140313_icomp_innovation_final
Transcription
140313_icomp_innovation_final
Case 39.740 – Google Submission to DG Comp on behalf of ICOMP 14 March 2014 ANALYTICAL FRAMEWORK FOR THE ASSESMENT OF INNOVATION EFFICIENCIES IN CASE 39.740 – Google I. INTRODUCTORY CONSIDERATIONS 1. This paper sets up an analytical framework for the assessment of innovation in Case COMP/39.740 – Google. We focus on one of the four concerns highlighted by Vice President Almunia in May 20121 in relation to the conduct of Google Inc. (“Google”): discrimination in vertical search or “search bias”. However, some of the conclusions reached in this paper are equally applicable to other conduct by Google. 2. Innovation and Article 102 TFEU is a vast topic. We narrow it down by reference to a number of findings that the European Commission (the “Commission”) has made in the course of the Google proceedings.2 i. Google is “dominant in the European Economic Area (“EEA”) both in web search and search advertising”, with a market share “well above 90% in web search in most European countries for several years now”. ii. There are “significant barriers to entry and network effects in both markets”. iii. Google’s exclusionary conduct has resulted in harm to competition to the detriment of consumers. According to the Commission, the conduct has led to “harm to consumers in terms of reduced choice, quality and innovation”. 3. Against this background, we set out an analytical framework for review of the interaction between innovation and the existence of a dominant position and the interaction between innovation and abuse. In our view, correctly ascertaining the importance of innovation requires analysis of the counterfactual, i.e., how much innovation would have taken place absent Google’s practices. The Commission does not appear yet to have conducted this analysis in the Google investigation. 1 See Almunia, J., “Statement of VP Almunia on the Google antitrust investigation”, SPEECH /12/372, European Commission, 21 May 2012. 2 See “Antitrust: Commission seeks feedback on commitments offered by Google to address competition concerns”, IP/13/371, European Commission, 25 April 2013 and “Commission seeks feedback on commitments offered by Google to address competition concerns – questions and answers”, MEMO/13/383, European Commission, 25 April 2013. II. INNOVATION AND MARKET POWER 4. We are not aware of any authors indicating or suggesting that the “new economy” cannot lead to the creation of dominant positions. That said, according to some authors, dominant positions in the co-called “new economy” are often temporary and fragile.3 In the words of Evans and Padilla: “[e]quating high market shares with dominance in the case of these “fragile monopolists” of the new economy is potentially very damaging. […] A better test of market power is contestability. If the market is contestable […] a firm with a high market share does not enjoy a position of dominance because potential entry imposes an effective constraint on its conduct”.4 5. Google’s position in the markets for search and search advertising cannot be characterised as “contestable” within any economically meaningful definition of this word.5 According to Shapiro, “[s]ales are contestable in the relevant sense if profitable sales shift toward the successful innovator”.6 Shares over 90%,7 in markets featuring 3 See, e.g., Schmalensee, R., “Antitrust Issues in Schumpeterian Industries”, AEA Papers and Proceedings, 2000, at p. 193. For a more recent re-statement, in the same vein see Teece, D., J., “The Nature of Competition in Regimes of Rapid Technological Change”, in Teece, D. J., Dynamic Capabilities & Strategic Management. Organizing for Innovation and Growth, Oxford, 2011 (“Teece 2011”), at pp. 233 ff. 4 See Ahlborn, C., Evans, D., and Padilla, “Competition Policy in the New Economy: Is Competition Law Up to the Challenge”, ECLR, at pp. 156 and 162. See further Almunia, J., “Abuse of dominance: A view from the EU”, SPEECH/13/758, Fordham Competition Law Institute Annual Conference, 27 September 2013 (“Almunia’s Fordham Speech”), according to whom “some claim that there is no need for antitrust intervention in high tech markets. […] The argument is that it is impossible for a company to become dominant – and to stay dominant in sectors where new products, platforms and services appear all the time. I’m not convinced by this argument. In fact, owing to some specific features of these markets, it can actually be easier for a company to hold a dominant position over time […]” (Vice President Almunia subsequently referred, in this regard, to network effects and switching costs) or Italianer, A., “Competition Policy in a Digital Age”, Innsbruck 47th Symposium, 7 March 2014. 5 The theory of “contestable markets” places the main emphasis on freedom of entry to, and exit from, a market. The minimum conditions for a contestable market to exist are an almost instantaneous entry and costless exit (see Baumol, W., J., Panzar, J., and Willig, R., Contestable Markets and the Theory of Industry Structure, Harcourt Brace Jovanovich, 1982; and Bishop, S., and Walker, M., The Economics of EC Competition Law: Concepts, Application and Measurement, Sweet & Maxwell, 2010, at para. 3.021). 6 See Shapiro, C., “Competition and Innovation. Did Arrow Hit the Bulls’s Eye”?, at Lerner, J., and Stern, S., (eds.), The Rate and Direction of Inventive Activity Revisited, University of Chicago Press, 2012 (“Shapiro 2012”), at p. 364. 7 In the words of Shapiro: “[w]e place less weight on [concentration as a proxy for lack of competition] than we did fifty years ago, but it certainly still has value, at least in properly defined markets. The recently revised [US 2010] Horizontal Merger Guidelines continue to use Herfindahl-Hirschman Index (HHI) thresholds with adverse competitive effects […] presumed likely for mergers that raise the HHI more than 200 and lead to a post-merger HHI greater than 2,500” (see Shapiro 2012, at p. 375). Moreover, by definition, these quasi monopolistic market shares would result in a big difference between the “market position of the dominant undertaking and its competitors” (see Enforcement Priorities, at para. 12). 2 high barriers to entry and network effects, during prolonged periods of time8 raise a strong presumption that the undertaking holding them is dominant in the affected markets9. The importance of network effects in these markets cannot be overstated. According to Hovenkamp, “one strong feature of many networks is path dependence, which means that once a format or technology is adopted and attains widespread acceptance, anyone offering an alternative technology or format faces significant market resistance”.10 The markets for search and search advertising are characterised not only by network effects11 but also by returns to scale, since a larger pool of users and their click data is necessary to enable search engines to provide relevant search results (both organic and paid). The ability to provide more relevant search results, in turn, attracts more users and increases the prices that the search engine can charge for search ads. Both network effects and returns to scale act as significant barriers to entry that deter competition against Google’s dominance in search. 