Commentary II: New Ideas about Regulating Platforms

August 31, 2021

Paul MacDonnell, executive director, Global Digital Foundation
 

GENERAL INTRODUCTION

This is the second of three short commentaries which, combined, will serve as a general introduction to, and discussion of, anti-trust / competition regulation of digital markets in Europe. Set against the backdrop of the EU’s proposed Digital Markets Act (DMA), they will serve as a primer for European policymakers and legislators who are unfamiliar with, yet interested in, the economic background and key issues at stake in this important area of emerging regulation. The first commentary, Digital Markets and Their Discontents, summarised key aspects of the growth of digital markets and the associated practices undertaken by some of their dominant platforms that have become the focus of dispute in the EU, and introduced the Union’s proposed Digital Markets Act (DMA) as both a response to perceived abuses and a significant contribution to the Union’s industrial policy in its own right.1 The second, New Ideas about Regulating Platforms, presented here, reviews platforms’ practices in the light of historical and recent antitrust / competition theory and key provisions in the proposed DMA. Finally, the third, Towards the Practical Regulation of Digital Markets, will propose practical steps for the regulation of the relationships between digital platforms and digital market participants in Europe. This project has been undertaken with the support of Google Europe whom we would like to thank.

Views expressed in this series are the author’s and neither those of Global Digital Foundation (which holds no corporate views) nor its supporters.

New Ideas about Regulating Platforms

PROTECTING THE STRUCTURE OF THE MARKET VS.CONSUMER HARM

In the 1970s competition policy in the United States switched from protecting market structures from ‘unfair’ competition towards avoiding consumer harm.2 The old ‘economic structuralist’ view had been that the consolidation and integration of businesses were conducive to anti-competitive behaviours such as price fixing, market division, blocking the entry of new competitors, and the abuse of bargaining power either through raising prices or degrading services. This view was successfully challenged by the American jurist Robert Bork, and by the free-market thinking of the Chicago and Austrian Schools of economics, on the basis that the regulations it led to protected producers at the expense of consumers.3 As a result, by the early 1980s, U.S. lawmakers and courts had changed their opinion about what constituted anti-competitive behaviour. They began to pay less attention to the structure of overall markets and to focus on price to consumers as the only valid metric for determining what constituted anti-competitive actions.

This new doctrine was not inconsistent with the EU approach to competition in so far as the Union aimed, with the 1993 Single European Act, to establish a single market for goods across all of its member states. However, Europe retained the older traditions of the social market economy—a belief that the social ends of policymaking must be given their due weight when making economic decisions.4 The advent of disruptive platform businesses since the 1990s has threatened this vision.

Recently in the United States there are signs of reversion to the earlier pre-Bork doctrine that the market structure embodies a degree of social capital which, under some circumstances, should be defended with antitrust law. As with Europe, the inspiration for this change is the dominance of technology platforms. Both states and the U.S. federal government have launched a series of probes and legislative initiatives aimed at curbing the power of technology platforms; and, in particular, the appointment of Lina Khan, a well-known critic of the Chicago School / Bork consumer-harm doctrine, as Chair of the Federal Trade Commission signals that the Biden White House is serious about wishing to see a change of direction.5 European competition regulators have welcomed these developments and signaled their belief that EU and U.S. competition policy will converge.6 It is against this backdrop that we must consider the DMA.

DISCONTENTS VIEWED IN THE LIGHT OF RECENT ANTITRUST / COMPETITION THEORY

HOW ARE DIGITAL PLATFORMS DIFFERENT?

The DMA is predicated on the idea that, as 'gate-keepers', some digital platforms are different in fundamental ways from regular businesses. Some commentators view these as businesses created solely by their executive founders and fueled through loss-making periods of growth by their investors for the eventual purpose of raising prices so that they can earn supernormal profits.7

Like all economic enterprises, platforms such as Amazon, Google and, especially, Facebook have, themselves, a social dimension which suggests that they are not the sole creations of their executive founders and investors. A key part of digital platforms’ infrastructure is the data willingly contributed by their users and customers and, in this sense, their users and customers can be seen as their co-creators. This social aspect of markets appears to be more explicitly true in relation to platforms than regular businesses and points to the properties that make some digital platforms dominant. Facebook, for example, is the dominant social network in the English-speaking world because its users know implicitly that its very dominance is what enhances its usefulness to them. If everyone is on one social network then that is the only social network to be on. Thus platforms, especially multi-sided platforms, embody characteristics that make them more useful as they become larger. With regard to the assertion that platforms aim to raise prices to consumers there is no evidence, as yet, that this has happened.7 Rather they have generated profits through scale and through charging their business customers for access (either informational or transactional) to their end customers or users. In some cases these latter charges have become a source of contention.

