Crediting Procompetitive Justifications for Digital Platform Defendants: Continued salience of a broad, efficiencies-focused approach

Ashley Ulrich*

Download a PDF of this article here.

Introduction

In 2020, digital platform businesses are everywhere. They help us park our cars, walk our dogs and find us places to stay when we travel. Hidden beneath the slick veneer of these platforms’ user interfaces, however, is a dizzying array of complicated mathematical and economics problems, ranging from auction rules to set prices for advertisements to recommendation engines programmed to make suggestions for new content. Big data, along with powerful algorithms and large user networks, have become critical competitive advantages, and indeed, barriers to entry for smaller startup firms. Along with success, however, comes litigation—digital platform businesses are increasingly being investigated and sued for antitrust offenses in the United States and around the world.¹ In October 2020, the Department of Justice and eleven state Attorneys General filed a civil lawsuit against Google alleging monopolization in the markets for digital search and search advertising, and less than two months later, the Federal Trade Commission and 46 state Attorneys General filed suit against Facebook for monopolization in the personal social networking market.²

The critical question that this Note asks is whether the current assessment of procompetitive justifications within rule of reason analysis is sufficiently broad and flexible to consider, and where appropriate, credit as procompetitive the range of ways that digital platform businesses enable efficiencies in the markets where they compete. As discussed more thoroughly in this Note, digital platform businesses may be responsible for a range of efficiency improvements, such as reducing the time and effort to find desired goods and services, reducing contracting costs, increasing product and services quality, increasing the rate of innovation, and improving allocative efficiency. To the extent that some of these efficiencies are enabled by or require a restraint to competition, it is critical for antitrust analysis to consider this balancing of effects. Otherwise, slight decreases or impediments to competition would make a defendant’s business model susceptible to antitrust challenge even if on net the conduct ultimately benefits competition and consumers.

In the twenty years since United States v. Microsoft, the first major antitrust case involving a digital platform defendant, the question of when to credit a procompetitive justification for any defendant, let alone a defendant operating in a highly technical and rapidly evolving industry, remains murky.³ In Federal Trade Commission v. Indiana Federation of Dentists (hereinafter “Indiana Federation of Dentists”), the Supreme Court suggested looking to see whether a defendant’s conduct demonstrates a benefit to competition, such as through a “creation of efficiencies.” ⁴ Yet this vague efficiencies-focused approach gives little guidance to lower courts regarding what conduct constitutes an efficiency improvement and when it should be credited. This is the approach that the D.C. Circuit followed in Microsoft and is used by a majority of circuit courts.⁵

Despite ambiguity in the Supreme Court’s doctrine, this Note confirms that the existing approach to crediting procompetitive justifications within rule of reason analysis is superior to alternatives suggested in legal scholarship or taken up by a minority of circuit courts because it is sufficiently broad and flexible to consider the range of ways that digital platform businesses enable efficiencies in the markets in which they compete. The existing efficiencies-focused approach permits defendants, including digital platform defendants, to assert a wide variety of justifications for their conduct, which they must then substantiate with record evidence. Further, the existing rule of reason analysis is well designed to reject justifications where a defendant does not carry its burden of proof, a plaintiff shows a less restrictive means to achieve the same benefit, or balancing shows that the anticompetitive harm from a defendant’s conduct outweighs its procompetitive benefit.

Whereas the breadth of what falls within improving competition through “creation of efficiencies”⁶ may feel too indeterminate to practically apply, this Note confirms that several historical and more recent cases demonstrate the contours of which efficiencies are likely to be cognizable. And, this Note confirms that courts have not struggled to apply the broad efficiencies-focused approach—or, more generally, rule of reason—to the conduct of digital platform defendants. Attempts to limit or cabin what can be credited as a procompetitive justification are worse paths forward and run the risk of leaving antitrust hostile to innovative technologies and business methods used by digital platform defendants.

This Note argues, however, that the existing efficiencies-focused approach could be improved if reviewing courts were to provide greater clarity about what criteria they are using when assessing procompetitive justifications. Courts can do so by applying the definition for “procompetitive justification” developed in Microsoft (itself an outgrowth of Indiana Federation of Dentists) as a series of steps for review: (1) assess if a procompetitive justification is cognizable, i.e., if it implicates an efficiency improvement; (2) determine if it is substantiated in fact; and (3) assess whether there is a less restrictive alternative to achieving the same result, should the plaintiff assert that such an alternative exists.⁷

Below, this Note outlines relevant background antitrust principles, including how courts generally apply rule of reason analysis and when they credit procompetitive justifications. This Note then summarizes some distinguishing economic characteristics of platform businesses and the efficiencies that they bring to bear in markets where they compete. Next, this Note reviews whether Microsoft presents a sufficiently clear and flexible standard by which to judge procompetitive justifications asserted by digital platform defendants. Finding that it does not, this Note proceeds to analyze and reject suggestions in legal scholarship that would alter or streamline review of procompetitive justifications for failing to adequately account for the efficiencies enabled by digital platform businesses. Finally, this Note presents a way forward that largely clarifies existing doctrine and demonstrates its continued flexibility and adaptability to assessing digital platform conduct. While the result is a broad, flexible, fact-driven approach to reviewing procompetitive justifications, this approach continues to produce the best and most consistent results.

I. Antitrust Enforcement and Procompetitive Justifications

Below, this Note outlines some background principles of antitrust review. Next, it summarizes when case law suggests that procompetitive justifications should be credited from a review of historical and modern cases.

A. Introduction to Antitrust Review and Principles

1. Overview to Antitrust Review

Antitrust cases are typically litigated under the Sherman Antitrust Act (“Sherman Act”) or the Clayton Antitrust Act (“Clayton Act”). Sherman Act Section 1 is most commonly invoked to prevent two rivals from working together to restrict output and raise prices, ⁸ while Section 2 is invoked to prevent a single firm, if shown to have monopoly power, from maintaining that power through anticompetitive means.⁹ The Clayton Act prevents mergers that are likely to negatively affect competition in a market.¹⁰ Federal agencies, state attorneys general and private plaintiffs may bring suit under the Sherman Act or Clayton Act.¹¹ Because this Note is most interested in how courts have assessed alleged procompetitive justifications for the conduct of digital platform defendants, the Sherman Act, and in particular, jurisprudence addressing Section 2 claims, is most relevant to this Note’s analysis.

The first step of a Sherman Act Section 1 case is distinguishable from a Section 2 case. If a plaintiff brings a case under the Sherman Act’s Section 1, the first step of the analysis requires her to show that there was an agreement between two parties to restrain trade.¹² A mere tacit understanding is likely to fail.¹³ If the plaintiff brings a case under Section 2, the first step of the analysis requires the plaintiff to show that the defendant has monopoly power in a relevant antitrust market.¹⁴ Monopoly power is “the power to control prices or exclude competition.” ¹⁵ The plaintiff may show monopoly power directly, through demonstrating that the defendant has profitably imposed higher prices, ¹⁶ or indirectly, through evaluating the defendant’s market share and barriers to entry in the relevant market in order to proxy its ability to profitably raise prices. ¹⁷ The relevant antitrust market is determined through defining a set of firms that offer products or services “reasonably interchangeable by consumers for the same purposes.”¹⁸ Barriers to entry include fixed costs like equipment, intellectual property protections or established contractual relationships.¹⁹ As a rule of thumb, a plaintiff should be able to show that a defendant has at least sixty percent market share to assert monopoly power. ²⁰

The second step to a Section 1 or Section 2 case is substantially the same: ²¹ a plaintiff must show that the defendant’s restraint is either unreasonable per se or unreasonable under a rule of reason analysis. ²² A restraint is per se unreasonable where it “always or almost always tend[s] to restrict competition and decrease output.”²³ Any agreement between rivals to restrict output or fix prices will fall into this category. ²⁴ Alternatively, a restraint is unreasonable under the rule of reason if the anticompetitive effect of the defendant’s conduct outweighs any credited procompetitive justification. ²⁵ Conduct reviewed under the rule of reason may take any of the following forms: an intrabrand restraint, exclusive dealing arrangements, a refusal to deal, tying arrangements or volume-based discounts.²⁶ Specific to refusal to deal cases, courts have sometimes reviewed anticompetitive conduct under a “sacrifice test,” where the court will evaluate whether the defendant’s conduct makes sense for some reason other than to restrict the opportunities of a rival.²⁷

Rule of reason is a multi-pronged, burden shifting test. At the first step, the plaintiff bears the initial burden to show that the defendant’s conduct has an anticompetitive effect.²⁸ Then, at the second step, the burden shifts to the defendant to show any offsetting procompetitive justification for its conduct.²⁹ If so shown, as a third step, the burden shifts back to the plaintiff.³⁰ The plaintiff gets a final opportunity to show that a less restrictive means to achieve the same benefit to competition was available,³¹ or that the anticompetitive effect of the defendant’s conduct outweighed the procompetitive effect.³² The fact-finder finally weighs these considerations and determines whether the defendant’s conduct is on net anticompetitive.³³

Conduct is determined to have an anticompetitive effect where it reduces social welfare through decreasing output, raising prices, lowering quality or otherwise harming competition; for example, imposing barriers to entry or squeezing out rivals through behavior that does not reflect competition on the merits.³⁴ Conduct that merely imposes “harm to one or more competitors will not suffice” to show an anticompetitive effect.³⁵ Instead, conduct “must harm the competitive process and thereby harm consumers.”³⁶

Just what constitutes a procompetitive justification is poorly defined in case law. The Supreme Court has described a procompetitive justification as “some countervailing procompetitive virtue—such as, for example, the creation of efficiencies in the operation of a market or the provision of goods and services.” ³⁷ What constitutes a cognizable efficiency will be more fully considered throughout this Note.

