The Executive Branch, Congress, and the federal enforcement agencies are focused on the competitive impact of large, allegedly dominant technology platform companies. The House’s Competition in Digital Markets majority report recommends major changes to antitrust law, but such changes are unlikely to be adopted. The Federal Trade Commission has adopted an expansive policy statement on the scope of what constitutes unfair methods of competition, but the policy statement appears inconsistent with developments in antitrust law over the past four decades and thus may not be sustainable. As an alternative to legislative change and revival of less credible theories of competitive harm, this article proposes that the antitrust enforcement agencies promulgate Platform Competition Guidelines that articulate a competitive effects analysis within the scope of the burden shifting approach applied in the horizontal merger guidelines.
By Bilal Sayyed[1]
I. INTRODUCTION
The majority staff report of the House of Representative’s recent Investigation of Competition in Digital Markets (“Majority Report”) recommends (i) that “Congress consider revitalizing the ‘‘essential facilities’’ doctrine … [and] overrid[e] judicial decisions that have treated unfavorably essential facilities and refusal to deal-based theories of harm.”[2] The majority report made several other recommendations to “rehabilitate monopolization law” because “over recent decades, courts have significantly heightened the legal standards that plaintiffs must overcome in order to prove monopolization.”[3]
According to the majority, Congress should also consider: (ii) extending the Sherman Act to prohibit abuses of dominance; (iii) creating a statutory presumption that a market share of 30 percent or more constitutes a rebuttable presumption of dominance by a seller, and a market share of 25 percent or more constitutes a rebuttable presumption of dominance by a buyer; (iv) revising the law of monopoly leveraging to not require “actually monopoliz[ing] (or dangerously threaten to monopolize) the secondary market; (v) revising the law to not require recoupment in predatory pricing or predatory buying cases; (vi) clarifying the law of tying, to confirm that “conditioning access to a product or service in which a firm has market power to the purchase or use of a separate product or service is anticompetitive”; and (vii), whether making a design change that excludes competitors or otherwise undermines competition should be a violation of Section 2, regardless of whether the design change can be justified as an improvement for consumers.[4]
In addition, the majority report believes Congress should consider (viii) overriding the Supreme Court case of Ohio v. American Express, by clarifying that cases involving platforms do not require plaintiff’s to establish harm to both sets of customers; (ix) overriding the district court case of United States v. Sabre Corp., by clarifying that platforms that are two-sided, or serve multiple sets of customers, can compete with firms that are one-sided; (x) clarifying that market definition is not required for proving an antitrust violation, especially in the presence of direct evidence of market power; and (xi) clarifying that false positives (erroneous enforcement) are not more costly than false negatives (erroneous non-enforcement), and that, in relation to conduct or mergers involving dominant firms, false negatives are costlier.[5]
Congress seems unlikely to adopt such an ambitious rewrite of the antitrust laws, and it is at least questionable whether the changes recommended by the Majority Report are likely to be welfare enhancing.
II. THE FTC’S NEW POLICY STATEMENT ON UNFAIR METHODS OF COMPETITION IS NOT SUFFICIENT TO DEVELOP THE LAW ON SINGLE FIRM CONDUCT AS APPLIED TO TECHNOLOGY PLATFORMS
As an alternative (or complement) to a revision of the statutory language of Section 2, the Federal Trade Commission seems poised to attack single firm conduct through a reinterpretation (to others, a reinvigoration) of the scope of its authority to prohibit unfair methods of competition. In November 2022, the Commission issued its interpretation of the scope of conduct prohibited as an unfair method of competition.[6] Recognizing, as the Supreme Court has, that the prohibition on unfair methods of competition is broader than the Sherman Act, the Policy Statement argues that Section 5 of the Federal Trade Commission Act prohibits conduct that (i) is a method of competition; (ii) may be coercive, exploitative, abusive, deceptive, predatory, involve the use of similar economic power, or is otherwise restrictive or exclusionary (under certain circumstances); and, (iii) tends to negatively affect competitive conditions.[7]
According to the statement, the size of the firm engaging in such conduct, its purpose, and its power, are relevant to a determination of the conduct’s legality.[8] Conduct is evaluated for its tendency to “generate negative consequences” including “raising prices, reducing output, limiting choice, lowering quality, reducing innovation, impairing other market participants, or reducing the likelihood of potential or nascent competition.”[9] Harm need not be show by quantifiable evidence.[10] A showing of market power or market definition is not required.[11] Section 5, according to the Commission, does not require that conduct be evaluated for its efficiency or justification, and a defense based on such will not generally be available.[12]
The Commission’s revised statement on Section 5 adopts, as unfair methods of competition, much of what the Majority Report indicated was appropriate for Congress to reconsider about the Sherman Act. The Commission advocates for finding illegal under a significantly looser set of conditions, conduct that may not violate Section 2 (in part because of efficiencies or cognizable business justifications): examples include (but are not limited to) de facto (and actual) loyalty rebates, tying, bundling, exclusive dealing, parallel conduct that may cause aggregate harm, leveraging of market power (not limited to monopoly power), discriminatory refusals to deal, practices that entrench market power, and conduct that, when considered in its entirety (or cumulatively), undermines competitive conditions.
