The Digital Services Act, the Digital Markets Act, and the New Competition Tool
November 10, 2020
This paper reviews ongoing deliberations surrounding the EU Digital Services Act (DSA) and Digital Markets Act (DMA) and discusses possible implications of upcoming legislation for European innovation, the business models of U.S. companies serving European customers, and global digital services standards.
Currently, it appears that, in lieu of the Digital Services Act, the European Commission will present a package of two legislative proposals: the DSA, intended to strengthen the EU single market and clarify digital services’ responsibilities and liabilities, and the DMA, which will tackle the economic power of large online platforms (LoPs). The DMA will include ex ante obligations (prohibiting and requiring certain behaviors before regulators have evidence of actual harm in the marketplace) for LoPs acting as “gatekeepers,” as well as the New Competition Tool.
Anticipated Contents of the DSA and the DMA
In late January 2020, the European Commission announced its intention to pass a DSA in the fourth quarter of the year as part of its Work Program 2020. Currently, the e-Commerce Directive, issued in June 2000, is the main document regulating digital services’ liability, and there is widespread agreement that it needs updating. The Commission aims to strengthen the EU Digital Single Market, clarify and increase digital services’ liability rules, and ensure a level playing field for small- and medium-sized businesses (SMEs). While the DSA package helps clarify the rules for most digital services, it could also give rise to enhanced obligations for LoPs and, very likely, U.S. champions. There are concerns that such an approach would not increase trust and safety online, but rather lead to a migration of illegal content, goods, or conduct to smaller, less regulated platforms, which might also have fewer resources to handle such problematic content.
Unapologetically, Commission officials have been clear that their overarching regulatory goal is to limit the market power held by LoPs acting as “gatekeepers”—a term that the Commission has yet to define. Margrethe Vestager, the executive vice-president of the European Commission for A Europe fit for the Digital Age, made the Commission’s view clear on October 29, just weeks before the final DSA and DMA proposals are expected to be released on December 2. “We’ve come to a point where we have to take action. A point where the power of digital businesses—especially the biggest gatekeepers—threatens our freedoms, our opportunities, even our democracy . . . So for the world’s biggest gatekeepers, things are going to have to change.”
As such, the DSA and the DMA will have a transformative impact across the board on large U.S. corporate tech champions and their business models of delivering internet and digital services to 500 million European consumers. Most believe that, at the very least, Google, Amazon, Facebook, and Apple will be targeted; however, the Commission has not made the scope of these two pieces of legislation clear. It is also unclear whether LoPs from China such as WeChat will receive the same scrutiny as the U.S. giants—despite WeChat being considered a “super-app” for its dominance across services as well as its massive and growing user base.
Unapologetically, Commission officials have been clear that their overarching regulatory goal is to limit the market power held by LoPs acting as “gatekeepers”—a term that the Commission has yet to define.
Defining Platforms and Intermediaries
A platform business facilitates the interaction of two or more groups, and an online platform uses the internet and digital tools to improve that interaction. A business-to-consumer (B2C) platform connects a business or businesses on one side of the market to consumers on the other side, a classic example being the yellow pages, which facilitates interaction between individuals and businesses. Online intermediaries include businesses that do not sell their own products or services but connect sellers and buyers. A business-to-business (B2B) platform connects businesses with other businesses to generate value. One such example is Easy Spares MarketplaceTM, a mobility industry marketplace jointly launched by Mirakl and Siemens Mobility and which offers materials for the mobility industry, from trains to components, from Siemens and other trusted suppliers.
In early June 2020, after months of stakeholder consultations and internal debate, the Commission released an inception impact assessment for each of two pillars of the original DSA. Pillar One focuses on ensuring trust and safety online by increasing responsibilities, obligations, and liabilities for digital services. Pillar Two deals with the establishment of an ex ante regulatory instrument to control the behavior of gatekeepers. According to press reports, this latter instrument will likely include prohibitions or restrictions of certain business practices that the Commission believes should be “blacklisted,” “whitelisted” obligations on platforms to modify their business practices to facilitate competition against themselves, and remedies that would be applied on a case-by-case basis to LoPs. More detailed impact assessments on each pillar were leaked in September. In the fall, the Commission also merged the second pillar with a separate legislative item, the New Competition Tool, to create the DMA—leaving the DSA with only the liability pillar.
As such, the DSA and the DMA will have a transformative impact across the board on large U.S. corporate tech champions and their business models of delivering internet and digital services to 500 million European consumers.
In the DSA Inception Impact Assessment, the Commission discusses potential new policies pertaining to the liabilities for digital services. This document outlines its objective of clearing up legal uncertainty via implementing a modern legal framework for digital services, strengthening the EU Digital Single Market, ensuring that digital service providers act responsibly to eliminate risks to their users in order to protect their civil rights and create a safe and trustworthy online environment. Notably, there is no agreement among stakeholders on what exactly these risks are. For example, it remains unclear whether the DSA will target only illegal content or take on harmful content that undermines respect for democratic values. Further confusing the picture, there is no common EU definition for illegal content, harmful content, or democratic principles and values, and each member state has its own sensitivities in defining those terms. Similarly, while the liabilities pillar also aims to cover the sale of counterfeit goods and services online, there is no clear definition of a counterfeit.
Based on press reports and conversations with informed observers, it appears that the Commission is leaning toward a revised liability regime for all services in the “internet stack.” New obligations would include: horizontal due diligence, which are basic obligations all platforms would be expected to meet; a notice-and-action mechanism to make platforms aware of any illegal content being offered or hosted and of their obligation to tell the poster that content has been removed; requirements to “know your business customer,” which obligate platforms to collect information on traders; requirements regarding transparent content moderation, which obligate platforms to report on measures taken, outcomes, removal rates, complaints, and use of automated tools for moderation; and cooperation with authorities.