Courts have also permitted a limited set of non-competition improving justifications to be balanced against the anticompetitive effects of a restraint within rule of reason analysis. In at least refusal to deal cases, the Supreme Court has permitted justifications it described as “valid business reason[s]” for the defendant’s conduct.³⁸ In Microsoft, the D.C. Circuit permitted the defendant’s justifications where they were found to be lawful extensions of copyright protection and technologically required to execute certain software functions.³⁹ Other circuit courts have permitted a limited set of justifications found to further social policy goals. In United States v. Brown University, the Third Circuit permitted the defendant’s justification that its restraint promoted access to higher education for students from lower income households.⁴⁰

2. Goals of Antitrust Enforcement

The goals of antitrust enforcement have been hotly debated since at least the 1960s.⁴¹ Some scholars have argued that antitrust enforcement serves to further economic efficiency and protect consumer interests in low prices and numerous, varied output.⁴² Others have argued that it serves to protect political and social values of dispersed control over limited economic resources. ⁴³ Still others have argued for equal opportunity in the marketplace, especially for smaller producers.⁴⁴ Some courts have articulated that antitrust enforcement serves to protect competition as a standalone virtue. ⁴⁵

Most commonly, courts have construed the goals of antitrust as serving to protect “consumer welfare,” which they have interpreted to mean an interest in low prices and numerous, varied output.⁴⁶ Several scholars have recently argued to expand or change courts’ focus on output and price effects in light of digital platforms’ business models, which they argue pass along low prices to consumers while monopolizing markets and harming smaller producers. ⁴⁷ Other scholars have argued to more broadly consider allocative efficiency or innovation rates as part of antitrust analysis.⁴⁸ Several federal agencies held a series of hearings on the continued efficacy of the “consumer welfare” standard in late 2018. ⁴⁹

B. When to Credit a Procompetitive Justification

1. Treatment of Procompetitive Justifications in Historical Cases

Several noteworthy historical cases have considered when to credit procompetitive justifications within rule of reason analysis. While these cases help to establish some principles and guidelines around when a justification might be credited or should be denied, no clear standard emerges by which to assess asserted justifications in future cases.

The earliest case to take up this question was Board of Trade of City of Chicago v. United States (hereinafter “Chicago Board of Trade”), wherein Justice Louis Brandeis outlined the contours of modern rule of reason analysis.⁵⁰ The defendant, a trade organization that oversaw grain trading in Chicago, had set certain rules for its members that limited after-hours price-making for a limited set of grain exchanges.⁵¹ While acknowledging that the rule was a price restraint, the Supreme Court considered whether it nonetheless “helped to improve market conditions,” finding that it did so through creating a market for “to arrive” grain, attracting a greater number of buyers and sellers to transact, and reducing information asymmetries about the prevailing market price.⁵² The Supreme Court also noted that the restraint helped to remove “risk” from private deal-making, as market participants could be sure that they would be able to find a deal-making partner.⁵³ Further, the restraint helped to improve allocative efficiency, as it ensured that country dealers and farmers—market participants least likely to have access to timely market information—could nonetheless determine the market price for trades.⁵⁴ Finally, the rule contributed to more efficient grain transportation routes. ⁵⁵ In sum, the Court permitted the restraint’s efficiency-improving benefits to be balanced against its anticompetitive harm, and the restraint overall was found not to unreasonably hinder competition.⁵⁶

In the more recent Broadcast Music, Inc. v. CBS, Inc. (hereinafter “BMI”) decision, the Supreme Court permitted defendants, licensing organizations for composers, to issue and set fees for blanket licensing agreements for copyrighted musical compositions.⁵⁷ Since 1914, licensing organizations have served as a clearinghouses for composers—the exclusive owners of the copyright to perform their work in public for profit—in order to negotiate licenses for third parties to perform these works and monitor compliance.⁵⁸ At the court of appeals, the defendant’s blanket licenses were found to be price fixing and illegal per se under the Sherman Act.⁵⁹ The Supreme Court agreed that the blank licenses involved price fixing “in the literal sense,” but the Court noted Congress’s intent for copyright owners to be able to vindicate their rights and the practical impossibility of forming contracts and monitoring compliance without blanket licenses.⁶⁰ The Court also described how the blanket licenses were valuable for reducing transaction costs in contract formation and monitoring, contributing to lower costs for buyers and sellers of composition performance rights.⁶¹ The Supreme Court determined that rule of reason analysis should apply in order to permit a reviewing court to consider these offsetting efficiencies from reduced transaction costs as procompetitive justifications for the blanket licenses, and the Court reversed and remanded the case to proceed below.⁶²

In National Society of Professional Engineers v. United States (hereinafter “Professional Engineers”), the Supreme Court affirmed the district court’s determination that a professional engineering society’s code of ethics prohibiting its members from bidding competitively for projects violated antitrust law even where the restraint was adopted to reduce the risk of low-quality engineering work from excessive price competition.⁶³ Justice John Paul Stevens wrote that the defendant’s conduct, a refusal to compete on price, had the same effect as price-fixing in terms of impeding “the ordinary give and take of the market place [sic]” to set prices and further deprived consumers of the ability to compare prices.⁶⁴ Because “[t]he Sherman Act reflects a legislative judgment that ultimately competition will produce not only lower prices, but also better goods and services,” the asserted justification of a quality improvement from a direct price restraint could not be credited.⁶⁵

2. Treatment of Procompetitive Justifications in Recent Cases

In the over 100 years since Chicago Board of Trade and 40 years since BMI, antitrust doctrine has not yet coalesced around a clear standard for when to credit a procompetitive justification within rule of reason analysis. In National Collegiate Athletic Association v. Board of Regents of University of Oklahoma (hereinafter “NCAA”), the Supreme Court suggested crediting conduct that “increase[s] output and reduce[s] . . . price,” “mak[es] possible a new product,” “widen[s] consumer choice,” or “enhance[s] public interest [in the relevant product].”⁶⁶ The decision’s reasoning, however, mostly focused on whether application of the per se rule was appropriate for the case, not when to credit a procompetitive justification.⁶⁷ The Supreme Court referenced these efficiencies sporadically throughout the opinion—it did not refer to them as a defined list by which to judge future asserted justifications.⁶⁸ In Indiana Federation of Dentists, the Supreme Court used a broad, efficiencies-focused approach to assess whether to credit a defendant’s asserted procompetitive justification, citing BMI, Chicago Board of Trade, NCAA, and Professional Engineers in doing so.⁶⁹ Yet, lower courts lack clear guidelines for which efficiencies to credit and when. In Indiana Federation of Dentists, the Supreme Court provided no further guidance than that lower courts should look for “some countervailing procompetitive virtue—such as, for example, the creation of efficiencies in the operation of a market or the provision of goods and services.”⁷⁰ The Supreme Court did not specifically reference the justifications discussed two years earlier in NCAA, nor did it state whether and which of these justifications might be credited within the efficiencies-focused approach it set out. Presumably, most of the categories discussed in NCAA would improve market competition through creation of efficiencies and be permitted, as they refer to conduct with the net effect to reduce price, increase output, or improve product quality and variety.⁷¹ And, the Supreme Court, by citing to BMI, Chicago Board of Trade, NCAA, and Professional Engineers, appeared to intend to encapsulate and carry forward the reasoning in those cases in the approach that it set out, suggesting the continued salience of those earlier cases for asserting procompetitive justifications.

As far as other categories of justifications that might be credited, in California Dental Association v. FTC, the Supreme Court suggested openness to crediting conduct shown to remove false or misleading claims from the market, although it did not do so to decide the case.⁷² In the vertical restraint context, the Supreme Court permits conduct that stimulates interbrand competition among manufacturers selling different brands of the same product by reducing competition among retailers selling the same brand.⁷³ The Supreme Court has suggested openness to crediting conduct to stimulate interbrand competition in monopolization cases as well.⁷⁴

The majority of circuit courts follow the broad, efficiencies-focused approach outlined in Indiana Federation of Dentists. The D.C. Circuit appeared to follow this approach in Microsoft, where it defined “procompetitive justification” as “a form of competition on the merits because it involves, for example, greater efficiency or enhanced consumer appeal,” although it did not specifically cite to Indiana Federation of Dentists in doing so.⁷⁵ The Second, Third, Seventh, and Ninth Circuits have followed this efficiencies-focused approach as well, citing to the definition for “procompetitive justification” outlined in Microsoft.⁷⁶ Yet, some circuit courts have also credited conduct with no direct effect on competition, such as conduct designed to execute a technical command, protect an intellectual property right, or increase diversity in higher education.⁷⁷ A minority of circuit courts have experimented with a “categorical” approach that only permits justifications that fall into predetermined categories. The Eleventh Circuit will credit a procompetitive justification if the defendant shows that its conduct has the effect to “reduce cost, increase output or improve product quality, service, or innovation.”⁷⁸ The Tenth Circuit will allow a justification if the defendant can point to “increasing output, creating operating efficiencies, making a new product available, enhancing product or service quality, and widening consumer choice.”⁷⁹

II. Economic Characteristics of Digital Platform Businesses

The following section summarizes some useful definitions and economic principles applicable to platform businesses. The section then describes market efficiencies that platform businesses often enable in the markets where they compete.

