In December 2020, the European Commission proposed an ambitious new set of rules, the Digital Markets Act (“DMA”),1 which could be applicable in 2023 and whose rationale leans heavily on innovation. In many places throughout the proposed DMA recitals, the Explanatory Memorandum included therewith and the Impact Assessment attached thereto, the Commission explains in greater detail how innovation is currently adversely impacted in the digital economy and how its proposal would improve the situation.2
General Relationship Between Competition and Innovation
At a general level, the Commission reiterates its core belief that there is a direct relationship between competition and innovation: more competition leads to more innovation.3 Similar statements are found throughout the decisions and soft-law instruments issued by competition agencies across Europe. One might fault the Commission for taking what looks like a strong Arrowian view4 of the relationship between competition and innovation, ignoring both the Schumpeterian analysis5 (which posits an inverse relationship) and the contemporary synthesis (which returns an inverted-U relationship) made by a cluster of authors around Aghion.6 In the context of competition law, authorities are usually dealing with markets where competition is rather diminished than excessive. It may then not matter much in practice whether they take an Arrow or Aghion theoretical perspective, since in all likelihood authorities are active on the upward-sloping part of the inverted-U, where Arrow and Aghion coincide.
While the DMA may also play out in the upward-sloping part of the inverted-U, it cannot be excluded that in some situations, competition is already strong, so that authorities could be acting on the downward-sloping part of the inverted-U, where it is no longer correct to assume that more competition will unavoidably foster innovation. Indeed, in the more elaborate parts of its analysis, the Commission shows that it is aware that the relationship between competition and innovation is not so simple. There are some trade-offs involved, especially once the specific features of so-called Core Platform Services, i.e. the digital services targeted by the DMA, are brought into the picture, namely: economies of scale and scope, network effects (compounded by multi-sidedness), lock-in, lack of multi-homing, vertical integration and data-driven advantages. Innovation can come from the platform itself – and the firm controlling it (the “gatekeeper” pursuant to the DMA) – or it can arise around the platform, typically driven by a firm using the platform to bring an invention to the market (a “user” pursuant to the DMA). The Commission is well aware that the DMA could reduce the innovation incentives of the platform owner; the question then becomes what is gained in return. Once the analysis reaches that level of sophistication, however, it starts branching out in different directions. As the following paragraphs show, two distinct innovation scenarios are bundled together in the DMA.
Scenario 1: Sustaining Innovation by Users Based on the Core Platform Services
As a starting point, the Commission recognizes that online platforms have proven to be innovation hotbeds, with innovation originating throughout the platform ecosystem, i.e. not just from the gatekeepers themselves, but also from platform users, businesses and individuals alike.7 Platform users innovate by introducing complementary products, for instance games or productivity apps for mobile operating systems. Yet the heady early days of the Internet are over. In the current context, the Commission points to evidence that the rise of powerful gatekeepers controlling the main online platforms leads to sub-optimal levels of innovation. Essentially, the innovation incentives of the gatekeepers and the users become misaligned, and the gatekeepers start to divert some of their efforts towards preventing or appropriating innovations brought about by platform users. In that case, certain courses of conduct by platform operators hinder business users and adversely affect their innovation incentives. Ultimately, these businesses refrain or are prevented from bringing innovative offerings to the market.
Under this scenario, the main concern is that the platform gatekeeper would go out of its way to control the flow of innovation around its platform. While blocking inventive offerings by users is certainly possible, a more likely course of conduct, witnessed in a number of cases already (from Microsoft to Google Shopping and Google Android),8 is that the gatekeeper would use bundling or self-preferencing to exclude the inventive user and appropriate the profits from the innovation via a competing offering of its own. A significant proportion of the obligations contained in the DMA are designed to address that concern. They include generic prohibitions against the use of non-public data to compete with business users or self-preferencing in search rankings, as well as an obligation to grant equal access to APIs and other interoperability features.9 In addition, the DMA proposal protects more specifically some neighbouring markets against gatekeeper conduct, namely identification services, apps and app stores and payment services.10
As far as innovation theory is concerned, this scenario involves what would typically be incremental innovation around the existing core platform service (as with the Novell and Sun server operating systems in relation to Windows Server Operating System in Microsoft). Furthermore, that innovation will also be sustaining innovation as it will remain within the value network or the architecture created by the platform. For instance, it would consist of apps running on a smartphone operating system platform such as iOS or Android; a specialized search engine (or another ancillary service such as maps, etc.) accessible through a general search engine such as Google; or a retail business hosted on an online retailing platform such as Amazon.
