Jump to Content Jump to Main Navigation
Big Data and Competition Policy by Stucke, Maurice; Grunes, Allen (1st June 2016)

Part III Why Haven’t Many Competition Authorities Considered the Implications of Big Data?, 8 Data-Driven Mergers Often Fall Outside Competition Policy’s Conventional Categories

Maurice E. Stucke, Allen P. Grunes

From: Big Data and Competition Policy

Maurice Stucke, Allen Grunes

From: Oxford Competition Law (http://oxcat.ouplaw.com). (c) Oxford University Press, 2015. All Rights Reserved. Subscriber: null; date: 02 July 2020

Agency agreements — Market power — Rights — Internet — Technology — National merger control

(p. 127) Data-Driven Mergers Often Fall Outside Competition Policy’s Conventional Categories

8.01  Another reason why the competition agencies have not fully considered the implications of Big Data is that data-driven mergers can defy the agencies’ conventions that work well for other mergers. This chapter explores three conventions: the categorization of mergers, the belief that similar products/services compete more fiercely than dissimilar products/services, and the focus on the services’ or products’ substitutability.

A.  Categorization of Mergers

8.02  Competition agencies categorize mergers as horizontal, vertical, and/or conglomerate.1 The competition agencies review and challenge horizontal mergers (ie between companies that directly compete) more often than vertical mergers (where the merging parties operate at different levels of the supply chain, with one party supplying a key input or being a key customer).2 Rarely do they challenge conglomerate mergers.3 Regardless of the category, the ultimate (p. 128) assessment is the same, namely whether the merger may substantially lessen competition or tend to create a monopoly (or monopsony). But the analytical framework differs depending on whether it is a horizontal, vertical, or conglomerate merger.

8.03  As we have seen, data-driven mergers can be horizontal. Facebook and WhatsApp offered competing texting apps. Google and Waze directly competed with turn-by-turn navigation applications for smartphones and tablets. Data-driven mergers can be vertical, as when TomTom, the leading manufacturer of portable navigation devices and navigation software, acquired Tele Atlas, a key supplier of navigable digital map data. At times, a merger can be both horizontal and vertical (such as when a company supplies and competes with the acquired company).

8.04  But data-driven mergers often defy the agencies’ three general categories. We saw this with Google’s acquisition of Nest Labs. The merger was not horizontal. Google never sold smart thermostats, carbon monoxide detectors, and surveillance cameras. Nor were Google and Nest Labs in a vertical relationship; neither company supplied the other’s industry with a key input or was a key customer. Nor was this a conglomerate merger in the antitrust sense. Neither company was ‘active in closely related markets (e.g., mergers involving suppliers of complementary products or of products which belong to a range of products that is generally purchased by the same set of customers for the same end use)’.4

8.05  When a merger like Google/Nest Labs falls outside the horizontal, vertical, or conglomerate categories, the agency lacks an existing guideline (or analytical framework) to assess the merger. It will likely close the investigation (if it even opened one). But that does not mean that these mergers are harmless. Several European Commission officials in 2015 identified a potential theory of harm:

In the digital economy, large sets of data (so-called ‘big data’) are becoming increasingly valuable as they reveal patterns of information that enable companies to understand user behaviour and preferences and improve (or target) their products and services accordingly. This makes the availability of ‘big data’ a significant competitive advantage for companies active in, for instance, targeted online advertising, online search, social networking services and software products. From a competition law perspective, a possible theory of harm is that combining the merging parties’ datasets could provide them with a competitive advantage, by helping them to improve the merged entity’s product or service post-merger in a way that competitors are unable to match.5

(p. 129) 8.06  A related candidate theory of harm is whether the merger increases the company’s market power in any market (such as search, online advertising, etc) by tipping the market to the merging party. As we shall see in Chapters 11–13, some industries, like the search engine market, have several data-driven network effects, where the scale and scope of data play a critical role. Although the merger could benefit consumers, it might also insulate the platform from competitive pressure by inhibiting smaller competitors’ growth and increasing entry barriers.

8.07  So even though some data-driven mergers, like Google/Nest Labs, fall outside antitrust’s conventional categories, the mergers could still harm competition and consumers. The competition agencies cannot ignore these outliers. But they currently lack an analytical framework for identifying which outlying merger deserves closer scrutiny.

