This Report, which can be found here, was written by a working group who came together to address specific problems arising from the digital platforms’ reach, scale, scope, and use of data. They examined concerns stemming from the market structure contemporary platforms have created, and to investigate their competitive behaviour, including the consequences of network effects that can create barriers to entry for new innovators and entrench incumbents.
The theme that runs throughout the report is the difficulty of entry into digital platform businesses once an incumbent is established. Whether the entrant is vertical or horizontal, has succeeded to some degree, is nascent, is a potential entrant, or is a large platform in an adjacent space, market entry improves consumer welfare by either providing more choice, different features, and a chance of higher quality, or creating a threat that spurs the incumbent to provide lower prices, higher quality and innovation, and to do so more quickly.
The Report is structured as follows.
A first section discusses the characteristics of digital markets.
Digital firms have been able to acquire significant market positions and preserve them over time. Many of the most innovative internet-derived digital markets, such as search engines, social networks, network operating systems, ecommerce, and ride-sharing, are highly concentrated and have been dominated by one or a few firms for a number of years. The lack of entry of competitors in these important markets—despite high profits—suggests either barriers to entry or exclusionary conduct, or both. From an economic perspective, there is no single new characteristic that would make competition in digital platforms different from more traditional markets. Rather, it is the coincidence of several factors at a scale that has not been encountered before that makes the problem unique and requires new analysis of market structure and market power.
Digital markets often have extremely strong economies of scale and scope due to low marginal costs and the returns to data. Moreover, they often are two-sided and have strong network externalities and are therefore prone to tipping – a cycle leading to a dominant firm and high concentration. Digital markets are prone to tipping for two primary reasons. First, because fixed costs play such an important role in digital markets, these markets feature especially large returns to scale. Second, many digital markets are driven by network effects that strengthen large incumbents and weaken new entrants. Given this, after tipping the competitive process shifts from competition in the market to competition for the market.
This combination of features means many digital markets feature large barriers to entry. The winner in these settings often has a large cost advantage from its scale of operations and a large benefit advantage from the scale of its data. An entrant cannot generally overcome these without either a similar installed base (network effects) or a similar scale (scale economies), both of which are difficult to obtain quickly and cost-effectively. The very consumers who are harmed by these digital platforms act in ways that reinforce their market power, ironically generating additional barriers to entry. Consumers do not replace the default apps on their phones, do not scroll down to see more results, agree to settings chosen by the service, and take other actions that may look like poor decisions if those consumers like to choose among options and experience competition. Often the actions needed to generate choice for the consumer seem trivial, such as a download and installation, opening another app, or a few clicks. Consumers make these “mistakes” because of inherent behavioural biases such as discounting the future too much and being too optimistic. Behavioural economics helps explain how frictions in decision-making, and the fact that consumers can be manipulated to take advantage of their biases, render consumers sticky – i.e. people are slower to move to a superior product than they would be absent the manipulation. This in turn makes demand less contestable and less favourable for an entrant.
A sub-section looks at the role played by data in entrenching incumbent firms.
The role of data in digital sectors is critical. Personal data of all types allows for targeted advertising to consumers, a common revenue model for platforms. The report shows that the returns to more dimensions and types of data may be increasing, which again advantages incumbents. if the dataset is large enough that it allows a company to make accurate inferences about a given population, the company will always benefit from having specific information about a given individual, allowing it to become more and more confident about what the consumer wants, and to better tailor its services and ads. Companies therefore have no incentives to stop looking for and accumulating new pieces of data, entrenching incumbents with large datasets vis-à-vis entrants with smaller databases. Since many digital markets have tipped, often there are only a few entrenched platforms able to gather this breadth of data.
Consumer data in the United States is not regulated in any way that gives useful control or privacy to consumers. According to the Federal Trade Commission, the market for data suffers from a lack of transparency Most consumers have no idea how much information is being collected about them, sold, and used to make a profit.37 One way in which digital platforms exploit their market power is by requiring consumers to agree to terms and conditions that are unclear, difficult to understand, and constantly changing.
Another subsection looks at the economics of free.