A. Platform Businesses

A platform brings together at least one group of users and facilitates the provision of products or services.⁸⁰ A platform might help an organization develop and innovate its internal business processes, or may be an external services offering for customers.⁸¹ Platforms allow users to innovate and interact more easily, or in ways practically not possible outside the platform, with the potential for increased utility and value creation.⁸² In a given market, participants may use a single or multiple platforms to purchase goods and services, the latter referred to as “multihoming.”⁸³ Whether users will seek out multiple platforms is a function of switching costs—where switching costs are high, users are likely to use one platform.⁸⁴

A digital platform is a technology-enabled platform that is built using software and may be offered over the internet.⁸⁵ Early digital platform providers made software and services for the personal computer, such as Microsoft’s Windows operating system and Netscape’s internet browser.⁸⁶ More recent digital platform providers make software and services for smartphones, such as Uber’s ridesharing application (“app”) and Airbnb’s home rental app.⁸⁷

Platforms are described as “two-sided” where changes in pricing for one group of users on the platform affects the participation rate of another group of users.⁸⁸ Two-sided platforms present unique pricing challenges, as the platform operator must choose not just the right price level but also the right price structure to maximize platform participation and returns.⁸⁹ Examples of platforms that connect two groups of users include credit card providers (merchants / cardholders) and newspapers (advertisers / readers).⁹⁰ Examples of digital platforms that connect two groups of users include computer operating systems (app developers / users) and ridesharing applications (drivers / users).⁹¹ Digital platform businesses are an increasingly popular way to transact goods and services.⁹²

While there are several ways to classify platforms, one helpful distinction is between innovation and transaction platforms. An innovation platform enables a range of products and services to be built on top of it through providing tools to third-party developers like application programming interfaces (“APIs”) and software development kits (“SDKs”).⁹³ A transaction platform helps match two groups of users together to enable a transaction, such as Airbnb matching renters and travelers or Uber matching drivers and riders.⁹⁴ Some economists have theorized a third group of platforms, audience makers, which aggregate large groups of users together in order to sell advertisements to them.⁹⁵

B. Economic Characteristics Specific to Platform Businesses

Platform businesses are distinguishable from other businesses due to the significance of network effects, increasing returns to scale, and upstart and ongoing pricing challenges. These effects persist whether a platform is digital or not but may be magnified on digital platforms.⁹⁶

Network effects refers to a positive feedback loop whereby having more users on a platform increases the value of participating in that platform.⁹⁷ With more users on a platform, there are more participants to interact or transact with, and the value that a user will assign to transacting through the platform grows.⁹⁸ Network effects may be direct or indirect. Network effects are direct where an increase in the number of participants from a single group of users increases the value of participating in the platform for all users.⁹⁹ A classic example of direct network effects is the landline telephone: as the number of households with connected telephones increases, the product is more valuable to all users, as they can all communicate with more friends and family members.¹⁰⁰ Network effects are indirect where the increase in the number of participants in one group of users on the platform increases the value to participating on the platform for some second group of users.¹⁰¹ A classic example of indirect network effects is a local farmer’s market: having more farmers take part in a local farmer’s market makes attending the market more valuable for shoppers.¹⁰² The same effect is true in the corollary: having more shoppers take part in a local farmer’s market makes attending the market more valuable for farmers.¹⁰³ Indirect network effects are common in digital platform businesses that bring together two groups of users.¹⁰⁴

Network effects are important for several reasons, including causing platform businesses to experience increasing returns to scale.¹⁰⁵ Whereas traditional businesses experience diminishing demand for their product at higher levels of output, the opposite is true for platform businesses.¹⁰⁶ With network effects, the value to transacting through the platform grows as the number of users on the platform increases, thereby driving up user demand to participate in the platform and willingness to pay for access.¹⁰⁷ Thus, as platforms increase in size and users increasingly demand to participate, platforms can charge a higher price for platform access.¹⁰⁸ Because network effects make platforms more valuable as they scale, platform providers are incentivized to try to cultivate network effects through growing user participation with strategic decisions about pricing, design, and marketing.¹⁰⁹

Network effects also contribute to two pricing challenges that prevail in two-sided platform markets: getting users to join the platform in the first place and maintaining an optimal price level and structure to maximize ongoing platform use. First, before users find value to transacting through the platform, there must already be existing users on the platform. An e-commerce buyer, for example, will only use a platform if there are already participating sellers, and a seller will only offer merchandise if there are participating buyers.¹¹⁰ To solve this conundrum—what is referred to as the “chicken-and-egg” problem—a platform provider typically has to induce at least one set of users to join the platform through offering a subsidy. ¹¹¹ Failure to adequately attract both sets of users at this initial stage will doom a platform.¹¹² Second, in the ongoing operation of a two-sided platform, a platform owner must consider not just the appropriate price level for the service overall, but also how to distribute that price across two groups of platform users.¹¹³ Often, two groups of platform users will have differentiated demand for platform services.¹¹⁴ This difference will determine which group of users will pay less or be subsidized during ongoing operation of the platform, whereas the other group will pay more or even bear the whole cost of receiving platform services.¹¹⁵ Overcharging one group of platform users comes at a significant cost—potentially causing the entire platform to unravel.¹¹⁶

C. Market Efficiencies Specific to Platform Businesses

Platforms improve market efficiencies where they bring together two groups of users more efficiently than if the two groups transacted directly.¹¹⁷ Platforms do so in several discrete ways, including reduced time and effort to find desired goods and services, reduced contracting costs, increased product and services quality, increased rate of innovation, and improved allocative efficiency. While these efficiencies are not necessarily specific to digital platforms, digital platforms’ ability to aggregate and analyze large data sets, provide recommendations using machine learning, and dynamically push out updates to users in real-time helps to capitalize on these efficiencies.

A topic not adequately addressed in the present Note is when platform efficiencies should be characterized as not merely improving market competition but also furthering consumer welfare. The former, for example, might be distinguishable from the latter where improvements to market competition merely increase upstream producer surplus without passing along benefits to consumers in terms of greater output and variety or lower prices.¹¹⁸ The former might also be distinguishable where only a discrete group of consumers is made better off, not consumers generally.¹¹⁹ Yet, many scholars would view an improvement to market competition as necessarily improving consumer welfare without these further showings—so perhaps this is merely distinction without difference.¹²⁰ For most purposes of this Note, “efficiencies” merely refers to improvements to market competition, not necessarily specific benefits flowing through to consumers. Where applicable, this Note considers what effects an efficiency might have on “consumer welfare” improvements or changes to “consumer surplus” through postulating what changes to product and services output, variety, and pricing are likely to result from certain platform efficiencies.

While not the focus of the present Note, many scholars have devoted considerable effort to cataloguing digital platform conduct likely to result in anticompetitive harm.¹²¹ Some of the most invidious conduct includes: creating closed standards and walled gardens;¹²² limiting data portability;¹²³ erecting barriers to consumer switching;¹²⁴ imposing algorithmic price discrimination;¹²⁵ and acquiring nascent rivals.¹²⁶ Some of the previously discussed economic characteristics of platform businesses may also have the effect of limiting expansion by rivals or entry by new competitors.¹²⁷ Network effects and increasing returns to scale, for example, may make a smaller firm’s offering inherently disadvantaged against a larger rival.¹²⁸ A larger rival’s strategy to maintain a closed ecosystem of services and prevent data portability may also ingrain existing leadership where network effects are strong.¹²⁹ As such, while there are many market efficiencies that platforms enable, a fact-based, holistic consideration of the net effect of platform activity (as rule of reason generally endeavors to undertake) is especially important.

1. Reduced Search Costs

Platforms, and especially transaction platforms, improve market efficiencies through reducing search costs to find a viable deal-making partner and improving the quality of matches between buyers and sellers, among other groups of users. In brick and mortar stores, search costs within a single location are limited, as a store’s physical size and layout limit the number of products that can be displayed.¹³⁰ However, comparison shopping across store locations and vendors can be challenging, requiring research of stores that sell a certain product or service and trips to several locations to assess that item.¹³¹ Without a platform intermediary, search costs can result in retailer markups for nearly identical goods, resulting in higher prices for consumers without any benefit to product quality.¹³²

Digital platforms like Amazon and eBay aid shoppers with tools like search functions and filters, which allow consumers to view a broad range of products,¹³³ and to quickly narrow their results by product attributes like price.¹³⁴ In doing so, platforms may reveal differing prices for comparable goods, enabling consumers to benefit from lower prices.¹³⁵ Platforms also offer algorithmic matching capabilities that help to promote search results that match a user’s preferences.¹³⁶ Further, network effects encourage the development of thicker markets, that is, the accumulation of large groups of users on both sides of a platform, which may help users discover products that they are interested in.¹³⁷ In sum, platforms can reduce a consumer’s search time and improve the quality of their results, as well as pass along lower prices.¹³⁸ To the extent that reducing search costs passes on lower prices to consumers and improving matches passes on better quality products and services, this market efficiency improvement also increases consumer surplus.

2. Reduced Contract Formation Costs

Digital platforms improve market efficiencies through reducing the costs of forming a contract.¹³⁹ Whereas Ronald Coase famously theorized that bargaining parties will be able to achieve an economically efficient outcome regardless of the original endowment of property rights between them,¹⁴⁰ in practice, transaction costs and information asymmetries hold up mutually beneficial deals.¹⁴¹ Platforms may reduce contract formation costs in several ways. First, platforms may lower contract formation costs through standardizing deal terms.¹⁴² A platform provider might require that all participants use certain default contracts or may set certain rules limiting negotiation, such as requiring price caps and imposing rules against surcharges.¹⁴³ Second, platform providers may take on tasks that reduce the risk of forming a deal on the platform. A platform provider may take on contract monitoring and oversight costs, or it may require that all users provide identity verification and only transact through the platform’s payment processing tools.¹⁴⁴ Third, platforms may guarantee deals by standing ready to accept returns or provide refunds.¹⁴⁵ On net, these activities may make contract formation easier and less risky than transacting outside the platform, allowing a greater number of deals to be realized. ¹⁴⁶ To the extent that these market efficiencies are passed on to consumers in the form of lower prices and greater output, they also improve consumer surplus.