B.  Belief that Similar Products/Services Compete More Fiercely than Dissimilar Products/Services

8.08  Outside of offences treated per se illegal (such as price-fixing), antitrust plaintiffs ordinarily must define a relevant market, where the goods (or services) are interchangeable with one another, but not with products (or services) outside the market. The conventional antitrust analysis is that similar products compete more closely against one another than dissimilar products. As the European Commission states:

A relevant product market comprises all those products and/or services which are regarded as interchangeable or substitutable by the consumer, by reason of the products’ characteristics, their prices and their intended use. A relevant product market may in some cases be composed of a number of individual products and/or services which present largely identical physical or technical characteristics and are interchangeable.6

8.09  Likewise, the US agencies note in industries with differentiated products, ‘some products can be very close substitutes and compete strongly with each other, while other products are more distant substitutes and compete less strongly. For example, one high-end product may compete much more directly with another high-end product than with any low-end product.’7 For many household goods this makes sense.

(p. 130) 8.10  One illustrative merger while we were at the US Department of Justice (DOJ) involved white bread.8 Although there was a high cross-elasticity of demand between the merging parties’ white bread, there was little substitution to other types of bread. Here, we see high functional interchangeability, as arguably white bread is roughly equivalent to wheat bread for the uses to which they are put, namely sandwiches. But purchasers (or their children) did not treat wheat or multi-grain bread as substitutes for white bread. White bread, the DOJ discussed,

differs significantly in product attributes from other types of bread, such as variety bread (e.g., wheat, rye or French) and freshly baked in-store breads, in taste, texture, uses, perceived nutritional value, keeping qualities, and appeal to various groups of consumers. These differing attributes give rise to distinct consumer preferences for each type of bread.9

Finding that a small but significant increase in the price of white pan bread by all producers would not be rendered unprofitable by consumers substituting to other breads, the DOJ defined the product market as white pan bread.

8.11  Standard product market definition looks at competition at the firm level; products may be further differentiated within a relevant market when some products are closer substitutes than others.10 One example is baby wipes. The DOJ alleged baby wipes as the relevant product market, but private label was not an effective constraint.11 Here, product market definition—either as premium baby wipes brands or baby wipes generally—is less critical, as the merging parties’ brands were each other’s most significant competitive constraint. At times, the competition authority simply defines the product market more narrowly, such as super-premium ice cream.12 Regardless, our overall point is that antitrust tends to focus on substitutability whether defining the product market or ascertaining unilateral effects within that market.

8.12  Substitutability played a key role in the Facebook/WhatsApp merger. The European Commission, according to several officials, recognized ‘how privacy could be (p. 131) regarded as a non-price parameter of competition which may be degraded by the merged entity post-merger’.13 The Commission also recognized that the degradation in privacy protection could infringe EU competition law (whether or not it also infringes the data protection rules).14 But this theory of harm is relevant only ‘where privacy is an important factor in the decision to purchase a product or service, i.e. a key parameter of competition’.15

8.13  This analysis would seem to make sense. It is consistent with how the agencies usually look at competition, which is to ask questions about substitutability. But the requirement that privacy be an ‘important’ factor in the decision to purchase or a ‘key’ parameter of competition would be too narrow, if construed to capture only mergers between two privacy-protective services, while ignoring mergers where one product is privacy-protective and the other is not (even though the merger would likely result in harm). Under the Commission’s reasoning, because Facebook Messenger and WhatsApp had different privacy protections, this, among other things, suggested that they were not close competitors.16 In reality, however, this analysis has flaws.

8.14  It does not follow that texting apps with substantially similar privacy policies compete more fiercely than ones with dissimilar privacy policies. For one thing, the most meaningful competitive distinction between the firms may be their privacy policies. If a customer is choosing between texting apps, she is not likely to say, ‘Oh Facebook has crummy privacy protections. Let me look at other apps that also have inferior privacy protections.’ Firms offer an array of privacy choices, but the fact that they offer privacy protection at all suggests that they realize that consumers value privacy. From an economic standpoint, privacy is a dimension on which they are competing, whether they offer a lot of protection for the data or a little.17

8.15  In choosing a texting app, the primary consideration, given the network effects, is whether their friends, family, and acquaintances use the app. WhatsApp and Facebook are the leading texting apps. The Commission noted how consumer communications apps featured a significant degree of ‘multi-homing’, that is, users who installed, and use, on the same handset several consumer communications apps.18 In particular, WhatsApp and Facebook Messenger were the two main consumer communications apps simultaneously used by the majority of the users in (p. 132) the EU. The interesting dynamic is that a very high percentage of WhatsApp users were already using Facebook’s social network. Facebook members could easily use Facebook Messenger. Facebook Messenger is integrated in the company’s social network. Moreover, the texting apps offered similar functionalities.19 Nonetheless, many Facebook social network subscribers opted for a texting app that afforded them significantly greater privacy protection than Facebook Messenger.20

8.16  So the difference in privacy protections, rather than suggest that the texting apps were complements, could easily reflect a critical competitive difference. Users who want to join a social network join the largest one, Facebook, because that is where their friends, family, and acquaintances likely are. But for texting apps, they had a choice between two equally sized competitors but with different privacy protections. Facebook Messenger and WhatsApp texting apps were not complements, in the sense that an increase in demand in one app increases the demand of the other (such as peanut-butter and jelly). Instead, Facebook sought users who spend more time on its texting app Messenger than WhatsApp. WhatsApp, to induce Facebook social network users to switch from Messenger, offered greater privacy protections.

