There is much to criticize about government meddling in the tech market. Tech has spurred productivity growth and created a bounty of new products in recent decades. Many of these products are free and digitization has by and large reduced costs compared to non-internet alternatives or predecessors. Arguments that the U.S. tech sector is not competitive disregard changes in prices across different economic sectors, while government regulatory barriers and labor intensiveness explain high prices in areas like education and housing.
Nonetheless, calls to expand antitrust are growing. One thrust is to promote digital startup competition. In recent years, competitive process scholars like Columbia’s Tim Wu and Lina Kahn (now appointees of the Biden administration), Yale’s Fiona Scott Morton, and former Federal Trade Commission economist Jonathan Baker have written books and articles suggesting that judicial emphasis on price and output in exclusionary conduct cases is particularly ill suited to the digital economy. They propose that courts alter how they weigh procompetitive and anticompetitive effects when evaluating vertical integration and vertical contractual restraints. They also suggest removing the below-cost and recoupment requirements for predatory pricing exclusion claims.
The competitive process school’s proposed reforms should be rejected because current law adequately prohibits unjustified dominant firm exclusionary conduct without unduly protecting inefficient competitors. The proposed changes would not improve the ability of courts to sanction unjustified exclusionary conduct that harms startups. Instead, the reforms might prevent dominant firms from achieving efficient scale and discourage new firms from aggressively pursuing price and quality competitiveness.
Balacing Analysis in Vertical Mergers, Vertical Restraints, and Vertical Ties
Balancing procompetitive and anticompetitive effects is the primary analytical tool used by courts in evaluating potentially exclusionary vertical mergers, vertical restraints, and product ties conducted or implemented by dominant firms. These balancing analyses usually do not take the form of an attempt to calculate the net economic effect of challenged conduct. Courts often consider whether a procompetitive benefit could be achieved by a less restrictive alternative. Additionally, defendants win close cases because of burden-of-proof issues. Therefore, balancing analyses consist of rough analyses of competitive and anticompetitive effects rather than precise attempts to quantify the conduct.
Competitive process scholars argue that balancing analyses focus too much on evidence of price harm and downplay the possibility of inefficient nascent competitors becoming efficiency leaders in the future. To fix this, they advocate adding legislatively mandated presumptions to the balancing analysis that would block more exclusionary conduct. Specific proposals given by Wu, Kahn, and Baker, and by a recent House Antitrust Subcommittee report, include altering doctrine by making a platform’s integration with an application (e.g., Amazon selling its own branded goods on its platform) prima facie evidence of anticompetitive behavior; increasing scrutiny of exclusive dealing, “most favored nation,” and data provision vertical contracts; and creating a presumption against approving vertical mergers if the target supplies at least 30% of the goods or services in its market.
The proposed reforms are a solution in search of a problem and ought to be rejected. A dominant firm’s conduct should not be prohibited unless its exclusionary consequences reflect the firm’s dominance as opposed to its efficiency, and could exclude a competitor of equal or superior efficiency. Applying this principle to a wide set of firms, industries, cost structures, and so on requires very careful attention to the specific facts of a case. One virtue of today’s law is that it understands that these and a host of other factors go into determining whether a certain practice with exclusionary consequences mainly reflects efficiency or dominance. Current law allows courts to address the specifics of each case without being burdened by the type of blunt presumptions the competitive process adherents propose.
Competitive process scholars argue that vertical exclusion warrants particular scrutiny when undertaken by a dominant digital platform because of the naturally monopolistic features of digital markets. For instance, Wu recently argued in the New York Times that Google’s acquisition of the mapping software company Waze should have been challenged and possibly blocked. According to him, a putative Google competitor with a better search platform might struggle to compete for market share unless it has good mapping software that it can integrate into search, as Google has done. Google’s mapping advantage might entirely deter the competitor from making the investments and raising the capital necessary to enter the search market.
Platform markets / These barriers could be reinforced by search’s platform market structure. Platform markets (also referred to as two-sided markets) are markets where businesses provide value, at least in part, by matching providers with consumers; think of Airbnb’s matching property owners and renters. Platform markets are often digitally based because a primary value of the internet is instantaneous processing of large sources of information. Thus, Amazon’s marketplace matches buyers and third-party sellers, Uber matches riders and drivers, and Google matches searchers with content. Platform businesses feature consumer network effects because either one or both sets of market participants benefit when more consumers use the platform. When customers have reason to use just one platform and switching costs are high, the market structure might encourage enduring monopoly even in the face of a superior competitor. By contrast, when switching costs are low and when it makes sense for consumers to use more than one platform, the market will support multiple firms, making it easier for new firms with competitive advantages to enter.