8 See, e.g., Enforcement Priorities, at para. 20, according to which one of the factors that the Commission takes into consideration in a finding of abuse is, “if the conduct has been in place for a sufficient period of time, the market performance of the dominant undertaking and its competitors”. 9 In Hoffman-La Roche the Court of Justice of the European Union (“CJEU”) clearly established that “although the importance of market shares may vary from one market to another, the view may legitimately be taken that very large shares are in themselves, and save in exceptional circumstances, evidence of the existence of a dominant position” (see e Case 75/76 Hoffman-La Roche v Commission [1979] ECR 461, at para. 41). In Akzo v. Commission, the CJEU referred to this passage from Hoffmann-La Roche and further clarified that “[t]hat is the situation where there is a market share of 50% such as that found to exist in this case” (see Case C-62/86 Akzo v Commission [1991], at para. 86). Indeed, in the AstraZeneca Case before the General Court, the General Court went as far as to indicate that the Commission could not disregard the importance to be attached to AstraZeneca’s very large market share throughout the relevant period of alleged abuse (see Case T-321/07 AstraZeneca ECR II-000 [2010], at para. 245). 10 See Hovenkamp, H., The Antitrust Enterprise. Principle and Execution, Harvard University Press, 2005 (“Hovenkamp 2005”), at pp. 279-280. See, further, e.g., Besen, S. M. and Farrell, J. “Choosing How to Compete: Strategies and Tactics in Standardization”, Journal of Economic Perspectives, Spring 1994, at p. 117; Farrell, J., and Saloner, G., “Installed Base and Compatibility: Innovation, product Preannouncement, and Predation”, 76 American Economic Review 940 and Katz, M. L., and Shapiro, C., “Technology Adoption in the Presence of Network Externalities”, 1986, 94 Journal of Political Economy 822. Similarly, Professor Teece emphasises that “one should expect path dependencies when there are increased returns of some kind. This will specially be true for […] information goods […]” (see Teece 2011, at p. 245). 11 Google’s dominance in search leads website owners to configure their websites in a way which maximizes the user-friendliness vis-à-vis Google’s web crawler technology and Google’s search ranking algorithms, which they do in order to increase their ranking in Google’s organic search results. This activity reinforces Google’s dominance, which in turn increases the value to websites of “optimizing” to Google’s web crawlers and search ranking algorithms. Likewise, Google’s dominance in search advertising leads online advertisers to adopt Google’s search advertising platform in order to maximize their exposure to potential customers. This activity reinforces Google’s dominance in search advertising, allowing it to charge higher prices. 3 6. Even accepting a theory according to which market shares can be downplayed as a proxy for market power in favour of the contestability of the market,12 the presence of barriers to entry would become key. In the Google case, the Commission has indicated that barriers to entry into the markets for search and search advertising are substantial and reinforced by network effects. Two key barriers to entry are: (i) the high costs of building and operating servers of sufficient scope to compete with existing horizontal search engines (costs that may well run in the tens or even hundreds of millions of Euros); and (ii) the costs of acquiring the minimum number of users needed to obtain the click data necessary to provide sufficiently relevant results. 7. The importance access to a sufficient body of users click data in order to provide relevant search results cannot be overstated. According to Shelanski, customer information is a “key input of production and a strategic competitive asset”. Consequently, “if a given consumer relies primarily on a single platform for […] searching the Internet, or any other service, then each such platform will probably have more information about consumers than its rivals will have. If a platform is the market leader, it will have more users than its rivals and, accordingly, more overall consumer information. This larger information set might enable the leading firm to make information-dependent product improvement that smaller rivals will be unable to replicate. […] depending on the point at which returns to additional customer information begin to diminish, large volumes of such competitive information can give a competitive advantage to a leading platform. The more modest the volume at which returns from additional information diminish, the more likely it is that multiple competing platforms can obtain the customer data they need to deliver a competitive product or service”. 8. Conversely, “if the benefits of additional information begin to decline only at a very high volume of information and if the relative differences in access to customer data matter”, .. “control of the largest share of customer data could contribute to market dominance”. These data, “may be a key commodity, exchanged between platforms and end users. Customers reveal information to platforms in exchange for services”.13 12 See, e.g., Teece 2011, at pp. 233 ff. 13 See Shelanski, H. A., “Information, Innovation, and Competition Policy for the Internet”, 161 U. Pa. L. Rev. 1663-1705, 2013 (“Shelanski 2013”), at pp. 1678, 1681 and 1687. 4 9. In any event, the following information is necessary for the purpose of assessing Google’s market power and any from its conduct: i. The point at which the returns deriving from gathering additional information diminish. ii. Who have been the genuine successful innovators in the affected markets and whether sales have shifted towards them. iii. The exact nature of barriers to entry into these markets and their contestability. iv. The exact market share of Google, the moment(s) when it increased, and the intertwining of these with the practices for which Google is being investigated. III. INNOVATION AND ABUSE A. General Considerations 10. According to Bork and Sidak, Google’s practices in search constitute a product improvement which “adds value to Google search from the perspective of both consumers and advertisers [whereas] Google’s critics have attempted to cast this innovation as a form of foreclosure”.14 11. This assessment overlooks two crucial questions: (i) whether Google’s search practices exclude competitors on the basis of a genuine product improvement and (ii) whether others, given the chance, would have delivered innovation that was substantially more valuable to consumers. More specifically, we understand that certain practices by Google such as algorithmic penalties have been clearly established by complainants.15 We understand that the Commission agrees that Google’s “favourable treatment” of its own services - that is, its practice of providing favourable ranking to Google’s own 14 See Bork, R. H., and Sidak, G., J., “What does the Chicago School Teach About Internet Search and the Antitrust Treatment of Google?”, report commissioned (as indicated at p. 8) by Google, 2012 (“Bork and Sidak”), at p. 1. Similarly, according a post by Google’s counsel: “[p]reventing Google from showing what it considers to be the most relevant, directly responsive results for a query would not promote competition. It would restrict competition and obstruct innovation” (see Graf, T., “The Google Commitments – Testing Substantive Theories Through Remedy Discussion”, 4 July 2013, available at http://kluwercompetitionlawblog.com/2013/07/04/the-google-commitments-testing-substantive-theoriesthrough-remedy-discussion/). 