Executive Summary:

Through the various prisms of economic theory, market data, policy, and law, the study reveals the complex links between innovation and market concentration, the key trends of, and obstacles to, innovation, and ways policy makers can promote innovation in modern digital markets.
Our first angle is a familiar one -- the theoretical economic literature on market characteristics and innovation. The general economic consensus is that by delivering technological improvements, and new products, services, and business models, innovation forms a central pillar to efficient markets and a key to future prosperity and economic growth. Innovation processes can stimulate dynamic markets, enhance citizens’ welfare, and help offset otherwise diminishing marginal returns.
The relationship between innovation and market dynamics, however, has been subjected to a range of theoretical assumptions. Under the Schumpeterian hypothesis, market concentration is understood to allow internalization of the rewards flowing from innovation efforts (increase monopoly rents). Firms innovate to escape competition. Under the Arrowian hypothesis, competitive pressure forms the key to investment in innovation, and that significant market power disincentivizes investment in further innovation. More recent scholarship notes the complex relationship between market concentration and innovation. Notable is the inverted U-shaped relationship between competition and innovation levels. Philippe Aghion and his co-authors suggest that an increase in competition (from an initial low position) increases the rate of innovation, but that high levels of competition decrease the rate of innovation. The reason for the inverted-U shape is that when there is not much competition, firms have little incentive to innovate. Increasing competition, accordingly, will increase the average innovation rate. But once competition is intense, increasing the competitive pressure further may result in a slower average innovation rate. In addition, other variables may impact the investment in innovation, including industry and company characteristics and the political/industrial dimension.
After this familiar angle, we examine how innovative are many markets today. Our second angle offers a macro view of the current level of innovation in the EU and US. It focuses on the supply of innovation – that is the extent to which companies invest in research and development of new products, systems, and processes. While the dynamism of many digital markets may suggest heavy investment in innovation, macro data give rise to concern. From high above, it appears that competition is below optimal levels in many US sectors, and to a lesser extent, in the EU. The data from this angle suggests that many markets are becoming more concentrated and less competitive. Profit margins are widening, with a few firms reaping a significant share. Innovation levels also appear sub-optimal. The reduction in competition in the US, one recent economic paper points out, also coincides with a decrease in labor’s share of profits, a slowdown in output and GDP, a decrease in the startup rate of new firms, due to higher barriers erected by incumbents, and an increase in wage inequality.
One noteworthy study is the OECD Digital Economy Outlook 2017. The OECD acknowledges information and communication technologies as enablers of innovation but notes emerging signs that business dynamism and entrepreneurialism are falling short of their potential. The OECD further notes that while small start-ups are better placed to seize new opportunities offered by digital technologies, access to capital and high finance costs may undermine this potential.

Across the Atlantic, innovation also appears to be lagging behind its potential. The head of the Council of Economic Advisers under the Obama administration similarly noted a slowdown in the creation of new businesses, with top firms capturing more market shares. Of concern are signs that higher returns to capital have not been associated with an increase in investment. Businesses in markets with rising concentration and less competition are investing relatively less.
Several 2018 empirical papers also reflect these disturbing trends. There is a significant increase of markups between prices and marginal costs of publicly traded firms in developed economies. The rise in measured markups is associated with increased market power and market concentration. In line with the inverted U-shaped relationship, an IMF study finds that high markups are correlated initially with increasing and then with decreasing investment and innovation rates. This non-monotonicity is more pronounced for firms that are closer to the technological frontier. More concentrated industries also feature a more negative relation among markups, investment, and innovation.
So from our macro view, we see that increased concentration levels and less competition are generally associated with greater profit margins, but not greater investment in innovation. In fact, indicators suggest a decrease in investment, in line with an inverted U-shaped relationship.
Our third angle looks at several emerging trends in the digital economy. Among the key characteristics, noteworthy are the use of data as a key resource for innovation in the digital economy and the ongoing investment in Big Data. Data acts as a significant engine for innovation, but can also act as a barrier in inhibiting entry and growth. We observe a positive feedback loop that may help powerful firms become stronger, as the weak get weaker. Beyond data, the exponential growth of the Internet and mobile communications has seen a proliferation of platforms that often act as intermediaries and as such occupy a central junction for users and service providers. The access to data on users and suppliers places platforms in a favorable position that, at times, act as gate keepers in industries characterized by network effects. Data-driven network effects, may at times, tilt the market in favor of a single winner, which thereby is significantly protected from competitive pressure.