8.17  Three Commission officials in their post-merger analysis of Facebook/WhatsApp offered a slightly different explanation:

… the Commission regarded privacy as one of many parameters of competition between consumer communications apps, the others being price, reliability of the service, functionalities offered, size of the underlying network, trendiness, etc. While the Commission recognised that an increasing number of users value privacy and security (as shown by the growing popularity of apps that offer increased security of communications, like Threema or Telegram), the majority of the consumer communications apps currently on the market do not compete (mainly) on privacy features (e.g. Facebook, Skype, WeChat, Line, etc.).21

8.18  The second argument, which the three Commission officials advanced, is that the majority of the consumer communications apps on the market did not compete (mainly) on privacy features. But this is perhaps too one-dimensional. Facebook’s privacy policy may have offered too little protection to consumers—a result of what Professor Farrell has called a ‘dysfunctional equilibrium’:

If firms perceive that few consumers shift their demand in response to actual privacy policies, then the firm’s incentives are to make its policy noncommittal and/or non-protective. … It would then be tempting to design disclosures so as not to (p. 133) really communicate the choice of policy, if it is possible to obfuscate for the minority of consumers while retaining the ability to claim that the policy was disclosed. Meanwhile, if consumers perceive that firms behave in this kind of way, they will not expect attentive reading of privacy policies to be a rewarding activity. These patterns of conduct and expectations would reinforce each other, which is what makes them a game-theoretic or economic equilibrium.

It is often difficult to escape a dysfunctional equilibrium if there are large numbers of players involved. A consumer can’t simply decide to start reading privacy policies—or rather, she can, but it won’t do a lot of good, since firms will still expect that few consumers do so, so the consumer is apt either to learn little (noncommittal or vague policies) or be confirmed in her wary cynicism (policies that reveal a lack of protection—the rational choice for the firm when it expects that few consumers’ behavior will be affected). A small firm, likewise, can’t simply decide to break out of the equilibrium by adopting more protective policies and clearer disclosures, because its demand won’t shift by much. … Escape from a dysfunctional equilibrium often, and probably here, requires action by large and powerful players, and/or concerted action by groups of players.22

8.19  Here, WhatsApp may have been the maverick seeking to disrupt this ‘dysfunctional equilibrium’ that benefitted the industry leader, Facebook, and the smaller texting apps, all of which opted for a revenue model that values the needs of behavioural advertising over users’ privacy needs. And, at least before the merger, WhatsApp may have been succeeding, at least based on evidence of its popularity.

8.20  Competition law reaches this theory of harm. The agencies consider whether a merger may lessen competition by eliminating a ‘maverick’ firm, ie, ‘a firm that plays a disruptive role in the market to the benefit of customers’.23 One traditional antitrust concern is when ‘one of the merging firms has a strong incumbency position’ (ie, Facebook) and the other merging firm ‘threatens to disrupt market conditions with a new … business model’.24 Here, WhatsApp sought to shake things up with a different business model that did not harvest users’ personal data to target them with behavioural ads. WhatsApp offered a different business model that strongly signalled how protecting users’ data and privacy was in its corporate DNA. This theory of harm is within antitrust’s framework.25

8.21  It seems, instead, that the Commission relied on the conventional antitrust theory that dissimilar products compete less fiercely. Products and/or services with nearly identical physical, technical, or privacy characteristics are lumped together as close competitors. Because the texting apps had different business models and privacy (p. 134) policies, this suggested that they were not close competitors. Although this heuristic may work with other differentiated goods (such as premium versus discount shampoos), its shortcomings are revealed where a key parameter of competition is privacy protection. Instead, the competition agency should consider whether the incumbent is seeking to eliminate a maverick that offers greater privacy protections, and thus the merger would likely lessen competition and harm consumers.