15 See, e.g., Bracha, O., and Pasquale, F., “Federal Search Commission? Access, Fairness and Accountability in the Law of Search”, 93 Cornell L. Rev., 2008, at pp. 1149, 1167, according to which “policy makers should at least consider restrictions on the ability of search engines to manipulate their results”. 5 services based on criteria other than relevance or other non-discriminatory factors - is not considered innovative or beneficial to the user 12. According to both the case-law of the CJEU and the Commission’s Guidance in this field, Article 102 TFEU considers as abusive, and prohibits, the conduct of a dominant undertaking consisting in acts which might potentially foreclose their competitors, thereby strengthening its position by using methods other than those which come within the scope of competition on the merits.16 13. According to Bork and Sidak, the two-sidedness of the market should prevent Google from providing a worse service in search, since it would lose customers in search advertising.17 However that assertion simply begs the question. In reality, Google has incentives to pursue exclusionary practices and run the risk that the outcome might be a degraded search experience for consumers. Furthermore, Google may degrade the consumer search experience in a variety of ways that are not obvious to consumers, e.g., by degrading the quality of organic search results in order to increase consumer clicks on paid search results or other results that Google has the ability to monetize. These incentives are unaffected by the two-sidedness of the market. According to Bork and Sidak, the latter would result in Google having to balance the increased returns which would derive from reducing its expenditure on innovation with the loss of revenues from advertisers resulting from fewer searches. However, the exact profit-maximising conduct for Google is heavily influenced by its super-dominance in search: given that would-be searchers have nowhere else to go (since competing search engines, as a result of Google’s exclusionary practices, do not have access to the volume of user clicks needed in order to provide equally relevant search results), search advertising revenue is 16 See, e.g., Case C-457/10 AstraZeneca AB and AstraZeneca plc v European Commission (“AstraZeneca”), judgment of 6 December 2012 [not yet reported], at para. 75 and Case C-202/07 P France Télécom v Commission [2009] ECR I-2369, at para. 106. See, finally, the Commission’s Guidance on its enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings, OJ C 45, 24.2.2009, p. 7–20, (the “Enforcement Priorities”), at para. 6, according to which “the emphasis of the Commission’s enforcement activity in relation to exclusionary conduct is […] ensuring that hold a dominant position do not exclude their competitors by other means than competing in the merits of the products and services they provide”. See further Rose, D., and Bailey, D., Bellamy & Child. European Union Law of Competition, Oxford 2012, at para. 10.053; González-Díaz, F. E., Snelders, R., Abuse of Dominance Under Article 102 TFEU, Claeys & Casteels, 2013, at paras. 3.4 and 3.5; O’Donoghue, R., and Padilla, J. A., The Law and Economics of Article 102 TFEU, Hart, 2013, at pp. 215 ff.; Faull, J., and Nikpay, A., The EC Law of Competition, Oxford, 2007, at para. 4.16; Wish, R., and Bailey D., Competition Law, Oxford, 2012, at pp. 198 ff. and Alison, J., and Sufrin, B., EU Competition Law, 2011, at pp. 378 ff. 17 See Bork and Sidak, at p. 10. 6 not much imperilled by lower quality. Consumers will continue to use Google given that Google benefits from enormous scale advantages based on the learning gained from previous search results and from the scale-driven network effects inherent to the search sector. According to Hovenkamp: “[t]he considerable advantages that accrue to an established market-dominating network create the possibility […] [for the undertaking] running the network [to] […] restrain innovation without causing excessive defections. The costs of being excluded from the network are simply too high”.18 14. In any event, there is no requirement to establish that the practice had actual anticompetitive effects: the establishment of likely or potential effects is sufficient for a finding of a breach of Article 102 TFEU.19 It is our understanding that complainants have submitted to the Commission abundant evidence of the existence of both these practices and of consumer harm.20 15. Consumer harm relevant for a finding of abuse obviously also includes harm to innovation.21 It should be highlighted, in this regard, that the Commission has indicated 18 See Hovenkamp 2005, at p. 281. In a similar vein, Professor Motta has indicated that exclusionary practices “should receive fresh attention”, because of two reasons, the second of which is the fact that “a growing share of today’s advanced economies is composed of sectors (for example computer software, Internet […]) that exhibit network and lock in effects. In such environments, entrants might find it very difficult to compete with incumbents, and particular attention should be paid to possible exclusionary practices” (see Motta, M., Competition Policy. Theory and Practice, Cambridge, 2004, “Motta 2004”, at p. 411). 19 See Case C-52/09 TeliaSonera Sverige [2011] ECR I-005277 (“TeliaSonera”), at para. 64; AstraZeneca, at para. 112 AstraZeneca, at para. 108. See further, a contrario, to Enforcement Priorities, at para. 20 (VI), according to which the “evidence of actual foreclosure” is only one of many elements which the European Commission takes into account when ascertaining whether there is an abuse. It should be noted that this is the case equally under Article 101 TFEU (see TeliaSonera, at para. 24. See further Case C-8/08 T-Mobile Netherlands BV and Others [2009] 2009 I-04529, at para. 38; Case C-501/06 GlaxoSmithKline Services and Others v Commission and Others ECR [2009] I-09291, at para. 63). 20 See, e.g., the Financial Times, according to which: “[f]ive out of every six items in the panels shown on a Google search made in America are more expensive than the same items from other merchants hidden deeper in the index, with an average premium of 34 per cent.” (see Waters, R., “Google’s showcased shopping found to come at a premium”, Financial Times, 24 November 2013). But the consumer harm deriving from Google’s practices does not only consist in increased prices. In the words of Shelanski: “[e]ven in the absence of any conventional practice or output effects, anticompetitive conduct or transactions could enable platforms to exercise market power to give consumers less of the good things – improved service, innovative products, and good privacy and data security policies – for which consumers might implicitly barter their information. […] any competitive effect analysis that nisses these two nonprice dimensions of platform market performance will be incomplete and could be biased toward underenforcement” (see Shelanski, at p. 1687). 