AMAZON

Amazon is a lens through which to examine the issues raised by some platforms’ practices. A single online store offering the widest range of goods, Amazon generates, by virtue of its size, a special value to its users—maximum choice and availability, low prices, and convenience. If the price of a product on Amazon is the same or less than anywhere else and a consumer already has an account with Amazon Prime, which offers next-day delivery on a wide range of goods, then there isn’t much incentive to register a payment card on another site to buy the same product without those benefits or to travel some distance to a shop just to buy it there. Amazon is aware that it is selling convenience and reliability as well as merchandise and services. When it launched Prime in 2007—which as well as quick delivery, includes movies, streamed music, free access to selected ebooks, and special offers—Amazon synergized a number of different consumer experiences into one combined offer. Consumers can order from a vast range of goods and other services and have them delivered in the shortest possible time. In this sense, Amazon is not just a retailer of merchandise, and a marketplace for third-party sellers: it is a services company that leverages and passes on to its customers—both businesses and consumers—the benefits of vast logistical power and almost unlimited choice. This enables it to provide its consumers and the consumers of its business customers with the kind of utility that only the wealthy could afford in bygone ages.

Amazon has, on occasions, used aggressive tactics to achieve this digital dominance. This includes forcing competitors to sell to it or face stronger competition. In 2010 the company pressured Quidsi, the owner of Diapers.com, Soap.com, and BeautyBar.com to sell to it for $445m.9 Writing about Diapers.Com, one Amazon executive wrote an email to staff saying: “We need to match pricing on these guys no matter what the cost.”10 In the case of Quidsi, Amazon was not targeting a company that used the Amazon platform but a competing e-commerce platform. Since that time Amazon has not achieved dominance in Quidsi’s markets. Furthermore, the company was never in a position to raise prices in what were Quidsi’s product lines without risk of competition and, so, from the point of view of the Bork consumer-harm doctrine, its acquisition of Quidsi presented no problem.

What about the fact that Amazon has, contrary to its own stated policies, used information about sales and consumer behaviour to launch products that compete with third-party sellers on its own Marketplace?11 Even here, from the point of view of the Bork doctrine, what Amazon is doing does not per se constitute consumer harm. Lina Khan has argued that the use of data provided by Amazon’s own sellers to launch competing products against them tests the saliency of the consumer-harm standard to destruction.12 Moreover, both the European Commission and some U.S. legislators agree that this practice shows a conflict of interest and, at the very least, is an anomaly that Bork and the free-market economists did not anticipate. It does, indeed, appear invidious that a company would, by its very participation on Amazon’s platform, hand Amazon the weapons to compete against it.

However, the case of an owner of a digital platform competing on that platform is relatively new to competition law and there are a number of questions to consider. First, there is the question of Amazon’s conflict of interest. Does Amazon owe a duty of care to its sellers not to compete against them under any circumstances? If Amazon had set up a company selling products in competition with a provider of manifestly inferior and overpriced goods on its own platform, who would they be harming? If Amazon were prevented from competing with sellers on its platforms couldn’t it just exclude some sellers altogether? If it couldn’t then wouldn’t that incentivize sellers to simply, at no or little cost, list products for sale on Amazon Marketplace in order to guarantee that Amazon wouldn’t compete with them? If the problem comes down to the invidiousness of Amazon’s use of data provided by the company’s participation on the platform to compete against it, then one can think of a number of targeted remedies short of outright banning Amazon from entering markets in competition with its own sellers. These could include: 1) a non-compete clause that would expire after a given time; 2) a ban on platforms using data supplied by a company to compete against it (proposed in the DMA); or, 3) preventing a platform from ranking favoured, or even its own, products ahead of cheaper third-party products (also proposed in the DMA).