C.  Substitutability of Data

8.22  Competition policy deals often with the cross-elasticity of demand, which ‘measures the responsiveness of the demand for one product to changes in the price of a different product’.26 If the price of one increased, all else remaining equal, would a sufficient number of consumers switch to the other product to defeat the price increase? Indeed, the agencies and courts determine the boundaries of a product and geographic market by the reasonable interchangeability or cross-elasticity of demand.27 Reasonable interchangeability and cross-elasticity of demand also help ‘evaluate the extent competition constrains market power and are, therefore, indirect measurements of a firm’s market power’.28 As the European Commission states, ‘[t]‌he higher the degree of substitutability between the merging firms’ products, the more likely it is that the merging firms will raise prices significantly.’29

8.23  Often the debate is whether products or services outside the proposed market are interchangeable with products inside the market (such as should the market for English muffins include bagels). Fatal to an antitrust claim is when the market is defined so broadly to include products that are not substitutable. As one example, consider what happened to the plaintiffs in Golden Gate Pharmacy Services, Inc v Pfizer, Inc, where the antitrust complaint was dismissed for failing to sufficiently allege a relevant product market.30 The plaintiffs alleged a relevant product market of ‘the pharmaceutical industry’, which included the ‘manufacture, sale, and innovation of all pharmaceutical products, prescription pharmaceutical products, non-prescription pharmaceutical products, brand name pharmaceutical products and particular pharmaceutical products and therapies specifically noted and identified (p. 135) by Pfizer and Wyeth in their annual reports’.31 The court dismissed the complaint because not all pharmaceutical products are interchangeable for the same purpose. ‘The failure to allege a product market consisting of reasonably interchangeable goods’ rendered the complaint ‘facially unsustainable’.32

8.24  At times, the competition agencies define the market as a cluster of products and services offered by the merging parties.33 With these products markets, the market reality is that companies to compete effectively must offer all or nearly all types of products and services in that cluster.34 No one contends that component products and services of the ‘cluster’ are interchangeable with each another. Even though the components are not interchangeable, the clusters themselves are.35

8.25  Consequently, substitutability plays a key role in competition law. Issues of market definition and market power often hinge on the degree of interchangeability, which ‘implies that one product is roughly equivalent to another for the use to which it is put; while there might be some degree of preference for the one over the other, either would work effectively’.36 As the FTC observed, the ‘identification of substitutes is at the core of product market definition’.37 And the primary purpose of defining a relevant market is ‘to facilitate the analysis of competitive effects of a transaction’.38

8.26  Data-driven mergers may involve issues of substitutability of the free products, such as the extent to which Waze’s navigation app was an effective substitute for Google Maps. Data-driven mergers or joint ventures, at times, involve issues of substitutability of the underlying data. Google and Waze, for example, both collected users’ geo-location data necessary in improving their mapping technologies. Microsoft and Yahoo! in their joint venture sought to combine data on their users’ search behaviour to improve the quality of the search results. Here we see how issues of substitutability can apply to both the free products and the underlying data collected.

(p. 136) 8.27  But one of the four critical ‘V’s of Big Data is the variety of data. Google in acquiring Nest Labs was not necessarily seeking data that was substitutable with the voluminous personal data that it already amassed through its search engine (Google), video platform (YouTube), browser (Chrome), mapping applications (Google Maps and Waze), email service (Gmail), and operating system for mobile devices (Android). The value in acquiring Nest Labs was not getting more of the same type of personal data Google already had, but rather other data, such as our offline behaviour in our homes, which Google in turn could incorporate in its existing free services and to increase advertising revenue.

8.28  Nor does it make sense, as we saw in Chapter 6’s discussion of Facebook/WhatsApp to treat all data as fungible, and roughly calculate how much data each company controls. So the real issue in data-driven mergers like Google/Nest Labs is not whether Google’s data is substitutable with the data collected from the thermostats and surveillance cameras. Instead, the issue is whether the merger, in enabling the firm to amass a greater variety of data, may likely increase the firm’s ability to maintain its dominant position, to leverage its dominance into another market, or to otherwise lessen competition in any side of the multi-sided market. Would a combined Google/Nest Labs reduce actual or potential rivals’ ability or incentive to compete, which in turn would reduce the competitive pressure on the merged entity? Suppose Google changes the privacy policy and starts using the data from Nest Labs’ home security cameras for its other operations. Google can better see what users are doing away from their computer, smartphones, or tablets. Google, with this data, could refine its user profiles for behavioural advertising, helping maintain its dominance in search advertising.

8.29  Likewise, suppose Facebook acquired a dominant health insurer and could, in that country, use the health data for behavioural advertising. Again, Facebook’s cluster of products and services are not a substitute to the products or services offered by the health insurance companies. Nor are Facebook’s and the health insurer’s databases substitutes. Nonetheless, the merger may substantially lessen competition by increasing entry barriers for any other platform seeking to provide an advertising-supported social network. It would lack many users’ medical history information, which is attractive to behavioural advertisers. From the user’s health records, texting, search inquiries, and interactions on its social network, Facebook has a superior view of the user’s interests, behaviour, medical concerns, and risks.