Additional transparency regulations are expected for advertisers and recommender systems, which suggest items to users. These obligations could apply to all online platforms, including those established outside (but with users in) the European Union, although it is not clear how the Commission would enforce them in such instances. Failure to apply liability obligations to firms that serve EU citizens but are not established in the European Union would create enforcement gaps, with implications for competition and safety. For example, Amazon would be within the scope of the DSA, but a company such as Alibaba would not. Obligations would not be uniform across all member states, as LoPs would have asymmetric, robust obligations—such as reporting and data access, audit and co-regulation, or access to advertisement1—proportionate to their potential to pose risks to society or businesses.
In a more business-friendly vein, the DSA Impact Assessment also recommends removing “legal disincentives for hosting services, including online platforms, to take additional, proactive measures to address illegal activities and information shared by their users.” Under current EU law, a platform is not liable for illegal content if it is unaware of hosting it, which has disincentivized platforms to voluntarily act on content. The proposal should clarify that “such measures do not, in themselves, take them outside the scope of the liability exemptions for online intermediaries.”2
The gatekeeper pillar follows from the Commission’s declared view that European antitrust regulators have been mostly unsuccessful in addressing the systemic problems that LoPs allegedly generate. There is concern that traditional European companies and industries are having trouble competing with large U.S. and Chinese online platforms, and that traditional competition policy is not equipped to deal with issues related to contestability, fairness, and innovation with respect to online platforms and digital markets.3
The inception impact assessment pertaining to ex ante regulations for “gatekeepers” outlines the Commission’s objective of improving competition and innovation across the EU Digital Single Market by regulating LoPs that act as gatekeepers in the digital services sector. The Commission also hopes to mitigate the imbalance of power between the gatekeeping platforms and their users to fight off the alleged monopolistic tendencies of the former. The Commission describes the situation as follows:
“A small number of large online platforms [LoPs] increasingly determines the parameters for future innovations, consumer choice and competition. Consequently, Europe’s estimated 10,000 online platforms are potentially hampered in scaling broadly and thereby contributing to the EU’s technological sovereignty, as they are increasingly faced with incontestable online platform ecosystems. This leads to a risk of reduced benefits from social gains deriving from innovation. These outcomes of platform dynamics may result in large-scale unfair trading practices and potentially reduce the social gain from innovation. Their impact is compounded by the opacity and complexity of the large online platform ecosystems, and the significant information advantage such platforms have over regulators.”
The Commission has identified three main problems stemming from LoPs that act as gatekeepers: (1) unfair practices by gatekeeper platforms vis-à-vis business users and competitors, (2) structural competition problems that undermine effective competition and market contestability, and (3) fragmented and ineffective institutional oversight and enforcement. Unfair practices include self-preferencing, anti-steering, unfair tying and bundling, and no or limited data sharing.4
Structural risks to competition could include market situations in which a firm takes advantage of online markets’ unique characteristics, such as network effects due to data accumulation or restrictions on multihoming, to dampen competition.5 Further lack of competition could occur when a market is failing to operate properly for structural reasons, due to high entry barriers for competitors or high switching costs for users.
Political momentum behind developing the proposed DMA, which includes the New Competition Tool, has been driven by widespread impatience among Commission officials, members of the European Parliament (MEPs), and constituents with evidentiary burdens and the slow process of established anticompetitive behavior under current European antitrust law and regulations. They see the “opacity and complexity of the large online platform ecosystems” as hindering effective enforcement. As their control over platform ecosystems grows, “gatekeepers” are viewed as an anticompetitive threat by virtue of their size, access to data, and their corresponding ability to leverage data gathered in one area of activity to improve, develop, or offer new services in adjacent markets. The Commission claims that the existing regulatory environment will result in “a small number of online platforms” determining “the parameters for future innovations, consumer choice, and competition” while preventing small, European online platforms from scaling. The Commission bases its views on the assumption that innovation will come only if new entrants and smaller companies are able to challenge large (mainly U.S.) online platforms. Through the DMA, the Commission seeks to create a fair trading environment and increase the innovative potential of online platform ecosystems in the EU Digital Single Market.
Based on conversations with informed observers, it appears that the Commission will pursue a new, ex ante regulatory framework for gatekeeping platforms. The framework would prohibit or require certain “blacklisted” or “whitelisted” practices and solicit regulatory intervention regarding less egregious, “graylisted” practices. A leaked Commission document laying out blacklisted, whitelisted, and graylisted practices suggests that the Commission is strongly considering restrictions on self-preferencing practices, bundling and tying practices, and gatekeepers’ exclusive use of certain data.6 The leaked document also proposes additional transparency and reporting requirements for gatekeepers, as well as data-sharing and access requirements. Some of the prohibited practices track with complaints the Commission has raised in antitrust proceedings against U.S. tech companies.
It is expected that the prohibitions of certain conduct and the imposition of proactive obligations will apply mainly to major U.S. platforms, not their smaller European or Chinese competitors, which may offer similar services.
In past years, competition enforcement in major, developed markets has trended toward a more economic and evidence-based approach focused on regulating anticompetitive conduct and promoting efficiency rather than on promoting a certain level of market concentration. Of particular concern to U.S. LoPs are the prohibitions and obligations that would limit LoPs’ ability to justify pro-competition, efficient, and welfare-enhancing conduct. Shifting the focus of intervention from ensuring efficiency to advancing market-structure objectives would push competition law in a historically rejected direction toward a structural approach (“big is bad”) that favors smaller, European competitors. While the leaked blacklist document outlines behavior that should be prohibited or required, it does so under a framework that only applies to “large” online platforms—which the Commission has yet to define. In other words, a company’s size will determine whether the new set of ex ante competition rules apply to it. That approach would ignore the dynamic competition that LoPs bring to the market, the consumer welfare generated by the existing framework, and the innovation and investment incentives necessary to generate future technological breakthroughs. The Commission’s decision to regulate based on as-of-yet undefined characteristics of firms instead of their behavior may lead to unjustifiable and arbitrary discrimination against U.S. companies in violation of World Trade Organization (WTO) rules, as discussed below.
These changes to the EU competition law system aimed at benefitting European incumbents and startups and should be examined to determine whether they are discriminatory, unfair measures of industrial protectionism.