THE RETURN OF ECONOMIC STRUCTURALISM

Both the U.S. Federal Government and the European Union are in unusual agreement that social good is the proper objective of competition policy and that this is as true for digital platforms as it is for traditional non-digital enterprises. In her capacity as a legal scholar Lina Khan has proposed a return to the ‘economic structuralism’ of the pre-1980s era—a restoration of an earlier antitrust paradigm whose test is not consumer harm but ‘the health of the market as a whole’. 13

Khan’s is the most comprehensive argument for a return to ‘economic structuralism’ in competition policy and, as such, her analysis can stand not just for the new U.S. perspective but for the thinking behind the EU’s Digital Markets Act (DMA). Her analysis rests on the empirical fact that, during the 1920s and 1930s, firms had been known to use below-cost selling to drive competitors out of markets so that they could raise prices later. She offers the case of Amazon’s ebook pricing as an example of deliberate below-cost selling, of the kind originally identified under the old paradigm, and argues that Amazon did just this when it entered the ebook market in 2007 with its Kindle e-reader and a price for New York Times bestsellers at just $9.99. Khan’s evidence includes the fact that, by 2009, Amazon had achieved market dominance and was selling 90% of all ebooks. However, her account ignores the fact that ebooks are part of a wider publishing market against which they technologically compete—just as printed books once competed with manuscripts and electric lighting with gas lamps. More problematic, however, is her assertion that Amazon’s eventual raising of ebook prices is evidence of anti competitive behaviour when, in fact, the company was forced to do this by the very publishing cartel it was trying to challenge. This cartel successfully imposed an agency-pricing model on ebooks whereby it was they, and not Amazon, that set, and continue to set, the floor-price of all in-copyright ebooks. 14 One consequence of this is that few of the vast savings that in-copyright ebook publishing generates are passed on to consumers. Because their cost is close to, or sometimes even greater than, their paperback counterparts and devices for reading them are so expensive, they have become a luxury product.

While the DMA is seeking to address seeming anomalies that arise within the platform economy (which the Bork doctrine does not appear to have anticipated), as a broad-based ex-ante measure it is vulnerable to challenge on a number of grounds. All competition law, including EU competition law, is tasked with identifying relevant markets when it considers competition cases. But defining relevant markets in competition law is not always easy and, when a particularly innovative product or service is in question, it can sometimes require an ability to see into the future, making the task impossible. The consumer-harm model reflects the acknowledgement of this difficulty and tends to rely on the use of price fluctuations as the only firm evidence that a particular ‘market’ even exists. More importantly the consumer-harm model recognises the ignorance at the heart of all markets which itself is a manifestation of the ignorance that arises from the distributed nature of information between individuals in society. 15 No one person or even government authority can ever know all of the ‘facts’, including motives, that affect the decision-making of individuals.16 The causes of price increases or falls can be numerous, indirect, and distributed across the globe. Human activity “functions through adapting itself to millions of facts which, in their entirety, are not known to anybody”.17 The economic structuralist models of the new FTC and the EU are concerned with treating the firm, and not the individual, as the most basic element of their analysis whereas the most basic element of the Bork / Chicago / Austrian views is the individual. And the Bork / Chicago / Austrian views treat the individual as unpredictable and, therefore, ultimately unknowable. At the heart of their conceptions of the market is an anti-authoritarian doctrine of free-will and individual choice. At its heart the DMA appears to view the market as a clockmaker views a clock—a mechanism whose moving parts can be understood and whose outcomes can be measured and adjusted so that it will work ‘better’. It even goes so far as to estimate the monetary benefit (€13 bn) that will result from its proposed intervention, something which the Bork / Chicago / Austrian proponents consider to be both practically and theoretically impossible. But treating the market as a clock entails the assumption that the free-will of consumers is either an illusion or less important. The distinction between the two perspectives is also analogous to that made by Richard Dawkins between the gene and the organism. The salient fact in understanding evolution is the survival not of the organism but of the gene.18 It is the cumulative desires of individuals and not the corporate objectives of associations or businesses (and not even governments) that determine the structure, survival, and evolution of social organisations.