8.30  Thus courts generally require the agencies to provide economic evidence that the merging parties’ products are ‘good substitutes’ and therefore in the same market.39 Data-driven mergers at times defy this logic. The value of the target firm’s data lies in its variety and non-substitutability with the data that the acquirer already (p. 137) possesses. Nonetheless, in acquiring the data, the firm can attain or maintain significant market power. Also the merger’s effect will not necessarily impact one particular side of the market or even one service: ‘As the data provided will typically be used for different purposes (e.g. across multi-sided markets), market concentration will need to be assessed in most cases across all sides of the market.’40 In effect, data-driven mergers may enable a super-platform, such as Google’s, to control key portals of data, which helps it attain or maintain its power across many products.

D.  Defining a New Category

8.31  Nothing in the US or EU competition law requires that the merging parties’ products or services be substitutes. The US Clayton Act requires sufficient evidence that the merger may substantially lessen competition or tend to create a monopoly. The Commission must assess whether or not a concentration would significantly impede effective competition, in particular as a result of the creation or strengthening of a dominant position in the common market or a substantial part of it.41 One analogy, perhaps, is the foreclosure theory used to challenge vertical mergers, whereby ‘actual or potential rivals’ access to supplies or markets is hampered or eliminated as a result of the merger, thereby reducing these companies’ ability and/or incentive to compete’.42

8.32  As one US court stated, ‘[t]‌he primary vice of a vertical merger or other arrangement tying a customer to a supplier is that, by foreclosing the competitors of either party from a segment of the market otherwise open to them, the arrangement may act as a “clog on competition” which “deprive[s] … rivals of a fair opportunity to compete”.’43 Another US court provided this helpful road map:

In a Section 7 [of the Clayton Act] analysis of vertical mergers, the competitive significance of a vertical merger results primarily from the degree, if any, to which it may increase barriers to entry into the market or reduce competition by (1) foreclosing competitors of the purchasing firm in the merger from access to a potential source of supply, or from access on competitive terms, (2) foreclosing competitors of the selling firm from access to the market or a substantial portion of it, or (3) forcing actual or potential competitors to enter or continue in the market only on a vertically integrated basis because of advantages unrelated to economies attributable solely to integration.44

(p. 138) A vertical merger’s competitive significance results primarily from the degree, if any, to which it may reduce competition by increasing entry barriers or rivals’ costs, foreclosing suppliers, customers, or inputs, or facilitate collusion.

8.33  One example involved Google’s acquisition of ITA Software, Inc.45 The vertical merger affected two markets: the downstream comparative flight search service market, which enables consumers to search online for flight prices, schedules, and seat availability on multiple airlines simultaneously (think Kayak or Orbitz), and the upstream market for airfare pricing and shopping systems which involve ‘a continuously-updated database of airline pricing, schedule and seat availability information, and a software algorithm used to search the database for flight options that best match consumers’ search criteria’ (think a search engine for flight data).46 Google was entering the comparative flight service market. ITA was the most popular airfare pricing and shopping system, given the speed at which its algorithms could find seat availability. The concern was that post-merger Google could foreclose or disadvantage its downstream rivals such as Orbitz, Kayak, and Microsoft’s Bing Travel. Google could refuse to renew existing ITA software contracts, refuse to enter into new ITA software contracts, enter into software contracts on less favourable terms than ITA would have, or degrade the speed or quality of ITA’s services offered to Google’s rivals.47 The DOJ also alleged that the merger would raise entry barriers in the downstream market, and that vertical expansion upstream was unlikely. Indeed, ‘Google looked at developing its own [airfare pricing and shopping] system as an alternative to acquiring ITA, but concluded it would take several years and require numerous engineers due to the complexity of the algorithms.’48

8.34  The DOJ’s consent decree required Google post-merger to continue licensing—and improve—ITA’s software, license a new flight search technology that ITA was developing, and erect a firewall at its business regarding the use of competitively sensitive information gained through ITA’s services.49 Access to airline seat and booking class data was a critical input to an airfare pricing and shopping system. To ensure that Google did not restrict access to this crucial data, the consent decree also limited Google from agreeing with an airline to restrict the airline’s right to share seat and booking class data with Google’s downstream rivals (unless an airline entered into exclusive agreement with a competitor). Google also had to make available to its downstream rivals any seat and booking class information it obtained for use in its new flight search service.

(p. 139) 8.35  Here the DOJ got the theory right. Moreover, as in Bazaarvoice, the DOJ recognized how data was a critical input for the algorithms. But its decree was incomplete. Google’s rivals feared that after purchasing the travel search engine, Google would use ITA data to feed details of specific flights directly to users of its search engine, rather than pointing them to Expedia’s and other competitors’ web pages.50 And that is what happened.

