benefits that this achieves. Mandating far-reaching policy changes in order to address sporadic and heterogeneous problems is thus unlikely to be the best solution.
Instead, the optimal policy notably depends on whether, in a given case, users and firms can coordinate their decisions without intervention in order to avoid problematic outcomes. A case-by-case approach thus seems by far the best solution.
And competition authorities need look no further than their own decisional practice. The European Commission’s decision in the Facebook/Whatsapp merger offers a good example (this was before Margrethe Vestager’s appointment at DG Competition). In its decision, the Commission concluded that the fast-moving nature of the social network industry, widespread multi-homing, and the fact that neither Facebook nor Whatsapp controlled any essential infrastructure, prevented network effects from acting as a barrier to entry. Regardless of its ultimate position, this seems like a vastly superior approach to competition issues in digital markets. The Commission adopted a similar reasoning in the Microsoft/Skype merger. Unfortunately, the Commission seems to have departed from this measured attitude in more recent decisions. In the Google Search case, for example, the Commission assumes that the mere existence of network effects necessarily increases barriers to entry:
The existence of positive feedback effects on both sides of the two-sided platform formed by general search services and online search advertising creates an additional barrier to entry.
A better way forward
Although the positive economics of network effects are generally correct and most definitely useful, some of the normative implications that have been derived from them are deeply flawed. Too often, policymakers and commentators conclude that these potential externalities inevitably lead to stagnant markets where competition is unable to flourish. But this does not have to be the case. The emergence of Zoom shows that superior products may prosper despite the presence of strong incumbents and network effects.
Basing antitrust policies on sweeping presumptions about digital competition – such as the idea that network effects are rampant or the suggestion that online platforms necessarily imply “extreme returns to scale” – is thus likely to do more harm than good. Instead, Antitrust authorities should take a leaf out of Ronald Coase’s book, and avoid blackboard economics in favor of a more granular approach.
One of my favorite stories in the ongoing saga over the regulation (and thus the future) of Internet search emerged earlier this week with claims by Google that Microsoft has been copying its answers–using Google search results to bolster the relevance of its own results for certain search terms. The full story from Internet search journalist extraordinaire, Danny Sullivan, is here, with a follow up discussing Microsoft’s response here. The New York Times is also on the case with some interesting comments from a former Googler that feed nicely into the Schumpeterian competition angle (discussed below). And Microsoft consultant (“though on matters unrelated to issues discussed here”) and Harvard Business prof Ben Edelman coincidentally echoes precisely Microsoft’s response in a blog post here.
What I find so great about this story is how it seems to resolve one of the most significant strands of the ongoing debate–although it does so, from Microsoft’s point of view, unintentionally, to be sure.
Here’s what I mean. Back when Microsoft first started being publicly identified as a significant instigator of regulatory and antitrust attention paid to Google, the company, via its chief competition counsel, Dave Heiner, defended its stance in large part on the following ground:
All of this is quite important because search is so central to how people navigate the Internet, and because advertising is the main monetization mechanism for a wide range of Web sites and Web services. Both search and online advertising are increasingly controlled by a single firm, Google. That can be a problem because Google’s business is helped along by significant network effects (just like the PC operating system business). Search engine algorithms “learn” by observing how users interact with search results. Google’s algorithms learn less common search terms better than others because many more people are conducting searches on these terms on Google.
These and other network effects make it hard for competing search engines to catch up. Microsoft’s well-received Bing search engine is addressing this challenge by offering innovations in areas that are less dependent on volume. But Bing needs to gain volume too, in order to increase the relevance of search results for less common search terms. That is why Microsoft and Yahoo! are combining their search volumes. And that is why we are concerned about Google business practices that tend to lock in publishers and advertisers and make it harder for Microsoft to gain search volume. (emphasis added).
Claims of “network effects” “increasing returns to scale” and the absence of “minimum viable scale” for competitors run rampant (and unsupported) in the various cases against Google. The TradeComet complaint, for example, claims that
[t]he primary barrier to entry facing vertical search websites is the inability to draw enough search traffic to reach the critical mass necessary to become independently sustainable.
But now we discover (what we should have known all along) that “learning by doing” is not the only way to obtain the data necessary to generate relevant search results: “Learning by copying” works, as well. And there’s nothing wrong with it–in fact, the very process of Schumpeterian creative destruction assumes imitation.
Neither perfect knowledge of the past nor complete awareness of the current state of the arts gives sufficient foresight to indicate profitable action . . . [and] the pervasive effects of uncertainty prevent the ascertainment of actions which are supposed to be optimal in achieving profits. Now the consequence of this is that modes of behavior replace optimum equilibrium conditions as guiding rules of action. First, wherever successful enterprises are observed, the elements common to these observable successes will be associated with success and copied by others in their pursuit of profits or success. “Nothing succeeds like success.”