Tying / Competitive process scholars acknowledge that some digital markets have a greater tendency toward sustained, inefficient monopolization. Nonetheless, they advocate new anti-dominant firm presumptions that would apply equally across the digital economy.
These reform proposals might undermine careful, fact-sensitive adjudication of exclusion cases. Consider Google’s acquisition of Waze and its proposed acquisition of Fitbit. If enforcement agencies challenge those acquisitions under current doctrine, a reviewing court would pay close attention to the efficiency benefits of the deals, namely the benefits to consumers of direct integration of real-time mapping or fitness tracking software with Google’s widely used email and search functions. Moreover, the court would note limits to the exclusionary effect. Waze’s and Fitbit’s technological advantages are their easy-to-use interfaces and automated methods of processing user data. In theory, those processes could be improved upon by a competitor solely with human capital inputs. A competitor with Google’s search might therefore need to invest some more capital to develop its own map and fitness platforms, but by no means does that make competing significantly harder. The search engine competitor could also forgo mapping and fitness integration and still survive. Integrating search with mapping and fitness functions certainly provides some consumer benefits, but they are not extreme relative to the alternative of using separate search, mapping, and fitness programs. Under the competitive process school’s proposed changes, however, a court would have to start its analysis of these proposed mergers by noting that the case involves a platform buying an app with a large share of its market. By placing large weight on that characteristic of the deal, the new doctrine might therefore minimize other facts that show barriers to entry were low and benefits to consumers important.
Similar drawbacks to increased doctrinal rigidity would characterize potential review of vertical exclusion by a dominant shopping platform like Amazon. Competitive process school adherents claim Amazon uses vertical contracts to further its market power in two ways. First, by employing most favored nation (MFN) clauses in contracts with third-party sellers, thereby requiring the sellers to give Amazon terms as good as any of the sellers’ other retailers, Amazon prevents competitors from offering targeted discounts on similar products. Second, Amazon uses its power to extract valuable data by contract from third-party sellers, which allows it to obtain product information, track consumer preferences, and thereby create its own product lines in high-demand areas. Amazon’s in-house products are produced efficiently because of scale, so demand elasticity may drive users to its platform. Under current law, the MFNs would receive more scrutiny as they have limited efficiency benefits when used by dominant retailers that do not advertise on a product-specific basis. Acquiring data from suppliers to compete in product markets would likely be upheld because of extremely important consumer benefits and limited exclusionary effects. If the rules were changed such that all vertical contracts by dominant platforms that offer preferential or non-preferential terms to certain suppliers or retailers were to be treated with special scrutiny, courts could end up reviewing MFNs and data provision contracts very similarly.
Perhaps the best example of the virtues of current doctrinal flexibility in balancing is United States v. Microsoft, decided in 2001. The case belies a reason for reform that on its face may be broadly appealing: namely, current balancing tests are overly concerned with short-term price competition and overlook innovators with the capacity to outcompete the incumbent in the future. In Microsoft, the D.C. Circuit found Sherman Act violations for exclusionary tying and vertical contracting. At issue was the Java programing language, which allowed software applications designed either by or for Microsoft’s operating system to run on any operating system. The Justice Department challenged many practices Microsoft allegedly employed to stymie Java’s development and market adoption, including conditioning its chip contracts with Intel and requiring hardware manufactures to only install Microsoft’s web browser, Internet Explorer. The D.C. Circuit found a violation because Microsoft did not offer a viable efficiency justification for those practices. Notably, the government never proved that Java improved price or quality relative to the Microsoft operating system. Instead, the Court emphasized that when a business practice has no justification other than exclusion, somewhat speculative future costs are enough to condemn it. Additionally, the D.C. Circuit reasoned that exclusionary barriers that provide no competitive justification are likely those that are the most difficult for an efficient competitor to overcome.
Competitive process scholars are well aware that the government won its case against Microsoft despite having no evidence of price or quality harm. They claim that the D.C. Circuit implied such harm from the innovation barriers and warn that courts that insist on more rigorous proof of exclusionary harm would not have reached the same result. This is a misinterpretation of the Microsoft opinion. Rather than exemplifying a less rigorous (and thereby more expansive) accounting of exclusionary costs, the crux of the opinion regards the lack of efficiency benefits. The Microsoft case thus shows that current doctrine will not tolerate limited exclusionary effects when no competitive benefit justifies them, but also will not exaggerate exclusionary consequences of efficiency-enhancing conduct. Adding presumptions about broad categories of conduct without regard for product and business strategy specifics risks judges moving away from the analytical virtues that Microsoft exemplifies.