Some platforms also improve market efficiencies through facilitating auctions, which help buyers and sellers to arrive at a market clearing price in real-time. ¹⁴⁷ While auctions have existed for centuries,¹⁴⁸ digital platform auctions are distinguishable because of the amount of data that they can aggregate about potential trading partners and the speed with which they can process matches.¹⁴⁹ Digital platforms can also facilitate a series of complex auction rules and conditional bidding schemes.¹⁵⁰ Further, digital platforms can integrate third party data about inventory characteristics to improve buyer knowledge of inventory quality.¹⁵¹ On net, auctions facilitated by digital platforms help buyers and sellers to quickly resolve ambiguities about inventory quality and pricing as well as execute a deal, thereby reducing transaction costs.¹⁵² To the extent that these market efficiencies are passed on to consumers in the form of lower prices, greater output and high quality goods, they also improve consumer surplus.

3. Self-Regulation and Improved Information Sharing

Platforms improve market efficiencies through acting as regulators for platform transactions. As discussed above, standard setting can reduce contract formation costs.¹⁵³ Standard setting on transaction and innovation platforms can further improve the quality of products and services available through the platform. Many platforms self-regulate through setting rules for who may join the platform and what types of conduct can take place on the platform.¹⁵⁴ These rules prevent fraud, distribution of malicious software, sale of counterfeit goods and other undesirable activities.¹⁵⁵ Many platforms also provide information back to users about other platform participants in the form of reviews or quality rankings.¹⁵⁶ This information sharing helps to protect users from less trustworthy deal-making partners and incentivizes good behavior on the platform.¹⁵⁷ Ultimately, users may be flagged and removed from the platform for conduct that violates platform rules.¹⁵⁸ On net, market efficiencies are improved because users can spend less time evaluating deal-making partners, more transactions can take place, and the transactions that take place will be for higher quality goods and services.¹⁵⁹ To the extent that users value higher quality products and services, albeit at somewhat higher prices, this efficiency improvement improves consumer welfare.

4. Increased Rate of Innovation

Innovation platforms improve market efficiencies by making it easier for third-party developers to reach the market with a product offering.¹⁶⁰ They do so by reducing the costs to create a viable product and enabling these products to launch at scale.¹⁶¹ Innovation platforms provide the ecosystem, infrastructure, or set of standards that enable third-party developers to interact with users in a common system.¹⁶² Microsoft’s Windows operating system, for example, allowed Netscape to offer its browser product to users without also having to create an operating system.¹⁶³ Innovation platforms may also provide tools to third-party developers that enable innovation, like software developer kits (“SDKs”).¹⁶⁴ SDKs reduce the costs for third-party innovation through providing a start-up bundle of code to build upon.¹⁶⁵ Innovation platforms may also provide open interface APIs, thereby facilitating compatibility between third-party applications and the platform.¹⁶⁶ Developers no longer have to write software to enable platform compatibility and can instead focus on developing their unique application offerings.¹⁶⁷ This compatibility with the platform also reduces costs that the developer would have had to absorb to market and distribute its application to consumers.¹⁶⁸ On net, innovation platforms improve market efficiencies by creating an ecosystem that brings together developers and users while reducing the cost for developers to create, market and distribute their applications. Because consumers benefit from a greater range of compatible product and service offerings, this market efficiency also improves consumer welfare.

5. Improved Allocative Efficiency

Platforms improve market efficiencies by dynamically adjusting the supply of platform services and pricing during periods of high demand. Platform providers create rules for the provision of platform goods or services that may increase availability in the face of increased consumer demand.¹⁶⁹ Ridesharing applications like Uber group together riders in shared cars when there is sufficient local demand, improving the efficiency of that driver’s service offering.¹⁷⁰ Platform providers may also create rules that dynamically adjust pricing with demand.¹⁷¹ Ridesharing applications impose “surge” pricing during periods of heavy demand, which likely reflects the real-time marginal cost of receiving a desired service when so many other users are simultaneously demanding the same service.¹⁷² A platform policy to dynamically update prices with demand may also improve allocative efficiency in future interactions by encouraging more price sensitive users to seek out services during periods of lower demand.¹⁷³ Thus, dynamic adjustment helps consumers to receive goods and services that they desire, but they may have to pay a higher price. Where this market efficiency improvement results in a net increase in the supply of a good or service, it represents a benefit to consumer surplus. To the extent consumers face price discrimination, it is not clear whether consumer welfare is improved.

III. Crediting Procompetitive Justifications for Platform Business Defendants

Below, this Note reviews the standard for crediting procompetitive justifications applied in the Microsoft case to determine if the decision provides a clear and consistent standard that may be used in other cases to assess the conduct of digital platform defendants. Finding it does not quite meet the mark, this Note assesses whether other approaches advanced in legal scholarship and some circuit courts provide a better way forward. Finally, this Note argues that the existing rule of reason analysis and the standard to credit a procompetitive justification where conduct is shown to improve market efficiencies remains the best approach. This Note substantiates this view through review of several recent cases addressing the conduct of platform business defendants where courts were found to appropriately apply rule of reason analysis and the broad efficiencies-focused standard with good results.

A. The Microsoft Case

The Microsoft decision is notable as the first court of appeals decision to apply rule of reason analysis to the conduct of a digital platform business exhibiting network effects.¹⁷⁴ As the D.C. Circuit described, before its decision, scholars were split as to whether platform businesses should be treated more harshly under antitrust review on a theory that network effects posed a barrier to entry for rivals or more leniently on a theory that incumbent companies operating in dynamic, high-technology markets were likely to see their market share eclipsed with technological change.¹⁷⁵ Most relevant to this Note, the D.C. Circuit’s decision expended significant time and effort to review several procompetitive justifications Microsoft asserted for its conduct. In its analysis, the D.C. Circuit provides a helpful working definition of “procompetitive justification” that has been adopted in several other circuit courts.¹⁷⁶ However, the Microsoft decision does not ultimately provide a very instructive standard by which to judge asserted procompetitive justifications in other cases.

In Microsoft, the United States Department of Justice and 20 states sued Microsoft alleging monopolization of the operating system market and attempted monopolization of the browser market.¹⁷⁷ In a significant victory for the plaintiffs, the district court found liability on the monopolization and attempted monopolization claims and granted a structural remedy to separate major lines of Microsoft’s business.¹⁷⁸ The D.C. Circuit largely upheld the district court’s finding of monopolization in the operating system market.¹⁷⁹ However, it reversed the district court’s finding of attempted monopolization, remanded for further review of plaintiff’s tying claims under the rule of reason, and remanded for further fact finding related to the remedy.¹⁸⁰ No structural separation ever took place, but Microsoft agreed to conduct remedies in a subsequent settlement with the Department of Justice and several states.¹⁸¹

Plaintiffs’ theory of the case in Microsoft was that the defendant undertook conduct to prevent the rise of Netscape’s Navigator, an internet browser, and Sun Microsystem’s Java, a middleware software product, in order to protect its market share for Windows, a personal computer operating system.¹⁸² Specifically, Plaintiffs argued that Microsoft was concerned that the rise of these products in adjacent markets would encourage the development of cross-operating system compatible applications, thereby reducing the value to operating systems like Windows.¹⁸³ As opposed to creating innovative updates and proprietary applications that might induce consumers to buy its operating system, Plaintiffs alleged that Windows endeavored to get original equipment manufacturers (“OEMs”) like Intel to refuse to provide Navigator pre-installed on users’ machines, the most popular sales channel for distributing browsers.¹⁸⁴ Microsoft also formed contracts with internet access providers to refuse to distribute or limit their distribution of Navigator.¹⁸⁵ To prevent the dissemination of cross-platform Java to application developers, Microsoft represented to developers that they could write applications using Microsoft’s custom version of Java that would be cross-platform compatible.¹⁸⁶ However, Microsoft only ever designed its Java version to be compatible with Windows.¹⁸⁷

Reviewing the district court’s finding that Microsoft had monopolized the operating system market, the D.C. Circuit affirmed at the first step of rule of reason analysis that Microsoft had market power in the operating system market.¹⁸⁸ At step two, the D.C. Circuit reviewed six categories of Microsoft conduct alleged to be anticompetitive and largely affirmed the lower court’s findings of anticompetitive effect.¹⁸⁹ At steps three and four, the D.C. Circuit reviewed Microsoft’s procompetitive justifications for its conduct as well as any proposed less restrictive alternatives or balancing from Plaintiffs in light of these claimed justifications.¹⁹⁰ On net, the D.C. Circuit largely upheld the findings of the lower court that Microsoft’s conduct was anticompetitive.¹⁹¹