The Commission views interoperability among online platforms as a means to establish a more competitive digital economy. This would allow users to choose between platforms more easily, enabling multihoming and preventing users from being locked into one platform due to strong network effects and a lack of competition. It should be recognized that interoperability can also be used by competitive rivals to free ride on the infrastructure and investments made by a platform operator in its proprietary ecosystem.
The Commission suggests that interoperability requirements could be one “tailor-made remedy” to address anti-competition and unfair practices by LoPs on a case-by-case basis.7 The leaked list of blacklisted, whitelisted, and graylisted practices confirms that the Commission views interoperability requirements as a component of its response to gatekeepers. For example, regulators could require gatekeepers to ensure business users, competitors, and “interested third-parties” have access to data provided, received, or generated while using the gatekeepers’ platform or service. The document also suggests that regulators could require gatekeepers to improve interoperability and data portability for business users and consumers seeking to use another platform or service. These provisions would build on the right to data portability within the EU General Data Protection Regulation, which grants users the right to receive data they provided an organization. Ostensibly, this would make it easier for competitors to enter the market by giving them access to the incumbent’s infrastructure and resources and by allowing users to move between online platforms and services more easily.
This view is shared by the European Parliament Committee on Internal Market and Consumer Protection (IMCO) and the European Parliament Committee on Legal Affairs (JURI), which suggests that mandated interoperability is likely to be among the specific remedies offered to the Commission by the DSA.
Government-Mandated Access to Data
The Commission’s approach to data as it relates to the DMA rests on the same assumptions listed above. In short, the Commission believes mandatory access to data for competitors is at the heart of unlocking competition between less established online platforms and dominant ones. In addition, the Commission could also require that data be made available to government authorities, researchers, and other third-party auditors. The leaked list of (potentially) unfair practices will require companies considered gatekeepers to increase other businesses’ and consumers’ access to data collected by the online platforms.
Under the blacklist/whitelist practices of the DMA’s proposed ex ante regulatory framework, gatekeepers are prohibited from using data collected on their platform unless this data is made accessible to other business users active in the same commercial activities, just as they are prohibited from using data received from business users for advertising services for any other purpose other than the advertising service. Gatekeepers must offer users the choice of combining off-site data with on-site data.
Under the graylist practices in the leaked document, gatekeepers: (1) cannot prevent third-party sellers from accessing essential information that gatekeepers collect on customers; (2) cannot restrict business users from accessing or using data that they provide, receive, or generate while using the gatekeeper’s platform; (3) must take the appropriate steps necessary to enable business users and consumers using their platform to use other platforms, either through enhanced portability of data or enhanced interoperability; and (4) must share “search and click” data related to search advertising at a “nominal price.”
Data-Sharing Challenges: GDPR, Innovation, and Other Costs
If the Commission mandates data sharing for LoPs, it risks creating conflicts with the General Data Protection Regulation (GDPR), which will be a serious concern for businesses large and small engaged in data exchange. The GDPR’s core principles of transparency, data limitation, minimization, storage limitation, and integrity and confidentiality all present challenges to data-sharing obligations that the DSA and the DMA might require. One solution may be to anonymize and encrypt user data subject to these acts’ sharing requirements. That solution, however, would mitigate the effectiveness of data sharing among large and small firms, as detailed individual user data is most valuable to online platforms and service providers and is key to improving services and innovation. The GDPR’s purpose-limitation, data-minimization, and storage-limitation principles may also undercut the competitive value of data sharing.
Per the purpose-limitation and data-minimization principles, if shared data can only be used for the purpose that it was originally collected, there is a risk that the original collectors (mainly LoPs) would maintain a significant first-mover advantage because competitors would be limited in how they could use the shared data. If data must be shared as soon as it is collected, and if it is collected with an expressed broad purpose, the first-mover advantage would be eliminated—but at the expense of discouraging competing firms from establishing their own data-collection operations.
The storage-limitation principle raises questions over how long non-anonymized personal data may be kept and made available to competitors. Smaller companies receiving user data from LoPs would need to ensure they have the tools and personnel to navigate the GDPR’s interaction with DSA-mandated data-sharing arrangements, which could require significant resources.
Mandated data sharing would also disincentivize both large and small companies from gathering and processing data in welfare-enhancing ways. Small firms now capable of free riding on large firms’ efforts would see little reason to invest in their own data-gathering capabilities. Forced to share their data, large firms would have fewer incentives to improve their existing data-gathering capabilities because the value of that data would not be exclusively theirs. Ironically, the Commission itself has stressed the value of gathering data for firms of all sizes but is embracing an approach that would discourage companies from doing so.
In addition to innovation costs, there are also likely to be costs associated with operationalizing such a data-sharing scheme. For data to be shared and useful, new infrastructure would need to be established, including agreements on uniform data formatting. These changes could create cybersecurity challenges, as large amounts of valuable user and business data would be pooled and made available to a variety of actors.
Stakeholder Reaction to the DSA Package
Companies and stakeholders have commented extensively on both Inception Impact Assessments. Regarding the DSA section, supporters note its potential to create legal certainty and deepen the European internal market while making liability and consumer-protection requirements more robust. Other voices have raised concerns over where the burden of liability and safety will fall and worry that illegal content would migrate to smaller, less regulated platforms.
New Liability and Responsibility Regulations
As discussed above, this pillar aims to remedy societal and economic risks generated by digital intermediary services; establish better supervision of digital intermediary services and harmonize regulation and cooperation across the European Union and its member states; and smooth out member states’ legal barriers preventing SMEs from scaling and growing cross-border operations. Stakeholders agree on the need for updated liability obligations, strong consumer-protection rules, and harmonized regulation across the member states.
Stakeholders expressed three primary concerns regarding this pillar: (1) how liability rules will impact fundamental rights such as freedom of expression, (2) how illegal and harmful content will be defined and treated according to their different associated risks, and (3) how the “proportionality principle” will be implemented without increasing burdens for SMEs who lack resources.