But the Commission’s view of Section 5 may not be enforceable. In the late 1970s and early 1980s, the Commission alleged violations of Section 5 that appellate courts rejected as insufficiently grounded in evidence,[13] inconsistent with a firm’s obligations,[14] and insufficiently defined as unfair and inconsistent with competition.[15]
Moreover, the Commission’s reading of Section 5 may prohibit efficient conduct. While the Commission believes Section 5 does not require an inquiry into the efficiencies associated with the alleged unfair competition, this view may not be sustained by the courts. The Supreme Court (and the lower courts) are clear that efficient conduct should not be subject to per se illegality, and that such efficiencies must be considered in determining the competitive effect of a practice. In Brunswick Corp., the Court found that antitrust injury was absent where a plaintiff alleged that an illegal acquisition threatened to bring a “deep pocket” parent into a market of “pygmies.”[16] In GTE Sylvania, the Court recognized the efficiency rationale of territorial and location-based restraints on intra-brand competition in support of competition at the inter-brand level and overturned the per se ban on certain vertical non-price restraints.[17] In BMI, the Court recognized the procompetitive rationale of a “blanket” music license as a reason to forego per se treatment of a horizontal agreement among the members of two music societies on license terms.[18] In Khan, the Court accepted the efficiency rationale of a vertical maximum price-setting arrangement, and overturned the per se ban on vertical, maximum price setting agreements.[19] In Leegin, the Supreme Court overturned the per se ban on vertical minimum price setting agreements because of the potential efficiencies associated with such agreements.[20]
Because the Commission’s interpretation of Section 5 is inconsistent with modern antitrust law, provides limited guidance, and because only the Commission, and neither the Department of Justice nor private plaintiffs can enforce Section 5, any attempt to expand or strengthen laws against anticompetitive unilateral conduct through an expansive interpretation of Section 5 will likely have only limited (if any) effect. An alternative approach to reinvigorating monopolization law should be considered.
III. THE AGENCIES SHOULD ISSUE “PLATFORM GUIDELINES” TO ADVANCE & DEVELOP ANALYSIS OF SINGLE FIRM CONDUCT BY ALLEGEDLY DOMINANT TECHNOLOGY PLATFORM COMPANIES
Developing and promulgating agency enforcement guidelines that provide an analytic framework for the evaluation of conduct by technology platforms would provide valuable guidance and could be a valuable contribution to the development of monopolization law.[21] Such guidelines are preferred to revisions, including substantial revisions, to the existing statutory framework because they are likely more flexible and avoid per se condemnation of certain conduct. For the same reason, guidelines are preferred to competition rules.[22] In this short article, I set out a framework and key elements that Platform Competition Guidelines (“PCGs”) should include.
The PCGs, as applied, should identify conduct that has the actual or potential effect of reducing market-wide output, quality, or innovation, and/or increasing firm or market-wide prices as compared to an alternative in which the conduct did not occur. – The PCGs should not condemn conduct that is pro-competitive or that is competitively neutral.
The Supreme Court has emphasized concerns about over-enforcement of the antitrust laws, without a firm theoretical or empirical basis. The PCGs must balance concerns about overenforcement and underenforcement, and should strive to minimize the sum of Type 1 errors (overenforcement) and Type 2 errors (underenforcement).
IV. OVERVIEW OF TECHNOLOGY PLATFORMS
Technology platforms are software-based businesses that provide services to multiple participants (consumers or other businesses) and often operate in multiple markets. Technology platforms may operate in several related markets. Markets in which technology platforms operate may have several adjacent and often, complementary markets. Technology platforms serve one or more distinct customer (or user) groups and allows for interaction between customers or customer groups. An important function of platforms is to attract both groups of customers in sufficient numbers or in a suitable balance to capture the value of positive (direct or indirect) network effects and to minimize negative network effects. Some—but not all—technology platforms operate “multi-sided transaction platforms” that facilitate transactions or other types of interactions between distinct groups of users. Some multi-sided transaction platforms offer multi-sided “marketplaces.” In some or all of their offerings, technology platforms may compete with non-platform firms. In many cases, firms build their business “on top” of a platform’s software.
Owners of technology platforms may be vertically integrated in input (upstream) and/or output (downstream) markets and provide platform services to itself as well as to others. Thus, a technology platform may also compete directly with its own customers by offering the same or similar products or services on its platform as its customers. The owner of a technology platform that also serves as a multi-sided marketplace may integrate and sometimes also be a “user” — typically a seller or content supplier — on one side of the platform, in addition to operating the platform itself.
Two common features of platforms are relevant to inquiries into a specific platform’s market power. The first is “multi-homing,” which occurs when a platform user connects to multiple platforms simultaneously or in relatively quick succession. For example, a consumer is multi-homing if she uses multiple ride-sharing apps in deciding whether to book a ride. Technology platforms may have an incentive to limit multi-homing, either through conduct that limits the ability of rivals to compete, or through acquisition of competing platforms. Multi-homing can occur on only one side of a platform or can occur on multiple sides. In the ride-sharing example, if both riders and drivers use multiple platforms, then both sides are multihoming, whereas if drivers tend to use a single platform, then only riders are multi-homing. The second is “switching costs,” which refer to the costs platform users must bear in switching from one platform to another. Switching costs are lower when platforms are interoperable. All else equal, lower switching costs imply that a platform has less ability to exercise market power or obtain (and maintain) monopoly power; firms have an incentive to increase switching costs.
V. POTENTIAL COMPETITIVE HARMS
A technology platform can harm competition in the platform market or the adjacent user markets. It can do so in the same ways that any business can harm competition: by entering into anticompetitive agreements; by monopolizing, attempting to monopolize, or conspiring to monopolize relevant markets (including by acquiring nascent or potential competitors); or by undertaking acquisitions that significantly lessen competition or tend to create a monopoly. The presence of direct[23] and indirect[24] network effects may make it difficult for a new entrant offering a competitively superior product to enter, expand, and successfully provide a competitive alternative to an incumbent platform.
