United States: Platform Economics and Mergers

Introduction

In 2021, the United States saw continued regulatory and legislative scrutiny of merger activity by large firms whose businesses can be analysed as two-sided or multi-sided platforms.^[2] Scrutiny of these businesses, and particularly of digital platforms, has been intensifying since President Biden issued an executive order on competition in July 2021. The executive order was intended to promote greater competition in several industries, including technology, and called for closer scrutiny of large companies’ acquisitions of nascent or potential competitors.^[3]

Following the executive order, legislators and regulators have taken steps to re-examine the statutes and regulatory framework that govern antitrust enforcement in ways that directly impact digital platforms:

Several bills were introduced in both houses of Congress addressing digital platforms, including the American Innovation and Choice Online Act, which seeks to regulate the largest ‘online platforms’;^[4] the Open App Markets Act, which is focused narrowly on payments in app marketplaces;^[5] and the Platform Competition and Opportunity Act of 2021, which would establish ‘certain acquisitions by dominant online platforms’ to be unlawful.^[6]

In September 2021, and because of strenuous objections from two of its five commissioners, the Federal Trade Commission (FTC) rescinded its support of the revised Vertical Merger Guidelines, which the FTC and the Antitrust Division of the Department of Justice (DOJ) had jointly issued in 2020. The three commissioners voting for the rescission – FTC Chair Lina Khan, Rebecca Kelly Slaughter and Rohit Chopra^[7] – noted that network effects and other aspects of digital markets result in a ‘broader set of tactics that firms may use to raise rivals’ costs’.^[8] While it did not officially withdraw its support, the DOJ stated at the time that it was conducting a review of both the Horizontal Merger Guidelines and the Vertical Merger Guidelines ‘to ensure they are appropriately skeptical of harmful mergers’.^[9]

Also in September 2021, the FTC published a retrospective analysis of acquisitions by selected technological platforms of small companies (i.e., transactions that were not required to be reported because they were below the statutory threshold).^[10] In discussing the report, Chair Khan remarked that it ‘captures the extent to which these firms have devoted tremendous resources to acquiring start-ups, patent portfolios, and entire teams of technologists—and how they were able to do so largely outside of our purview.’^[11]

In January 2022, the FTC and the DOJ (the Agencies) jointly issued their Request for Information on Merger Enforcement (RFI), soliciting public comments in 15 areas that the agencies were re-examining to better understand ‘how the agencies can modernize enforcement of the antitrust laws regarding mergers.’^[12]

In addition to the RFI, the Agencies hosted ‘listening forums’ on the ‘firsthand effects of mergers and acquisitions’, including a session on technology mergers that was held on 12 May 2022.^[13] In this session, both Assistant Attorney General Jonathan Kanter of the DOJ and FTC Chair Khan expressed a concern with incumbent platforms acquiring nascent competitors, among other concerns that led them to suggest the need to revise the guidelines ‘to better fit a modern economy’.^[14]

This focus on digital platform competition is rooted in the economics of two-sided platforms, which can present features that need to be incorporated into antitrust analysis. In this Chapter, we first provide a brief background on the economics of platforms. We then use the themes outlined in the FTC and DOJ RFI to explore implications for digital markets when analysed as two-sided platforms, including summarising some of the commentary received to date by the agencies.

Brief overview of the economics of two-sided platforms

Two-sided platforms are businesses that enable and encourage two groups of users to connect to and interact with each other.^[15] These businesses may follow a variety of models, but they share two key features:

• They act as matchmakers between user groups,^[16] making it easier for users on both sides to find each other (described by economists as reducing ‘search costs’) and lowering the costs of any associated interaction (described by economists as reducing ‘transaction costs’).^[17] For example, the operator of a ride-sharing platform first matches a rider seeking travel from point A to point B with a driver willing to make that trip and subsequently facilitates the payment of the associated fare.
• They experience indirect network effects,^[18] meaning that the value of the platform to one group of users is connected to the presence of users on the other side of the platform. This means that platform operators must nurture both sides of the platform simultaneously.

Two-sided platforms can be observed throughout the economy. For example, payment cards and newspapers are classic examples of two-sided platforms. Still, as the economy has become more and more digital in the past 20 years, it has led to an increasing prevalence of (digital) platform businesses,^[19] driven by the abundance of data and increased processing power that enable improved matching.^[20] Commonly cited examples of two-sided digital platforms include online marketplaces and sites related to the sharing economy, from ride-sharing apps to vacation rentals.