These trends can directly impact citizens’ welfare and choices. They may facilitate control over the users’ interface. Moreover, they may enable providers to affect the use of, and access to, competing services, increase friction in switching to alternatives, reduce awareness of outside options, and promote the platform’s own services. Through the use of personal data and advanced algorithms, platforms and their suppliers may control to a greater extent the digital paths seen and used.
So, what is the price we might pay if a dominant platform suppresses some types of innovation? Our fourth angle examines the implications of sub-optimal innovation levels. At least two perspectives emerge. The first, being narrow, acknowledges that many markets today may not be as innovative as their current potential, but views this as a transitory state. Policy decisions today can be used to affect future levels of investment in innovation and help optimize markets for innovation. The second, wider perspective, offers an evolutionary view. The level and nature of innovation, being path dependent, may not necessarily return to their natural state. Under this evolutionary perspective, current impediments to innovation can affect not only future levels of innovation but also the types of innovation. Basically, some types of innovation may be lost forever. As a result, today’s policy decisions affect not only future levels of investment, but also the paths for innovation and the nature of innovation. This view puts greater responsibility on policy makers to preserve competitive portals, which can have a crucial role to play in the shaping of tomorrow’s innovation.
Given these potential stakes, what can policy makers do to promote innovation in the digital economy? We offer policy makers three key perspectives – the supply of innovation, the demand for innovation, and the nature of innovation.
Our fifth angle assesses the variables that affect the supply of innovation. The supply of innovation, as we synthesize from the literature, will likely depend on four key variables: market contestability (markets need to remain contestable for innovation to flourish), appropriability (the extent to which a firm can capture the value created by its innovation and protect the competitive advantage associated with it will increase the incentive to innovate), synergies (for instance, the combination of complementary assets necessary to engage in R&D will enhance the ability to innovate), and the nature of innovation.
From this angle we can see the complex relationship between market structure, and the levels and nature of innovation. No optimal ratio exists among these variables to increase the supply of innovation. Some degree of market power, at times in some industries, serves as an incentive that stimulates innovation (appropriability). But while greater concentration might result from a firm’s welfare-enhancing innovation, one cannot say that increasing market concentration, by itself, will necessarily spur welfare-enhancing innovation. As we also see, innovation can continue to occur in heavily concentrated markets, but the nature of innovation might change. For example, open systems, relying on user-driven innovations, might slowly close after a few firms dominate the industry. Users, rather than develop and modify products and services for their own use, rely instead on the dominant firm’s innovations. Finally, the primary beneficiaries from the innovation might change. Innovation may simply reinforce the dominant platform’s power and user lock-in.
After viewing how these key variables can affect the supply of innovation, we switch perspectives to explore the demand for innovation. Using five stages in the users “innovation-adoption process,” we identify how large platforms can influence the demand for, and rate of adoption of, different Adoption of several key technologies in the past took decades. The good news is that with dominant platforms, the adoption rate for some technologies can be shortened to years, if not months. But just as powerful platforms can help users through the five stages in deciding to adopt an innovation, they can increase barriers in one or more of these stages, thereby impeding the technology’s adoption. Tactics used to thwart an innovation’s adoption may include limiting the potential user’s exposure to competing technologies, the use of defaults to take advantage of status quo bias, or the use of data-advantages to reduce users’ likelihood of adopting competing products or technologies. Among our examples is how Google and Apple successfully thwarted for years ad blocking technology for smartphones.
The insights from this sixth angle illustrate how a powerful gatekeeper can influence users’ adoption of innovations. As a result, one should not solely focus on contestability, appropriability, and synergies that affect the supply of innovation. Policy makers must also consider the pathway of innovation from the angle of user adoption of that technology. Dominant firms can reduce the demand for, and adoption of, technologies, even when markets are contestable, synergies exist with other innovative products, and the dominant platform does not seek to appropriate any gains from that technology.