As noted earlier, a technology platform that also serves as a multi-sided marketplace may integrate and sometimes be a “user” — typically a seller or content supplier — on one side of the platform, in addition to operating the platform itself. Such integration, on its own, typically does not necessarily harm competition and may be procompetitive, in that the platform offers a competitive alternative to third-party sellers or content providers using the platform.
However, in some circumstances, a technology platform that both operates the platform and competes in an adjacent market may have the incentive and ability to exclude a competitor, in full or in part, sufficient to harm competition.[25] Different theories of harm require analysis of the competitive conditions in the various relevant markets. A platform may harm competition in the “core” platform market by engaging in anticompetitive conduct that may either (a) limit the ability of a rival platform to compete, or (b) prevent a firm in an adjacent market from growing or achieving scale, if the firm in the adjacent market is or may in the future become a competitive threat in the core platform market. An investigation of this theory of harm would typically require analyzing current and/or anticipated future competitive conditions in the core platform market.
A technology platform may also harm competition in an adjacent market directly: for example, by engaging in anticompetitive conduct to acquire or maintain market power in that market, or by entering into agreements that restrain competition and harm consumers in the adjacent market. An investigation of this theory of harm would typically require analyzing current and/or future competitive conditions in the adjacent market, which may differ from competitive conditions in the core “platform” market.
If the technology platform is vertically integrated – where the platform acts as a seller or content provider on the platform it operates, a rival may also be a customer of the platform (rather than the operator of a separate, competing platform). In this context, the Agencies should assess whether harm to an unintegrated rival is connected to harm to competition, and therefore whether and how the platform’s conduct affects consumer welfare. This may be shown by evidence that the platform has raised its unintegrated rival’s costs, with the effect that competition in the market in which the unintegrated rival competes is reduced, or through some other coherent economic theory of harm.
VI. POTENTIAL COMPETITIVE BENEFITS
Platforms can create economic benefits by reducing transaction costs, i.e. by facilitating transactions between actual and potential trading partners. In some cases, platforms, by creating a market, enable transactions to occur that otherwise would not have occurred. In other cases, platforms disrupt existing marketplaces by enabling the sale of goods, services, and content at lower prices and/or higher quality. They may also allow customers to find goods, services, and content they would not otherwise have been able to find. The popularity of many large platforms is likely related to these benefits.
Conduct such as vertical integration — which can occur via entry, merger, or contract –may appear to be restrictive, but also can offer procompetitive benefits. Such practices can improve competition, by, for example, reducing transaction costs, eliminating double marginalization, and otherwise aligning companies’ incentives in a way that benefits consumers.
Platform operations often involve economies of scale. Such markets may be most efficiently organized with just a few large suppliers. Acquisitions of competing platforms or assets used to create, operate or expand platforms, exclusive contracts with users, or other actions that increase the volume of business flowing through a platform can allow the platform to benefit from scale economies and become a lower cost or otherwise more efficient competitor.
Technology platforms may engage in conduct or enter into various agreements that generate efficiencies and benefit platform users or are otherwise justified. For instance, certain conduct and agreements may enable the platform to better satisfy user preferences (including conduct that benefits one set of users but potentially harms another), to reduce transaction costs between users, to reduce double marginalization, or otherwise to enable firms to offer goods or services that are cheaper, more valuable, or brought to market faster than would otherwise be possible. Accordingly, conduct that produces anticompetitive effects may nevertheless be justified on the ground that it is procompetitive, in that it “is indeed a form of competition on the merits because it involves, for example, greater efficiency or enhanced consumer appeal.” There are certain types of unilateral conduct in which a technology platform might engage that, without more, will tend not to harm competition and/or will tend to produce efficiencies. These include (1) vertical integration by entry; (2) product design decisions; and (3) above-cost price reductions. These categories of behavior are typically competitive responses and attempts to deliver value to consumers.
Other conduct may plausibly have mixed effects, such as contractual arrangements between a platform and another firm that provides a complementary product or service. Such vertical restraints can yield an anticompetitive effect, but they also have the potential to create cognizable efficiencies, such as allowing firms to streamline production, inventory management, or distribution, or create innovative products in ways that would have been difficult to achieve without a contractual arrangement that includes restrictive terms.
VII. SPECIAL FACTORS IN COMPETITIVE EFFECTS ANALYSIS OF CONDUCT BY TECHNOLOGY PLATFORMS
Each platform business is different, as are the markets in which such businesses compete. Nevertheless, some factors are present in many platform markets, and these factors may be relevant to the competitive impact of the conduct of technology platforms.
A. Data
Although the use of data to make competitive decisions is not unique to technology platforms, firms that operate technology platforms often have access to very large data sets. The competitive significance of data may vary significantly from one case to another in light of the nature or the relevant market or markets; the nature of relevant business models in those markets; and the source, content, nature, breadth, utility, and availability of competitively significant data.
Consistent with the foregoing, the utility and competitive significance of data may vary by market, by time, and by characteristics specific to the data. Different data or data sets may have different and varied characteristics. Some data may be ubiquitous, easily replicable, readily substitutable, or of limited or transitory value. Other data may provide significant and hard-to-replicate competitive advantages for the firms that own or control them. A need for data may impede expansion or entry of an actual or potential competitor, particularly when there are few or no commercially available alternatives to the data, and when access to such data is competitively important in some way.