Microsoft’s alleged anticompetitive conduct included: (1) contract provisions in licensing agreements with OEMs that prevented OEMs from removing any desktop icons and folders, altering the startup boot sequence, or otherwise altering the Windows desktop appearance;¹⁹² (2) technologically binding Internet Explorer to Windows by commingling code, making Internet Explorer software irremovable in the “Add/Remove Programs” utility, and overriding a user’s choice of a non-Internet Explorer browser;¹⁹³ (3) licensing Internet Explorer to internet access providers (“IAPs”) for free on the condition that IAPs commit to promoting and distributing Internet Explorer as their exclusive, compatible browser;¹⁹⁴ (4) licensing Internet Explorer to internet software vendors (“ISVs”) for preferred support, early integration with new Windows versions and other technical information in exchange for ISVs making Internet Explorer their default browsing software;¹⁹⁵ (5) promoting a Java middleware version that would be compatible with Windows and third-party operating systems and inducing developers to use this middleware in designing applications while never adding cross-platform compatibility;¹⁹⁶ and (6) other course of conduct claims.¹⁹⁷ Microsoft’s procompetitive justifications for its conduct were, respectively, that: (1) it was merely “exercising its rights as the holder of valid copyright[s]”;¹⁹⁸ (2) greater technical integration “is highly efficient and provides substantial benefits to customers and developers”;¹⁹⁹ (3) it was merely attempting “to keep developers focused upon its APIs”;²⁰⁰ and (4) licensing and distribution agreements were “part of a multifaceted set of agreements” between parties.²⁰¹

The D.C. Circuit’s treatment of Microsoft’s asserted procompetitive justifications is notable for several reasons. First, the D.C. Circuit helpfully defined a “procompetitive justification” as “a nonpretextual claim that [the defendant’s] conduct is indeed a form of competition on the merits because it involves, for example, greater efficiency or enhanced consumer appeal.”²⁰² This definition comports with the Supreme Court’s reasoning in Indiana Federation of Dentists, which advised crediting a justification where it led to “creation of efficiencies in the operation of a market or the provision of goods and services.”²⁰³ It also aligns with BMI, which suggested crediting a justification if its effect was to “increase economic efficiency and render markets more, rather than less, competitive.”²⁰⁴ Yet, this definition does not quite amount to an applicable standard, as it provides no guidance as to which efficiencies to credit and when.

The D.C. Circuit’s analysis of Microsoft’s asserted justifications provides some more guidance on how to apply this definition to the facts of a given case. Reviewing Microsoft’s justifications, the D.C. Circuit seemed to be looking for two criteria: whether Microsoft had asserted a cognizable justification—i.e., a justification that demonstrated some improvement in market efficiency or consumer appeal, and whether Microsoft had substantiated this justification with record evidence. In application of the first criteria, the D.C. Circuit swiftly rebutted Microsoft’s asserted justifications of attempting “to keep developers focused upon its APIs” and forming licensing and distribution agreements that were merely “part of a multifaceted set of agreements” between parties because neither claim demonstrated an improvement to competition like advancing an efficiency. The D.C. Circuit noted that the first claim was “not an unlawful end, but neither is it a procompetitive justification for the specific means in question here”²⁰⁵ and the second claim was wholly “irrelevant” to the inquiry.²⁰⁶ Yet, the D.C. Circuit provided no affirmative statement regarding what conduct might be credited as efficiency enhancing, just that these two claims clearly missed the mark. In application of the second criteria, the D.C. Circuit rejected Microsoft’s claim that greater technical integration “is highly efficient and provides substantial benefits to customers and developers” because it neither “specifie[d] nor substantiate[d]” these claims with record evidence.²⁰⁷ Again, the D.C. Circuit did not elaborate on what types of evidence might be necessary or helpful in making out a claim. Finally, the D.C. Circuit permitted some limited justifications where Microsoft had narrowly asserted that its conduct was a valid exercise of its intellectual property rights,²⁰⁸ and where integration of Internet Explorer and Windows code was shown to be functionally necessary.²⁰⁹

In sum, it is possible for courts deciding subsequent antitrust cases to try to apply Microsoft’s definition of procompetitive justifications as a standard by which to assess asserted justifications. A court could do so by looking for two criteria: (1) whether the defendant had asserted a cognizable justification—i.e., a justification that demonstrated some improvement in market efficiency or consumer appeal; and (2) whether the defendant had substantiated this justification with record evidence. A plaintiff would then be permitted to argue that a less restrictive means to achieve the same benefit was available and that on net, the defendant’s conduct resulted in a greater anticompetitive effect than procompetitive benefit. More helpful to courts reviewing future cases, however, would have been a clearer statement from the D.C. Circuit that it meant for its definition to be applied in this way. It would also have been helpful for the D.C. Circuit to specifically list and demonstrate use of the criteria that seemed to be animating its reasoning. Further, subsequent cases would have been improved with more guidance regarding which categories of efficiencies might be cognizable and what types of evidence might be helpful to substantiate a claim. Finally, future cases would have benefited from greater consideration of less restrictive alternatives in the Microsoft decision. However, because the Plaintiff did not offer any such rebuttals to the Microsoft claims found to be cognizable, there was nothing for the appellate court to review.²¹⁰

Ultimately, the reasoning of Microsoft, while helpful for setting out a definition for “procompetitive justification,” fails to offer a very instructive framework or standard to use when determining whether to credit asserted procompetitive justifications arising out of the operation of digital platform businesses.

B. Rejecting Proposals for Reforming Review of Procompetitive Justifications Advanced in Legal Scholarship and Minority Circuit Courts

Several scholars have noted that when to credit a procompetitive justification within rule of reason remains poorly defined in case law.²¹¹ Whereas these scholars agree that a more uniform approach could improve the predictability of antitrust outcomes and reduce error risks, they diverge on what approach to adopt.²¹² Three proposed approaches for when to credit a procompetitive justification include: (1) where the conduct solves a market failure; (2) where the conduct improves the competitive process; or (3) where the conduct falls within some predetermined set of categories for permitted justifications.²¹³ None of these approaches, however, adequately considers the broad range of ways that digital platform defendants enable efficiencies in the markets where they compete. Thus, these approaches do not offer an improvement over the existing efficiencies-focused approach used in most circuit courts.

1. Market Failure Approach

One approach that antitrust scholars like John M. Newman have proposed is to credit a defendant’s justification where it alleviates a market failure, i.e., improves some previously inefficient allocation of market resources.²¹⁴ A market failure might be the result of hold-ups to deal-making like high transaction costs, coordination challenges and information asymmetries.²¹⁵ It may also be the result of costs being borne by only a few market participants, as with free-riding.²¹⁶

Newman’s market failure approach aligns best with Supreme Court decisions that have focused their analysis of procompetitive justifications on economic effects. Justice Brandeis’s reasoning in Chicago Board of Trade, which permitted balancing a restraint’s anticompetitive effect against evidence showing that the same restraint “helped to improve market conditions” through increasing the number of trading partners, reducing transaction costs, and lessening risk from failed transactions, could be considered as aligned with this approach.²¹⁷ In the vertical restraint context, the Supreme Court credits conduct designed to limit free-riding among retailers, a market failure, as procompetitive.²¹⁸ More recently, in Ohio v. American Express, the Supreme Court suggested openness to crediting justifications shown to reduce negative externalities like free-riding outside the vertical restraint context.²¹⁹ No Supreme Court decision has ever required a defendant, however, to go further than showing that an alleged procompetitive justification improved market conditions and show that it actually alleviated a market failure.

Newman’s approach is deficient for several reasons and would not work well to assess the conduct of digital platform defendants. First, Newman’s definition of “market failure”— “the relevant market produces outcomes that are less efficient than they might be”—is broad and indeterminate, making application of the approach difficult in practice.²²⁰ The definition seems to encompass any market that falls short of perfect competition. In doing so, the approach provides no benefit or guide to practitioners regarding when a procompetitive justification should be credited. Second, Newman’s approach does not fit well to a circumstance where a platform owner develops an entirely new way to match buyers and sellers or offer a product or service. Platform providers like Uber and Airbnb, which conceptualized novel ways to deliver traditional services like taxi and limousine rides and short-term apartment rentals, would be barred from seeking procompetitive justifications for restraints that flow out from making these platforms operational, like setting rules limiting direct negotiations between users on the platform over ride price (Uber) or limiting direct communication for booking and billing (Airbnb).²²¹ Failure to credit justifications from creating a new product or significantly improving how an existing service is offered would seem in derogation of the Supreme Court’s holding in BMI.²²² Further, Newman’s approach might not credit efficiencies from defendant conduct that merely improves competitive conditions without demonstrating a market failure, such as platform conduct to reduce search and contract formation costs or to set rules for online auctions.²²³ This is despite similar cost reduction and regulatory-type behavior being credited for improving efficiencies in Chicago Board of Trade.²²⁴ Newman’s approach also does not appear to capture platform efficiencies from increased innovation and improved allocative efficiency, which platform businesses help to promote.²²⁵ This position stands in opposition to that of several circuit courts, which specifically require recognition of innovation improvements as a cognizable efficiency,²²⁶ or others that have suggested openness to recognizing innovation improvements.²²⁷ In sum, in attempting to streamline courts’ approach to assessing procompetitive justifications, Newman removes too many categories of efficiencies that should be credited, especially for digital platform defendants.