21 Competition law speaks frequently in terms of price increases or restrictions of output as a shorthand for a variety of instances of consumer harm. See, e.g., the Commission’s Guidelines on Horizontal Mergers, which expressly equate a reduction of innovation to a price increase (see Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ C31, 5.2.2004, p.5, at para. 8). 7 its concerns that Google’s practices in search “unduly [divert] traffic away from Google's competitors in specialised search towards Google's own specialised search services. […] [and that] this may reduce competitors' incentives to innovate in specialised search”.22 Shelanski considers that digital platforms raise competitive concerns mainly “related to innovation and customer information that may warrant increased antitrust scrutiny of their activity”.23 Moreover, Vice President Almunia noted in January 2014 that “the prospects for economic growth are hampered if dominant firms are allowed to raise barriers to entry and prevent or constrain the innovations of rivals. The case that first comes to mind is our on-going investigation of Google”.24 16. Authors emphasising the importance of competition in innovation, such as Teece, stress the importance for antitrust enforcement to prevent practices which result in innovation losses.25 In the words of Shapiro, “exclusionary practices […], if not checked by antitrust law, can make current monopoly power more durable by deterring innovative entrants”.26 17. This analysis should take into account not only Google’s alleged innovations but also the innovation that would have taken place in the market absent the abuse. President Almunia put it: As Vice “antitrust enforcement should not prevent Google from improving its services if it wishes to do so, but should at the same time allow innovative products from rivals to be able to compete on their merits”.27 This should involve allowing sales to shift to the more innovative undertakings, which is exactly what Google is preventing with its conduct including in vertical search. We will set out this in more detail below, where we discuss the innovation counterfactual in the Google case. 18. As the General Court confirmed in the Microsoft case, an abuse becomes particularly harmful when it forecloses those competitors of the dominant undertaking which are 22 “Commission seeks feedback on commitments offered by Google to address competition concerns – questions and answers”, MEMO/13/383, European Commission, 25 April 2013. 23 See Shelanski 2013, at p. 1667. In fact, Shelanski (see p. 1663) expressly adopts a position according to which the social costs of antitrust under-enforcement on the Internet are “at least at high as those of overenforcement”. 24 See Almunia, J., “Competition policy for the post-crisis word: a perspective”, SPEECH/14/34, Bruges, 17 January 2014 (“Almunia’s Bruges Speech”). 25 See, e.g., Teece 2011, at pp. 233 ff. 26 See Shapiro 2012, at p. 401. 27 See Almunia’s Bruges Speech. 8 likely to constitute the more credible competitive threat in the next generation of products.28 Leveraging dominance into neighbouring / nascent markets becomes a particularly harmful practice in terms of innovation particularly given that, as highlighted by Teece, “as a general principle, novelty comes about by changing and combining existing artefacts and structure”.29 19. Likewise, Google’s enduring dominance in multiple inter-connected EU markets30 has in all likelihood exacerbated the exclusionary effect of the abusive behaviour Google has engaged in. In the words of Shelanski, “a digital platform monopolist controls its own product or service as well as access to a much broader universe of products and services. […] As one set of authors puts it […] Google has become the main interface for our whole reality. […] the key point about bottleneck discrimination is that of ability: Google’s scale and scope […] appear to give it substantial power in a wide range of markets”.31 20. It is noteworthy to recall here that, according to the Commission’s Enforcement Priorities, “the higher the market share and the longer the period over which it is held, the more likely it is that it constitutes an important preliminary indication of the existence of a dominant position and, in certain circumstances, of possible serious effects of abusive conduct”.32 B. Assessment of Innovation Under an Efficiencies Defence 21. Once conduct is prima facie abusive, the balance of its alleged positive and negative effects in relation to innovation only takes place in the context of the “efficiencies defence”.33 We have seen no evidence that Google has sought to establish such a 28 See Case T-201/04 Microsoft ECR II-3601 [2007], at para. 1089. 29 See Teece 2011, at p. 242. 30 See, e.g., the Commission Decision of 18 February 2010 in Case COMP/M.5727 – Microsoft / Yahoo! Search Business, at para. 112, where the Commission stated that “Google is by far the platform that enjoys the highest market shares. In 2009, Google has market shares above 90% in most national markets whereas Microsoft and Yahoo both have market shares below 5%.” 31 See Shelanski 2013, at pp. 1676-1677. 32 See Enforcement Priorities, at paras. 15 and 20 (I). 33 Note that this would be inconsistent with the Commission’s approach regarding Article 101 TFEU (where no balance of the positive and negative effects of the conduct is made under Article 101(1), such balance being made only under Article 101(3) TFEU (see Guidelines on the application of Article 81(3) of the Treaty, Official Journal C 101, 27/04/2004 P. 0097 – 0118, at para. 30). See further Case T-112/99 Métropole télévision (M6) [2001] ECR II-2459, at para. 74, where the General Court held that it is only in (Cont'd on next page) 9 defence, given that it has continued to deny its dominance. However, Google is very unlikely to prevail in establishing an innovation-based efficiencies defence for its abusive conduct. More specifically, for a prima facie abusive practice to benefit from an efficiencies defence under Article 102 TFEU, four cumulative requirements need to be met.34 22. First, the efficiency gains must have been brought about by the conduct in question. 23. Second, the conduct should be “indispensable to the realisation of those efficiencies: there must be no less anti-competitive alternatives to the conduct that are capable of producing the same efficiencies”. 24. Third, the efficiency gains must “outweigh any negative effects on competition and consumer welfare the conduct might have in the affected markets”. In other words, the harm to competition and consumers caused by Google’s abusive practices could only be justified if were completely and unequivocally offset by the efficiencies demonstrated to derive from the said conduct in the affected markets.35 Moreover, this balancing exercise (Cont'd from previous page) the precise framework of Article 81(3) that the pro- and anti-competitive aspects of a restriction may be weighed. See further González-Díaz, F. E., Snelders, R., Abuse of Dominance Under Article 102 TFEU, Claeys & Casteels, 2013, at paras. 