The analysis of competitive effects of conduct (including acquisitions) should consider whether the relevant data is easily replicable and/or whether rivals or entrants can compete effectively without the data. Data may be competitively significant for one or more reasons, including the nature (e.g. historical, real-time, etc.), breadth and depth (e.g. varied, voluminous, etc.), utility (e.g. basis for better analytics or new products and services, etc.), and availability (e.g. costly, time-intensive, and hard-to-replicate alternatives, etc.) of the data. The process by which a technology platform collects, uses, and shares data may also inform the antitrust analysis.
In markets where data or data-derived products and services is a key differentiator, a technology platform’s access to a distinctive or competitively significant dataset (or to large amounts of competitively significant data) may allow it to engage in conduct to protect its position in a market. Evidence that a technology platform selectively denies rivals access to data to prevent smaller rivals from efficient distribution channels or sufficient sales to operate at sufficient scale is relevant to competitive effects analysis. Similarly, agreements that prohibit customers from sharing data are relevant to a competitive effects analysis.
In some cases, particularly cases involving mergers or acquisitions, there may be a relevant market for certain types of data or for access to certain types of data.[26] Anticompetitive transactions or other practices that affect such markets may substantially lessen competition or tend to create a monopoly or may constitute anticompetitive conduct.
A platform owner that also operates business units that compete with other users of the platform may, by virtue of owning the platform, obtain competitively sensitive information from these users. The access and use of such information by a platform owner’s business units that compete with other users, including any related agreements providing for the transmission of such data by platform users, may result in reduced competition that harms consumers.
B. Network Effects & Multi-Sidedness
The existence, scope and strength of direct and indirect network effects are factual questions relevant to the competitive effects analysis. Network effects exist on a continuum, where at one endpoint on the continuum, any indirect network and feedback effects are strong and at the other endpoint the effects are weak. The strength and degree of direct or indirect network effects may be different on different sides of a firm’s technology platform, and the network effects may be positive or negative. The Agencies should be mindful that the degree and nature of network effects can change, sometimes rapidly, in response to a new technology or business model. They should take account of information regarding the presence and strength of network effects in evaluating technology platform conduct. The presence of network effects does not insulate a technology platform from antitrust scrutiny, and the Agencies may determine that a technology platform that benefits from positive (direct or indirect) network effects has engaged in, or is engaging in, conduct in violation of the antitrust laws.
Strong indirect network effects may make it more likely that an incumbent technology platform can profitably and effectively engage in anticompetitive or exclusionary conduct. In these circumstances, a new entrant technology platform must attract sufficient numbers of customers on both sides of the platform to create value, and the indirect network effects operating across customer groups can make entry more challenging than if indirect network effects were not present.
Multi-sided platforms can benefit from “feedback effects.” Indirect network effects can sometimes be self-reinforcing. If indirect network effects are positive in multiple directions, then these effects can build upon one another. Using the example of a computer operating system, more consumers leads to more software applications, and more software applications leads to more consumers, yielding a “virtuous cycle” of growth for the platform. Feedback effects can also be negative, in that losing users on either side can lead to a “vicious cycle,” whereby the platform’s scale decreases quickly.
The nature of competition may be different on each side of a multi-sided technology platform. The Agencies should consider whether competition on one side of the platform makes it less likely that a technology platform can engage in anticompetitive or exclusionary conduct on other sides of the platform. Competition on one side of a technology platform can be insufficient to prevent an unlawful exercise of market power, especially in the face of significant market power on one side. In assessing conduct directed only or primarily at one side of the platform, the Agencies should consider the degree of competition on other sides of the platform, and any other competitive constraints on a technology platform.
C. Non-Price Competition
Anticompetitive effects may manifest themselves in price effects, non-price effects, or both. Non-price effects include, among other things, effects on output, quality, variety, and innovation. An enforcement action may be based on actual or anticipated price effects only, actual or anticipated non-price effects only, or both.
In some matters involving platforms, effects on non-price competition may be especially important because price effects may be difficult to detect or to measure. Many platform businesses provide a product or service to one set of users at a nominal price of zero, which does not change over time. In this context, in detecting whether such users have been harmed, it may be appropriate to focus the competitive effects analysis on non-price effects.
Even when a product or service is provided for a nominal price of zero, effects on competition and welfare may manifest, in whole or in part, in changes in output, variety, the rate of innovation, and/or quality (including terms of service to users and businesses, or privacy and data security practices). It may be appropriate to challenge practices or transactions that harm consumers solely through such non-price effects.
With regard to effects on output, a practice may, or be reasonably likely to, harm competition even when market-wide output is increasing over time (or when nominal price is declining or unchanging). The relevant comparison is between the actual or anticipated effects of the challenged conduct and actual or anticipated scenarios where the conduct has not occurred or does not occur. Such actual or anticipated scenarios where the challenged conduct does not occur may also involve market-wide output increasing.
With regard to effects on innovation, the Agencies should consider how a challenged practice or transaction affects or may affect innovation, compared to an alternative in which the practice or transaction did not occur. Markets in which platform businesses compete — when competitively healthy — may be characterized by intense innovation. Accordingly, effects on innovation may be, and in some instances, must be a significant focus in matters involving technology platforms.
Anticompetitive effects on innovation could take the form of a reduced incentive to continue existing product-development efforts or reduced incentives to develop new products. A competitive effects analysis should consider whether the practice or transaction enables or may enable innovation to occur that would otherwise not take place. Both incumbents and entrants, and likewise both large firms and small firms, engage in innovation that increases welfare. As in the analysis of effects on other factors relevant to competition, the focus should be on market-wide effects on innovation, of which effects on individual firms’ incentives and abilities to innovate is a part.
The analysis of effects on innovation may include both quantitative and qualitative evidence. Reliable quantitative evidence regarding how a business practice affects innovation may not be available or of limited utility. Quantitative evidence should not be necessary to determine that a firm’s conduct has produced harm to innovation.