8.36  In the Google/ITA merger, Google was not yet in the market. Its comparative flight search service was forthcoming. One problem is that the vertical foreclosure theory generally presupposes an existing (or imminent) buyer–seller relationship between the acquirer (and its competitors) and the acquired (and its competitors).51

8.37  Another possible problem, mainly in the US, is that agencies’ guidelines for vertical mergers are dated. The US competition agencies, while updating the horizontal merger guidelines five times,52 have not updated their vertical mergers guidelines since 1984.53 The US agencies infrequently rely on them. Moreover, the last time they litigated a vertical merger was in 1979.54 One reason is that there is not the same level of consensus on vertical mergers as horizontal mergers. As one US court stated, ‘respected scholars question the anticompetitive effects of vertical mergers in general’.55 Among the scholars cited were Robert Bork (‘Antitrust’s concern with vertical mergers is mistaken. Vertical mergers are means of creating efficiency, not of injuring competition. … [The] foreclosure theory is not merely wrong, it is irrelevant.’56) and Herbert Hovenkamp (‘Of all mergers, vertical acquisitions are the most likely to produce efficiencies and the least likely to enhance the market power of the merging firms.’57).

8.38  Nonetheless, as Professors Michael Riordan and Steven Salop observed 20 years ago, many economists have moved beyond the Chicago School; in ‘apply[ing] the newer methodology of modern industrial organization theory to more (p. 140) realistic market structures’, they find that vertical mergers can have anticompetitive effects.58 Likewise, the competition agencies’ theories have evolved, incorporating post-Chicago School theories. Two antitrust practitioners noted that the agencies’ consent decrees focus on

(1) whether the post-merger entity will have the incentive to foreclose downstream competitors from obtaining key inputs by, inter alia, raising those competitors’ costs of obtaining such inputs—an inquiry to which barriers of entry are not only relevant but critical; (2) whether the merger will result in customer foreclosure; and/or (3) whether the merger will facilitate collusion in the upstream or the downstream product market.59

8.39  Moreover, in considering whether the merger may substantially lessen competition, the competition agency can consider the increased risk of other anticompetitive behaviour, such as ‘monopoly leveraging’.60 Alternatively, the merger can be seen as monopoly maintenance (if it helps insulates the product with market power) or attempted monopolization (if it affects the adjacent market).

8.40  Nonetheless, some courts and competition agencies may be reluctant to condemn data-driven mergers under a vertical foreclosure theory. This is especially so when ‘data’ appears to be ubiquitous and when the competition authorities have not developed the tools to assess these types of anticompetitive risks.


1  European Commission, Guidelines on the Assessment of Non-Horizontal Mergers Under the Council Regulation on the Control of Concentrations Between Undertakings [2008] OJ C 265/07, para 3 (‘EC Non-Horizontal Merger Guidelines’), http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:C:2008:265:0006:0025:en:PDF (‘Two broad types of non-horizontal mergers can be distinguished: vertical mergers and conglomerate mergers.’).

2  Ibid, para 11 (‘Non-horizontal mergers are generally less likely to significantly impede effective competition than horizontal mergers.’).

3  Ibid, para 92 (acknowledging ‘that conglomerate mergers in the majority of circumstances will not lead to any competition problems,’ but ‘in certain specific cases there may be harm to competition’); Deborah Platt Majoris, ‘Merger Enforcement at the Antitrust Division’, Speech given at KPMG/Chicago School of Business Mergers and Acquisitions Forum, Chicago, 27 September 2002, http://www.justice.gov/atr/speech/merger-enforcement-antitrust-division-0 (‘[A]‌ntitrust should rarely, if ever, interfere with any conglomerate merger. Under only the rarest conditions will a conglomerate merger, unlike a horizontal or vertical merger, likely give the merged firm the ability and incentive to raise price and restrict output. Conversely, conglomerate mergers have the potential as a class to generate significant efficiencies.’).

4  EC Non-Horizontal Merger Guidelines, above note 1, para 91.

5  Eleonora Ocello, Cristina Sjödin, and Anatoly Subočs, ‘What’s Up with Merger Control in the Digital Sector? Lessons from the Facebook/WhatsApp EU Merger Case’, 1 Competition Merger Brief (February 2015): p 6.

6  European Commission, ‘Form CO Relating to the Notification of a Concentration Pursuant to Regulation (EC) No 139/2004’, at EU Competition Law: Rules Applicable to Merger Control, Situation as at 1 April 2010, p 59 (‘Form CO’), http://ec.europa.eu/competition/mergers/legislation/merger_compilation.pdf.