3.116 ff and O’Donoghue, R., and Padilla, J. A., The Law and Economics of Article 102 TFEU, Hart, 2013, at pp. 286 ff. It has been argued that balancing of positive and negative effects is not necessary where the exclusionary practice is indispensable to achieve the alleged efficiencies. However, in the Microsoft case the Court appeared to reject this theory (which also finds no support in the subsequent Post Danmark Ruling of the CJEU) and indicated that the fact that tying enabled software developers and Internet site creators to be sure that Windows Media Player was present on PCs worldwide was “precisely one of the main reasons why the Commission correctly took the view that the bundling led to the foreclosure of competing media players from the market” (see Case T-201/04 Microsoft v Commission [2007], ECR II-3601, at para. 1151; see further O’Donoghue, R., and Padilla, J. A., The Law and Economics of Article 102 TFEU, Hart, 2013, at p. 289). In any event, even accepting this theory, Google would still need to establish to the requisite standard of proof that its conduct is necessary for achieving its alleged innovation. 34 See Case C-209/10 Post Danmark [NYR], at para. 41. See further Enforcement Priorities, at para. 30. See Rose, D., and Bailey, D., Bellamy & Child. European Union Law of Competition, Oxford 2012, at para. 10.063; Jones, A., and Sufrin, B., EU Competition Law, Oxford, 2014, at pp. 389 ff; González-Díaz, F. E., and Snelders, R., Abuse of Dominance Under Article 102 TFEU, Claeys & Casteels, 2013, at paras. 3.116 ff.; Whish, R., and Bailey, D., Competition Law, Oxford, 2012, at pp. 211 ff and O’Donoghue, R., and Padilla, J. A., The Law and Economics of Article 102 TFEU, Hart, 2013, at pp. 285 ff. 35 A similar solution is reached under the EU rules on merger control (where the EU Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ C 31, 5.2.2004, p. 5–18, at para. 79, provide that any efficiencies should “in principle, benefit consumers in those relevant markets where it is otherwise likely that competition concerns would occur”; see further COMP/M.4000 Inco / Falconbridge, at para. 544; Drauz, G. and Jones, C., EU Competition Law. Mergers and Acquisitions, Claeys & Casteels, 2012, at para. 4.673 and Levy, N., European Merger Control Law: A Guide to the Merger Regulation, LexisNexis, 2011, at § 15.02[4][a]) and under Article 101 TFEU (see the Commission’s Guidelines on the application of Article 81(3) of the (Cont'd on next page) 10 should include the assessment of any negative effects which might derive from Google’s practices. The Commission is familiar with this exercise, which is conducted relatively frequently in the domain of merger control. In the words of Levy, “[i]n markets characterised by technological innovation, it may be necessary to assess how the merger will affect the competitive pressure to innovate”.36 Indeed, if the Commission is capable of determining ex ante the likely effects of a transaction in terms of innovation during a review period of under six months, it should also be able to ascertain ex post the positive and negative effects of a practice by a dominant undertaking during a case in which the investigation has been on-going for several years. 25. Fourth, the conduct in question “does not eliminate effective competition, by removing all or most existing sources of actual or potential competition”. i. According to the Commission “rivalry between undertakings is an essential driver of economic efficiency, including its dynamic efficiencies in the form of innovation. Where there is no residual competition and no foreseeable threat of entry, the protection of rivalry and the competitive process outweighs possible efficiency gains”.37 This requirement applies particularly to innovation-related alleged efficiencies. ii. In any event, marginalised competition is tantamount to no effective competition. In the words of Dolmans and O’Donoghue: “[a] dominant firm could always allow one or two small rivals to remain on the market as marginalized competitors (sometimes referred to as ‘bonsai’). But the mere (Cont'd from previous page) Treaty, Official Journal C 101, 27.04.2004, p. 97-118, at para. 43, providing only an exception for “related markets”, in which case benefits achieved on separate markets can be taken into account “provided that the group of consumers affected by the restriction and benefitting from the efficiency gains are the same”, see further Faull, J., and Nikpay, A., Faull & Nikpay. The EC Law of Competition, Oxford, 2007, at para. 3.409). Last but not least, the General Court has indicated that abusive practices could never be justified by benefits to the dominant undertaking (see Joined Cases T-191/98 to T-214/98 Atlantic Container Line AB v Commission [2003] ECR II-3275, at para. 1114). 36 See Levy, N., European Merger Control Law: A Guide to the Merger Regulation, LexisNexis, 2011, at § 9.02[5], referring, e.g., to COMP/M.2922 Pfizer / Pharmacia and a variety of other pharmaceutical mergers. For a non pharma-related merger precedent where the European Commission took into consideration the alleged effects of the transaction in innovation, in order to require from the parties substantial divestitures see COMP/M.6458 UNIVERSAL / EMI, e.g., at para. 887. The effects of the transaction in innovation have also played a key role in the Decision of the US DOJ to successfully challenge Bazaarvoice Inc.’s acquisition of rival PowerReviews (prohibition upheld by a Federal District Court on 10 January 2014). 37 See Enforcement priorities, at para. 30, IV. 11 presence of a competitor does not mean that no elimination of competition has occurred. […] effective competition does not mean the mere presence of one or more niche rivals”.38 iii. Moreover, the Commission has also clearly indicated its reluctance to apply an efficiencies defence to quasi-monopolies like Google. In the words of the Commission, “exclusionary conduct which maintains, creates or strengthens a market position approaching that of a monopoly can normally not be justified on the grounds that it also creates efficiency gains”.39 26. Pursuant to the judgment of the General Court in the Microsoft Case and the Commission’s Enforcement Priorities, the burden of proof of the alleged innovation efficiencies would fall on the dominant undertaking.40 In addition, Google’s attempt to justify its abusive practices on the basis of an innovation-based efficiency defence would need to meet the required relevant standard of proof. According to the Commission, “the dominant undertaking will generally be expected to demonstrate, with a sufficient degree of probability, and on the basis of verifiable evidence” that all the cumulative conditions are satisfied.41 27. Accordingly, it is very unlikely that the alleged innovation deriving from Google’s abusive conduct would meet the applicable EU requirements for an efficiencies defence but if such a defence were advanced it would need to be reasoned, supported by verifiable compelling evidence and rigorously tested. 38 See Dolmans, M., O’Donoghue, R., and Loewenthal, P.-J., “Article 82 EC and Intellectual Property: The State of the Law Pending the Judgement in Microsoft v. Commission”, Competition Policy International, Spring 2007. Indeed, Dolmans and O’Donoghue continue “it is well established in the economics literature that there is a significant risk of falsely concluding that no harm to competition has occurred merely because rivals have not fully exited” (quoting Krattenmaker, T., and Salop, S., “Anticompetitive Exclusion: Raising Rival’s Costs to Achieve Power Over Price”, 1986, 96 Yale L.J. 209). 