The origins of economic research into two-sided platforms can be traced back to early work on network effects or network externalities, which focused on ‘system competition’ between alternative combinations of components, such as hardware and software.^[21] Starting in the early 2000s, and alongside the rise of the internet, economists’ growing understanding of network effects led them to the recognition that ‘[m]any if not most markets with network externalities are two-sided’ and ‘must “get both sides of the market on board,”’^[22] giving rise to the field we now know as platform economics.

Platform economics initially focused on pricing strategies,^[23] and specifically on how the characteristics of demand on each side of the platform, and the structure of indirect network effects, create different pricing incentives on each side of a platform. Because demand on one side of the platform affects demand on the other side, prices cannot be set in isolation; instead, different prices may be charged to different groups of consumers. As an example, a credit card issuer may charge merchants for each transaction completed with the card, while offering rewards to, rather than charging, consumers using the card.

The study of pricing within a single platform naturally led to the study of pricing when there is competition among platforms, including when customers on one or both sides of the platform use multiple platforms, known as multi-homing.^[24] For example, consider drivers that simultaneously accept rides from both Uber and Lyft, or riders that maintain both Uber and Lyft accounts and switch between them regularly.^[25]

Another question analysed by platform economists is whether the prevalence of network effects leads platform competition to a ‘winner-takes-all’ situation, or alternatively results in ongoing competition among multiple platforms.^[26] Intuitively, the presence of network effects may give an advantage to platforms that enjoy early success and, therefore, acquire an even larger number of users, but many factors have been found to reduce this early advantage, including the possibility of multi-homing, the structure of the network and the possibility that the network effects are fragile.^[27] Furthermore, the success of social media platform start-ups such as Snapchat and TikTok demonstrate that, at least in some instances, even competitors with large, established platforms face significant competition from new entrants.

Finally, research in economics and strategy has focused on the ways in which platforms can compete over time or dynamically. Topics have included incentives to innovate and improve quality.^[28] Inquiries into innovation and quality focus not just on how a platform competes through innovation of its own services, but also on how the platform can affect its own participants’ incentives to innovate and compete. For example, platforms that operate online marketplaces can create ratings systems that encourage sellers to improve their offerings,^[29] and innovations in operating system features can induce video game developers to improve graphics and play experience.^[30]

As we discuss in the next section, while the Agencies’ RFI raises broad topics that affect mergers in businesses of all types, it also addresses whether the unique characteristics of two-sided platforms, and of digital businesses specifically, warrant special treatment or whether they can be addressed in the broader context of traditional merger review.

Platform-related themes in the RFI

The RFI is part of a broader effort in which the Agencies seek to ‘modernize enforcement of the antitrust laws regarding mergers’^[31] and update both the 2010 Horizontal Merger Guidelines and the 2020 Vertical Merger Guidelines. The RFI appears to be driven by a belief that merger enforcement has been too lax: the Agencies describe themselves as being ‘particularly interested’ in any aspects of competition that current guidance may ‘underemphasize or neglect’ and seek ‘specific examples of mergers that have harmed competition’.

The RFI is structured as a series of questions on 15 topics, spanning the purpose and goal of the merger review process, the types and sources of evidence that Agencies should use in their analysis of merger outcomes and the types of analysis that can identify those outcomes, and themes arising in specific areas of merger review: potential and nascent competition, monopsony power and labour markets, innovation and intellectual property, and digital markets. These topics are not separated into horizontal or vertical mergers, and one question is whether this distinction should be revisited.

As at the time of writing, the RFI has received more than 1,900 comments,^[32] which are expected to inform updated guidelines by the end of 2022.^[33] It drew comments from a broad range of stakeholders, including academics, practitioners, industry groups and think tanks, as well as individuals.

The questions on digital markets highlight themes that are common to the antitrust analysis of two-sided platforms, from the possibility of ‘tipping’ to the treatment of multiple sides in market definition. In this section, we focus on these digital market questions, zooming in on those that are most closely related to two-sided platforms. Most of these questions tap into existing debates about the appropriate treatment of two-sided platforms in general and digital platforms in particular. For each set of questions, we first present the questions as stated in the RFI, and then we discuss the relevant economics and present some of the commentary from academics, practitioners and other commentators.