7  US Department of Justice (DOJ) and Federal Trade Commission (FTC), Horizontal Merger Guidelines, 19 August 2010, s 6.1 (‘US Horizontal Merger Guidelines’), http://www.justice.gov/atr/public/guidelines/hmg-2010.html.

8  United States v Interstate Bakeries Corp & Continental Baking Co, Civil Action No 95C 4194, 1995 WL 803599 (US Dist Ct (ND Ill), 1995).

9  Ibid, Competitive Impact Statement, 21 July 1995, http://www.justice.gov/atr/case-document/competitive-impact-statement-132.

10  European Commission, Guidelines on the Assessment of Horizontal Mergers Under the Council Regulation on Control of Concentrations Between Undertakings [2004] OJ C 31/03, para 28 (‘EC Horizontal Merger Guidelines’), http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52004XC0205%2802%29.

11  United States v Kimberly-Clark Corp & Scott Paper Co, Case No 3:95 CV 3055-P (US Dist Ct (ND Tex), 1995), Complaint, paras 36–7, http://www.justice.gov/atr/case-document/complaint-141.

12  FTC, ‘FTC to Challenge Nestle, Dreyer’s Merger: Agency Will Allege $2.8 Billion Ice Cream Deal Violates Antitrust Laws’, Press Release, 4 March 2003, https://www.ftc.gov/news-events/press-releases/2003/03/ftc-challenge-nestle-dreyers-merger.

13  Ocello et al, above note 5, p 6.

14  Ibid.

15  Ibid.

16  Facebook/WhatsApp (Case Comp/M.7217), Commission Decision C(2014) 7239 final, 3 October 2014, paras 102, 107.

17  Keith Waehrer, ‘Online Services and the Analysis of Competitive Merger Effects in Privacy Protections and Other Quality Dimensions’, Working Paper, 12 January 2016, http://waehrer.net/Merger%20effects%20in%20privacy%20protections.pdf.

18  Facebook/WhatsApp, above note 16, para 105.

19  Ibid, para 104.

20  Ibid, para 105. Between December 2013 and April 2014, ‘between [20–30]% and [50–60]% of WhatsApp users already used Facebook Messenger and between [70–80]% and [80–90]% of WhatsApp users were Facebook users and were therefore already within the reach of Facebook Messenger. Conversely, over the same period 60% to 70% of Facebook Messenger active users already used WhatsApp.’ Ibid, para 140.

21  Ocello et al, above note 5, p 6.

22  Joseph Farrell, ‘Can Privacy Be Just Another Good?’, 10 J on Telecomm & High Tech L (2012): pp 251, 258–9.

23  US Horizontal Merger Guidelines, above note 7, s 2.1.5.

24  Ibid.

25  Ibid, s 6.4.

26  US Horticultural Supply v Scotts Co, 367 F App’x 305, 309 (US Ct of Apps (3rd Cir), 2010); EC Horizontal Merger Guidelines, above note 10, para 29 n 38.

27  Form CO, above note 6, p 39 (market definition depends on ‘substitutability, conditions of competition, prices, cross-price elasticity of demand or other factors relevant for the definition of the product markets (for example, supply-side substitutability in appropriate cases’)); United States v Archer-Daniels-Midland Co, 866 F2d 242, 246 (US Ct of Apps (8th Cir), 1988).

28  Archer-Daniels-Midland Co, above note 27, p 246.

29  EC Horizontal Merger Guidelines, above note 10, para 28.

30  Golden Gate Pharmacy Services, Inc v Pfizer, Inc, 433 F App’x 598 (US Ct of Apps (9th Cir), 2011).

31  Ibid, p 599.

32  Ibid.

33  United States v Philadelphia National Bank, 374 US 321, 356–7, 83 S Ct 1715, 1736–8 (1963) (relevant product market consisting of various kinds of credit products and services (eg checking accounts and trust administration) offered by commercial banks); Federal Trade Commission v Sysco Corp, Case No 1:15-CV-00256 (APM), 2015 WL 3958568, p *10, 12 (US Dist Ct (D DC), 23 June 2015) (FTC alleging, and the court accepting, that the relevant product market was ‘broadline foodservice distribution’; what was relevant for the court’s consideration was not any particular food item sold or delivered by the merging firms, ‘but the full panoply of products and services offered by them that customers recognize as “broadline distribution’ ”).

34  United States v Grinnell Corp, 384 US 563, 572, 86 S Ct 1698, 1703 (1966) (‘We see no barrier to combining in a single market a number of different products or services where that combination reflects commercial realities.’).