39 See Enforcement priorities, at para. 30, IV (in fine). 40 See Case T-201/04 Microsoft v Commission [2007] ECR II-3601, at para. 688. See further Enforcement priorities, at para. 31. It should be noted that this would be consistent with the rules applicable under 101 TFEU (see Council Regulation No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty, OJ L 1, 4.1.2003, p. 1–25 (“Regulation 1/2003”), at Article 2, according to which: “[t]he undertaking or association of undertakings claiming the benefit of Article [101(3) TFEU] shall bear the burden of proving that the conditions of that paragraph are fulfilled”). 41 See Enforcement Priorities, at para. 30. See further González-Díaz, F. E., and Snelders, R., Abuse of Dominance Under Article 102 TFEU, Claeys & Casteels, 2013, at paras. 3.119. Moreover, the case-law in relation to the “objective justification” defence requires the dominant firm to produce sufficiently “firm evidence” (see Case T-57/01 Solvay v European Commission [2009] ECR II-4621, at para. 334). 12 IV. THE COUNTERFACTUAL IN ARTICLE 102 TFEU CASES 28. Were one to follow Bork and Sidark, one might conclude that in Article 102 TFEU cases when assessing innovation one needs only to look at the activities of the dominant undertakings. However, that is not the case: there needs to be a balancing exercise taking into account not only whether the conduct constitutes genuine innovation (something which needs to be established to the requisite standard of proof) but also on the innovation foregone as a result of the exclusionary conduct. 29. In other words, in a similar manner to an Article 101 TFEU investigation,42 an Article 102 TFEU case requires an adequate analysis of the counterfactual, i.e., the determination of the performance of the affected market(s) absent the practice. It should be stressed, in this regard, that the Commission already provides in its Enforcement Priorities that its “assessment will usually be made by comparing the actual or likely future situation in the relevant market (with the dominant undertaking’s conduct in place) with an appropriate counterfactual, such as the simple absence of the conduct in question or with another realistic alternative scenario, having regard to established business practices”.43 Moreover, the Commission has already conducted an analysis of the counterfactual in an Article 102 TFEU Case in its Decision in Case Telekomunikacja Polska.44 30. Two considerations come to the fore in this regard: i. Even though an adequate analysis of the counterfactual is relevant both in terms of the existence of a prima facie abuse and in the context of an efficiencies defence, it is only under the latter that a balancing exercise can be made between the positive and negative effects of given conduct. ii. While some economists disagree as to the relationship between market structure and innovation, the presumption should be that it is the smaller players, and not a quasi-monopolistic undertaking, which are more likely to be / have been innovative. This is consistent with both traditional antitrust theory and with the findings of economists specialised in innovation. In the words of 42 See Case 56/65 Société Technique Minière [1966] ECR 337. See further Guidelines on the application of Article 81(3) of the Treaty, Official Journal C 101 , 27/04/2004 P. 0097 – 0118, at para. 17. 43 See Enforcement Priorities, at para. 21. 44 See Case COMP/39525 Telekomunikacja Polska, at paras. 812 – 872. 13 Teece: “I believe that the debate over whether to favor competition over monopoly (as the market structure most likely to advance innovation) was won long ago in favor of some form of rivalry / competition”.45 V. INNOVATION AND REMEDIES 31. It would appear that the fact that these markets are innovation-intensive has played a role in the Commission’s choice of the remedy for Google’s abusive conduct. More specifically, according to Vice President Almunia, “given the specific characteristics of these markets”, it is preferable for the Google proceedings to end through an Article 9 Regulation 1/2003 commitments Decision than an Article 7 Prohibition Decision. In the words of Vice President Almunia “[the] alternative of adversarial proceedings would take many years, with many uncertainties, and would not have the same immediate impact. It would also not necessarily deliver a better outcome for consumers given the specific characteristics of this market”.46 32. As a matter of law, Google’s abusive conduct needs to be subject to effective remedies pursuant to the general principle of effectiveness of EU law enshrined in multiple Recitals of Regulation 1/200347 and the Alrosa ruling of the CJEU.48 “Quick” 45 See, e.g., Teece 2011, at pp. 235 and 236 and Niels, G., Jenkins, H., and Kavanagh, J., Economics for Competition Lawyers, Oxford, 2011, at p. 172. See further Motta 2004, at p. 57, according to whom “[a]lthough theoretical and empirical research on the link between market structure and innovation is not conclusive […] [there is] a ‘middle ground’ environment, where there exists some competition but also high enough market power coming from the innovative activities [which would] […] be the most conducive to R&D output. […] The result is that some intermediate levels of competition might be optimal for innovations and productive efficiency, a result which is […] confirmed in empirical studies that do find an inverted U relationship between competition and innovation. […]”. Moreover, Motta unambiguously adds: “[t]he only sound and robust conclusion we can derive from analyses like the one above is that a monopoly […] is worse than competitive market structures because it fails to stimulate dynamic efficiency”. In a similar vein, in the domain of merger control, Katz and Shelanski, recommend the resort to a “rebuttable presumption of harm [to innovation]”, in case of a “merger to monopoly” (see Katz, M., and Shelanski, H., “Mergers and Innovation”, Antitrust Law Journal 74:1-85, at p. 4). 46 See Almunia, J., “Statement on the Google Investigation”, SPEECH/14/93, Brussels 5 February 2014 (“Vice President Almunia’s Speech on 5 February 2014”). For a similar position see “Antitrust: Commission obtains from Google comparable display of specialised search rivals- Frequently asked questions”, MEMO/14/87, 5 February 2014. Vice President Almunia appears to have changed its position as to remedies in high-tech sectors, given that as recently as in September 2013 he had stated that, in relation to “the relative merits of commitment and prohibition decisions”, that “there is no principle that can direct our thinking one way or the other” (see Almunia’s Fordham Speech), a statement which, while clearly incorrect, is different from the Commission’s current preference for a commitments solution. 47 See Regulation 1/2003, at Recitals1, 2, 3, 5 and 8. 48 See Case C-441/07 P European Commission v Alrosa Company Ltd [2010] ECR I-05949. 14 enforcement is only desirable, and acceptable, as a matter of law, if it amounts to “effective” enforcement. 