Should there be dedicated guidelines for digital markets?

From the RFI:

How, if at all, should the guidelines’ analysis of mergers in digital markets differ from mergers in other markets? How should markets be defined in the case of mergers in the digital sector where products and services undergo rapid change? How should the guidelines address prospective competitive harms in rapidly evolving markets?^[34]

This first set of questions in the RFI relates to whether digital markets are sufficiently different from other markets to merit special guidelines, focusing on the possible challenges associated with market definition when products change rapidly through innovation, and allowing for the possibility that harm may be prospective in digital markets. During the listening forum on technology, AAG Kanter mentioned the ‘unique economics of the internet’, and Chair Khan referred to ‘dynamic and novel issues’ in digital markets.^[35]

The comments received reflect disagreement among commentators over the need for special treatment of digital businesses.

Those in favour of a different approach appeared to worry that digital businesses have certain unique features that make traditional merger analysis harder to implement or that may require novel theories of harm. For example, they suggest that traditional approaches to market definition may result in overly narrow markets that lead to ‘overlooking’ potential anticompetitive effects,^[36] or that the prevalence of products with no monetary price in digital platforms renders the Guidelines’ primary focus on the analysis of price effects ineffective.^[37] Others, however, suggest that in digital markets in particular, it is necessary to consider the impact of an acquisition across multiple markets.^[38]

Those against such distinction point to the imprecision of defining a digital market when digital businesses are highly diverse and many industries offer products that may have digital components, or when digital offerings may compete with physical offerings,^[39] questioning the need to single out digital markets. They also argue that Guidelines should be organised around analytical concepts, and that the analytical concepts that apply to digital businesses are not unique to the digital world and may apply more broadly.^[40]

Even commentators who favour a special focus on digital markets acknowledge that the factors that arise in digital markets are not necessarily new, but instead that the ‘degree to which they exist in digital markets’^[41] may require special attention. A middle-of-the-road approach may be to update the Guidelines to reflect the type of competitive dynamics that arise in digital markets, while stopping short of creating separate guidelines.

Whether through specific guidelines for digital markets or through the analysis of platform economics more generally, the remaining questions raise relevant economic issues, which we address next.

Network effects and potential tipping

From the RFI:

How should the guidelines analyze mergers in markets subject to tipping toward oligopoly or monopoly, such as may result from significant network effects? How should the nature and timing of enforcement strategy differ in markets subject to tipping?^[42]

This set of questions focuses on the possibility that an industry might ‘tip’ into a concentrated structure because the magnitude of network effects for businesses operating in the industry, and that merger enforcement in such industries might need to be modified to take such risks into account, including by having different ‘timing’, where timing appears to refer to the idea that protecting nascent competition may require deployment of antitrust enforcement before there is evidence or likelihood of actual harm.

The concept of tipping reflects the winner-takes-all hypothesis discussed in the section on platform economics. Under this hypothesis, indirect network effects within a platform could grow sufficiently large such that the leading platform will ‘pull away from its rivals in popularity once it has gained an initial edge’.^[43]

Theoretical research in platform economics has pointed to the possibility of tipping in markets that exhibit indirect network effects.^[44] A commonly cited theoretical conclusion is that competition under such circumstances can be ‘for the market’ rather than competition for share within the market.^[45] This does not necessarily lead to a sustained absence of competitive constraints for incumbents, as entrants may seek to challenge the existing incumbents by offering a differentiated or improved product.

More recently, the theoretical inevitability of tipping has been challenged. In the context of digital platforms, scholars have pointed to factors such as the fragility of incumbent network effects when they are not tethered to a hardware platform,^[46] as well as to the importance of how the network is structured and whether the network effects are local,^[47] challenging the idea that markets with two-sided platforms are inevitably prone to tipping and highlighting the possibility that there could be ongoing competition within the market.