So what, for example, are Amazon’s relevant markets? It is certain that even Amazon doesn’t know all of them. The retailers that Amazon competes with may, or may not be, part of the company's relevant markets. Amazon itself may not even be an active or direct participant in certain markets which, nevertheless, are affected by its innovations and which, therefore, should be considered as relevant to its activities. The company is competing, of course, with businesses that sell the goods and streaming services that it is also selling. But in its synergies, and especially the synergies of its Prime service, Amazon is competing with much more than major brick and mortar retailers. Its rapid delivery competes with private and public transport. Fewer trips to the shops are needed. And, for reasons of both its range of services, and the logistical capability and technologies that support their delivery and pricing, Amazon competes with property that is near major retail centres by enhancing the utility of living anywhere where customers can usefully avail of these benefits. In other words, digital markets are innovations; and what makes them innovative is their ability to indirectly affect non-adjacent and, even, remote, markets in ways that we cannot predict or even truly understand. If we are to compare markets to models taken from the natural sciences then it is to the indeterminate world of quantum theory and not the clockwork universe of Newton and Descartes that we must allude.

CONSIDERING THE DMA IN DETAIL

The DMA proposes to take action against practices that have proved controversial in Europe and, increasingly, the United States. In Europe, policymakers and legislators are concerned about the immediate practical impact of these practices and their long-term implications for an economy that relies on small enterprises to a much greater extent than the United States. Not only do they believe that traditional legal remedies are too slow to address these anxieties, they want to bind digital platforms into a broader industrial policy.

That the text of the DMA has been drafted by DG CONNECT and refers to ex-ante rules, proposed in the EC's communication, ‘Shaping Europe’s Digital Future’, points to a possible ambiguity of purpose. Is the DMA industrial or competition policy? A case can be made for the regulatory targeting of some of the practices that are in the DMA’s sights but, as an all-encompassing ex ante regulation, is it pretending to have knowledge that no one can know? 19 Below we consider the details of the DMA’s key provisions—contained in articles 5 and 6.

1. Preventing platforms from acting, without consent, to combine personal data and use it to sign users into additional platform services.

If this is a matter of platforms offering an enhanced experience then consent should not present much of a barrier. This provision is also consistent with the General Data Protection Regulation (GDPR).

1. Preventing platforms from favouring their own products over other business users.

and

1. Preventing platforms from using business users' contributed data or that contributed by their end users' or any other generated data from either to compete with business users.

2 and 3 are, of course, aimed at the larger 'gate-keeper' platforms. Here we return to the question. Is it always bad for platforms to compete against products they carry? Why should a seller of an inferior product be immune from competition from the platform itself? Knowledge about the inferiority of a product may exist only on the platform. 3 appears to exclude the use of all data, including consumer reviews for competition. That the use of sales-data analytics could be banned seems defensible on the grounds that it could be deemed an inalienable property of firms that generate it. There is no benefit to firms in being forced by platforms’ terms and conditions into giving it away. Such a ban would not seem to be inconsistent, or very inconsistent, with the Bork consumer-harm doctrine. However, there is a much weaker case for banning the use of customer reviews which are, by definition, public.

CONCLUSION

The DMA is not narrowly targeted at a few specific practices. It is the business end of a broad EU vision of regulation, reporting, and control aimed at taming and integrating platforms into a specifically European vision of ordoliberalism—the doctrine that the state is best placed to design the rules that will maximise the market economy’s ability to bring about social good. 20 The implicit starting point of the DMA is the jettisoning of any idea that consumer harm should be a test for competition policy.

But before we reject the consumer-harm doctrine we should be sure that we have a better idea. It would certainly be helpful if rules about competition and digital platforms were decided by courts and based on the weighing of evidence and consideration of the larger economic picture. However, we must acknowledge that the European courts have proved no match for the anomalous nature of some of the practices under discussion. Furthermore, their decision-making abilities are too slow for the rapid evolution of the digital economy.