Digital platforms are characterised by free services. “Free” is not a special zone where economics or antitrust do not apply. Rather, a free good is one where the seller has chosen to set a monetary price of zero and may set other, non-monetary, conditions or duties. It is possible that a digital market has an equilibrium price that is negative; in other words, because of the value of target advertising, the consumer’s data is so valuable that the platform would pay for it. As a result, barter is a common way in which consumers pay for digital services. They barter their privacy and information about what restaurants they would like to eat in and what goods they would like to buy in exchange for digital services.
However, in principle, that information has a market price that can be analysed. It is not easy to see if the value of any one consumer’s information is exactly equal to the value of the services she receives from the platform. However, many digital platforms are enormously profitable, and have been for many years, which suggests that in aggregate we do know the answer: the information is more valuable than the cost of the services. The economics literature has modelled this setting and is able to define a data mark-up. At the same time, the current inability to use both positive and negative prices for digital goods means that the policy discussion cannot focus on dollars alone as the unit of cost. Rather, digital platforms should be analysed using both price and quality. “Quality-adjusted price” is a metric often used by economists in this situation. If a platform’s price is fixed at zero and the quality of the service improves, then its quality-adjusted price has fallen. Conversely, if a platform’s price remains zero but its quality falls, its quality-adjusted price has risen. When the price is fixed at zero, it is possible to track quality-adjusted price over time: the movement in quality accurately reflects quality-adjusted price.
The next section discusses the various problems and harms caused by digital platforms.
There are many well-known problems that follow from lack of competition, including higher prices, less innovation, and lower quality in all its forms. Market power, consumer biases and an ad-supported platform model can generate significant consumer harms.
First, market power in advertising markets will result in mark-ups paid by advertisers. Ad-supported platforms’ high mark-ups provide a powerful reason to try and keep users online for another minute in order to show more ads. These profits push platforms to design their firms around “engagement”—an obsession with keeping users on their system for as much time, and with as much attention, as possible. The financial incentive created by a large mark-up can lead to anticompetitive behaviours as well as exploitative ones. Platforms may seek to reduce interoperability and awareness of outside options. For example, platforms may exclude certain services or increase friction in accessing third parties’ services. High search and switching costs are used to “lock in” users and reduce the ability of competitors to access those users. Platforms may adopt strategies to reduce multi-homing to obtain more market power over their users.
Secondly, while behavioural economists have studied consumer biases and firm responses in offline markets, these are swamped by what digital businesses can learn by using high-dimensional, large datasets to explore every nook and cranny of consumers’ many behavioural shortcomings and biases in real time. Framing, nudges, and defaults can direct a consumer to the choice that is most profitable for the platform. A platform can analyse a user’s data in real time to determine when she is in an emotional “hot state” and then offer targeted sales. These tactics reduce the quality of the zero-price content the user experiences on the platform. With big data and machine learning, firms are able to understand and manipulate individual preferences at a scale that goes far beyond what is possible in traditional markets. This capability is qualitatively new. The environment is characterised by extreme asymmetries of information and analytical capacity between the platform and the user. This enables firms to charge higher prices (for goods purchased and for advertising) and engage in behavioural discrimination, allowing platforms to extract more value from users where they are weak.
Furthermore, digital platforms are able to engage in sophisticated exclusionary strategies. In addition to de novo entry, platforms fear disintermediation by a partner or complement. A platform that has total control of demand due to control over framing of consumer choices, policies for complements, and technical standards can steer customers to content and complements of most benefit to it. Platforms understand that in some settings they can obtain higher margins if they either, make all of the necessary complements themselves, or, position themselves as a mandatory bottleneck between partners and customers. In particular, digital platforms are often very careful to maintain complete control over the user relationship so that they do not face any threat of disintermediation from a complement. These technological and policy choices can be used to reduce the possibility of successful entry by direct competitor. Other strategies such as exclusive contracts, bundling, or technical incompatibilities can also be used by platforms to restrict entry of competitors. Some of these strategies could be violations of existing antitrust law.