33. As a matter of economics, while we agree with Vice President Almunia that “the purpose of antitrust enforcement should be to allow consumers and users to benefit from competition on the merits as soon as possible”,49 this statement begs the question of whether an Article 9 solution will, quickly or otherwise, allow consumers and users to benefit from competition on the merits and innovation. We are of the view that this will not be the case for the following reasons: i. While it is true, as stated by the CJEU in its TeliaSonera Ruling that “particularly in a rapidly growing market, Article 102 TFEU requires action as quickly as possible”, this is in order “to prevent the formation and consolidation in that market of a competitive structure distorted by the abusive strategy of a [dominant] undertaking”.50 Once the distortion has already taken place the focus has to be on correcting it, and it would be a perverse outcome if the delay in the enforcement against the abusive behaviour of an overwhelmingly dominant undertaking were to constitute a reason for accepting ineffective commitments. ii. It is also important to take into account whether “innovators can appropriate [a fair share] of the social benefits [which] their innovations have caused”. As Shapiro also indicates “appropriability can greatly affect innovation incentives”.51 However, none of the remedies which the Commission is willing to accept favours the appropriability of innovation by parties other than Google. On the contrary, the remedies the Commission is willing to accept in relation to discrimination in search have resulted in and will result in 49 See Vice President Almunia’s Speech on 5 February 2014. 50 See TeliaSonera, at para. 108. 51 See Shapiro 2012, at p. 388. The concept of “appropriability” is relevant for all four of Google’s abuses but becomes particularly relevant in the context of scraping. By appropriating content produced by third parties, Google is depriving them of their share of the benefits of their innovations, whereas, as also indicated by Carlton and Perloff, the lack of appropriability discourages research (see Carlton, D., W., and Perloff, J., M., Modern Industrial Organization, Pearson, 2005, at pp. 532 ff. Last but not least, Director General Alexander Italianer has indicated that “[w]hen intervening in high tech markets, it is important […] to strike a careful balance so as not to undermine undertaking’s incentives to invest and innovate” (see Italianer, A., “Innovation and Competition”, Fordham Competition Law Institute Conference, 21 September 2012; see further Italianer: “Innovation and Competition Policy in the IT Sector: The European Perspective, Beijing, China, 26 June 2012). 15 competing vertical search sites unduly being deprived from the revenues deriving from their innovation. This is, among other reasons, because (i) the so-called “paid rival links” results in Google’s competitors having to hand over a significant portion of the revenues from their innovations over to Google; (ii) because many entities which are or can become vigorous competitors of Google will not qualify as “paid rival links” (most notably, subdomains of competing search engines and SME’s not having the resources to prevail in Google’s auctions); and (iii) because only a principled remedy that simply forbade Google from discriminating would be forward looking and likely to increase the space in Google’s SERP available for organic results, which would enable rivals to attract clicks without having to pay anything to Google. As a result of all of which, both innovation and investment in innovation will be discouraged. iii. Uncertainty as to the effects of the proceedings cannot be invoked as a ground for choice of a remedy, particularly after years of investigation. Again, in the words of Shapiro: “[i]t appears to me that a […] misleading ‘complexity proposition’ has taken root and threatens to become the conventional wisdom, namely, […] [that which suggests that] the relationship between competition and innovation is so complex and delicate that there should be no presumption that the elimination of product market or R&D rivals will diminish innovation”.52 iv. Last but not least, we would disagree with any view that antitrust intervention in “new markets” involves the need for refraining from imposing fines for clearly abusive conduct of a dominant company. When imposing a fine of over a billion euros on Intel, the Commission indicated that “Intel limited consumer choice and stifled innovation by preventing innovative products for which there was a consumer demand from reaching end customers. Such practices deter innovative companies which might otherwise wish to enter and compete in the market. By ordering Intel to end its abusive practices, competition on the […] market will play out on the merits with the effect that 52 See Shapiro 2012, at p. 388. 16 innovation to the benefit of the consumer can flourish”.53 The Commission’s position in the Intel case contrasts with that of the current proceedings. It should be recalled, in this regard, that the Commission is clearly of the view that Google is dominant.54 The Commission is also of the view that Google has engaged in four types of exclusionary conduct none of which are particularly novel but which have resulted in harm to innovation and consumers.55 Under these circumstances, the lack of fines for conduct which is clearly abusive will send a clear signal of under-deterrence as to Google’s conduct and it will discourage both innovation by potential new entrants and investment by venture capitalists into any market to which Google might leverage its overwhelmingly dominant position. v. Similarly, although it is true that the Commission has relatively recently accepted commitments offered by Microsoft, Apple and IBM, it should be stressed that the Microsoft browser commitments were only accepted after the issuance of a formal Statement of Objections and once several fines had been imposed on Microsoft.56 VI. CONCLUSIONS 34. While the relative prominence of innovation as a competitive driver might be relevant to an analysis of the compatibility of alleged innovative conduct with Article 102 TFEU, we do not believe this requires a fundamental change in the analytical framework used in order to determine whether given conduct constitutes an abuse.57 In the words of Porter: 53 See “Antitrust: Commission imposes fine of 1.06 billion euros on Intel for abuse of dominant position; orders Intel to cease illegal practices - questions and answers”, MEMO/09/235, Brussels, 13 May 2009. 54 See the press conference of Vice President Almunia on 5 February 2014, available at http://ec.europa.eu/avservices/video/player.cfm?ref=I086152. We have previously referred to the Commission’s findings in relation to Google’s enduring market shares and the high barriers to entry and network effects that, according to the Commission, characterise these markets. 55 See “Antitrust: Commission seeks feedback on commitments offered by Google to address competition concerns”, IP/13/371, European Commission, 25 April 2013 and “Commission seeks feedback on commitments offered by Google to address competition concerns – questions and answers”, MEMO/13/383, European Commission, 25 April 2013. 56 See “Antitrust: Commission accepts Microsoft commitments to give users browser choice”, Press Release IP/09/1941, Brussels, 16 December 2009. 