Predatory Pricing
Competitive process school adherents also propose changes to the rules governing predatory pricing. They argue that current doctrine’s below-cost and recoupment requirements are particularly ill-suited to digital markets and advocate more doctrinal fluidity. They prefer an approach where plaintiffs could make out a predatory pricing case based on allegations of the defendant’s strategy and market dynamics alone, and defendants could possibly rebuff that prima facie case by providing a non-predatory strategy that explains the pricing behavior.
Those proposed reforms ought to be rejected. Digital market dynamics do not make it materially easier for a dominant firm to exclude a more efficient competitor via predatory pricing. The same network effects and low marginal costs that minimize the losses taken by a dominant digital firm attempting predation also reduce the costs its competitor must bear to fight back. Capital available to potentially profitable digital firms also makes it more difficult for a dominant firm to successfully predatorily price out an equally efficient competitor. Reforms would thus only succeed at encouraging frivolous lawsuits at best and may deter beneficial price competition or lead to firms being held liable only for their superior efficiency at worst.
Current law views predatory pricing claims as especially suspect because low prices most often reflect efficiency. The predatory pricing test, which comes from Brooke Group v. Brown & Williamson Tobacco (1993), attempts to avoid condemning large firms for their scale or efficiency alone by requiring a plaintiff to show both that the defendant firm is pricing below cost and that the firm has a significant likelihood of recouping its losses via subsequent monopoly pricing. The below-cost and recoupment factors together serve to disambiguate a dominant firm’s scale-related superior efficiency from that firm’s taking temporary losses for the sole purpose of removing a competitor and charging monopoly prices.
According to the competitive process school, the below-cost and recoupment requirements are insufficient to capture all harmful predatory pricing, especially in tech-related cases. Because of network effects and the inputs needed to produce digital products, marginal costs decline rapidly in many tech businesses. Therefore, the dominant firm may be able to exclude a nascent competitor in the process of becoming an efficiency or quality leader without pricing below cost. Consider a hypothetical monopolist in a market for specialized office management or productivity software. The monopolist has high fixed costs — namely the human capital needed to develop the software, address systemic technical issues, design updates, and the like — and low marginal costs, especially because its product is well known and it can thus devote less time to sales, enterprise relationships, or tailoring to new customers’ needs. A new firm entrant will likely have higher marginal costs than the dominant firm because of higher sales and account management expenses, as well as a greater need to offer tailoring or other enhanced services to differentiate itself. The dominant firm could thus undercut the new entrant while still pricing above costs. Its ability to do so would reflect network effects and first-mover advantage more so than pure efficiency, and it could recoup losses by ending its targeted discounts once the nascent competitor leaves the market.
Competitive process school scholars also argue that the current doctrine is insufficient for preventing a first-moving firm from obtaining a dominant position in a new digital platform industry. A firm’s position as the dominant digital platform in an industry yields a host of benefits, they argue, such as opportunities for profitable vertical integration and data-driven price discrimination. Notably, these benefits are not limited to supra-competitive pricing, which is the touchstone of a recoupment analysis. Therefore, the first mover in a nascent digital industry may price well below a maximizing level to attract as many users to its platform as quickly as possible. Once it achieves dominant status in its initial market, the firm then has the scale and revenue needed to vertically integrate and price discriminate.
These tactics increase profit without monopoly pricing and protect the dominant firm from future challengers, according to the competitive process scholars. In a 2017 law review article, Kahn argues that Amazon used predatory pricing to achieve dominant platform status; the article received favorable citations in the New York Times and the Washington Post as well as from the House Antitrust Subcommittee. Kahn also argues that Uber is using predatory pricing with the aim of becoming the dominant rideshare platform, which in her view explains its high market capitalization despite unprofitability.
Competitive process scholars are incorrect that low marginal costs, network effects, and benefits of dominant platform status justify predatory pricing liability without below-cost pricing and recoupment via monopoly pricing. First, a steeply declining marginal cost curve from network effects cuts both ways: it may allow an established firm to price far below the profit-maximizing level for an extended period, but it may also provide a “safe harbor” for a startup. Once the startup reaches the market share where marginal costs become negligible, its bleeding from the price war will stop while the established firm will have forgone significant revenue. In this regard, predatory pricing seems less likely to be successful in a digital market than in a traditional industrial market where marginal costs are significant even at the most efficient scale. The low-marginal-cost safe harbor will allow a new entrant that has true potential to compete with the dominant player on price and quality to secure the capital market funding needed to scale even as it takes initial losses.