35  Philadelphia National Bank, above note 33.

36  US Horticultural Supply, above note 26, p 309, quoting Allen-Myland, Inc v International Business Machines Corp, 33 F3d 194, 206 (US Ct of Apps (3rd Cir), 1994).

37  In re Promedica Health System, Inc, 2012 WL 1155375 (FTC, 22 March 2012).

38  Ibid.

39  DSM Desotech Inc v 3D Systems Corp, 749 F3d 1332, 1340 (US Ct of Apps (Fed Cir), 2014).

40  OECD, Data-Driven Innovation for Growth and Well-Being: Interim Synthesis Report, October 2014, p 59, http://www.oecd.org/sti/inno/data-driven-innovation-interim-synthesis.pdf.

41  EC Non-Horizontal Merger Guidelines, above note 1, para 1.

42  Ibid, para 18.

43  Geneva Pharmaceuticals Technology Corp v Barr Labs Inc, 386 F3d 485, 511 (US Ct of Apps (2d Cir), 2004) (internal quotation marks omitted).

44  Yankees Entertainment & Sports Network, LLC v Cablevision Systems Corp, 224 F Supp 2d 657, 673 (Dist Ct (SD NY), 2002); Fruehauf Corp v FTC, 603 F2d 345, 352 (US Ct of Apps (2d Cir), 1979).

45  United States v Google Inc, Case No 1:11-cv-00688 (US Dist Ct (D DC) 2011), Complaint filed 8 April 2011 (‘Google Complaint’), http://www.justice.gov/file/497686/download.

46  Ibid, para 24.

47  Ibid, para 38.

48  Ibid, para 42.

49  United States v Google Inc, Case No 1:11-cv-00688 (US Dist Ct (D DC) 2011), Competitive Impact Statement filed 8 April 2011, http://www.justice.gov/file/497671/download.

50  Richard Waters, ‘How Diller’s Google Prediction Came True’, Financial Times, 15 April 2015, http://www.ft.com/cms/s/0/121f171e-e2f8-11e4-aa1d-00144feab7de.html.

51  Brown Shoe Co v United States, 370 US 294, 323, 82 S Ct 1502, 1522 (1962) (‘Economic arrangements between companies standing in a supplier-customer relationship are characterized as “vertical”.’); EC Non-Horizontal Merger Guidelines, above note 1, para 4 (‘[v]‌ertical mergers involve companies operating at different levels of the supply chain’).

52  DOJ, Merger Enforcement, http://www.justice.gov/atr/merger-enforcement.

53  DOJ, Non-Horizontal Merger Guidelines, 14 June 1984, http://www.justice.gov/atr/non-horizontal-merger-guidelines.

54  James A Keyte and Kenneth B Schwartz, ‘Getting Vertical Mergers Through the Agencies: “Let’s Make A Deal” ’, Antitrust (Summer 2015): p 10, citing Fruehauf Corp v FTC, 603 F2d 345 (US Ct of Apps (2d Cir), 1979).

55  Alberta Gas Chemicals Ltd v EI du Pont de Nemours & Co, 826 F2d 1235, 1244 (US Ct of Apps (3rd Cir), 1987).

56  Robert Bork, The Antitrust Paradox (New York: Free Press, 1978), pp 226, 237.

57  Herbert Hovenkamp, ‘Merger Actions for Damages’, 35 Hastings LJ (1984): pp 937, 961.

58  Michael H Riordan and Steven C Salop, ‘Evaluating Vertical Mergers: A Post-Chicago Approach’, 63 Antitrust LJ (1995): pp 513, 515.

59  Keyte and Schwartz, above note 54, p 13.

60  Monopoly leveraging ‘involves a firm using its market power in one market to gain more market share in another market, other than by competitive means’. AstroTel, Inc v Verizon Florida, LLC, Case No 8:11-CV-2224-T-33TBM, 2012 WL 1581596, p *5 (US Dist Ct (MD Fla), 4 May 2012); see also Verizon Commc’ns Inc v Law Offices of Curtis V Trinko, LLP, 540 US 398, 415, 124 S Ct 872, 883 (2004) (‘monopoly leveraging’ theory requires a ‘dangerous probability of success’ in monopolizing a second market and anticompetitive conduct). In Europe, a firm with a dominant position on one market cannot engage in anticompetitive practices on a second market, even if the company does not enjoy a dominant position on the second market. Tetra Pak v Commission [1994] ECR II-755, [1997] 4 CMLR 726 (CFI), aff’d [1996] ECR I-5951, [1997] 4 CMLR 662 (ECJ); Scott M Kareff, ‘Tetra Pak International SA v. Commission (Tetra Pak II): The European Approach to Monopoly Leveraging’, 28 Law & Pol’y Int’l Bus (1997): pp 549, 550.