Innovation Matters: Competition Policy for the High-Technology Economy, Richard J. GILBERT

Richard J. Gilbert

This section selects books on themes related to competition laws and economics. This compilation does not attempt to be exhaustive but rather a survey of themes important in the area. The survey usually covers publication over the last three months after publication of the latest issue of Concurrences. Publishers, authors and editors are welcome to send books to for review in this section.

The idea that American markets are too concentrated is now widespread. The suggestion was initially supported by neo-Brandeisian advocates, who especially blamed “big tech” and “big pharma” for benefiting from too much market power. They generally perceive this situation as the result of lax merger review policies having allowed for anticompetitive acquisitions. While such ideas are controversial, potential shortcomings of antitrust enforcement in these sectors are now debated within the entire antitrust community. In his book Innovation Matters: Competition Policy for the High-Technology Economy, Richard J. Gilbert acknowledges that competition law enforcement is partly flawed in the high-tech industry. This, he believes, is evidenced by a lower rate of start-up creation and a rise in market concentration. To tackle these issues, Gilbert does not discard the current framework. Rather, he proposes moving from an innovation-centric perspective to current price-centric enforcement.

In his analysis, the author highlights the importance of elements that are often overlooked, such as innovation dynamism, spillovers or firms’ technological capabilities. He combines a large diversity of sources to reach his conclusions, ranging from economic and managerial theories to findings of empirical studies and in-depth analyses of case law. He denounces some traditional analytical tools for failing to properly encompass innovation aspects. For instance, defining relevant product markets ignores, by nature, potentially upcoming products.

The book firstly presents the general theoretical framework of interactions between competition and innovation incentives before zooming in on mergers: theoretical and empirical research is produced, supplemented by case studies. This is followed by a focus on single-firm conducts affecting innovation, addressed by analysing the Microsoft and the Google Search cases. Both having been brought in the European Union (EU) and in the United States (US), Gilbert compares the decisions reached in both regions. Before concluding, the author tackles the issue of designing an antitrust policy for standards.

To illustrate the contrary dynamics shaping innovation incentives, the author firstly engages in an extended review of existing theoretical frameworks on innovation incentives.

He begins with the two influential analyses on the links between competition and innovation by Joseph Schumpeter and Kenneth Arrow, which draw opposite conclusions. For Schumpeter, market power stimulates innovation; it prevents wasteful duplication of expenditure, facilitates access to external capital and enables stabler platforms for R&D investments. On the contrary, according to Arrow, a monopolist has profit flows at risk from innovation, which is not the case for a competitor. This potential “replacement effect” of old technologies by new innovative ones discourages monopolists from investing and innovating.

For Gilbert, both these models overlook many crucial features. For instance, phenomena such as buyout prospects or portfolio effects are ignored. If B, the innovation, is a complement to A that the company already sells, then the company may lower the price of A to increase the sales of B, internalising the gains and increasing consumers’ welfare. This is known as the “Cournot effect,” and highlights that a monopolist can decide to innovate for a variety of reasons. Besides, incentives also depend on whether the innovation can be shared (as opposed to winner-takes-all), or whether it is drastic (as opposed to incremental). Technological lags between competitors equally play a decisive role. Moreover, spillovers and the cumulative dimension of innovation are omitted. Follow-on innovations are central to an economic ecosystem dynamism.

Managerial and organisational theories can also usefully integrate the scope of analysis. For instance, successful firms tend to maintain strategies that have worked in the past and focus on consumers’ immediate needs. They might therefore overlook waves of radical innovations. It is also more likely that an undertaking will innovate if doing so does not require organisational adaptations.

Turning to empirical studies, Gilbert shows that while results are sometimes contradictory, they reflect the diversity of features involved in innovation processes.

Having set the general analytical framework, the author zooms in on merger policy. He provides reading guides for understanding the potential effects that a merger can cause and analyses the effects of different types of remedies in the long run to assess their efficiency innovation-wise.

Gilbert distinguishes three types of mergers, requiring different kinds of antitrust inquiries. The first one is “product-to-project.” Firm A produces a product that might overlap with the potential product that firm B is currently developing. At present, challenging such an acquisition requires proving that the unmerged undertaking would have managed to enter the market—for Gilbert, this is too high a burden. A non-negligible probability that the competitor would have gone through with its innovation should suffice. Absent credible efficiency justifications, there is no opportunity cost of enforcement, even if the innovation does not work out in the end. On the contrary, clearing the merger can lead to terminating or significantly delaying a project and/or reducing price competition.

The two other types of mergers are “project-to-project” mergers (both firms are engaged in investing towards similar applications) and overlapping R&D capabilities (the projects may not be concrete yet but undertakings could have the same goal). For those, Gilbert lists a series of questions that should be answered systematically in merger reviews. For instance, do a few undertakings still compete in the relevant R&D market? Do any of the merging parties increase their profits at risk because of the merger? How appropriable are the new products? How intense are spillover phenomena in this industry (the lower the riskier)? Interestingly, he also asks whether the acquisition was a necessary prospect to incentivise the acquired firm in the first place. This illustrates his aspiration to better consider buyout effects on innovation. While often coined as “killer acquisitions,” they also incentivise innovation. Many start-ups do not attempt to compete, nor do they want to; they are created with the intent to be bought from the start. Restricting possibilities for such acquisitions could lead to a reduction rather than an increase in innovation.