We often assume increasing innovation levels improves our collective well-being. But does it? Our seventh angle looks beyond the veneer of innovation. From this vantage point, we consider how characteristics of the digital economy may impact the nature of innovation. Because dominant platforms can promote some innovations, while thwarting other innovations that threaten their dominance or business model, one might ask whether innovations are always good? Does increasing the level of innovation necessarily increase overall welfare? Not always.
In examining the nature of innovation, we describe three categories of innovation: positive, negative, and mixed. We explore several examples of this negative innovation. Firms employ these innovations to maintain or obtain monopoly power without benefitting consumers. At times, they use this negative innovation to transfer wealth from consumers to themselves, or to exclude competitors. From this angle, we consider how changes in market characteristics may impact the nature of innovation and the possibility of it being exploitative, exclusionary, or cannibalistic.
Several key takeaways emerge from this perspective. First, the nature of innovation may take a path that runs against societal goals and benefits a few at the expense of many. Second, increasing the overall level of innovation will not necessarily increase overall welfare. Third, while policy makers generally do not want to chill the incentives to innovate, some types of innovation should be chilled. Fourth, policy makers cannot assume that market forces or regulators will generally deter negative innovation. Some types of negative innovation may be kinds of innovation.

beyond the scope of antitrust, privacy, or consumer protection law. Even when they aren’t, enforcers may be overly deferential to the claimed innovation. Finally, developing the tools to determine when innovation is positive, negative, or mixed, what conditions foster the myriad forms of negative innovation, and implementing policies to deter negative innovation will be critical.
Thus, the goal for policy makers is to not simply increase the overall level of innovation, as that will not necessarily increase overall welfare. Ideally, the regulatory framework would reduce firms’ incentive/payoffs to engage in negative innovation, while promoting (or at least not chilling) their incentive to invest in innovations that generally promote overall welfare. So what is the recipe to achieve this balance?
We would caution policy makers about anyone peddling a simple recipe. In our final part, we review several of the available policy instruments used to facilitate innovation. Inevitably, the level, nature, and direction of innovation may be influenced by a variety of regulatory policies, including in the digital economy, privacy, consumer protection, competition and state aid, education, taxation, intellectual property, access to capital, and property law. Thus, boosting positive innovation requires a comprehensive policy approach.
With these challenges in mind, we explore the benefits and limitations of several available policy and enforcement measures. We consider it preferable to focus future intervention on ex-ante measures – aimed at creating a regulatory and economic landscape, which helps open the competitive portals for positive and welfare-enhancing mixed innovation. Even then, one should be aware that any form of intervention night influence the identity of the winners and losers of tomorrow. Ex-post, case-by-case intervention should be limited primarily to instances when actions by companies are clearly in breach of existing legal regimes such as competition, privacy, consumer protection, or intellectual property laws.
Whenever reaching into the tool box, policy makers should assess the challenges and risks associated with intervention or the lack of it. It therefore underscores the need for a measured and careful approach. On one hand, excessive intervention comes at a cost and could chill innovation, hinder disruptive positive innovation, undermine investment, increase the burden on smaller operators, and determine the likely winners and losers. On the other hand, non-intervention should not be seen as benign, as it too reflects a policy decision on the likely winners and losers under the status quo, and may be detrimental to welfare-enhancing innovation. The goal is to optimize the preconditions for welfare-enhancing innovation, while accounting for the legal, business, technological, and market environments.
Ultimately, there is no easily available recipe for policy makers on how to promote the good forms of innovation, while deterring the bad. Indeed, one key takeaway from this paper is that such a recipe is illusive, and would likely need to be continuously updated. In a rapidly evolving environment, the task is far from simple. Nonetheless, the seven angles outlined herein can help policy makers refine their tools to promote welfare-enhancing innovations. But in this pursuit, it is worth reminding ourselves that innovation, like competition, is not an end to itself. It is simply one of many means to promote overall well-being. Citizens may sacrifice innovation, at times, to further other, more important, values, including privacy and autonomy, necessary for our well-being.