Recent research has also pointed to the limits in the power of network effects, which do not always grow with the total size of the network, but may instead depend on whether the ‘right’ participants join the network (a phenomenon that is sometimes referred to as local network effects).^[48] It is also possible that, at some point, positive indirect network effects from increased size are outweighed by negative externalities from ‘congestion’, which limits a platform’s optimal size.^[49]

How does theoretical economic research relate to the RFI? The RFI appears to start from the premise that two-sided digital platforms are competing for the market rather than within the market, in the sense that a traditional antitrust analysis may not correctly identify as being in the same product or geographic market the full set of competitive forces faced by two merging parties. If this were the case, it is possible that the established approach to merger analysis, which focuses on product and geographic markets in which the merging parties currently overlap, would fail to flag certain mergers of current competitors.

In a similar vein, the focus on the nature and timing of enforcement strategy appears to originate from the concern that firms that are benefitting from having achieved this tipping point would want to acquire potential competitors before they enter the same market.

This view was espoused by AAG Kanter during the listening forum:

If given room to grow, the unique economics of the internet and digital autonomy mean that new and innovative company [sic] can rapidly disrupt existing markets, challenging incumbents, and enhancing consumer choice. But if we allow dominant firms to buy up or block these nascent competitors before they get to scale, we will lose out twice.^[50]

Similar concerns are raised in other regulatory efforts. The three bills discussed in the first section of this Chapter note a 50 million user threshold for the platforms they seek to regulate, which can be interpreted as a reference to network effects and a de facto definition of a tipping point. In this same vein, the Platform Competition and Opportunity Act explicitly forbids acquisitions if the asset or stock issuer is a ‘nascent or potential competit[or] to the covered platform.’^[51]

Responses to the RFI revealed mixed views on the need for special attention to nascent competitors and on the need to establish special mechanisms for the analysis of businesses with strong indirect network effects. Commentators presented both sides of the economic debate on the strength and durability of network effects.

Commentators in favour of creating a special analytical framework for mergers of businesses with strong network effects, particularly digital platforms, argued that dominant incumbent platforms enjoy durable market power and prevent entry by new firms.^[52] Consistent with Chair Khan and the FTC’s retrospective study of acquisitions, these commentators also argued that digital platforms have adopted a ‘serial acquisition strategy’ of smaller competitors.^[53] A common message was that the risk of tipping may call for increased enforcement.^[54] These positions were summarised succinctly by one commentator:

[E]nforcers and courts need to pay particular attention to potential and nascent competition in digital platform markets. They create markets prone to tipping, where a small, new, or potential competitor may play an outsized role. To protect competition in these markets, it’s especially important to recognize the harms of acquisitions of potential or nascent competitors and block mergers that might be allowed in other types of markets.^[55]

Other commentators noted that more recent research has pointed to the fragility of network effects in many digital platform settings and point to TikTok’s rivalry with Facebook as an example of an ‘upstart’ effectively challenging an incumbent platform,^[56] arguing that there is no particularly significant risk and that transactions should be evaluated on a case-by-case basis. One commentator on this side of the debate offered the following summary:

The propensity of network effects to tip markets and the resulting harm are both exaggerated. Networks effects are just one of the many forces shaping market evolution. Dominance can be unstable and short-lived with digital platforms, and several other forces can counter the positive reinforcement of network effects. . . . The strategy and timing of merger enforcement should not depend on the possibility of tipping.^[57]

It is interesting to note that in addition to the questions discussed in this section, the RFI has a section entirely dedicated to whether the guidelines should change standards or propose new approaches to address ‘potential and nascent competition’ more generally, across all industries. These are not focused on digital markets or platforms, so we will not discuss them here, and instead turn to the next topic: transaction platforms.

Are transactions platforms different?

From the RFI:

How should the guidelines evaluate mergers in two-sided simultaneous transaction platform markets? What are the competitively-relevant differences between two-sided simultaneous transaction platforms and other kinds of multi-sided platforms?^[58]

These questions address a potentially significant distinction between different kinds of two-sided platforms, calling out platforms that enable simultaneous transactions as potentially meriting differential antitrust analysis.