It is necessary to clearly define illegal and harmful content (if the DSA does cover harmful content) at the EU level, as the former will be subject to more stringent regulation than the latter. Without clear definitions, digital service providers and intermediaries will be held to opaque and unreasonable standards. Confusion about what constitutes illegal content could lead service providers and intermediaries to wrongfully crack down on some types of content, which would harm fundamental rights such as freedom of expression and opinion. Currently, the definitions of illegal and harmful content vary based on national law, and ambiguous definitions of either term in the DSA would create an unpredictable regulatory environment for all digital service providers in Europe.
Because companies vary in size and capacity to moderate content, a one-size-fits-all approach to a new liability regime could result in smaller platforms and intermediaries facing a larger regulatory burden compared to LoPs. Therefore, the Commission appears to be leaning toward an asymmetric approach in which LoPs have greater obligations—such as reporting, data access, and auditing—because of their influence and reach.
The Commission aims to create a safe and trustworthy online environment and avoid unintended consequences, namely the fragmentation and migration of illegal content, goods, or conduct to smaller, less regulated platforms. Thus, it also needs to address the other side of the equation: how new liability requirements will impact SMEs that have less capacity to monitor content and activity on their platforms. The Commission does need to extend its proportional approach from LoPs to SMEs.
As mentioned above, stakeholders have also expressed the need for the Commission to remove policies that disincentivize companies from being aware of the content they host, a concern the Commission seems to have taken on board.
Ex ante Framework for LoPs
With regard to the gatekeeper section, the Commission expresses the laudable goal of reducing regulatory fragmentation and promoting consistency among member states, which will allow SMEs to compete across a more coherent European digital market. Many companies and organizations agree with the Commission that, on top of transparency and interoperability requirements, additional regulation is needed to ensure competition, promote fair markets, and incentivize innovation.
However, three major concerns exist regarding the current ex ante regulatory proposals, as evident in the public comment submissions: (1) a lack of clarity in definitions and scope of the framework, (2) the risks of regulatory overlap, and (3) the risks of reducing the benefits that LoPs can provide. Additionally, stakeholders have voiced concerns that an ex ante framework would result in the Commission undertaking competition policy by regulation, which would rob targeted companies of due process, potentially harm consumer welfare, and set a dangerous precedent for other countries considering using competition policy hand in glove with industrial policy.
Advisory Reports in European Parliament
On October 20, the European Parliament approved non-binding reports from the Committee on Internal Market and Consumer Protection (IMCO), the Committee on Legal Affairs (JURI), and the Committee on Civil Liberties, Justice, and Home Affairs (LIBE), which included various views and recommendations for the DSA and DMA. The IMCO report includes calls for platforms and online intermediaries to: (1) improve their efforts for online consumer protection; (2) establish a European notice-and-action mechanism, which does not currently exist at the EU level; and (3) respect the right of consumers to be informed if a service is enabled by artificial intelligence (AI). The report also advocates establishing an ex ante regulation to prevent market failures caused by big platforms’ anti-competitive behavior rather than remedy it after the fact.
IMCO called for including specific rules regarding targeted advertising and micro-targeting based on personal data collection and for enforcing them more strictly, which would encourage advertising that does not require extensive tracking of users’ interaction with content. The committee also recommended avoiding a “one-size-fits-all” approach. Instead, the DSA should distinguish between economic and non-economic activities and between “different types of digital services hosted by platforms rather than focusing on the type of the platform.” For example, a social media platform such as Facebook should not receive different treatment than a media-sharing platform such as YouTube, because they are different platforms. Especially as they increasingly offer a range of additional services, it becomes hard to parse them as distinct types of platforms. Instead, the DSA should focus on the services offered, of which there could be considerable overlap.8 In the committee’s view, all digital service providers established in third countries should be required to adhere to DSA rules when directing their services to consumers or users in the European Union.
The IMCO report further recommended that platforms and online intermediary services improve their efforts to detect and remove false statements and that online marketplaces inform consumers once a non-compliant product they purchased has been removed from the site. Another suggested provision, the “know your business consumer” principle, would require platforms to monitor and stop fraudulent companies from selling their illegal and unsafe products and content using their services.
The JURI report called for harmonized standards across the European Union for how platforms should moderate content and employ notice-and-action mechanisms. Users should have options for redress if their content is taken down, and those proceedings should occur before an independent judiciary and not a private entity. The MEPs on the committee recommended that enforcement be carried out by an EU body or network of national authorities, and that they be given the power to monitor platforms and impose fines if liability rules are breached. In line with the views of most stakeholders, JURI also called for the Commission to clearly distinguish illegal content from harmful content. The committee recommended new rules regarding targeted advertising, including giving users the option to opt out of being served ads based on personal data and the ability to use online services anonymously. This recommendation is controversial: there has been no assessment of its impact, and it could threaten SME growth by choking off a key revenue stream.
As did the Committee on Legal Affairs, LIBE recommended that the DSA clearly distinguish between illegal and harmful content, and that users be able to opt out of micro-targeted ads that make use of user data and algorithms to curate content, which the committee claims is in line with GDPR obligations. The committee also warned that automated tools for analyzing content are imperfect and may result in content being taken down unnecessarily. And, as the other two committees did, LIBE recommended that liability rules be harmonized at the EU level.
Once the European Commission publishes the final DSA and DMA proposal, the European Parliament and the European Council of Ministers will have to agree on the proposal, then adopt the legislative act, and task each national government with adapting laws to achieve the newly established goals.
What Is Driving the DSA Package?
Observers have posited that the global economy has entered a “superstar era” in which the ability to scale up differentiates firms that thrive from those that fail, with these successful firms reaping disproportionate gains. It is striking the degree to which European firms have fallen behind international competitors in the global digital services market over the last fifteen years. In 2020, Europe was home to about 20 percent of the world’s top 500 companies, compared to the United States’ share of 24 percent and China’s share of 25 percent. In 1995, 30 percent of the world’s top 500 companies were from the European Union, equal with the United States and well above China, which had just two companies in the top 500. Europe’s percentage share of large global companies in the top decile dropped from about 50 percent in 1995 to just 16 percent in 2016. Today, despite a market of over 500 million consumers and a sophisticated workforce, Europe lacks indigenous global digital platforms of significant size, in part due to the fragmentation of the EU Digital Single Market.