D. Monetizing Platforms and Business Models
Technology platforms may use different business models and may monetize their products and services in various ways. This may include earning revenue through the sale of advertising, or by charging fees to users on one or more side of the platform for access to the platform. Some platforms may use a mix of strategies to earn revenue, including charging fees to users and selling advertising opportunities.
Platforms can experiment with different approaches to earn revenue. They may change their approaches over time for various reasons, such as responding to new competitive conditions and expanding their product or service offerings. Innovation in business models may be as important and as valuable as innovation in technology, but just as it may benefit consumers, it may also harm them. It may be appropriate for the Agencies to investigate such changes to determine whether the changes reflect or are part of a platform’s effort to impair the competitive process. For example, it may be appropriate to investigate whether a platform’s change to its approach to earning revenue is a mechanism through which the platform can raise the cost of using a competitor’s platform.
VIII. BURDEN SHIFTING APPROACH TO THE ANALYSIS OF COMPETITIVE EFFECTS
The Agencies should follow a structured but flexible analytic framework to identify the competitive effect of conduct (including acquisitions) by technology platforms. The Agencies should challenge only that conduct that has demonstrable anticompetitive harms caused by that conduct that are not outweighed by conduct specific, cognizable, and verifiable procompetitive benefits caused by that conduct. When required to balance harms and benefits, the Agencies should consider whether the platform could achieve similar benefits through less restrictive or exclusionary alternatives. In evaluating whether any anticompetitive harms outweigh any anticompetitive benefits, the focus must be on the state of competition with, as compared to without, the practice under review.
To avoid determining competitive effects by “characterization” of conduct, the analytic framework should consist of 3-to-5 steps, set forth below.
- First: The Agency must identify a plausible basis for alleging the conduct has had or will have an anticompetitive effect. It should identify the significance, magnitude, and likelihood of that effect.
- Can the Agency demonstrate that the firm has market power or monopoly power? Conduct that may be benign or procompetitive when engaged in by a firm — even a firm operating a technology platform — that does not possess significant market power, may have anticompetitive effect when engaged in by a platform with significant market or monopoly power. [27]
- Anticompetitive effects may manifest themselves in price effects, non-price effects, or both. Non-price effects include, among other things, effects on output, quality, variety, and innovation. An enforcement action may be based on actual or anticipated price effects only, actual or anticipated non-price effects only, or both.
- Harm to rivals must be connected to harm to competition to support an enforcement action.[28] Conduct by a platform that harms a rival or a third party (or a group of similarly situated rivals or third parties) is typically insufficient by itself to support a challenge to the practice. Such harms may be part of a body of evidence that supports a challenge to the practice, however, as harm to a rival can be consistent with a viable theory of harm to competition.
- Second: If the Agency makes the requisite showing of harm, the Respondent may rebut this showing by establishing that its existing or threatened future market power or monopoly power is not durable because of customer or competitor response.
- Ease of entry or ease of significant or material expansion will often defeat a claim of future harm associated with existing market or monopoly power.
- Third: Alternatively, the Respondent may identify efficiencies or procompetitive benefits linked to the conduct. The Respondent has the burden of showing the significance, magnitude, and likelihood of the procompetitive effect.
- Whether such a justification for conduct is convincing will depend upon facts specific to the conduct or transaction under review or consideration.
- The Agencies should evaluate any efficiency claim to determine whether it represents a non-pretextual claim that the conduct is a form of competition on the merits: perhaps it involves or creates, for example, greater efficiency or enhanced consumer appeal, or a lower cost structure. Conduct that is reflective of a “lower cost structure” is typically competition on the merits.[29]
- The identification of efficiencies or other procompetitive benefits is not sufficient to rebut the Agencies’ claim (and proof) of exclusionary or predatory conduct, but in the presence of durable monopoly power, cognizable efficiencies are necessary to rebut the Agency’s proof of exclusionary or predatory conduct.
- If a platform offers procompetitive justifications for its conduct, the claimed benefits must be related to the conduct under review, i.e. “conduct-specific,” in addition to being verifiable and cognizable. The general economic benefits the platform has created will typically not be relevant to the determination whether the conduct is anticompetitive. Rather, the platform must explain and show, with evidence, how the specific conduct at issue benefits (or does not harm) competition.
- Fourth, if market power or monopoly power is durable and that there is a plausible basis (and credible evidence) for both the harm alleged and for non-pretextual procompetitive justifications for the conduct, the Respondent must show that there is no less restrictive alternative that would allow the relevant efficiencies to be obtained. Such alternatives must be practical, not merely theoretical.
- Fifth, if the Respondent makes such a showing, the Agency should determine, and be prepared to prove that, on balance – on “net” –the conduct is harmful or beneficial to competition.[30]
- The Agencies should compare the likelihood and magnitude of anticompetitive effects with the likelihood and magnitude of efficiencies to determine the likely or actual overall effect. As the expected harm of the conduct or agreement increases, the required offsetting benefits should also increase.
- In analyzing harms and benefits, it is necessary to consider the scope and strength of the evidence of actual or likely effects. Mere assertions of potential affects or a business justification are insufficient. However, neither the Agency nor Respondent must identify and weigh each anticompetitive and procompetitive effect with specificity and precision. Such analysis may not be possible or efficient in an individual investigation.