Gilbert then analyses the impact of different kinds of remedies in the long term. He refers to cases in which unconditional challenges were decided. To review these mergers, enforcers used analytical tools such as relevant innovation markets and took into consideration capability features. In his opinion, these decisions were successes: entities kept investing afterwards or were bought in alternative operations that did not raise anticompetitive concerns.

The author next investigates structural remedies and draws insightful conclusions for antitrust policy. He shows that recipients can fail to make full use of divested assets because they are not prepared, interested, or informed enough. While such remedies can be extremely effective, the author emphasises that success depends on distinctive features. First, assets should be in a late stage of development and include any related necessary elements, such as related intellectual property, specialised manufacturing equipment and key personnel. Moreover, the recipient should show a strong record in relevant R&D activities. Enforcers must consider such decisive elements when designing structural remedies.

Lastly, Gilbert challenges the common idea that the behavioural remedy of compulsory licensing undermines innovation incentives for products that are easily copied. In most cases, he argues, such an effect does not materialise. Usually, merged companies maintain their patent rates while recipients increase their own. Such a tool must, however, be handled with care; over-resorting to it could end up disincentivising innovation.

These chapters provide the reader with a convincing and extensive analytical framework, while encompassing theoretical and empirical dimensions. The author highlights therein the powerful and contradictory forces shaping innovation and delivers concrete policy tools. Gilbert points at some of the deficiencies of the present framework. Difficulties to delimitate relevant innovation markets or overly high burdens of proof requirements, incompatible with the inherent degree of uncertainty of innovation-related matters, cause enforcement to be blind both to potential efficiencies likely to be passed on to consumers and to potential harms that are not acknowledged.

The subsequent chapters use a comparative approach to learn lessons from the Microsoft and the Google Search cases that were adjudicated in both the US and the EU and in which single-firm conducts were at stake.

In the Microsoft case, American and European enforcers found the company to be liable. In the US, Microsoft was considered guilty of having monopolised the market for licensing Intel-compatible PC OS. In the EU, the company was held to have abused its dominant position by unlawfully tying Windows Media Player to the OS and making its own products incompatible with rival ones. Microsoft was required, inter alia, to disclose all necessary interoperability information. After a few years, these remedies appeared to fail in both the US and the EU. The author uses this example to highlight the difficulty of regulating such remedies. Enforcers must be highly specific when defining the scope of information that must be made available by the incumbent. Otherwise, the company is likely not to fully cooperate and competitors cannot make effective use of the information they receive. Moreover, with this case, Gilbert emphasises the importance of protecting nascent competition, however likely it is to succeed. Innovation follows a path that is difficult to predict in any case so antitrust enforcement must keep the door open for potential evolutions.

In the Google Search case, Gilbert presents the different approaches to product-design evolutions in European and American antitrust law. He concludes that both analytical frameworks are incomplete. In the US, Google’s decision to change its ranking algorithm to better match users’ demands, or so it said, was considered an improvement to the product—no further competitive assessment was undertaken. In the EU, Google was found to have abused its dominant position: its behaviour risked foreclosing competitors and reduced competition incentives. The Commission paid little attention to whether the change was an improvement for consumers.

Under the cover of evolution in design, undertakings may purposefully exclude rivals from markets. To distinguish between legitimate and anticompetitive situations, Gilbert proposes a “truncated rule of reason.” He thus differentiates between substantial and modest innovations. Substantial innovations unaccompanied by any exclusionary conduct separable from the modified product or technology should be presumed lawful, whereas modest innovations should be analysed under a full rule of reason. As such, under exceptional circumstances, a “modest” modification could be proven anticompetitive even without exclusionary conduct accompanying it. Gilbert acknowledges the complicated operationalisation of the rule given that distinguishing between substantial or modest evolutions, while ambiguous, becomes crucial.

According to the author, these cases reveal flaws in both American and European reasonings. The US system should envision the possibility of less restrictive alternatives and scrutinise the validity of business justifications. Absent technological merit or legitimate pro-competitive business reasons, the exclusionary effects should have been considered as offsetting benefits. And in the EU benefits for consumers should at least have entered the scope of analyses.

Before concluding, Gilbert addresses the complex relationship between antitrust and standards, which have contradictory effects on innovation. He suggests some avenues to improve the adequacy of the framework. For instance, he underlines that liability rules have sometimes allowed owners to receive compensation unrelated to economic contribution of patented technologies. He therefore proposes to apportion the value of a standard among the patents necessary to practice it: if a hundred patents are required, then each should account for 1% of the standard’s economic value.

This book is illuminating in delivering a thorough analysis and providing extensive detail to support the argument. Ranging from merger reviews to current shortcomings in standard policies, the study is comprehensive in its survey of antitrust effects on innovation. No argument is made without presenting a technical account of sub-possibilities where the mechanisms at stake could be reversed. Such complexity renders the design of policy recommendations difficult, yet Gilbert meets this challenge. He delivers very practical analytical tools as well as general provisions that ought to be kept in mind in enforcing innovation-related antitrust. This book should be put in the hands of policymakers, enforcers, and academics alike. It establishes a new research agenda for both empirical and theoretical enquiries on the topic of innovation and competition from an interdisciplinary perspective.

PDF Version



Isaure d'Estaintot, Innovation Matters: Competition Policy for the High-Technology Economy, Richard J. GILBERT, February 2021, Concurrences N° 1-2021, Art. N° 98940, pp. 269-270

Publisher MIT Press

Date 14 July 2020

Number of pages 336

Visites 75

All reviews