This gap is especially evident in the technology sector. Europe is home to only three tech companies in the Fortune Global 500, compared to 12 from the United States, five from China, and six each from Taiwan and Japan.9 Europe lags behind Asia and the United States in terms of investment in technology as well. In 2019, European tech companies received $34.8 billion in capital investment, compared to $62.5 billion invested in Asian companies and $116.7 billion invested in U.S. ones. However, the scale of investment in European tech companies has risen for the past five years—total investment in 2015 was just $15.3 billion. In 2019, 174 European tech companies had a valuation of over $1 billion, compared to just 13 in 2010. Of those 174 companies, 99 are backed by venture capital. The most investment in 2019 went toward financial technology, enterprise software, health, energy, transport, food, and marketing—sectors in which, excepting enterprise software, European companies have long been globally competitive. Of note, capital invested in UK tech companies has grown from $4.8 million in 2015 to $11.1 million in 2019, double that of capital invested in Germany and France and roughly triple that invested in Sweden.
This suggests that—contrary to the Commission’s pessimistic view of the status quo—investors see some promise in European tech companies, even though Europe clearly trails the United States and Asia as a place to invest. Diverging views between the Commission and the private sector over Europe’s competitiveness in the technology sphere are apparent in survey data as well. The data suggest that Europe’s tech ecosystem has seen limited recent success despite increasing regulation from the Commission, and that new regulation is not necessary to make Europe a more attractive investment location. Most of the European technology sector’s founders, employees, and investors see startups as most impacted by regulatory burdens. Only one in five European tech company founders believe European policymakers respond to startups’ concerns and views, and over 40 percent believe regulation makes it more difficult to found a company in Europe and scale it. Just 32 percent believe that the “direction” of European tech regulation is positive for the European tech ecosystem.
However, a majority of founders agree that European startups should receive preferential treatment from European policymakers, and that regulation is an opportunity for Europe to gain a competitive advantage in attracting investment and talent. Nearly 60 percent of investors and venture capitalists believe European tech startups can compete equally globally, while just 38 percent of European policymakers and regulators share that view. Similarly, European policymakers and regulators are more pessimistic than investors about whether European tech companies will gain ground relative to the United States and China over the next ten years, and they are more pessimistic about the future of European technology in general.
The Decline of Europe’s Position as Leader in the Global Digital Economy
Europe’s overall relative decline in innovative digital services is characterized by several trends. First, European startups have struggled to scale up into major companies, especially in the digital and health sectors. Europe’s success in turning startups into “unicorns” (private startups valued at $1 billion or more) has occurred at about half the rate as in the United States because investors are relatively scarce, as is their capital. Second, as digital technologies become a driver of performance for global firms, Europe has become less central to international digital trade flows due to the virtual absence of European companies in the digital platform space. Third, Europe has a declining share of global research and development (R&D), particularly in the digital sector. Only two-thirds of digital potential was reached by European firms compared to their U.S. counterparts—in other words, European firms are far less digitized overall than U.S. firms.10 Research and development spending by European software and computer services firms was roughly 8 percent of the global total, compared to 11 percent for Chinese companies and 77 percent for U.S. companies.
It is evident, and indeed clearly expressed by European officials, that the new DSA measures are designed to stem Europe’s innovative decline and increase European global competitiveness, but there is little empirical evidence that shackling global competitors who sell to European consumers and businesses with new restrictions will achieve the intended result of getting EU firms to grow Europe’s rate of innovation and global competitiveness in digital services. By constraining how the most innovative online service providers and intermediaries can operate in Europe, the Commission is decreasing the chance that the next disruptive and innovative breakthrough will occur in Europe.
Analysis of the Economic Dynamics Surrounding LoPs
Gatekeepers: The Benefits of LoPs
The Commission’s underlying assumptions about the impact large online gatekeepers have on innovation, and the digital ecosystem deserve scrutiny.
First, online platforms support SMEs and businesses through a variety of ways. LoPs increase the ease with which startups can learn about, share, and profit from good ideas; provide application programming interfaces and software development kits; and lower costs of intermediation and ease restrictions on market entry, thus creating otherwise nonexistent markets by allowing sellers to tap into a huge customer base.
LoPs also make it easier for new firms to gain an immediate online presence and access new sources of funding, such as through crowdfunding sites. Non-digital SMEs can also benefit from LoPs: a local restaurant can make use of payment systems they offer and easily appear on a map and reviews app, for example. By connecting developers with more users and sellers with more buyers, LoPs also facilitate greater competition among firms. This results in increased market efficiency, lower costs, and more innovation.
Second, LoPs provide large economic benefits domestically and internationally. The Organization of Economic Cooperation and Development (OECD) explains that greater market access for retailers through these platforms allows businesses to contribute more to GDP and that greater competition across markets leads to lower prices, greater production and consumption, and higher growth. Additionally, the LoPs contribute to GDP growth through their independent businesses.
LoPs likewise boost international trade by allowing SMEs who previously engaged only in domestic commerce to expand into foreign markets. For example, according to the OECD, “Approximately 300,000 third-party sellers who participated on Amazon’s Marketplace platform exported goods from the United States to other countries in 2017. . . . Most of them were smaller, domestic businesses that rely on Marketplace’s digital presence outside the United States to reach foreign buyers.” This game-changing access to customers around the globe extends to SMEs in developing countries as well, as LoPs promote economic development abroad.
Third, LoPs encourage internal innovation via two main channels: product development and business models. Regarding product development, large companies are increasingly investing more in R&D relative to smaller companies—in 2014 alone, nine U.S. platforms received 11,585 patents. The difference between mean annual R&D spending of large and small companies grew from less than $20 million in the 1980s to almost $120 million in 2017. This is likely because successful innovation often requires large, upfront, fixed costs that small firms with low-profit margins are unable to finance.
Online platforms also accelerate innovation by tapping into a potentially unlimited pool of external innovators, or an “innovation ecosystem.” For example, Apple’s App Store is an innovation machine that produces content at a speed and scale that Apple could not achieve by itself. In 2019, estimated billings and sales facilitated by the Apple App Store ecosystem in Europe was $51 billion (€46 billion).