- Finally, where remedies are required, they should be designed to: (i) address the competitive harm from the conduct or transaction; (ii) fit the facts of the case and characteristics of the relevant market, which requires a close and logical nexus between the theory of harm and the remedy; (iii) focus on preferred and time-tested approaches, though novel remedies may be appropriate in some contexts; and (iv) preserve efficiencies to the extent such efficiencies are consistent with effective relief. The party proposing the remedy has the burden of showing the remedy meets these criteria.
- Remedies should preserve or restore competition and prevent or correct the exercise of market power that has resulted in harm to competition. To the extent possible, remedies should preserve efficiencies associated with the prohibited conduct, where such remedies are consistent with effective relief. Remedies that may be appropriate include: (i) divestiture or separation; (ii) a prohibition on, or pre-consummation notice of, mergers or acquisitions; (iii) compulsory licensing, including the licensing of data sets or intellectual property; (iv) interoperability requirements; (v) non-discrimination requirements; (vi) corrective actions; and (vii) monetary equitable remedies.
The proposed approach is a competitive effects analysis, not simply a reasonableness test.[31] While this framework approximates a structured rule-of-reason analysis, much of Section 2 case law is neither a full rule of reason analysis nor a full competitive effects analysis.[32] Prior to the modern era, cases (particularly at the Supreme Court) often stopped at Step 1 – the identification, often mere characterization, of conduct as exclusionary or predatory. In the modern era, district courts and appellate courts too often stop at Step 3 – the identification of a competitive justification – and avoid the balancing or analysis of harms and benefits.
IX. KEY PRINCIPLES OF A COMPETITIVE EFFECTS ANALYSIS
Analyzing the competitive effects of conduct by technology platforms (or any firm) is a fact-specific process through which the Agencies, guided by their experience, apply a range of existing tools – analytical tools – to the available evidence. The competitive effects analysis should be focused on and limited directly on how the conduct (including acquisitions) may affect price, output, quality, and innovation.[33] The analytic focus of this inquiry must be on the actual or reasonably likely effects of the conduct, not its label. While characterizing a platform as an “essential facility” or a platform’s conduct as “an exclusive deal,” a “refusal to deal,” a “product design decision,” or “self-preferencing” may be helpful in identifying relevant prior case law or whether the conduct is engaged in by market participants without market power, such labeling should not be the focus of the inquiry.[34]
The Agencies should consider effects on each class of platform participants that are proximately affected by the conduct under review, including in markets upstream or downstream, or operating in a complementary relationship to the relevant market, and with respect to both inputs and intermediate and final products (or services). As part of this analysis, the Agencies should consider the nature of the relationship and interaction between the different sides (or different sets of users) of the platform, if a platform is operating a multi-sided business.
The analysis of the potential anti- and procompetitive effects of the conduct at issue should include, where possible, both short-term and long-term effects. A platform’s effort to prevent a nascent technology or nascent competitor from maturing into a viable competitive alternative constitutes a valid theory of harm even if the process of maturation may take some time or is uncertain. Similarly, the fact that any benefits caused by the conduct may not accrue in the near term should not preclude a determination that such benefits are procompetitive. Competitive conditions in a market or industry may change over time. Meaningful competitive constraints may emerge in a market that currently lacks such constraints; and conversely, the existence of meaningful competition today does not necessarily mean that such competition will, or is likely to, persist in the future. Such facts, if available, are germane to the evaluation of the complained of conduct. The evidentiary standard required to show anti- or procompetitive effects should be symmetrical.
The Agencies should not be required to construct a hypothetical alternative marketplace or otherwise show with specificity how competition would have occurred had the firm (the alleged dominant firm or monopolist) not engaged in the putatively anticompetitive conduct, nor to determine whether the conduct is anticompetitive. The focus of the inquiry is only on the connection between the conduct under review and any harms and benefits it may produce, in comparison to competition absent the conduct.
Nascent or potential competition may be important in markets where technology platforms are present, or, more generally, in dynamic markets – markets characterized by rapid change (or the potential for rapid change). Accordingly, the Agencies should be attentive to conduct that has the effect of excluding or harming nascent or potential competitors, including acquisitions that remove a nascent competitive threat.[35] However, they must also recognize that antitrust does not condemn competition on the merits even when that competition harms a nascent or potential competitor.
The Agencies should not be required to evaluate or determine whether a new entrant is offering a competitively superior alternative to the incumbent in order to show harm from exclusionary or predatory conduct or in conjunction with a merger; the focus of the analysis should be on whether the platform has engaged in conduct that harms competition, such as by foreclosing the new entrant from key distribution channels or key customers such that the new entrant’s costs are raised and the incumbent faces less disciplining competition to the detriment of consumers.
[1] Senior Competition Counsel, TechFreedom, and formerly the Director of the Office of Policy Planning, Federal Trade Commission. The author thanks Steve Salop for his helpful comments, and Derek Moore, Elyse Dorsey, and others in the Office of Policy Planning,” for many conversations on and analysis of the application of the antitrust laws to the conduct of platform firms. All errors and positions in the paper are of the author and not necessarily of Salop, Moore, and Dorsey.
[2] Majority Staff Report and Recommendations, Investigation of Competition in Digital Markets, Subcommittee on Antitrust, Commercial, and Administrative Law of the Committee on the Judiciary of the House of Representatives (Oct. 2020), at 336 (hereinafter “Majority Report”). The Majority Report defines the essential facilities doctrine as “the legal requirement that dominant firms provide access to their infrastructural services or facilities on a nondiscriminatory basis.” Id.
[3] Id. at 334.
[4] Id. at 334-337.
[5] Id. at 337.