So on the one hand, I find the hand wringing about Microsoft’s “copying” Google’s results to be completely misplaced–just as the pejorative connotations of “embrace and extend” deployed against Microsoft itself when it was the target of this sort of scrutiny were bogus. But, at the same time, I see this dynamic essentially decimating Microsoft’s (and others’) claims that Google has an unassailable position because no competitor can ever hope to match its size, and thus its access to information essential to the quality of search results, particularly when it comes to so-called “long-tail” search terms.
Long-tail search terms are queries that are extremely rare and, thus, for which there is little user history (information about which results searchers found relevant and clicked on) to guide future search results. As Ben Edelman writes in his blog post (linked above) on this issue (trotting out, even while implicitly undercutting, the “minimum viable scale” canard):
Of course the reality is that Google’s high market share means Google gets far more searches than any other search engine. And Google’s popularity gives it a real advantage: For an obscure search term that gets 100 searches per month at Google, Bing might get just five or 10. Also, for more popular terms, Google can slice its data into smaller groups — which results are most useful to people from Boston versus New York, which results are best during the day versus at night, and so forth. So Google is far better equipped to figure out what results users favor and to tailor its listings accordingly. Meanwhile, Microsoft needs additional data, such as Toolbar and Related Sites data, to attempt to improve its results in a similar way.
But of course the “additional data” that Microsoft has access to here is, to a large extent, the same data that Google has. Although Danny Sullivan’s follow up story (also linked above) suggests that Bing doesn’t do all it could to make use of Google’s data (for example, Bing does not, it seems, copy Google search results wholesale, nor does it use user behavior as extensively as it could (by, for example, seeing searches in Google and then logging the next page visited, which would give Bing a pretty good idea which sites in Google’s results users found most relevant)), it doesn’t change the fundamental fact that Microsoft and other search engines can overcome a significant amount of the so-called barrier to entry afforded by Google’s impressive scale by simply imitating much of what Google does (and, one hopes, also innovating enough to offer something better).
Perhaps Google is “better equipped to figure out what users favor.” But it seems to me that only a trivial amount of this advantage is plausibly attributable to Google’s scale instead of its engineering and innovation. The fact that Microsoft can (because of its own impressive scale in various markets) and does take advantage of accessible data to benefit indirectly from Google’s own prowess in search is a testament to the irrelevance of these unfortunately-pervasive scale and network effect arguments.
We have just uploaded to SSRN a draft of our article assessing the economics and the law of the antitrust case directed at the core of Google’s business: Its search and search advertising platform. The article is Google and the Limits of Antitrust: The Case Against the Antitrust Case Against Google. This is really the first systematic attempt to address both the amorphous and the concrete (as in the TradeComet complaint) claims about Google’s business and its legal and economic importance in its primary market. It’s giving nothing away to say we’re skeptical of the claims, and, moreover, that an approach to the issues appropriately sensitive to the potential error costs would be extremely deferential. As we discuss, the economics of search and search advertising are indeterminate and subtle, and the risk of error is high (claims of network effects, for example, are greatly exaggerated, and the pro-competitive justifications for Google’s use of a quality score are legion, despite frequent claims to the contrary). We welcome comments on the article, and we look forward to the debate. The abstract is here:
The antitrust landscape has changed dramatically in the last decade. Within the last two years alone, the United States Department of Justice has held hearings on the appropriate scope of Section 2, issued a comprehensive Report, and then repudiated it; and the European Commission has risen as an aggressive leader in single firm conduct enforcement by bringing abuse of dominance actions and assessing heavy fines against firms including Qualcomm, Intel, and Microsoft. In the United States, two of the most significant characteristics of the “new” antitrust approach have been a more intense focus on innovative companies in high-tech industries and a weakening of longstanding concerns that erroneous antitrust interventions will hinder economic growth. But this focus is dangerous, and these concerns should not be dismissed so lightly. In this article we offer a comprehensive cautionary tale in the context of a detailed factual, legal and economic analysis of the next Microsoft: the theoretical, but perhaps imminent, enforcement action against Google. Close scrutiny of the complex economics of Google’s technology, market and business practices reveals a range of real but subtle, pro-competitive explanations for features that have been held out instead as anticompetitive. Application of the relevant case law then reveals a set of concerns where economic complexity and ambiguity, coupled with an insufficiently-deferential approach to innovative technology and pricing practices in the most relevant precedent (the D.C. Circuit’s decision in Microsoft), portend a potentially erroneous—and costly—result. Our analysis, by contrast, embraces the cautious and evidence-based approach to uncertainty, complexity and dynamic innovation contained within the well-established “error cost framework.” As we demonstrate, while there is an abundance of error-cost concern in the Supreme Court precedent, there is a real risk that the current, aggressive approach to antitrust error, coupled with the uncertain economics of Google’s innovative conduct, will nevertheless yield a costly intervention. The point is not that we know that Google’s conduct is procompetitive, but rather that the very uncertainty surrounding it counsels caution, not aggression.
RSS - Posts