Resume: At the end of the year 2020, Professor Nicolas Petit published an important book in the Oxford University Press. The book, entitled “Big Tech and the Digital Economy”, is already considered by academics and practitioners as a major work. The book brings a significant and brilliant contribution to the ongoing discussions on the concentration of digital markets and the domination of certain major digital companies.  In his study, Professor Petit introduces a new and original theory to understand the framework within which competition takes place between large tech companies. According to the author, “big tech” should not be defined as monopolies. In reality, they would be described as “moligopolies”. This change in terminology is far from insignificant. Because in the world of “moligopolies”, competition would be fierce. What are the implications for competition policy of such a paradigm shift?[1] Professor Petit gives some answers.

The book is available at the following link: https://global.oup.com/academic/product/big-tech-and-the-digital-economy-9780198837701?cc=us&lang=en&

 

1° Nicolas, first of all, I would like to sincerely express our gratitude for accepting this interview. Competition forum is sincerely delighted to have the opportunity to exchange with you about your new book: Big Tech and The Digital Economy. My first questions are quite simple. When did you decide to write this book and why did you take up the pen? As you well know, there is a huge amount of academic literature on the topic. What is the benefit of writing a book?

The idea came three years ago upon reading a fictional dialogue in Slate.  The two columnists asked what would happen if Google and Apple went to (actual) war? I found this funny. I often compare competition to war. But then I realized that there was more depth to the metaphor. If pundits compare Google and Apple to belligerents at war with each other, why do antitrusters spend their time talking about them as monopolists?

2° The central theory of your book, which is not exclusively dedicated to antitrust, is that Big Techs are “moligopolies”, not “monopolies”.  Can you explain this neologism in a few words? Can you also briefly describe the tools you used to come to this conclusion?

Moligopoly describes the situation of a multiproduct firm that holds a monopoly position in a market and at the same time takes competitive pressure from oligopolistic interaction in other markets. This configuration is ubiquitous in digital markets. Single product firms are the exception; diversified organizations are the norm.

Why do digital firms diversify? Efficiency is key. The data structures and infrastructures that constitute the backbone of digital markets can be combined and recombined almost endlessly to support value creation and growth. Economists often stress that digital markets have substantial economies of scope.

But business strategy also explains digital firms diversification. Incumbents know that new entrants will not challenge their market position directly. From a new entrant perspective, monopolies are too competitive. Incumbents enjoy extreme returns to scale on the supply and demand side. This state of affairs incentivizes new entrants to select an “indirect entry” path that consists in supplying highly differentiated products, complements, or “new combinations” of existing offerings. If incumbents want to limit survival threats, they must therefore partake in “broad spectrum” competition, and enter multiple adjacent markets at the same time.

This does not tell us the direction of diversification. That is, we cannot predict “where digital firms diversify”. Here, the logic followed appears more comparable to how venture capital firms invest. An economist colleague once told me: the set of diversification trajectories is in principle infinite, but the number of viable new products or adjacent markets is finite. The result is that digital firms tend to take bets by entering (and exiting) a large range of markets, even if the odds of success are minimal: payment, video, podcasting, healthcare, gaming, etc.

 

3°In the chapter on the moligopoly, you demonstrate that the competition between the GAFAMs (named MAGNAFs in the book) is very heterogeneous. Put simply, the competitive pressure that Apple would face would not be the same as the competitive pressure that Netflix would face. However, in the academic literature, particularly in antitrust literature, the expression GAFAM is used profusely. The expression suggests that we should look at these firms with the same magnifying glass. As you described in your book, it is a false impression. Hence, shouldn’t the acronym GAFAM be abandoned? In your opinion, does the use of this expression participate in giving a truncated vision of the reality of digital markets?

I would be happy to keep the GAFAM acronym (or another one), but with a less judgmental twist.

Two of the most salient features of most firms covered by the acronym are that they (i) hold large value/output positions in capital/antitrust markets; and (ii) are top R&D spenders. Most scholars, policymakers, and pundits focus on the first aspect and give short shrift to the second. I do not understand this. R&D behavior is relevant to market competition. Indeed, many first rate economic theorists like Hicks, Baumol, Leibenstein, and Fisher have stressed that what defines a monopoly is inertial behavior. In digital markets, observed levels of activity, risk-taking, and anxiety are incompatible with the quiet life of monopolists. Four of the five GAFAM appear in the top 10 of the latest European R&D ranking of the world’s top 2500 companies. Bottom line? Either GAFAM do not deserve the monopoly moniker, or some of them deserve to be looked at as an new type of efficient monopolists.