[6] Federal Trade Commission, Policy Statement Regarding the Scope of Unfair Methods of Competition Under Section 5 of the Federal Trade Commission Act (hereinafter UMC Policy Statement) (Nov. 10, 2022), https://www.ftc.gov/system/files/ftc_gov/pdf/P221202Section5PolicyStatement.pdf. Commissioner Wilson issued a dissenting statement. Dissenting Statement of Commissioner Christine S. Wilson Regarding the Policy Statement Regarding the Scope of Unfair Methods of Competition Under Section 5 of the Federal Trade Commission Act (Nov. 10, 2022), https://www.ftc.gov/system/files/ftc_gov/pdf/P221202Section5PolicyWilsonDissentStmt.pdf.
[7] UMC Policy Statement at 8-9.
[8] Id.
[9] Id. at 10.
[10] Id. at 11.
[11] Id.
[12] Id. at 10-11.
[13] Boise Cascade Corp. v. FTC, 637 F.2d 573 (9th Cir. 1980) (holding that in the absence of evidence of overt agreement to utilize a pricing system to avoid price competition, the Commission must demonstrate that the challenged pricing system has actually had the effect of fixing or stabilizing prices; without such effect, a mere showing of parallel action will not establish a Section 5 violation, and declining to adopt the Commission’s suggestion that industry-wide adoption of an artificial method of price-quoting should be deemed a per se violation of Section 5).
[14] Official Airline Guides v. FTC, 630 F.2d 920 (2nd Cir. 1980) (monopolist, as long as he has no purpose to restrain competition or to enhance or expand his monopoly, and does not act coercively, retains the right to determine who it will deal with).
[15] E.I. DuPont De Nemours & Company v. FTC, 729 F.2d 128 (2nd Cir. 1984) (“Commission owes a duty to define the conditions under which conduct claimed to facilitate price uniformity would be unfair so that businesses will have an inkling as to what they can lawfully do rather than be left in a state of complete unpredictability.” Id. at 139. Commission’s “rulings and order appear to represent uncertain guesswork rather than workable rules of law.” Id.) See also FTC v. Abbott Laboratories, 853 F. Supp. 526, 536 (D.D.C. 1994) (Where the Department of Agriculture chose not to contest an improper procurement of infant formula for the special supplemental food program for woman, infants and children, Abbott “cannot be faulted and accused of unfair trade practices for failing to sue a U.S. Government territory.”)
[16] Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc. 429 U.S. 477, 487 (1977).
[17] Continental T.V. Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977).
[18] Broadcast Music v. CBS, 441 U.S. 1 (1979).
[19] State Oil v. Khan, 522 U.S. 3 (1997).
[20] Leegin Creative Leather Products v. PSKS, Inc., 551 U.S. 2705 (2007).
[21] The agencies undertook a similar effort in drafting the 2000 Competitor Collaboration Guidelines. See Remarks of Robert Pitofsky, Chairman, Federal Trade Commission, Joint Venture Guidelines: Views from One of the Drafters (Nov. 11, 1999), available at https://www.ftc.gov/news-events/news/speeches/joint-venture-guidelines-views-one-drafters. See also Ernest Gellhorn & William E. Kovacic, Analytical Approaches and Institutional Processes for Implementing Competition Policy Reforms by the Federal Trade Commission (Dec. 12, 1995) (“The lesson of these hearings to us is that the FTC’s most important assignment today is to modernize antitrust rules, to concentrate enforcement where serious systemic blocks to competition may exist, and to lessen the burden as well as improve the transparency of its enforcement efforts.”), https://www.ftc.gov/system/files/documents/public_statements/418071/951212comppolicy.pdf.
[22] This article takes no position on whether the FTC has the authority to issue competition rules.
[23] Direct network effects exist when a single user’s desire to use a network is a function of the number or identity of users from the same set of users who also use the network.
[24] Indirect network effects are a feature of multi-sided platforms. Such effects exist when the demand for the network for a user from one set of users is a function of the number and identity of users from a different set of users who also use the network.
[25] See, e.g. Thomas G. Krattenmaker & Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’s Costs to Achieve Power over Price, 96 Yale Law Journal 209 (1986); Steven C. Salop & David T. Scheffman, Raising Rivals Costs, 73 Amer. Econ. Rev. 267 (1983); and Susan Athey and Fiona Scott Morton, Platform Annexation, 84 Antitrust Law Journal 677 (2022). See also the theories of harm regarding mergers of firms offering complementary products in the 2020 Department of Justice and Federal Trade Commission Vertical Merger Guidelines and the Federal Trade Commission’s Commentary on Vertical Merger Enforcement (December 2020). The FTC has withdrawn from the Vertical Merger Guidelines and rescinded the Commentary on Vertical Merger Enforcement (but not the matters discussed within).
[26] See, e.g. Complaint, CoStar Group, No. 9398 (Nov. 30, 2020) (relevant market was “internet listing services advertising for large apartment complexes), https://www.ftc.gov/system/files/documents/cases/d09398complaintpublic.pdf; Fidelity Nat’l Fin., Inc., No. 9385, 2019 WL 4461620, at *7 (F.T.C. Sept. 5, 2019) (relevant market was “title information services”—the provision of access to title plant information); Corelogic, Inc., No. C-4458, 2014 WL 2331024, at *1 (F.T.C. May 20, 2014) (relevant market was “national assessor and recorder bulk data”); Nielsen Holdings N.V., No. C- 4439, 2014 WL 869523 (F.T.C. Feb. 24, 2014) (national syndicated cross-platform audience measurement services); Dun & Bradstreet Corp., 150 F.T.C. 144, 146 (2010) (relevant market was “kindergarten through twelfth grade educational marketing data”); Reed Elsevier NV, No. C-4257, 2009 WL 1639519, at *2 (F.T.C. June 1, 2009) (relevant market was “electronic public records services for law enforcement customers”); Complaint at 12, FTC v. Hearst Trust, No. 1:01CV00734, 2001 WL 36080059 (D.D.C. Apr. 5, 2001) (relevant market was “integratable drug data files, and/or one or more subsets”); Complaint at 2, Fidelity Nat’l Fin., Inc., No. C-3929 (F.T.C. Feb. 25, 2000) (relevant market was “the provision of title information services”—the provision of selected information contained in a title plant (a collection of records and indices regarding the ownership of and interests in real property), https://www.ftc.gov/sites/default/files/documents/cases/2000/02/fidelitycmp.pdf; Automatic Data Processing, Inc., No. 9282, 1996 WL 768219, at *5 (F.T.C. Nov. 13, 1996) (relevant market of, among others, “salvage yard inventory data for estimates”).