New and improved business models are one of LoPs’ most influential innovations, as these can have far-reaching effects across industries. LoPs often introduce “disruptive innovations” that do not just add on to existing practices within an industry but drastically change the way in which an industry functions. For example, the disruptive innovation of Uber permanently changed the taxi market, and Spotify, a Swedish company, changed the music industry.11
Do LoPs Promote or Prevent Competition?
While the effects of LoPs on SME competition have been heavily debated, evidence suggests that LoPs engage in a different form of competition than traditional models reflect and which drives them to constantly engage in innovation and R&D.
A New Form of Competition
When a new company enters a market with a disruptive innovation, technology, or business model, that company threatens existing firms. This is particularly prevalent across LoPs. However, the OECD explains that, in the long term, disruptive innovation is beneficial, as less efficient firms exit and are replaced by more efficient firms. This process increases consumer welfare and raises productivity—as long as it occurs within a competitive environment. This begs the question: are LoPs operating within the constraints of a fair and competitive market?
LoPs may simply engage in a new form of competition, called “Schumpeterian rivalry,” in which a succession of temporary monopolies displace one another through innovation. This is not a traditional model of competition in which several companies contend with each other mainly over price. Rather, this dynamic competition is based on innovation and occurs gradually.
Earlier evidence explains that obtaining scale is much less difficult than sustaining scale. As a result, LoPs face competition in the form of potential innovators who will come in and displace the incumbent platforms. LoPs are thus encouraged to attract third-party sellers, retain users, create new markets, and invest in R&D to stay ahead of their competition.
Concerns surrounding firms’ monopolistic tendencies are also overblown for several reasons. First, while larger firms may have a larger market share, the “constant pressure to reinvest revenues in order to keep up with the pace of innovation keeps them from exercising much market power.” In fact, from 1972 to 2012, industry concentration was positively correlated with productivity and output growth but not with price changes. This belies the Commission’s view that concentration—in this case, “gatekeeping”—inevitably produces anticompetitive outcomes. Monopolies or market concentration caused by innovative “superstar” firms or economies of scale are linked with falling prices and increased output—not rising prices, less consumer choice, or less consumer welfare.
Second, increased firm size and industry concentration can result in lower industry-wide costs, as economies of scale are essential. For example, a European study found that “high-tech firms’ capacity to increase the level of technological knowledge depended on their size. In these cases, size did not lead to excessive profit margins.”
Competition with SMEs
Despite the many benefits that LoPs offer for SMEs, small companies are increasingly caught in a “small size trap” in which scaling up to medium-sized and large companies is becoming more difficult. Until 2000, 15 to 20 percent of small companies became medium-sized or large companies each year. By 2017, half as many scaled up.
However, some SMEs continually break through the trap and grow to be large companies. The Harvard Business Review attributes the success of scaled-up SMEs relative to those stuck in the “small size trap” to higher investment in intangibles, more debt incurred to finance investments, fewer annual losses, higher asset bases, and younger startup age. These factors, however, are relatively independent of LoPs, suggesting that the EU’s proposed regulation will not address the root causes of low startup scale-up in Europe.
Through cooperative interoperability, LoPs enable SMEs to scale faster and with fewer resources. SMEs are able to leverage tools and services offered by LoPs to tap into massive existing user bases with relative ease. Often times, particularly for technology companies, the ability to scale up can make or break a company. In return, LoPs have continued to innovate to attract new businesses to their platforms, creating a virtuous cycle of innovation. Schumpeterian competition exhibited by LoPs has resulted in SMEs only needing to tap into a few LoPs to achieve scale. Spotify’s partnership with Facebook was significant for the then-upstart, Europe-based music platform because it allowed Spotify to reach a massive amount of users in one move. Spotify increased its attraction to users by partnering with other platforms such as Uber and PlayStation. Regulation that aims to disrupt Schumpeterian competition in order to increase the number of sizeable platforms competing concurrently would create a fractured ecosystem for SMEs and raise scalability costs. As scalability is king, these higher costs would reduce investment in SMEs and discourage entrepreneurship. Furthermore, fractured platform ecosystems can reduce competition between SMEs. Firms could opt to appear on only a few platforms they believe will provide the largest returns at the lowest additional cost, leaving would-be competitors to operate with less competition on the remaining platforms.
In terms of startups that aim to compete with LoPs, establishing a distinct set of additional requirements for platforms that reach a certain size could discourage small ones from scaling. That dynamic would limit investment into startup and small platforms and would make acquisitions of SMEs—a key exit strategy for investors and startups—less attractive. New rules targeted at platforms of a certain size may not be necessary to combat overwhelming network effects that allegedly lead to winner-take-all situations. SMEs can differentiate themselves by providing distinct services, superior quality, or catering themselves to a particular audience. For example, the neighborhood social media platform Nextdoor has seen significant user growth and has expanded its offerings to include advertising and a user-to-user marketplace. Nextdoor’s growth occurred despite the social media platform space, apparently being saturated and dominated by the likes of Facebook and Twitter.
B2B and Traditional European Companies
Even though the Commission seeks to ensure space for European platforms to play to their traditional strengths, such as in the B2B and industrial space, the Commission’s analysis and proposals thus far have focused on B2C platforms and lack clarity as to how B2B platforms fit into the proposed regulation. Despite some similarities, B2B platforms operate differently than B2C platforms and have different business models. Competition concerns associated with B2C platforms are less likely to arise in the context of B2B platforms. B2C platforms tend to have short supply chains, needing only to connect a user with a service—such as a rider with a driver or a reader with an advertisement. Industrial B2B platform supply chains are longer and have more players with different roles. Each of those players holds much more power over the platform than individual B2C users. A single rider cannot determine the viability of a ridesharing platform, but a single supplier may hold significant sway over a B2B platform if its output is otherwise hard to come by.
B2B e-commerce is projected to reach at least twice the size of the B2C e-commerce space by the end of 2020 and dominate e-commerce growth through 2025. Almost half of B2B businesses currently sell products online, and almost 90 percent are expected to do so by 2025. In addition, some B2C companies are beginning to jump into both industrial and e-commerce B2B space.