There is growing evidence that conglomerate digital platforms are in an advantaged position to stop or block entry by more focused rivals when compared to traditional businesses. Companies like Alphabet, Amazon, and Facebook operate in multiple business verticals (for example, mail maps, and search), collecting different dimensions of data on a consumer (for example, identity, location, and purchase intent) which give faster intelligence on competitive threats and new chinks in the platform’s competitive armour. This gives the platform an advantage over a rival entrant considering the same set of opportunities, and increases their abilities to exclude such rivals. A rival platform with similar economies of scope, data insights, and installed base may be a more formidable entrant. If large digital platforms have both the incentive and ability to purchase and block entrants that compete with them, or might compete in the future, the question is whether they have done so. The evidence that platforms have bought a series of potential competitors in recent years is anecdotal but fairly robust. The evidence that platforms have blocked potential entrants is likewise anecdotal. More formal research in this area is essential.
Lastly, insufficient competition and entry result in harms to investment and innovation. There is significant theoretical and empirical research that concludes that anticompetitive creation or maintenance of market power will cause a reduction in the pace of innovation. The lessening or blocking of innovative entry is of particular concern given its value to consumers. Engaging in successful innovation is certainly both feasible and common for a large platform with its enormous collection of data and other assets. These companies routinely spend large sums on R&D, launch new products and services, and are more able than other competitors to derive superior insights into how they should innovate based on the data collected from aggregating demand and advances in machine learning and advanced data analytics. However, the relevant counterfactual is whether the pace of innovation would be faster if platforms faced more robust competition
The subsequent section identifies some possible solutions.
The findings of this report suggest that rapid self-correction in markets dominated by large digital platforms is unlikely. Data sharing, full protocol interoperability, non-discrimination requirements, and the unbundling of content from a platform are all tools that the regulator, in conjunction with the antitrust authority, could apply and monitor over time in order to restore competitive markets. Antitrust enforcement is a possible response, but economists and lawyers will have to develop tools to explain to courts the role of behavioural biases in the creation of market power and in their effect on the quality of content. Just to give an example, the existence of zero money prices means that measurement of quality will be critical. The law needs better analytical tools to take into account the impact of potential and nascent competitors and competition. Market definition will vary according to what consumers are substituting between, whether there is competition on the platform between complements, or competition between platforms, or competition between a platform and potential or nascent competitors regarding possible future markets. Existing doctrines such as vertical foreclosure, predatory pricing and may need to be rethought in light of digital platform’s importance and business models. Perhaps most importantly, antitrust law might be revised to relax the proof requirements imposed upon antitrust plaintiffs in appropriate cases or to reverse burdens of proof.
While US antitrust law has long been flexible in combatting anticompetitive conduct, enforcement better suited to the challenges of the Digital Age may therefore require new legislation. This report suggests the establishment of a specialist competition court to hear all private and public antitrust cases which would allow judges to develop some expertise. However, because technology platforms present the enforcement challenges detailed above, even effective enforcement may not be enough to generate competitive digital markets in a timely fashion. Therefore, the report suggests that Congress should consider creating a specialist regulator, the Digital Authority, tasked with creating general conditions conducive to competition. The Digital Authority could have a mandate to create “light touch” behavioural nudges when they will make markets more competitive. An example of a regulation that would enhance competition is data portability. The Digital Authority could set up rules that allow users to easily port their data from one service provider to another and monitor compliance. The Digital Authority may also promote open standards in such areas as micro-payments and digital identities. Should Congress request it, the Digital Authority could oversee a mandate for interoperability in any market where market power has become entrenched and threatens long-term harm to competition
The committee also suggests separating out some types of regulation that will apply to virtually all market participants while other regulation will apply only to companies with bottleneck power. “Bottleneck power” describes a situation where consumers primarily single-home and rely upon a single service provider, which makes obtaining access to those consumers for the relevant activity by other service providers prohibitively costly. The Digital Authority could routinely collect data on digital transactions and interactions, with an emphasis on data from businesses with bottleneck power. These data – made public to the extent possible – would allow policy makers and researchers to assess the performance of the sector. Furthermore, some regulations could apply only to firms that meet the Digital Authority’s definition for bottleneck power. Because the cost of false negatives is high and there is uncertainty, the public interest requires the DA to take a more interventionist approach in these settings. The Digital Authority could have merger review authority over even the smallest transactions involving digital businesses with bottleneck power because nascent competition against these entities is very valuable for consumers. Non-discrimination rules could protect against a complement that is a potential competitor of the platform itself, or one that operates only on the platform as a rival provider of content.