[27] A firm is generally considered to have market power when it can profitably raise price above what would occur in a competitive market by restricting output below competitive levels (i.e. “supracompetitive” prices). See, e.g. Nat’l Collegiate Athletic Ass’n v. Bd. of Regents of the Univ. of Okla., 468 U.S. 85, 109 n.38 (1984) (citing, inter alia, Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 27 & n.46 (1984). A firm is generally considered to have monopoly power when it can control prices or exclude competition. U.S. v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 391 (1956); Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297, 307 (3d Cir. 2007) (citing U.S. v. Grinnell Corp., 384 U.S. 563, 571 (1966)). Issues that can arise in cases featuring technology platforms that may inform an analysis of relevant markets and market power include: (i) whether, and to what extent, firms that offer goods and/or services for sale through brick-and-mortar outlets compete with firms that offer goods and/or services for sale online; (ii) whether a single platform’s marketplace can constitute a relevant antitrust market; and; (iii) whether platform users face unreasonably high costs to switch to a competing platform, effectively locking them in to use the dominant platform.
[28] Harm to a competitor can be associated with harm to competition, particularly in concentrated markets with few rivals. For example, if the conduct at issue effectively raises the rival’s costs, thereby enabling the platform profitably to raise price and/or reduce quality compared to an alternative in which the conduct did not occur, it may be anticompetitive. Harm to rivals also can be the product of competition on the merits. For example, a linear, non-predatory price reduction that has the effect of transferring sales from a rival to the price cutter may harm the rival but is consistent with competition on the merits. Because the antitrust laws protect competition, not competitors (See Brown Shoe Co. v. United States, 370 U.S. 294, 320 (1962)), competition on the merits—even when practiced by a firm with market power—can harm rivals without harming the competitive process.
[29] Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 223 (1993). In Brooke Group, the Supreme Court adopted a framework to evaluate predatory pricing claims to avoid chilling competitive price discounting, and in part out of a concern regarding the administrability of such claims. There, the Court required the plaintiff to show that the allegedly unlawful price was below an appropriate measure of cost and that competitive conditions were such that the discounter could recoup its investment in below-cost pricing.
[30] See Ohio v. Am. Express Co., 138 S. Ct. 2274, 2284 (2018) (identifying approach under Section 1 of the Sherman Act); United States v. Microsoft, 253 F.3d 34, 58-59 (D.C. Cir. 2001) (per curiam) (en banc) (identifying approach under Section 2 of the Sherman Act); United States v. Baker Hughes 902 F.2d 981, 982-83 (D.C. Cir. 1990) (Thomas, J.) (identifying approach under Section 7 of the Clayton Act).
[31] The Department of Justice rejected a competitive effects analysis in its since withdrawn report on Single Firm Conduct. See U.S. Dep’t. of Justice, Competition and Monopoly: Single Firm Conduct Under Section 2 of the Sherman Act (2008) (“Although focusing analysis on the effect of consumer welfare is appropriate, the Department does not believe that using an effects-balancing test as a general standard under Section 2 is likely to maximize consumer welfare. The Department believes that it is better for long-run economic growth and consumer welfare not to incur the costs and errors from attempting to quantify and precisely balance procompetitive and anticompetitive effects as required under this test.” Id. at viii.).
[32] Compare to the alternatives discussed in the DOJ’s Competition and Monopoly Report, including the general standard concepts of no-economic sense, profit sacrifice, equally efficient competitor and disproportionality, as applied to specific categories of potentially exclusionary (or predatory) conduct.
[33] Where the Agencies can identify and administer the analysis of conduct on factors other than price, output and innovation that, based on factual analysis, are important competitive variables, the effect of firm conduct on such factors can be incorporated into the competitive effects analysis. However, the Agencies should proceed cautiously in expanding the relevant factors for competitive effects analysis and should do so on a case-by-case basis.
[34] Caribbean Broadcast Sys. v. Cable & Wireless, 148 F.3d 1080, 1087 (D.C. Cir. 1998) (“‘Anticompetitive conduct’ can come in too many different forms, and is too dependent upon context, for any court or commentator ever to have enumerated all the varieties.”); U.S. v. Microsoft, 253 F.3d 34, 58 (D.C. Cir. 2001) (“[T]he means of illicit exclusion, like the means of legitimate competition, are myriad.”).
[35] See, e.g. Bilal Sayyed, Actual Potential Entrants, Emerging Competitors, and the Merger Guidelines: Examples from FTC Enforcement 1993-2022 (Dec. 2022), available at https://techfreedom.org/wp-content/uploads/2022/12/Actual-Potential-Entrants-Nascent-Competitors-and-the-Merger-Guidelines-Examples-from-FTC-Enforcemen.pdf.