2. Competitive Process Approach

A second approach that some antitrust scholars including Gregory J. Werden have argued for is a “competitive process” approach, whereby procompetitive justifications would be assessed holistically in conjunction with anticompetitive effects to determine whether conduct on net helps or harms competition.²²⁸ Werden cites to the decision of Professional Engineers to support his view, where Justice Stevens wrote about rule of reason analysis: “[T]he purpose of the analysis is to form a judgment about the competitive significance of the restraint.”²²⁹

Some Supreme Court cases have suggested that procompetitive justifications should be analyzed broadly in terms of their effects to competition. In Chicago Board of Trade, Justice Brandeis wrote that “[t]he true test of legality is whether the restraint imposed is such as merely regulates and perhaps thereby promotes competition or whether it is such as may suppress or even destroy competition.”²³⁰ In NCAA, Justice Stevens wrote that the “ultimate focus of the inquiry,” referring to rule of reason analysis, “is whether or not the challenged restraints enhance competition.”²³¹ Further, the Court wrote that “the criterion to be used in judging the validity of a restraint on trade is its impact on competition.”²³²

However, the competitive process approach is deficient in several ways and would not work well to assess the conduct of digital platform defendants. First, the competitive process approach does not clearly delineate what conduct tends to help as opposed to harm competition. It speaks generally about harms to competition and the “competitive process,” but does not clearly delineate what conduct would be permissible.²³³ Were this approach adopted, digital platform defendants would have no greater clarity about whether their conduct conforms with the law and instead may be even more befuddled about what is required of them. Second, it is not obvious that the competitive process approach would permit platform owners that develop an entirely new way to match buyers and sellers or offer a product or service, as in the examples of Uber and Airbnb above, to claim a precompetitive justification.²³⁴ Scholars supporting this approach have not specified whether the improvement to competition must take place in an existing antitrust market or if creating a new market altogether, as in markets for app-based delivery of ride-sharing services or short-term rentals, could be credited. Yet, the Supreme Court’s holding in BMI strongly suggests that offering a new product or service to market or significantly reducing transaction costs should be credited as a procompetitive justification.²³⁵ Relatedly, it is not clear whether a single firm’s improvement of product and services delivery on its own platform through regulatory behavior could be credited if there is no showing that competition improved in the market more broadly.²³⁶ Third, it is not clear how this approach would view platform efficiencies like improving innovation and resource allocation.²³⁷ On one hand, these efficiencies reduce the costs of market entry and ensure that resources flow to their highest valued use. And yet, because these efficiencies pose a more indirect improvement to competition than effects to prices and output, it is not clear that the competitive process approach would permit crediting these efficiencies. Finally, the approach muddles rule of reason’s multi-pronged, burden-shifting framework. Instead of first asking whether a plaintiff met the requisite showing of anticompetitive harm and then determining if a defendant can show an offsetting procompetitive justification, the approach considers all the evidence together. In addition to being contrary to accepted antitrust practice, the approach would obscure whether a case is more properly resolved under a per se analysis, forestalling consideration of procompetitive justifications altogether.

3. Categorical Approach

Some scholars have argued for a categorical approach, where justifications would only be credited if they fall into certain predetermined categories.²³⁸ The Tenth and Eleventh Circuits have adopted this approach, permitting justifications that have led to “increasing output, creating operating efficiencies, making a new product available, enhancing product or service quality, and widening consumer choice”²³⁹ or that “reduce cost, increase output or improve product quality, service, or innovation,”²⁴⁰ respectively.

The categorical approach has several positive attributes. Practically, it is the easiest approach to apply and might help reduce the time and expense to litigate antitrust cases under rule of reason, which courts have noted to be considerable.²⁴¹ Further, the included categories are fairly broad, spanning economic effects like reduced cost and increased output as well as more abstract considerations like improvements to innovation and widening consumer choice.²⁴² The Supreme Court has not yet ruled on whether a categorical approach to assessing procompetitive justifications would be permitted or useful.

The categorical approach should, however, be rejected for several reasons. First, it is overly formulaic and likely to lead to errors where categories are construed too narrowly or too broadly.²⁴³ This is worrisome for digital platform defendants experimenting with new ways to transact and offer services to customers, as they may unexpectedly find themselves unable to assert a justification. Whereas some platform efficiencies like reduced search and contract formation costs are likely to be credited, it is not clear that self-regulatory conduct fits into a category used by the Tenth or Eleventh Circuit.²⁴⁴ This is despite a long history of similar cost reduction and regulatory-type behavior being credited for improving efficiencies dating back to Chicago Board of Trade.²⁴⁵ Further, the categories of justifications adopted in jurisdictions that apply this approach are not uniform and are likely to lead to confusion. Both the Tenth and Eleventh Circuit permit increasing output and improving product or services quality to be credited, but only the Tenth Circuit permits creating operating efficiencies and making a new product available, and only the Eleventh Circuit credits reducing cost and improving innovation.²⁴⁶ These types of incongruities are problematic for digital platform defendants that often serve a national market and thus face potentially conflicting legal treatment. Digital platform defendants would also be better served by an approach that consistently recognized improvements to innovation and allocative efficiency as cognizable procompetitive justifications.²⁴⁷ Finally, it is not clear that the Supreme Court would ever adopt a categorical approach for crediting procompetitive justifications. The Supreme Court has consistently used very broad language to describe what conduct can be recognized as a procompetitive justification.²⁴⁸

C. PROPOSAL: Broad consideration for the ways that platforms improve market efficiencies with some clarification but no explicit change to how procompetitive justifications are assessed within rule of reason

No radical departure from the existing efficiencies-focused approach for assessing procompetitive justifications within rule of reason analysis is necessary or helpful for antitrust cases involving digital platform defendants. The existing efficiencies-focused approach to review asserted procompetitive justifications is sufficiently broad and flexible to consider the range of ways that digital platform defendants enable efficiencies in the markets where they compete. As demonstrated below in recent cases implicating digital platforms, courts already consider a broad range of defendants’ asserted efficiencies and appropriately consider whether the defendant has carried its burden to substantiate these efficiencies with record evidence. Further, existing rule of reason analysis is well-calibrated to permit only restraints well-tailored to achieving efficiencies by requiring that no less restrictive alternative to achieving that benefit exist. Finally, existing rule of reason analysis already contemplates net effects to competition by weighing the anticompetitive effects of a restraint against the procompetitive benefits in its final balancing requirement.

Although the Supreme Court’s statement in Indiana Federation of Dentists regarding when courts should credit a procompetitive justification—when it is shown to benefit competition through “creation of efficiencies”—is vague, this remains the best approach to guide lower courts, because it is sufficiently broad and flexible to consider the range of efficiencies that a defendant may assert and support with record evidence.²⁴⁹ This approach also incorporates preexisting jurisprudence from Chicago Board of Trade, BMI, NCAA, and Professional Engineers, all of which lend further specification to which justifications might be credited. Conduct that “help[s] to improve market conditions” by increasing the number of trading partners and deals, reducing information asymmetries, removing risk, and improving allocation of resources may be permitted by analogizing to the facts of Chicago Board of Trade.²⁵⁰ Conduct that is necessary to overcome extreme transactional costs and to vindicate an intellectual property right may be permitted under BMI.²⁵¹ And, conduct that “increase[s] output and reduce[s] . . . price,” “mak[es] possible a new product,” “widen[s] consumer choice,” or “enhance[s] public interest [in the relevant product]” may be compared to the categories of efficiencies that the Supreme Court considered in NCAA.²⁵² Conduct that represents a refusal to compete on price or that so disrupts “the ordinary give and take of the market place” to set prices cannot be considered procompetitive in the application of Professional Engineers.²⁵³ The majority of circuit courts are right to follow the approach of Indiana Federation of Dentists and apply a broad standard that considers a wide range of cognizable efficiencies instead of limiting review to a pre-set category of efficiencies, as the minority of circuit courts have tried to do.²⁵⁴

One improvement that reviewing courts can adopt is to take the definition for assessing procompetitive justifications provided in Microsoft (itself an outgrowth of Indiana Federation of Dentists) and try to operationalize it as a series of steps for review: (1) assess if a procompetitive justification is cognizable, i.e., if it implicates an efficiency improvement; (2) determine if it is substantiated in fact; and (3) assess whether there is a less restrictive alternative to achieving the same result, should the plaintiff so assert that there is one.²⁵⁵ As demonstrated in the cases below, most courts consider these three elements in some manner, but like the D.C. Circuit in Microsoft, they do not tend to clearly describe their process.²⁵⁶ More strictly following a series of steps like those outlined above would help businesses and legal practitioners to understand if an asserted justification failed for not being cognizable, not being supported by record evidence, or not being sought out through the least restrictive means to do so. Thus, cases could be more instructive to those in the field regarding what type of conduct will be recognized as efficiency improving and what evidence is helpful to substantiate a claim. Cases might also be more instructive for reflecting what types of alternatives a party may need to undertake before imposing a restraint with certain deleterious effects to competition. Finally, this approach might streamline review on appeal to the extent that parties can stipulate that only one aspect of the lower court’s decision is being contested.

One additional improvement in cases involving digital platform defendants is for practitioners, not courts, to more specifically enunciate the efficiency being asserted and to try to substantiate that efficiency with more fulsome record evidence. As described above, the range of potential efficiencies that a platform defendant may assert is broad: improving market efficiencies through reducing costs, improving quality, facilitating information sharing, speeding innovation and improving allocative efficiency, among others.²⁵⁷ Yet, the defendant must show that a given restraint actually furthers an asserted efficiency—it is not sufficient to be granted a procompetitive justification that a platform’s ordinary business operations achieve this result without the restraint. The cases involving Apple and Sabre below contemplate this circumstance: in both cases, the court was right to reject outright or uphold the jury’s determination rejecting the defendant’s asserted justification for not providing record evidence that showed that the restraint itself furthered an efficiency.²⁵⁸ Otherwise, the restraint is merely anticompetitive conduct undertaken by an entity that would otherwise produce efficiencies in the market where it competes. The latter circumstances should not be afforded any special deference or consideration in antitrust review.