But there is a more general issue. I find it a tad bit irresponsible to sweep GAFAM’s colossal R&D efforts under the rug. The social gains produced by R&D investments are enormous. In a recent paper, Jones and Summers estimate that 1 USD spent on innovation investments generates between 4 and 20 USD in social returns. Given Governments’ limited ability to channel more taxpayers’ money towards expensive R&D programs, we as citizens should welcome that corporate organizations take over this task. This is a safe policy because the costs of R&D failure are ultimately borne by shareholders, not citizens.

 

4°When you address the economics of Big Techs, you use several known concepts such as network effects. You also insist on the notion of “tipping effect”. This notion, to my knowledge, is quiet unfamiliar to legal scholars.  Could you quickly explain what it means and what implications it has for Big Tech competition?

At a high level, a tipped market is one in which a product, service, or technology has “won the game.” The economic literature generally discusses tipping as a variation of the single supplier situation found in monopoly markets.

But there are richer definitions of tipping that insist on performance. A common understanding is that tipping occurs when a large enough installed base has been gained so that new users are more likely to join the product/service with the leading installed base regardless of the superiority of rivals. When that tipping position has developed, we shall observe reduced incumbent investment, inertial behavior, and a low degree of effort. The incumbent firm benefits from what Nobel Prize winner Jean Tirole has called in a different context “lucky demand conditions.” If this behavioral understanding is right, then disaggregated R&D, employment, and marketing data might allow a fact finder to draw inferences of tipping. We might, for instance, see firm-level differences between Google and Apple in app stores, even though both services are currently being discussed in monopoly terms.

One last remark. Tipping is a healthy process. Fierce competition “for” the market takes place during the tipping phase. Decision-makers should think twice before they adopt a prophylactic policy against tipping. During the tipping phase, firms experiment, make demand expansion expenditures (e.g., marketing), and hire. Tipping also leads to efficient outcomes by removing duplication, fragmentation, and uncertainty that delays entry of developers and adoption. Antitrusters know all this perfectly well. Most efforts undertaken by standard-setting organizations (SSOs) with the active support of policymakers consist in using cooperation (not competition) to generate “tipped” technology applications, systems, or products.

5° In Chapter 5 of your book, where you examine competition law in the US and the EU, you indicate that competition law should not seek to promote rivalry in digital markets. That there is equivalent pressure on these markets to encourage innovation and prevent the laziness of monopolists. But this pressure is not exactly a rivalry pressure. Then, you suggest a paradigm shift, if not a revolution. By the way, you are perfectly aware of this. According to you, Competition authorities should promote uncertainty, not rivalry. Is this not to empty competition law of its substance and putting an end to more than sixty years of ordoliberalism in Europe?  Moreover, by departing from traditional antitrust principles, do you not fear that competition law will no longer be seen by firms as a constraint? In my view, your proposal is very similar to saying that the digital market can self-regulate. Isn’t it the return of the “laissez-faire” doctrine? Isn’t it a paradoxical message at a time when digital markets are highly concentrated?

I am not a big fan of “laissez-faire”. But I am even less a fan of “faire anything.” Today, no rational observer of the policy that emerges can say what the EU wants to achieve (beyond bringing to an end excessive regulatory fragmentation at Member State level). Leaving aside the fairness rhetoric, the EU keeps making rote reference to the promotion of “competition.” In more recent texts, this is further specified as “contestability”. But these concepts are too general to be useful. There is more than one way to promote competition, and they are not necessarily consistent. The EU has so far not articulated if it means more competition by differentiation, commodification, disintermediation, or recombination.