57 It should be stressed that this is also the position of, among others, (i) the CJEU (see TeliaSonera, at para. 111, according to which: “the fact that the markets are growing rapidly and involve new technology, requiring high levels of investment is not, as a general rule, relevant to establishing whether the […] practice at issue constitutes an abuse within the meaning of Article 102 TFEU”); (ii) Judge Posner, (Cont'd on next page) 17 “[i]n periods of transition, such as the one we have been going through, it often appears as if there are new rules of competition. But, as market forces play out […] the old rules regain their currency”.58 35. Indeed, authors emphasising the importance of competition in innovation, such as Teece, stress the importance for antitrust enforcement to prevent practices which result in innovation losses. In the words of Shapiro, “exclusionary practices […], if not checked by antitrust law, can make current monopoly power more durable by deterring innovative entrants”.59 36. While some authors are of the view that “new economy” markets are inherently unstable, we are not aware of any economic theory according to which sustained market shares over 90% in markets characterised by high barriers to entry and network effects is not indicative of the existence of substantial market power. Indeed, it is likely that this market power will be akin to that of a monopoly which, as provided in the TeliaSonera Ruling of the CJEU, would exacerbate the effects of any abuse committed by Google.60 37. In any event, those authors proposing that market shares should lose importance as indicators in “fast moving markets” such as, allegedly, would be the case of those under investigation in these proceedings, focus on the “contestability” of these markets. A correct analysis of contestability involves a correct determination of barriers to entry, which the Commission has already indicated are “substantial” and reinforced by network effects in these markets. Among the network effects we would highlight those deriving from the scale advantages deriving from Google’s leading position gathering information, a key input of production and a competitive asset in these markets. (Cont'd from previous page) according to whom “antitrust doctrine is sufficiently supple, and sufficiently informed by economic theory to cope effectively with the distinctive-seeming antitrust problems that the new economy presents” (see Posner, R., Antitrust Law, 2nd Edition, The University of Chicago Press, 2001, at p. 256); and (iii) Vice President Almunia, according to whom “[…] dynamic industries are not immune to careful antitrust scrutiny, nor should the basic antitrust principles modified” (see Almunia, J., “Abuse of dominance: a view from the EU”, SPEECH/13/758, Fordham’s Competition Law Institute Annual Conference, 27 September 2013). 58 See Porter, M. E., “Strategy and the Internet”, On Competition, Harvard University Press, 2008, at p. 101. 59 See Shapiro 2012, at p. 401. 60 See TeliaSonera, at para. 81. 18 38. Certain practices by Google such as algorithmic penalties have been clearly established by complainants. These clearly constitute abuses since they cannot be characterised as “competition on the merits”, irrespective of whether they are a part of an arguably overall genuinely innovative search engine or, for that matter, self-servingly characterised as “product improvement”. 39. This is supported by the Commission’s finding that Google’s conduct “may reduce competitors' incentives to innovate in specialised search”. Adequately ascertaining the innovation lost as a result of Google’s abuses should be a key element when determining the legitimacy and gravity of Google’s conduct. As the General Court confirmed in the Microsoft case, an abuse becomes particularly harmful when it forecloses those competitors of the dominant undertaking which are likely to constitute the more credible competitive threat in the next generation of products.61 40. Correctly ascertaining the importance of innovation would have also required analysing the counterfactual, namely, how innovation would have taken place absent Google’s practices and, more specifically, the innovation losses deriving from Google’s exclusionary conduct. 41. Google’s argument according to which its exclusionary practices in search constitute an innovation-based efficiency should not prevail. The burden of proof would be on Google to establish that: (i) its exclusionary practices are strictly necessary to achieve the alleged innovation; (ii) a balance of the losses and gains deriving from the conduct clearly results in the latter “outweigh[ing] any negative effects on competition and consumer welfare the conduct might have in the affected markets”; and that (iii) the exclusionary practice does not result in the elimination of “effective competition”. In other words, the harm to competition and consumers, including harm to innovation caused by Google’s abusive practices could only be justified if were completely and unequivocally offset by the innovation allegedly deriving from the said conduct in the affected markets.62 Moreover, the Commission had clearly indicated in its Enforcement 61 See Case T-201/04 Microsoft ECR II-3601 [2007], at para. 1089. 62 A similar solution is reached under the EU rules on merger control (where the EU Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ C 31, 5.2.2004, p. 5–18, at para. 79, provide that any efficiencies should “in principle, benefit consumers in those relevant markets where it is otherwise likely that competition concerns would occur”; see further COMP/M.4000 Inco / Falconbridge, at para. 544; Drauz, G. and Jones, C., EU Competition Law. Mergers and Acquisitions, Claeys & Casteels, 2012, at para. 4.673 and Levy, N., (Cont'd on next page) 19 Priorities that it is unlikely for an efficiencies-based defence to prevail where the exclusionary practices reinforce a monopoly-like position such as that which Google enjoys in search. 42. “Quick” enforcement is only desirable to the extent it is effective. The only way to “quickly” restore competition in online search and search advertising is for the Commission to issue an Article 7 Decision imposing a sufficiently high fine, to deter Google from further testing ways in which to breach Article 102 TFEU. * * * (Cont'd from previous page) European Merger Control Law: A Guide to the Merger Regulation, LexisNexis, 2011, at § 15.02[4][a]) and under Article 101 TFEU (see the Commission’s Guidelines on the application of Article 81(3) of the Treaty, Official Journal C 101, 27.04.2004, p. 97-118, at para. 43, providing only an exception for “related markets”, in which case benefits achieved on separate markets can be taken into account “provided that the group of consumers affected by the restriction and benefitting from the efficiency gains are the same”, see further Faull, J., and Nikpay, A., Faull & Nikpay. The EC Law of Competition, Oxford, 2007, at para. 3.409). Last but not least, the General Court has indicated that abusive practices could never be justified by benefits to the dominant undertaking (see Joined Cases T-191/98 to T-214/98 Atlantic Container Line AB v Commission [2003] ECR II-3275, at para. 1114). 20