It is, therefore, understandable that the European Commission should wish to intervene and, considered as discrete targeted measures, some of its proposals could likely be implemented without harm to the economy or consumers. In particular there is a case for securing, for businesses, their inalienable property rights to set conditions over the use of their sales data unfettered by the terms and conditions of dominant gate-keeper platforms. But the DMA—presented as the result of consultation with interest groups and representative organisations, some of which believe that digital platforms have placed them at a competitive disadvantage—goes beyond addressing anomalies in this way. It proposes the regulation of platforms in a manner that attempts to redesign and repurpose them as servants of a wider economic agenda. This leaves us with three outstanding questions. First, will the DMA protect incumbents’ interests at the expense of consumers? Second, is the DMA an easy political ‘win’ for the EU when its real challenge is the stalled EU single market (especially in services) and the lack of pan-EU capital markets of the kind that are necessary to finance the growth of precisely the kind of enterprises its policymakers are seeking to constrain? And, finally, does the DMA’s conception of the market as a clock downgrade the role of citizens and consumers to the point where it poses a threat to economic freedom?

End

NOTES

1. European Commission (2020). Digital Markets Act. COM(2020) 842 final, 2020/0374(COD). Available: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:52020PC0842&from=en. Michaels D. and Kendall B. (2021). 'U.S. Competition Policy Is Aligning With Europe, and Deeper Cooperation Could Follow'. Wall Street Journal, 15th July, 2021; and Espinoza J. (2021) 'EU tech policy is not anti-American, says Vestager'. Financial Times, 20th June, 2021; and Stacey K. and Politi J. (2021). 'Joe Biden styles himself as a 21st century ‘trust buster’'. Financial Times, 13th July, 2021.

2. Khan L. (2017). ‘Amazon's Antitrust Paradox’. Yale Law Journal. 126:710-805.

3. Bork R. H. (1968) The Antitrust Paradox: a policy at war with itself. New York: The Free Press; Hunt S. D. and Arnett D. B. (2001). 'Competition as an Evolutionary Process and Antitrust Policy'. Journal of Public Policy & Marketing. Spring 2001, Vol. 20 Issue 1, p15-26.

4. Ducci, F. (2020) ‘Introduction’ in Natural Monopolies in Digital Platform Markets. Cambridge: Cambridge University Press (Global Competition Law and Economics Policy). p. 93.

5. Michaels D. and Kendall B. (2021) ‘U.S. Competition Policy Is Aligning With Europe, and Deeper Cooperation Could Follow’. Wall Street Journal. 15th July, 2021; McKinnon J. D. (2021). 'App Store Competition Targeted by Bipartisan Senate Bill'. Wall Street Journal. 11th August, 2021; McKinnon J. D. and Kentall B. (2020). 'States to Move Forward With Antitrust Probe of Big Tech Firms'. Wall Street Journal. 19th August, 2020.

6. Michaels D. and Kendall B. (2021).

7. Khan L. (2017). pp. 713, 717, 746, 760, 779, 787, 788, 791, 799; Barwise P. and Watkins L. (2018). 'The Evolution of Digital Dominance' in Moore M. and Tambini D. (Eds.) Digital Dominance: The Power of Google, Amazon, Facebook, and Apple. Oxford: Oxford University Press. p.27.

8. Khan L. (2017); Ducci (2020). p.93.

9. Khan L. (2017); Daly M. (2020). 'Amazon Emails Show Effort to Weaken Diapers.com Before Buying It'. Bloomberg. 29th July, 2020.

10. Daly M. (2020).

11. Mattoli D. (2020). 'Amazon Scooped Up Data From Its Own Sellers to Launch Competing Products'. Wall Street Journal. 23rd April, 2020.

12. Khan (2017: 754).

13. Khan L. (2017). p.716.

14. Grady C. (2019). 'The 2010s were supposed to bring the ebook revolution. It never quite came'. Vox. 23rd December, 2019.

15. Hayek F. A., (1982), 'Rules and Order', in Law, Legislation and Liberty. London: Routledge. P.12-13.

16. Popper K. R. (1982). The Open Universe: An Argument for Indeterminism. London: Routledge. P.18-21.

17. Hayek (1982). p. 13.

18. Dawkins R. (2016), The Selfish Gene, 4th Edition, Oxford, Oxford University Press. p. 14.

19. Bentata P. (2021). 'Regulating “gatekeepers”: predictable “unintended consequences” of the DMA for users’ welfare'. SSRN; Portuese A. (2021). The Digital Markets Act: European Precautionary Antitrust. ITIF. 24th May, 2021.

20. Ducci (2020). p. 93.