As a normative matter, the Commission cannot be faulted for intervening to safeguard the ability of third parties to carry out incremental innovation around the core platform services. Incremental innovation is especially prevalent in the digital economy, and it can greatly contribute to consumer welfare. Starting with Microsoft, competition enforcement in the EU has protected incremental innovation in the digital economy, although this has never been explicitly stated.11 In that respect, the DMA proposal merely extends the innovation policy choices made in competition law enforcement.
Whereas the first scenario is laid out in the Commission Impact Assessment and translated in the list of obligations and prohibitions applicable to gatekeepers, it does not entirely fit within either the “contestability” or “fairness” objectives defined in the DMA proposal. Regarding fairness, the proposal indicates that it should be understood as a contractual imbalance between the respective rights and obligations of gatekeeper and user.12 This definition does not fil well with the first innovation scenario which is more about equality of competitive opportunity.13 As for contestability, the DMA proposal almost always defines it in relation to core platform services, in line with economic theory where contestability is a redeeming feature of monopolistic markets.14 However, the first innovation scenario, as regards both the analysis and the remedial obligations, has little to do with the contestability of core platform services: rather, it is about containing gatekeeper power and preventing it from adversely affecting neighbouring markets in the ecosystem of the core platform service.
Scenario 2: Innovation in the Core Platform Services: Frontal Competition or Disruption?
When it comes to contestability on the core platform services, the DMA proposal does not clearly explain the link with innovation. The Impact Assessment mentions that the gatekeepers divert their resources away from R&D and towards M&A, in order to compete “for the market.” At the same time, it is known that “a significant amount of innovation is driven by disruptive firms,” so that the DMA proposal should seek “to protect the competitive process by which disruptive firms challenge the status quo.”15 These passages hint at disruptive innovation, but do not fully develop the analysis. For one, the Commission misunderstands disruptive innovation by linking it with competition “for the market.” In the case of gatekeepers within the meaning of the DMA, competition for the market would be very difficult to achieve, since gatekeepers have fully exploited the characteristics of core platform services to build a quasi-unassailable position. There is little hope for a new search engine to outperform Google, for a competing social network to oust Facebook or for an alternative online commerce and retail platform to outcompete Amazon.
The more likely scenario is not frontal competition, but rather sideways competition, where a core platform service is side-lined and made less central for users (competition “on the market”).16 Such sideways competition usually involves disruptive innovation in the technical sense – as introduced by Christensen and then updated by Gans17 – namely an innovation where the incumbent firm is caught off-guard and punished despite doing what made it successful. Disruptions are never frontal assaults: they involve a shift in the value network binding consumers to a given product space, or in the dominant architecture used by suppliers on that space. We witnessed a number of disruptions in the digital economy in recent years, usually with positive implications for competition policy. So it is that Google heralded the rise of Internet-centric computing, which turned client operating systems (such as Windows) into a sideshow. Then Facebook turned a social media platform into an alternative portal to search engines such as Google, limiting the impact of Google’s dominance. The rise of smartphones – led by the iPhone – not only reshuffled the market for mobile devices but also made computers less central, thereby reducing the impact of dominant positions in CPUs, for instance.