B2B e-commerce platforms help businesses manage orders, products, customers, revenue, shopping carts, and payments. B2B industrial platforms can be defined as “a collection of shared infrastructures, systems and processes, where the value of the platform is based on data sharing and data exploitation. . . . Data from industrial production processes, product usage and operations - the ‘data footprint’ of manufacturing, logistics and aftersales services - play an important role in Industrial B2B platforms.”
Amazon has recently begun a move into the e-commerce B2B space using a similar model to its B2C operation—a platform where sellers can connect with buyers. Like its B2C operation, Amazon takes a cut of sales, and sellers have little control over how buyers use the platform. Shopify, Alibaba, and OpenCart have also made moves into the B2B e-commerce space.
Amazon, Google, and Microsoft are also major players in cloud services, which underlie many industrial B2B platforms and the Internet of Things (IoT), including industrial IoT applications. However, because of the range of operations, an industrial B2B platform would need to engage in to dominate a range of industries, the industrial B2B market also remains fragmented.
In Europe, certain industries are ramping up online industrial B2B platforms. Over half of the top companies in Germany are turning to platform models, and most are focused on B2B or a mixed model. B2B industrial online platforms are most effective if they are multisided and enable direct interaction among multiple participants, which allows for scalability via network effects. IoT platforms that enable automation are the best example. They connect machinery from different manufacturers and allow third parties to utilize data to enable services that are part of the industrial ecosystem.
The Commission should ensure that its final regulations do not result in regulatory burdens on B2B platforms that fundamentally operate differently than other digital service providers. Otherwise, B2B platforms may be held liable for illegal content posted on a B2C platform that it supports or enables through backend services or infrastructure. In that case, the B2B platform would have little to no control over the content posted on the B2C platform and would be unduly burdened by enforcement or regulation. Cloud service providers, for example, are legally and technically limited in their ability to address content—including illegal content—stored by their customers or users. The DSA should instead focus on digital intermediary services that facilitate dissemination of content, products, or conduct to the public at the direct request of the content provider or business user and have the technical ability to remove specific pieces of content.
Confusion over B2B and B2C liability and responsibility requirements could dissuade new B2B platforms from entering the industry due to either overly burdensome regulations or concerns over lack of clarity. On the other side of the equation, imposing new liabilities on B2B service providers may make them less willing to partner with SMEs that do not have the capacity to police the content they host.
Ex Ante Pillar
If the DMA’s application to B2B platforms remains unclear, both U.S. and European B2B platforms will be left wondering if they will face ex ante regulation in Europe for accumulating market power and cementing themselves as popular platform service providers.
Given how fragmented the B2B platform ecosystem is, regulation appears to be premature and may discourage new entrants or discourage platforms from expanding, which would hurt competition in the long term. For comparison, U.S. companies accounted for 68 percent of the market capitalization of the world’s top 100 platforms, while European companies accounted for just 3 percent and companies in the Asia-Pacific for 27 percent. About two-thirds of the top 100 platforms provide B2C services, and five companies—Apple, Microsoft, Amazon, Facebook, and Google’s parent company, Alphabet—together account for roughly 35 percent of the market cap of the top 100 global platforms. This suggests that the B2C space is relatively saturated and less fragmented than the B2B space. However, while Apple, Microsoft, Amazon, Facebook, and Alphabet provide B2B services, they do not dominate that space as they do for B2C. It is simply too early to know if current B2C competition dynamics will occur in the B2B space.
For example, data interoperability or data portability might not be acceptable for B2B platforms or their customers—both of which may view their data as proprietary. That data could reveal business secrets, intellectual property, or know-how.
The DSA Package and the WTO
Would new ex ante requirements for online platforms that meet certain characteristics, such as firm size, user base, or market share, be consistent with Europe’s obligations under the WTO General Agreement on Trade in Services (GATS) not to discriminate against foreign services or service suppliers?
The proposed DSA and DMA target “large online gatekeepers”—a category apparently designed to target foreign service suppliers. These firms provide services covered by commitments in the European Union’s GATS schedule, such as advertising, data processing, database services, market research, retailing and mail ordering, and computer and related services.
Discrimination under the GATS occurs if two like services or service suppliers are treated differently. The Panel in EC–Bananas III found that “to the extent that entities provide . . . like services, they are like service suppliers.” The panel in Canada–Autos followed this reasoning in determining that because there were no like Japanese and Canadian service suppliers, there we no like services being supplied.
Panel reports and the Appellate Body of the WTO have not determined the likeness of service suppliers based on the characteristics of the firms supplying services, for that approach would make it nearly impossible to find that any two “like” service suppliers. The panel and the WTO would have to grapple with how to draw lines based on firm size, technology used to provide the service, ownership structures, or other distinctions. A characteristics-based approach to determining the likeness of services suppliers would shrink the scope of the national treatment obligation.
In Argentina–Financial Services, the Appellate Body wrote, “We consider that the concept of ‘likeness’ of services and service suppliers under Articles II:1 and XVII:1 of the GATS is concerned with the competitive relationship of services and service suppliers. This is consonant with the Appellate Body’s understanding of ‘likeness’ in the ambit of trade in goods.” Consider the companies and services targeted by the DSA and the DMA. One could claim that the environmental focus of German search engine Ecosia or the privacy features of French search engine Qwant makes them unlike Google’s search engine. But that claim rings hollow. At their core, each delivers web search services regardless of its ancillary offerings. A car is a car regardless of its gas mileage or whether it has an electronic advanced driver-assistance system on board. And two different grocery store chains would likely be considered like even if one offered curbside pickup and the other did not. In our view, the size of search companies such as Ecosia and Qwant—or the size of their user base—also does not differentiate them from Google, Bing, DuckDuckGo, or any other search competitor for purposes of a “like services” definition.