D. Application to recent cases involving digital platform defendants

From a review of recent antitrust cases involving digital platform businesses as defendants,²⁵⁹ courts have so far stepped up to the challenge of broadly considering possible efficiencies associated with the operation of these businesses, while also ensuring that a justification is only credited where it is supported by adequate evidence and no less restrictive alternative to achieving this benefit exists. This result supports this Note’s conclusion that no radical change to the existing approach for assessing procompetitive justifications is necessary but that some greater uniformity and precision in how the approach is applied would be helpful.

1. Realcomp II, Ltd. v. FTC

Realcomp, an association of approximately 14,000 real estate agents and brokers, provides two critical services to its members: access to its MLS database and advertising services.²⁶⁰ Realcomp maintains the largest MLS database in Michigan, with detailed information about available property listings and contact information for brokers representing home sellers.²⁶¹ The Realcomp database is only available to Realcomp association members.²⁶² Realcomp also provides an advertising service for members that allows them to share information about active listings on the MLS database to certain pre-approved public-facing websites.²⁶³ All brokers, whether full-service or limited service, pay the same fee to become Realcomp members and access its services.²⁶⁴

The FTC accused Realcomp of violating the FTC Act for two practices.²⁶⁵ First, the FTC challenged Realcomp’s policy of refusing to share information about certain exclusive agency listings and other nontraditional listings in its MLS database to public-facing websites as part of advertising services that it offered to association members.²⁶⁶ Second, the FTC challenged Realcomp’s practice of preventing exclusive agency and other non-traditional listings from being included in the default search settings for its MLS database.²⁶⁷ The FTC alleged that these practices were designed to insulate traditional, higher fee brokers from competition with new low-cost, limited-service brokers and individual buyers and sellers.²⁶⁸

Realcomp claimed two procompetitive justifications for its conduct. Realcomp claimed that its conduct reduced “free-riding” by non-traditional brokers and individuals, who allegedly did not pay full fare for the benefits of property advertising, listing, and search services that they received from Realcomp.²⁶⁹ The defendant also claimed that its conduct helped to reduce a “bidding disadvantage” that buyers using traditional brokers faced when bidding against buyers with low-cost brokers.²⁷⁰

In the initial FTC Commission decision, the Commission succinctly but comprehensively reviewed and rejected both of the platform defendant’s procompetitive justifications and applied the right law to do so. Citing Indiana Federation of Dentists, the Commission stated that its review would look for “some countervailing procompetitive virtue – such as, for example, the creation of efficiencies in the operation of a market or the provision of goods and services.”²⁷¹ Outlining its approach, the Commission wrote that it would assess: “whether those purported justifications are legitimate (i.e. ‘cognizable’ and ‘plausible’); whether they are supported by evidence in the record; and whether the restraints they impose are a reasonably necessary means to achieve a legitimate, procompetitive end.”²⁷² As such, the Commission’s approach was tightly aligned with the process for evaluating procompetitive justifications outlined in Microsoft and proposed by the Note as a best practice.²⁷³ Analyzing Realcomp’s first procompetitive justification of reducing free-riding, the Commission had no trouble analogizing to the vertical restraint context and determining that such an efficiency could be cognizable.²⁷⁴ The Commission then promptly rejected this claim on a finding that no record evidence showed that the restraint in fact improved market outcomes; instead, all users, traditional and low-cost brokers, had to pay the same access fees to use the Realcomp platform, so “there was no ‘free ride.’”²⁷⁵ The Commission also did not struggle to determine that Realcomp’s second claim for solving a “bidding disadvantage” among types of brokers merely masked an attempt by traditional brokers to prevent low-cost brokers from passing on cost-savings to customers.²⁷⁶ Thus, this claim failed to demonstrate a cognizable efficiency improvement.²⁷⁷ The Sixth Circuit ultimately confirmed the Commission’s reasoning and result.²⁷⁸

In sum, both the Commission and the Sixth Circuit demonstrated no difficulty in assessing procompetitive justifications asserted by a digital platform defendant under the broad efficiencies-focused approach outlined in Indiana Federation of Dentists and Microsoft. The Commission considered both whether the alleged justifications were cognizable efficiency improvements and if they were supported by record evidence. It also clearly outlined its reasoning to do so. The Sixth Circuit was right to confirm this reasoning and result.

2. US Airways, Inc. v. Sabre Holdings Corp

In Sabre, an airline brought suit against the owner of a global distribution system (“GDS”), a computer reservation service that matches airlines and other travel providers with travel agents looking to book travel plans for mostly corporate clients.²⁷⁹ In its 2011 contract with Sabre, US Airways paid Sabre a booking fee of at least $3.41 per US Airways flight segment booked on the Sabre platform.²⁸⁰ In the period 2006 to 2012, Sabre paid more than $1.2 billion in incentive fees to travel agents to encourage them to use its platform.²⁸¹

US Airways claimed that Sabre had violated the Sherman Act through several contract provisions in the agreement between the companies.²⁸² The provisions prevented the plaintiff from offering discounts or better benefits on tickets booked through non-Sabre sales channels or imposing surcharges on any tickets booked through Sabre’s platform.²⁸³ The provisions also prevented US Airways from inducing travel agents or customers to directly purchase tickets from the airline, i.e., to circumvent the Sabre system.²⁸⁴

Defendant’s claimed procompetitive justifications for its conduct were twofold. First, Sabre claimed that certain contract provisions enabled customers to see the lowest-priced fares across all GDS networks.²⁸⁵ Thus, consumers could more easily comparison shop for flights across airlines using the Sabre platform.²⁸⁶ Sabre claimed that these contract provisions also encouraged efficient booking practices by travel agents, as they reduced the time to make and change reservations.²⁸⁷ Second, Sabre claimed that its contract provisions were necessary to prevent US Airways from undermining the agreement between the parties and offering a better deal to rivals, as well as to prevent travel agent customers from being induced to use another platform to book lower cost flights after having researched those flights on Sabre’s platform.²⁸⁸ That is, certain provision were necessary to deter free-riding on the platform services.²⁸⁹

At the district court, a jury heard evidence from Sabre supporting its alleged procompetitive justifications as well as from the plaintiff regarding less restricting alternatives Sabre could have pursued.²⁹⁰ Sabre presented evidence that its contract provisions led to increased competition among airlines by enabling travel agents to shop more efficiently among multiple airlines and compare prices for fares.²⁹¹ US Airways presented evidence that Sabre could have separately charged for searching and booking flights, as well as permitted airlines to pass on a surcharge to customers booking through Sabre where other platforms charged lower fees to the airline.²⁹² US Airways argued that these alternatives would have adequately remedied Sabre’s free-riding concerns arising from travel agent booking practices while also permitting greater competition among airlines by allowing them to pass along savings from lower booking fees to customers.²⁹³ The jury ultimately determined that Sabre unreasonably restrained trade through its contract provisions, demonstrating that the jury either was not convinced by Sabre’s procompetitive justifications or was swayed by US Airways’ presentation of less restrictive alternatives.²⁹⁴

While the district court played a more limited role in assessing the record evidence in Sabre than in Realcomp because there was a jury sitting as factfinder, where relevant, the court nonetheless demonstrated acuity in evaluating the platform defendant’s asserted procompetitive justifications. Reviewing Sabre’s motion for summary judgment, the district court found that Sabre’s contract provisions could demonstrate benefits to competition from comparison shopping and more efficient booking methods, i.e., they were cognizable, and thus would be assessed by the jury.²⁹⁵ Analogizing to vertical restraint cases, the district court described that “[c]ontract provisions that result in a better or more efficient product to meet consumer demand are procompetitive.”²⁹⁶ The court also found Sabre’s free-riding claims to be cognizable.²⁹⁷ Although the district court did not give a clear statement as to what standard it was using to evaluate procompetitive justifications offered by the defendant, its reasoning reflected that it was looking for justifications that advanced a market efficiency and that were supported by record evidence, an approach aligned with Indiana Federation of Dentists and Microsoft.²⁹⁸

Later, when reviewing Sabre’s motion for a judgment as a matter of law under Federal Rule of Civil Procedure Rule 50(b), the district court upheld the jury’s determination that the defendant’s justifications, while cognizable on their face, may not have been convincingly substantiated, failed on a showing of less restrictive alternatives, or else failed to balance strong anticompetitive effects from Sabre’s conduct.²⁹⁹ The court described the process for evaluating procompetitive justifications as one that asks: “[W]hether ‘there [is] strong evidence that the challenged practice creates substantial efficiencies by reducing participants’ costs or improving product or service quality.’”³⁰⁰ Although the court cited Phillip Areeda and Herbert Hovenkamp’s antitrust treatise for this standard instead of Indiana Federation of Dentists or Microsoft, this approach largely aligns with the broad efficiencies-focused approach promoted in those cases.³⁰¹ One distinguishable aspect is that neither Indiana Federation of Dentists nor Microsoft required a showing of “strong evidence” that a challenged practice creates “substantial” efficiencies, merely a showing of “some countervailing procompetitive virtue—such as, for example, the creation of efficiencies in the operation of a market or the provision of goods and services” or “a nonpretextual claim that its conduct is indeed a form of competition on the merits because it involves, for example, greater efficiency or enhanced consumer appeal.”³⁰² That said, the Sabre court did not appear to apply a heightened standard of review to assess the procompetitive justifications asserted by the defendant.³⁰³ Rather, it found that Sabre presented evidence sufficient for the defendant to carry its burden, but that the jury as factfinder was within its discretion not to credit Sabre’s witnesses or arguments, to find that there were least restrictive means to achieve the same efficiencies, or to determine that on balancing, Sabre’s conduct was anticompetitive.³⁰⁴ The district court’s evaluation of procompetitive justifications was not challenged on appeal to the Second Circuit.³⁰⁵

In sum, the district court adequately assessed procompetitive justifications asserted by a digital platform defendant under a broad efficiencies-focused approach. The district court, where appropriate, considered whether the alleged justifications were cognizable efficiency improvements and if they were supported by record evidence. Although the clarity and precision of its reasoning could be improved, the district court’s ultimate conclusions were appropriate.