Second, firms that achieve dominant status in a digital platform market may gain advantages related to vertical integration and price discrimination, but these advantages do not include excluding a more efficient competitor. Vertical integration usually yields efficiency benefits from reduced transaction costs and production process improvements. To the extent that vertical integration excludes for reasons other than efficiency, it should be handled under the rules governing vertical conduct and not predatory pricing. Price discrimination enabled by a dominant firm’s data access is not inherently exclusionary because it gives no advantage over an upstart competitor. This is because the upstart competitor does not need data to know its potential customers are existing customers of the dominant firm. Consider Amazon’s home-furnishings competitor Wayfair. Amazon may be able to use consumer data to identify people who buy furniture on its platform and offer them discounts, but Wayfair also knows (assuming Amazon is a monopolist) that all possible consumers interested in its platform are Amazon customers, allowing it to target them equally well.
Wayfair and Amazon’s competitive relationship is, moreover, emblematic of the importance of the below-cost and recoupment requirements (the recoupment requirement especially) in preventing false positives in the context of digital platform competition. Wayfair has successfully established itself as an online platform connecting buyers and sellers of furniture. It has been able to do so despite Amazon’s structural advantages because of vast interest among capital market participants in possible successful competitors to Amazon’s online business. This support allows Wayfair to price and expand aggressively despite its current inferior scale. Additionally, Wayfair has deeper vertical relationships with its suppliers than Amazon. Because of this advantage, Wayfair can outcompete Amazon on matters such as quality control and user friendliness.
Assume, hypothetically, that Amazon’s response to Wayfair’s competitive innovation was to vertically integrate and attempt to cultivate loyalty to its own platform. Amazon then prices furniture below its initial costs to attract an initial customer base and provide proof of concept as it considers how much to scale. Amazon’s initial low-cost pricing, mixed with its platform dominance, would serve as prima facie evidence of a violation if predatory pricing law had no recoupment requirement and incorporated default platform skepticism. Given that Amazon could not test the waters with this loss-leading strategy without paying enforcement-related costs, it might decide that full vertical integration simply is not worthwhile. And if Wayfair knows it can hinder Amazon’s integration by filing suit, it might choose that route instead of drawing on the capital market to improve its own production and supplier relationships. The recoupment requirement thus ensures that the dominant firm can engage in legitimate business strategies associated with improving efficiency and that upstart competitors use available means to improve cost in response to such competition.
Conclusion
Exclusionary conduct by a dominant firm presents difficulties for courts because both procompetitive and anticompetitive conduct may cause exclusion. The notion that only exclusionary conduct that stems from dominance and not efficiency should be unlawful, while a useful heuristic, only goes so far. Conduct often excludes both because of a firm’s dominance and because of its efficiency.
Given that difficulty, the federal courts have developed doctrine to ensure that dominant firm conduct that is beneficial on the whole is not condemned. Regarding vertical exclusion, balancing analysis promotes a rigorous accounting for efficiency benefits and a fact-specific inquiry that considers matters like capital availability, switching costs, and product differentiation among suppliers to determine the height of the exclusionary barriers. As for predatory pricing, the below-cost and recoupment-via-monopoly-pricing requirements help differentiate an established, dominant firm temporarily taking losses to bankrupt a competitor from legitimate low-price strategies related to efficient scaling.
Competitive process school adherents argue that the law is not strong enough to prevent dominant digital firms from excluding new competitors that have the potential to outcompete in the future. The growth path of innovative new firms is, of course, an important policy concern, but current law does a very good job of protecting these firms from unjust exclusion without saving inefficient firms in the process.
Nascent competitors may need to take risks and act aggressively to compete with the dominant digital firms because of the current exclusion doctrines. But such competitive pressure encourages innovation and benefits consumers. By contrast, an antitrust law that unduly shields young firms would undermine the competitive process rather than promote it.
Readings
- “Amazon’s Antitrust Paradox,” by Lina Kahn. Yale Law Journal 126(3): 710–806 (2017).
- “Antitrust and Innovation: Welcoming and Protecting Disruption,” by Giulio Federico, Fiona Scott Morton, and Carl Shapiro. Innovation and the Economy 20: 125–190 (2019).
- Big Business: A Love Letter to an American Anti-Hero, by Tyler Cowen. St. Martin’s Press, 2019.
- “Google, You Can’t Buy Your Way Out of This,” by Tim Wu. New York Times, October 22, 2020.
- The Antitrust Paradigm: Restoring a Competitive Economy, by Jonathan Baker. Harvard University Press, 2019.
- The Curse of Bigness: Antitrust in the New Gilded Age, by Tim Wu. Columbia Global Reports, November 13, 2018.