If disruptive innovation does not involve frontal competition and blindsides incumbents, can these incumbents do anything to avert it? Possible defensive strategies include trying to prevent potential disruptors (to the extent they can be detected) from gaining a foothold – as Microsoft did when it saw the threat emerging from Netscape in the 1990s – or acquiring potentially threatening firms to throttle any disruption.18
By now it has become clear that the Facebook/Instagram and Facebook/WhatsApp acquisitions were textbook cases of the latter strategy, as the recent US state-level and FTC antitrust case against Facebook indicate.19 The DMA proposal picks up on strategic acquisitions with its obligation to inform the Commission about intended concentrations but it is a relatively weak provision.20 In combination with the new Guidance on Case Referrals to the Commission,21 it could become a clever workaround to the notification thresholds. However, even if the Commission could eventually review these strategic acquisitions, they remain a blind spot in the current merger control law. Many questions remain unaddressed, such as which theory of harm would justify blocking such acquisitions, and which standard of proof should apply (balance of probabilities or balance of harms).
As for the defensive exclusionary strategy, even if the disruption analysis is not developed in the DMA, some of the obligations could help to keep gatekeepers vulnerable to disruption: they include the obligation to keep advertising markets transparent, to provide data to users and to allow for data portability, to refrain from MFN clauses and steering and to offer access to search engine data to third-party search engine.22 It is apparent from the rationale and the wording of these obligations that they are meant to support frontal competition with the core platform service in question. Yet these obligations can also provide cover for a disrupting innovator to come close enough to the gatekeeper, so that the disruptor can use its position as a stepping stone to shift the value network or the dominant architecture.
By way of illustration, the obligations relating to data portability, or to the availability of data generated by and through the activities of business users,23 could be used to enable a frontal competitor to Google, Facebook or Amazon – however unlikely – to survive by feeding on the data obtained from such gatekeepers via its users. But these obligations – depending on how they are specified in practice – could also be used to funnel data to an innovative entrant that would try to disrupt the value network or the dominant architecture: hypothetically, business users may for instance transfer the data they obtained from Amazon to a provider of a platform dedicated to second-hand sales and trades, local sourcing or ethical sourcing, that could disrupt online retail if successful.
For the sake of completeness, it should be added that disruption could also conceivably come from an innovator in a complementary product around a core platform service (Scenario 1), where this product would evolve from a mere complement into a disruptive offering. Such was the case, for instance, for Netscape’s original web browser, which started as a complement to an operating system such as Windows (enabling it to open up to the World Wide Web). Later it became an existential disruptive threat, triggering an anti-competitive reaction from Microsoft in order to protect its position. The obligations that surround Scenario 1 can therefore protect not just sustaining innovation on a core platform service, but also open up a path to disruption.
DMA and Innovation
Putting it all together, the DMA would be meant to affect innovation as follows. First, it seeks to preserve the innovation incentives of the users of core platform services, so that they can deliver innovative offerings of the incremental type, by way of complementary products within the platform ecosystem and thus of sustaining innovation (Scenario 1 above). Second, as far as the contestability of core platform services is concerned, the DMA proposal seems focused on enabling frontal competition with the gatekeepers (Scenario 2 above). It is questionable whether this is a realistic perspective, and how significant an innovation this would generate. Third, it could be argued that, in trying to safeguard innovation Scenarios 1 and 2, the DMA proposal is also opening some paths for inventive rivals to attempt disruption, either by starting from a complementary offering or by staking a market position close to the gatekeeper and using it as a springboard.
Thus, the DMA proposal leaves as many paths open as possible to innovative firms. This may be the best policy choice given the inherent unpredictability of innovation.24 Fairness (defined as equality of opportunity) and contestability (defined as reducing entry barriers on and around core platform services) would then be shorthand for a regulatory objective of keeping markets open and competitive as much as possible. Seen from that angle, the DMA would fit within a revamped version of the ordo-liberal tradition that still underpins much of EU competition law and economic regulation.25 This ordo-liberalism for the 21st century would have dynamism and innovation at its core, as the main reason why markets should be kept open and competitive.26 Seen from that angle, the DMA will not kill but rather promote the diversity – and hopefully the level – of innovation in Europe.
* Professor of Law and Innovation, Montreal University.
+ Professor of EU Law, Namur University and Academic Co-Director, CERRE.
1 Proposal of the Commission of 15 December 2020 for a Regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act), COM(2020) 842. For a description and an analysis of the Commission proposal, see A. de Streel & P. Larouche, The European Digital Markets Act proposal: How to improve a regulatory revolution, Concurrences, 2021/2, 43.