3. Other relevant recent cases

As noted above, the courts have yet to fully resolve many cases involving digital platform businesses as defendants.³⁰⁶ Two further recent cases suggest, however, that existing rule of reason analysis and its efficiencies-focused standard for assessing procompetitive justifications is sufficient to meet the challenge of reviewing conduct undertaken by digital platform defendants.

In United States v. Apple, Inc., the Second Circuit correctly determined that no procompetitive justifications applied to the conduct of a digital platform for e-book sales because the defendant, Apple, had participated in a horizontal price-fixing conspiracy with book publishers to try to gain market share in the e-books market.³⁰⁷ In dicta that briefly considered the merits of Apple’s procompetitive justifications—to encourage market entry by publishers to challenge Amazon’s market dominance and innovation benefits from collaboration with publishers to improve Apple’s tablet products—the Second Circuit correctly recognized which of Apple’s procompetitive justifications were not cognizable, and which were cognizable but not substantiated with record evidence. The Second Circuit did not give a clear statement as to what standard it was using to reject Apple’s procompetitive justifications but it appeared to be looking for justifications that advanced a market efficiency and that were supported by record evidence, an approach aligned with Indiana Federation of Dentists and Microsoft.³⁰⁸

Regarding Apple’s first justification, the use of anticompetitive conduct to disrupt an existing market cartel, the Second Circuit properly determined that the claim was not cognizable, as such conduct is never a permitted justification.³⁰⁹ The Second Circuit also correctly determined that Apple’s second justification, improving innovation in tablet products, was cognizable but not supported by record evidence.³¹⁰ Apple did not provide any evidence to establish a connection between the conduct at issue and the benefit that it alleged—it could point to no new innovation that it had undertaken with book publishers nor an explanation for why a price-fixing arrangement would be necessary to further this result.³¹¹ Thus, the court came to a sensible conclusion using a broad, efficiencies-focused approach for assessing the defendant’s asserted procompetitive justifications that required justifications to be both cognizable and supported by record evidence. Whereas the clarity and precision for what standard was being applied and what steps were being taken to reach this result could be improved, the ultimate conclusion was appropriate.

In Viamedia, Inc. v. Comcast Corp., the Seventh Circuit declined to resolve the merits of an antitrust case implicating the conduct of a high technology platform, Comcast, on a motion for summary judgment and reserved rule of reason analysis to be undertaken by the factfinder.³¹² In dicta that considered the merits of Comcast’s procompetitive justifications for tying and refusing to deal with a rival in the market for spot cable television advertising placement, the Seventh Circuit cited to the Microsoft decision’s definition for “procompetitive justification” to guide analysis upon remand to the court below.³¹³ Thus, the Seventh Circuit intended for review of procompetitive justifications to consider both whether a justification was cognizable and if it was supported with record evidence. The court further expounded on what justifications might be cognizable by citing to Phillip Areeda and Herbert Hovenkamp’s antitrust treatise, which advises crediting procompetitive justifications where conduct results in “higher output, improved product quality, energetic market penetration, successful research and development, cost-reducing innovations, and the like.”³¹⁴ This approach aligns well with the categories of justifications discussed in NCAA, and is also not intended to be exhaustive, in accord with the broad, efficiencies-focused approach of Indiana Federation of Dentists.³¹⁵ While not deciding the issue, the Seventh Circuit expressed skepticism that Comcast would be able to substantiate its alleged procompetitive justifications for tying and refusing to deal with a rival with record evidence.³¹⁶ As such, the Seventh Circuit adopted a broad, efficiency-focused approach for reviewing the defendant’s alleged procompetitive justifications that well encapsulated the leading cases for what justifications are cognizable as well as required that any justifications be substantiated in fact. Its analysis was clear and easy to follow upon remand to the court below.

Two other recent decisions at the Supreme Court and courts of appeals challenge whether courts are properly evaluating digital platform defendants’ procompetitive justifications. These cases are, however, distinguishable for involving mixed cases of antitrust and intellectual property licensing or for not reaching the merits of a full rule of reason analysis.³¹⁷

In Federal Trade Commission v. Qualcomm, Inc., the Ninth Circuit afforded significant and surprising weight to Qualcomm’s explanation for tying intellectual property licensing agreements to chip supply agreements to reduce transaction costs from multi-level licensing agreements.³¹⁸ The Ninth Circuit’s reasoning suggested a single firm’s cost reduction alone might be sufficient to rebut an allegation of anticompetitive harm or that it might be recognized as a procompetitive efficiency despite no improvement to market competition.³¹⁹ Yet, the Ninth Circuit’s reasoning reflected special deference to Qualcomm’s licensing preferences given the particular facts of the case—a mixed case of antitrust and intellectual property rights.³²⁰ The court wrote that “the rules of contract and patent law are better equipped to handle commercial disputes” between technologically advanced companies party to these licensing agreements, suggesting that the latter concern, not antitrust reasoning, was motivating its result.³²¹

In Ohio v. American Express, the Supreme Court suggested openness to crediting American Express’s asserted procompetitive justifications for preventing merchants from steering customers toward lower cost payment options at the customer’s point of purchase: improving interbrand competition among credit card providers and reducing free-riding among merchants that hold themselves out as accepting American Express.³²² This was despite the fact that American Express’s anti-steering provisions reflected a direct restraint to negotiations over price, presumably foreclosed by Professional Engineers, and evidence presented in the district court and not reversed on appeal that American Express had raised prices to merchants 20 times during the relevant five-year period.³²³ The Supreme Court did not reach the merits of this determination, finding instead that the plaintiff’s failure to show harm in a single market containing merchants and cardholders resolved the case.³²⁴ Four justices joined Justice Breyer’s dissenting opinion, which favored finding anticompetitive harm at the first step of rule of reason analysis and remanding the case to the Second Circuit for a determination of offsetting procompetitive benefits.³²⁵ The dissent also expressed strong resistance to crediting American Express’s asserted procompetitive justifications.³²⁶

Conclusion

In sum, the Supreme Court’s statement in Indiana Federation of Dentists as to when to credit a defendant’s asserted procompetitive justification—where it is shown to benefit competition through “creation of efficiencies”—remains the best path forward for lower courts.³²⁷ This approach permits defendants, including digital platform defendants, to assert a wide variety of justifications for their conduct, which they must then substantiate with evidence demonstrating that an improvement to market efficiencies has actually taken place. While somewhat vague, this approach means that platform defendants have a chance to assert efficiency improvements from reduced time and effort to find desired goods and services, reduced contracting costs, increased product and services quality, increased rate of innovation, and improved allocative efficiency as procompetitive justifications. In addition, earlier Supreme Court decisions including Chicago Board of Trade, BMI, NCAA, and Professional Engineers have helped to delineate what types of justifications may be credited, including conduct that: increases output, reduces price, makes possible a new product, expands consumer choice, or enhances consumer interest in the relevant product.³²⁸

Further, existing rule of reason analysis is well designed to permit only justifications well-tailored to achieving an asserted benefit to competition to meaningfully affect the outcome of a case. The factfinder is within its discretion to reject a justification where a defendant does not carry its burden of proof, a plaintiff shows a less restrictive means to achieve the same benefit, or balancing shows that the harm from the conduct’s anticompetitive effect outweighs its procompetitive effect. Thus, whereas the body of efficiencies that might be credited as a procompetitive justification is broad, very few will be determinative of an antirust claim.

Finally, the existing approach appears to be working well as it is applied to cases involving digital platform defendants. A review of recent cases implicating these defendants confirms that lower courts have sufficient guidance for how to apply the broad, efficiency-focused standard for assessing procompetitive justifications from Indiana Federation of Dentists and Microsoft as well as how to integrate these results into fuller rule of reason analysis. And, the current approach is superior to other ways forward suggested in legal scholarship and experimented with in a minority of circuit courts, which are poorly calibrated to considering the range of ways that platforms enable efficiencies in the markets where they compete.

One improvement suggested by this Note is for courts to provide greater clarity when assessing procompetitive justifications regarding what criteria they are using to determine whether to credit a given efficiency. Courts can do so through applying the definition for “procompetitive justification” developed in Microsoft as a series of steps for review: (1) assess if a procompetitive justification is cognizable, i.e., if it implicates an efficiency improvement; (2) determine if it is substantiated in fact; and (3) assess whether there is a less restrictive alternative to achieving the same result, should the plaintiff assert that there is one.³²⁹ Courts that more strictly follow a series of steps like those above would help businesses and legal practitioners to understand if an asserted justification failed for not being cognizable, not being supported by record evidence, or not being sought out through the least restrictive means. Thus, cases could be more instructive to those in the field regarding what type of conduct will be recognized as efficiency improving and what evidence is helpful to substantiate a claim. Cases might also be more instructive for reflecting what types of alternatives a party might need to undertake before imposing a restraint with certain deleterious effects to competition.