Let me put the point differently. In my book, I say that indirect entry is the ordinary mode of contestability in digital markets. And I suggest that observed levels of indirect entry are quite high. Should a competition decision-maker be satisfied by this state of affairs and turn a blind eye on digital markets? Not necessarily. A case might be made that digital markets undersupply direct entry with substitutes, in a sort of public goods argument. If this is the case, antitrust intervention can help. But whether antitrust intervention is desirable rests ultimately on the empirical question whether the benefits of intervention outweigh its costs. In digital markets, market concentration produces supply and demand-side efficiency; hence, the answer to the question cannot be predetermined. A case-by-case investigation of market power failures in the particular case is required. One that moves us beyond the high-level observations of structural concentration, increasing returns, and incumbency advantages reported in many excellent reports commissioned by agencies these past years.

Unfortunately, only a few people seem willing to walk this extra mile in Brussels. A sense of urgency permeates the policy discussion. Several well-organized, well-funded, and well-resourced groups in the antitrust industry and beyond have managed to rally powerful voices in the press, the academia, and policymakers to the view that a precautionary approach – I have previously talked of an incipiency attitude– was the only way forward. As if a monopoly apocalypse was before us…

 

6° In your book, you quote a famous phrase from the economist Sir John Hicks: “The best of all monopoly profits is a quiet life”. Does this sentence equally apply to “moligopolies”? If not, which catchphrase would be most relevant for moligopolies?

I think I answered this before. Allow me the “quiet life” of not answering again.

 

7° More seriously, you criticize severely the European Commission’s methodology on the delimitation of relevant markets.  In particular, you attack the market share criterion. For you – I quote a passage from the book – “perhaps, the dominant market share of Google and Facebook is the result of persistent “competition seriously at work”” ; “As we often tell students : the point that Facebook has occupied a dominant share over a decade is irrelevant”.  If I understand correctly, you consider that the market share criterion should be abandoned. Instead, you propose, among other things, to make the departure between tipped markets and untipped markets.  Why would such a reform be relevant?

Market share analysis is a tool, not an end. The evaluation of market power should “recognize competition where, in fact, competition exists” and “should include all significant competition even though that competition differs in form or nature” (Transamerica Computer Company Inc. v IBM,481 F. Supp. 965). Drawing a line between tipped and untipped markets opens up a possibility to uncover competitive dynamics that are impossible to see under a market share analysis. If the goal of the antitrust inquiry is ultimately to understand how markets work, why should we dispense with a method that raises diagnosis accuracy at little cost?

 

8° I can’t help but ask you a question about fake news. You seem to consider that fake news does not currently raise competition concerns. But you report in the book that 62 per cent of Americans get news on social media. Does it seem unimaginable to you that a firm like Facebook would ever use its platform to denigrate a potential competitor? Couldn’t such behaviour be qualified as abuse of dominance?

Most legal systems have laws against denigration. These laws have broad applicability, including in economic matters. This suggests that all firms, not just those with market power, can and do denigrate. And indeed, many cases of denigration brought before courts involve firms without market power.

With this background, I guess the answer to your question ought to be that a dominant firm that denigrates its rivals is not abusing any “special responsibility”, power or influence that stems from its market position, just the ordinary responsibility bearing on any natural or legal person to abstain from slandering. There should thus be no liability under EU competition law.

 

9°Finally, I would like to have your opinion on a recent event affecting one of the GAFAMs: Google. What do you think of the complaint recently filed by the Department of Justice? Does it mark the ideological victory of the neo-structuralists? Do you believe it is a sign of a potential “Europeanisation” of antitrust law in the United States?

At this stage, this particular case is still a complaint, so it’s hard to talk of a victory. That said, the resemblance with the 2018 EU Commission decision in Google Android is unmistakable.

One factor that keeps confusing me is that often, in an antitrust case, the company paying “multibillion dollars” to another is the victim of monopoly, not its perpetrator. Put simply, the pocket in which the money goes is where the market power is. Why does this matter? On (legitimate) demand from Apple, Google, as well as many other firms, pays “Traffic Acquisition Costs” (TACs) to be a default on Apple’s products. As public data shows, Google has relentlessly tried to decrease the share of TACs paid to Apple. When this consideration is brought to the table, Google’s payments look less like costs of exclusion, and more like costs incurred because of competition from Apple in handsets, and incomplete Google vertical integration.

 

Interview by Walid CHAIEHLOUDJ

 

 [1]The interview was conducted before the Commission proposed new rules for online platforms. For this reason, the questions do not address the potential impact of Digital Service Act and Digital Market Act on the digital economy.