Regardless of the relationship between services and service suppliers, the fact that the European Union is targeting certain U.S. companies to assist its own companies concedes that competition and substitutability already exist between the services supplied by U.S. gatekeepers and those supplied by European companies, as well as the service suppliers themselves. This de facto discrimination may give rise to a WTO challenge. It could also lead to calls for a Section 301 investigation by the United States Trade Representative akin to that undertaken in response to France’s digital services tax proposal, which similarly de facto targets U.S. tech companies.
The documents discussed here, both official and leaked, lead observers to conclude that the European Commission is well on its way to embracing a decidedly anti-United States regulatory assault against large online digital platforms. Recognizing that there is a debate in the United States over the ability of antitrust law to moderate the behavior of ascendant companies in the digital economy, Europe, for its part, seems to be set on rushing to enact a raft of measures despite little empirical evidence that the new regime of tech regulation will be effective in achieving Europe’s objective of engendering “tech sovereignty.” Ex ante regulations are unusual, require labor-intensive enforcement, and are “poorly fitted for sectors that are rapidly evolving.” As such, the DSA package poses a real threat of suppressing innovation in Europe.
The digital economy, in which large actors often compete against each other as much as they lead and assert their own business models, has nurtured powerfully effective forces of competition. These firms are fundamentally transforming how individuals pursue personal goals and interests and how business is done globally. From what can be observed, the new digital ecosystem is enabling and nurturing innovation in Europe, allowing European entrepreneurs and businesses to grow and prosper as they improve the lives of EU consumers.
Since 2010, Amazon has invested €55 billion ($64 billion) in Europe, hired 115,000 permanent employees, and supported 200,000 jobs by allowing independent businesses to sell on Amazon Marketplace. Additionally, 990,000 companies and creative businesses have built their businesses with Amazon Marketplace, Amazon Web Services, and Kindle Direct Publishing. In January 2020, Facebook estimated that around 25 million businesses in Europe—mostly small ones—use Facebook services each month and that Facebook apps helped European businesses generate roughly €208 billion ($244 billion) in sales and create 3.1 million jobs in 2019. The apps also helped European businesses generate €40 billion ($47 billion) of export sales outside the European Union. In 2019, estimated billings and sales facilitated by the Apple App Store ecosystem in Europe were $51 billion. In May 2020, Microsoft announced it is investing $1.5 billion in Italy, including plans to build a regional set of data centers there. This investment could generate more than 10,000 jobs and over $9 billion in direct and indirect economic benefits. Additionally, Microsoft intends to create new digital training programs for local businesses (focusing on SMEs), offering Microsoft’s cloud services to roughly 500,000 companies and startups. It will also add new locations in the “Polish Digital Valley,” Greece, and Austria, where Microsoft has committed to training roughly 120,000 Austrians by 2024. This envisioned cloud services region is expected to generate up to $2.1 billion in new revenue in Austria over the next four years.
Europe, for its part, seems to be set on rushing to enact a raft of measures despite little empirical evidence that the new regime of tech regulation will be effective in achieving Europe’s objective of engendering “tech sovereignty.”
LoPs underpin a dynamic digital ecosystem that is supporting the success of many diverse businesses in Europe. Disassembling this ecosystem with the proposed DSA and DMA package threatens to have the unintended effect of dampening healthy job creation and harming Europe’s ability to achieve its “tech sovereignty” goals of ramping up innovation, entrepreneurship, and global competitiveness in the digital economy.
Restraining only U.S. online platforms in Europe effectively raises barriers to entry for U.S. exporters, even as it shields EU industry incumbents with import substitution and other protections borrowed from industrial policy. In the face of a comprehensive political decision to target five U.S. companies, it can be expected that the U.S. Congress and administration will be concerned that not only are U.S. companies being marked for stricter regulation, but they are also being restrained while Chinese companies such as Alibaba, WeChat, and Tencent will apparently be free to pursue their business objectives in Europe.
Under the new competition policy being contemplated in the DMA, the Commission seems intent on moving dramatically away from widely accepted antitrust enforcement principles, in which accused firms are afforded due process and the opportunity to offer evidence in defense of an allegation of dominant behavior. Targeting a specific group of companies identified by their size and success in the European market seems a violation of normal judicial protections. The effect would be to nullify corporate rights, including the legal precept that a law or regulation should address a specific behavior, not an individual or an arbitrarily defined group of companies. By eliminating due process for even a subset of practices, the Commission emboldens other countries that have a history of using competition law to discriminate in ways that favor domestic firms.
Moreover, there is a huge risk that a WTO-violating regulatory assault on large U.S. online platforms would add another major issue to the distressingly long list of intractable transatlantic trade disputes that includes the Boeing–Airbus rivalry, digital services taxes, U.S. national security tariffs on steel and aluminum, and the EU courts’ invalidation of the Privacy Shield framework for data flows. Such an assault would further complicate the seriously ailing, $7 trillion transatlantic economic relationship.
Of course, it is difficult to offer solid conclusions until the Commission publishes its DSA and DMA package proposal and submits draft legislation to the European Parliament and the member states. But we expect that, by then, many elements of the regulatory framework and legislation will be essentially locked in without the Commission having fully considered their impact on innovation and economic opportunities for SMEs in Europe. Our goal in this paper is to discuss how ambitious the Commission aspires to be in its regulatory reach—and how unprecedented and counterproductive these policies could prove to be.
Meredith Broadbent is a senior adviser (non-resident) with the Scholl Chair in International Business at the Center for Strategic and International Studies in Washington, D.C.
The author would like to thank William Reinsch, Jack Caporal, Jasmine Lim, Victoria Meyer, Ally Brodsky, and Jacob Kopnick for constructive comments and research support, and all participants in CSIS' roundtable on the Digital Services Act.
This report is made possible through the generous support of the Computer & Communications Industry Association (CCIA).
This report is produced by the Center for Strategic and International Studies (CSIS), a private, tax-exempt institution focusing on international public policy issues. Its research is nonpartisan and nonproprietary. CSIS does not take specific policy positions. Accordingly, all views, positions, and conclusions expressed in this publication should be understood to be solely those of the author(s).
© 2020 by the Center for Strategic and International Studies. All rights reserved.
Please consult the PDF for references.