The notion of a special two-sided platform called a ‘simultaneous transaction’ platform was put forth by the US Supreme Court in its Ohio v. American Express (Amex) decision on anti-steering provisions in credit cards.^[59] Amex analysed credit cards as two-sided platforms connecting buyers and merchants and highlighted the simultaneous nature of interactions between the platform and each side:

Because the interaction between the two groups is a transaction, credit-card networks are a special type of two-sided platform known as a “transaction” platform. The key feature of transaction platforms is that they cannot make a sale to one side of the platform without simultaneously making a sale to the other.^[60]

While the Supreme Court focused on simultaneity, the economic literature that originated the term transaction platform focused on the presence of an observable interaction (or transaction) between both sides of the platform.^[61] Importantly, the definition of transaction platforms does not require that this interaction be a monetary transaction, but that the different sides of the platform interact directly with each other, through the platform. A transaction platform differs from a newspaper, which could be seen as bringing together readers and advertisers (therefore operating as a two-sided platform) who never interact directly with each other through the platform.

Amex and the underlying economic literature upon which the decision relies suggest that this distinction directly informs the approach to market definition in the antitrust analysis of two-sided platforms. The main question is whether an antitrust analysis should define a single market that combines the services provided to both sides of the platform, or to define two related markets, and whether the answer depends on the platform being a transaction platform.

Amex and proponents of defining a single market for transaction platforms view the platform as providing the service of facilitating interactions between both sides:

[D]efining a single market implies defining the market for services to a transaction. The product that is offered is the possibility to transact through the platform. . . . Candidate substitute products are not only other platforms which offer, to both sides, the possibility to transact but also non-intermediated transactions.^[62]

In Amex, the Supreme Court ruled that both sides of the platform should be included in a single market for credit card transactions.^[63] The Supreme Court also found that finding a price increase on one side of the platform was not sufficient to show anticompetitive effects in the relevant market, which should consider joint price and joint output.^[64]

The Amex decision has proven to be controversial, with certain antitrust scholars and practitioners embracing the approach taken by the Supreme Court,^[65] and others questioning whether the definition of a single market may miss relevant competitive dynamics on either side of the platform.^[66]

The controversy surrounding Amex is echoed in the commentary on the RFI, with certain commentators arguing for limiting the application of a single market definition in platform settings, without addressing the question of whether a platform is a transaction platform, and others arguing that such a single-market approach is often necessary and appropriate for transaction platforms.

Commentators arguing in favour of limiting the Amex approach argue that there should be no special consideration given to transaction platforms, and that antitrust authorities should always separately consider competitive pressures on each side of the platform rather than defining a combined market.^[67] Many commentators directly acknowledged their view that following the approach delineated in Amex may be ‘problematic’,^[68] and that new guidelines may need to depart from the approach or limit its application to narrow circumstances related to competition being uniform across both sides of the platform, rather than whether the platform is a transaction platform.^[69]

On the other side of the debate, certain commentators supported the need to understand transaction platforms as providing a single service that connects consumers on both sides of the platform, and therefore analyse market definition and competitive effects by defining a single market,^[70] with some commentators emphasising the need to follow Amex in vertical mergers.^[71]

The question of whether all or some two-sided platforms are seen as transaction platforms, and therefore as providing the service of intermediating interactions, can have wide-ranging consequences on the antitrust analysis of a proposed transaction, and it is likely that any updated guidelines will provide additional guidance on the Agencies’ interpretation of the law and description of their current practice. The next set of RFI questions digs deeper into the specifics of merger analysis for two-sided platforms.

Choosing the right analytical framework

From the RFI:

What are the appropriate indicia of market power in complex and multi-sided markets? Are traditional market definition approaches reliable frameworks for assessing the existence and magnitude of market power in these markets? Are other tools as effective or more effective than market definition in those contexts?^[72]

This group of RFI questions suggests that the characteristics of industries with two- or multi-sided platforms may require a re-examination of the antitrust toolkit available to regulators when assessing market power. The questions suggest two separate concerns: one is whether existing approaches that originate in the analysis of one-sided markets must be adapted to apply to settings with two-sided platforms; the other is whether there are new tools or approaches that would be ‘more effective’ in these settings.