2 Impact Assessment Report of the Commission Services on the Proposal for a Regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act), SWD(2020) 363.
3 DMA Impact Assessment, para.279.
4 Arrow, “Economic Welfare and the Allocation of Resources for Invention,” in R. Nelson, The Rate and Direction of Inventive Activities: Economic and Social Factors, Princeton University Press, 1962.
5 Schumpeter, Capitalism, Socialism and Democracy, New York: Harper and Brothers, 1942.
6 Aghion, Bloom, Blundell, Griffith & P. Howitt, “Competition and Innovation: An Inverted-U Relationship,” 120 Quarterly Journal of Economics, 2005, 701.
7 Jacobides, Cennamo & Gawer, “Towards a theory of Ecosystems,” Strategic Management Journal, 2018, 2255; Cennamo & Santaló, “Generativity tension and value creation in platform ecosystems,” Organization Science, 2019, 447.
8 Commission Decision of 24 March 2004, Case 37 792 Microsoft confirmed in appeal by the General Court in Case T-201/04, Microsoft v. Commission, EU:T:2007:289. Commission Decision of 16 December 2009, Case 39 530 Microsoft (Explorer); Commission Decision of 27 June 2017, Case 39 740 Google Search (Shopping); Commission Decision of 18 July 2018, Case 40 099 Google Android.
9 DMA Proposal, art.6.1(a), 6.1(b), 6.1(f).
10 DMA Proposal, art.5(e) covers identification services, art.5(f), 6.1(b), 6.1(c), 6.1(e) extend to apps and app stores, and the definition of “ancillary service” at art.2.14 expressly includes payment services.
11 Larouche, “The European Microsoft Case at the Crossroads of Competition Policy and Innovation,” 75 Antitrust Law Journal, 2009, 933.
12 DMA Proposal, art.10(2a).
13 Equality of competitive opportunity seeped from Article 106(1) TFEU case-law into Article 102 TFEU analysis, in the wake of Cases C-280/08P, Deutsche Telekom EU:C:2010:603.
14 Baumol, Panzar & Willig, Contestable Markets and the Theory of Industry Structure, Saunders College Publishing/Harcourt Brace, 1982.
15 DMA Impact Assessment, para. 280, 282-3, 322; also Rec. 17 of the proposal.
16 Larouche, Platforms, “Innovation and Competition on the market,” CPI Antitrust Chronicle, February 2020; Petit, Big Tech and the Digital Economy: The Moligopoly Scenario, Oxford University Press, 2020.
17 Christensen, The Innovator’s Dilemma, Boston: Harvard Business School Press, 2007; Gans, The Disruption Dilemma, MIT Press, 2016.
18 Federico, Scott Morton & Shapiro, “Antitrust and innovation: Welcoming and protecting disruptions,” in J. Lerner & S. Stern (eds), Innovation Policy and the Economy, University of Chicago Press, 2019, 125.
19 https://www.ftc.gov/news-events/press-releases/2020/12/ftc-sues-facebook-illegal-monopolization Also, Argentesi., Buccirossi, Calvano, Duso, Marrazzo & Nava “Merger Policy in Digital Markets: An Ex-Post Assessment,” CESifo Working Paper No. 7985, 2019.
20 DMA Proposal, art.12
21 Commission Guidance of 26 March 2021 on the application of the referral mechanism set out in Article 22 of the Merger Regulation to certain categories of cases, O.J.  C 113/1.
22 Resp. DMA Proposal, art.5(g) and 6.1(g), art.6.1(h), art.5(b), art.5(c), art.6.1(j).
23 Resp. DMA Proposal, art.6.1(h), art.6.1(i).
24 Kerber, Competition, Innovation, and Competition Law: Dissecting the Interplay, 2017, available on SSRN.
25 Eucken, The Foundations of Economics: History and Theory in the Analysis of Economic Reality, Springer, 1992; Gerber, Law and Competition in the Twentieth Century Europe, Oxford University Press, 1998.
26 de Streel, Should Digital Antitrust be Ordo-liberal?, Concurrences, 1-2020, 2.