Research in platform economics acknowledges challenges with applying existing tools to two-sided platforms. For example, researchers have highlighted the difficulties with the application of the hypothetical monopolist test based on a small but significant increase in price (SSNIP)^[73] because of the need to account for indirect network effects that create feedback loops. Researchers have also emphasised the point that ‘the standard quantitative tools of merger analysis can’t be used mechanically when multi-sidedness is important’,^[74]giving as examples the need to adapt critical loss analysis to contexts with zero prices on one side of the platform, and the need to include indirect network effects in merger simulations. More recently, researchers have proposed adaptations of analysis of upward pricing pressure (UPP) to settings with two-sided platforms, finding that not doing so may overstate UPP.^[75]

Responses to these questions were broad, consistent with the broad scope of the inquiry. Most commentators seemed to agree with the need to adjust antitrust analysis to the realities of any industry, but while some commentators noted that ‘[c]omplex and multi-sided markets are not so special’,^[76] others proposed specific ways in which traditional analyses may need to be adjusted. For example, certain state enforcers commented on the need to adapt structural presumptions to settings with network effects, suggesting that firms in such settings may be able to exercise market power at lower shares,^[77] and to create analogues of the SSNIP test as alternatives for markets in which prices on one side of the platform may be non-monetary, such as by considering potential increases in attention or data requirements from customers, rather than increases in price.^[78] Other suggestions are to de-emphasise the need for a SSNIP test and to focus instead on a holistic approach, focusing on qualitative evidence of substitutability that identifies products or services that are ‘reasonably interchangeable’.^[79]

Overall, commentators did not seem to embrace a wholesale change away from traditional antitrust analysis, but instead supported the careful application of existing principles to two-sided platforms specifically.

One factor unique to two-sided platform settings is the concept of multihoming. The RFI included a set of questions specifically related to multi-homing, which we address next.

The role of multi-homing

From the RFI:

How should the guidelines account for multihoming or interoperability? To what degree does multihoming or interoperability offset competitive concerns in actual practice?^[80]

The final set of RFI questions that we analyse in this chapter is related to multihoming and interoperability, which is an important factor in the analysis of competition between platforms.

As presented in the RFI and some of its responses, the discussion of tipping and competition for the market suggests that users choose a single platform on which to interact. The concept of multi-homing presents an alternative to this assumption.

Multi-homing is present when users can easily access multiple platforms within a given industry. Theoretical economics literature shows that the nature of competition among two-sided platforms depends on the extent to which one or more of the sides are able to multi-home, although there is no single answer, and the consequences of multi-homing depend on how each model is set up and what kind of multi-homing it can accommodate.^[81]

The ability to multi-home is related to the ease of switching among competing platforms, and academics have argued that increased ease of multi-homing increases competition.^[82] Interoperability between platforms has a similar effect to multi-homing, in that it makes switching easier and increases competition.^[83]

Commentators have differed regarding the weight to put on considerations of multi-homing when analysing mergers in industries with two-sided platforms. They have pointed to the potential mitigation of anticompetitive effects driven by the availability of multi-homing,^[84] with some putting forth the idea that a platform may acquire a company in a different market to reduce competition by making multi-homing more difficult or preventing interoperability.^[85] This would be the case, for example, if this target company was in the business of facilitating multi-homing, for example, by facilitating sellers’ maintenance of storefronts in competing online marketplaces. Supporters of this view note a need for the merger guidelines to be ‘especially wary of potential mergers to make multi-homing more difficult’.^[86]

While acknowledging the relevance of multi-homing and interoperability, commentators suggested that it is ‘too specialized for inclusion in the Guidelines’ and too subjective and unlikely ‘to provide meaningful insights into a merger’s effect on competition’.^[87] These commentators pointed to the need to understand the magnitude of the importance of multi-homing and interoperability in any given setting, which requires understanding consumer behaviour. In this view, the possibility of multi-homing or interoperability is just one additional factor in antitrust analysis and does not rise to the level of requiring inclusion in the guidelines.^[88]

The debate on multi-homing captured in comments to the RFI echoes that of other questions: whether the guidelines need to include specific mention of phenomena that may be particularly relevant in two-sided platforms or instead provide a general conceptual framework that can be adapted to the specifics of the transaction under analysis.

Conclusion

The increased focus on antitrust enforcement in technology has extended to the arena of mergers and, in particular, to the merger guidelines, as evidenced by the RFI and listening forum. The two main dimensions of debate seem to be the degree to which merger enforcement should look beyond established practices (e.g., scrutinising mergers when products overlap) and the degree to which the economics of two-sided markets, and of two-sided digital platforms in particular, need to be addressed specifically within the guidelines, or whether they can be addressed by a general conceptual framework that can be applied across industries and market structures. Whatever the outcome, the upcoming year is likely to bring renewed scrutiny of acquisitions in the digital space.

Notes