When it comes to regulating data and the growing power of tech companies, Europe is the global test case. While the EU is the model for other countries, it is also a cautionary tale of the unintended consequences of applying broad regulatory fixes to a rapidly evolving landscape, and the recently proposed Digital Markets Act, which targets platforms, is no exception. New rules on how data can be used and combined, and how digital platforms direct consumers, may end up stifling competition and innovation — a bad outcome for both consumers and companies. To avoid this outcome, regulators should follow four guiding principles: 1) preserving business model innovation should be the top priority; 2) regulators should focus on why ecosystems are competitive, not on who is winning; 3) stay focused on fostering market contestability in adjacent segments; and 4) regulators should hold companies accountable, but not tell them what to do.
More than ever, the perception is that search engines, social networks, and cloud services providers act as “gatekeepers,” controlling access to their core and connected digital services. As the power of these platform companies has grown, so too has concern over how they choose to wield it, and we are on the cusp of a new era of regulation — an inflection point that will impact business leaders, civil society, and government regulators. If done well, this new regulatory framework could provide clarity and guidance about what behavior is lawful (if not desirable), and balance concerns of access, competition, and privacy. There’s a real risk, however, that this new regulation will boil down to a coarse list of bans — of questionable efficacy — that can chill business opportunities.
Right now, Europe looks like the test case for what the regulatory response to platforms’ power. The EU’s General Data Protection Regulation (GDPR) and the recently announced Digital Markets Act (DMA) are setting the standard for countries around the world — and in some cases offering a cautionary tale. A concern with the European approach is that it might be too blunt, with the risk of constraining value creation of all sorts of platforms that might fall under gatekeeper status, doing little to promote competition, and in some cases producing unintended consequences that actively harm competition.
Consider the GDPR, which was implemented in 2018, and aimed to give users more control over their data. The regulation made it harder for firms to collect, store, and analyze customer data in the name of protecting privacy — a tradeoff between consumer privacy and possibly higher quality and competition in services, which also raised the cost of doing business. The consequence of this is that startups and small firms have found themselves at serious disadvantage against large firms: Research shows that GDPR led to an increase in relative concentration in the web technology vendor market by 17%, and that websites are now 15% less likely to share personal data with small web technology providers in favor of bigger providers. GDPR has also negatively impacted growth of AI startups with customers in Europe, and reduced investments in European tech startups.
The DMA, which could be adopted by 2022, similarly threatens to produce stifling unintended consequences through its broad approach. The regulation aims to rein in the power of the digital economy’s gatekeepers — defined as those providing core platform services, acting as important “gateways” to consumers, or having a “significant impact” on the market — and to ensure “fair” and “open” digital markets. The problem is that the definition of a gatekeeper is overly broad, does not account of the different nature of those services, and might miss on the key aspects of competition and innovation dynamics.
The Problem of Unintended Consequences
Gatekeepers would be subject to a number of obligations, but three are particularly at risk of creating unintended consequences which could stifle innovation.
1. Gatekeepers’ data practices will be closely monitored, including the amount of information they gathered on third-party businesses from their direct control over user data. But this might come at the cost of undermining service innovation and competition among those services, leaving users with fewer — or more expensive — options.
Giving users more control over their data can solve potential privacy and data security problems, but it can stifle competition in surprising ways. Users often place more trust in large players (which can invest more in data security) and refrain from sharing their data with small players. As a result, there’s ultimately less competition across similar platforms.
2. There will be limitations on how platforms can use and combine large datasets, which will make it harder for complementary businesses to work together and force some companies to invest in duplicate data infrastructures. Here, the focus on size misses the point — it’s how the data is used that matters. Competition suffers when data and algorithms are used to direct consumers towards products that are more profitable for the platform, but inferior to other alternatives. The DMA prescriptions would do little in this regard, but they may harm legitimate uses of big data sets.
Consider the case of Booking.com. By collecting and aggregating different types of data from users and service providers such as online travel agencies and hotels, Booking.com has managed to create an entire digital travel ecosystem, with both traditional services such as hotel reservations and car rentals as well as new digital services, such as personalized offers and direct control over discounts and rewards. In doing so, it uses data as both an input and the product. This kind of business will likely suffer under the new rules — the whole business model, which involves aggregating and combining data, is under questioning and possibly on the chopping block, but there’s little reason to think that would foster greater competition.
3. Gatekeepers also would be forced to refrain from “self-preferencing” — treating their own services and products more favorably — and maybe even spin off parts of their business. This could mean that a digital company that orchestrates its own ecosystem could not sell goods or services in its ecosystem.
There are three main problems with limiting how platforms can offer services. First, the record of such types of separation has not been generally positive in banking, commerce, or airlines. Second, platforms use these first-party products and services as a critical strategic lever to differentiate their platforms and compete against larger or dominant platforms, and to re-direct innovation effort of complementors towards other, less-developed niches that benefit users. All of this tends to create more choice for consumers. Third, while this restriction is designed to target the Big Tech companies, this policy could hinder business model innovation and digital transformation of incumbents in traditional sectors who are increasingly adopting platform-based business models and building their own ecosystems to remain competitive.
Four Principles Regulators Should Follow
Each of these obligations is in danger of stymying growth and innovation throughout the entire platform economy, while failing to address actual problems. We articulate four guiding principles we believe regulation of digital markets should follow to balance the need to regulate with protecting innovation.
First, preserving business model innovation should be the top priority. Platforms represent a new and evolving way of organizing economic activity and coordinating independent multiple actors to expand the set of value propositions to customers. Banning critical levers of the platform business model through an overly restrictive DMA would put a break on business model innovation for all firms; mandating specific design choices will constrain particular innovation opportunities. It’s likely that neither will promote competition and benefit consumers. Instead, regulators should focus on particular problematic activities by platforms — such as undue restrictions to access their marketplaces or “biased” recommendation systems that may unduly favor their own products and services against those of competitors — and aim to write restrained policy that allows platforms to continue to grow and innovate.
Second, regulators should focus on why ecosystems are competitive, not on who is winning. Does the mere presence of a gatekeeper in a core service create structural impediments to competition? Current regulatory thinking seems to assume that reducing gatekeeper power will consequentially unlock competition, but this might not happen. To keep platforms competitive and innovative, the regulatory focus should be on specific anticompetitive practices that create structural barriers to competition between ecosystems — not the dominance of a few specific gatekeepers.
Consider the importance of data. Accruing more data makes platforms better able to serve customers, but it also can lock customers into a given platform because that’s where all of their data is stored. One regulatory solution to this problem would be to make data more portable — allowing customers to use it on more than one platform — rather than try to partition platforms or dictate which business a given platform should or shouldn’t be in. That way, the benefits of a successful ecosystem can be shared more broadly by allowing “porting” of services across competing ecosystems while leaving the platform still in control over its own ecosystem. Some platform scholars refer to this as “in situ information exchange,” where data, instead of being transferred to a competitor’s interface, is used at the location it is collected.
Third, it’s important to stay focused on fostering market contestability in adjacent segments. To avoid one extreme — platforms being able to restrict market access at will — regulatory proposals seem to go the other extreme, imposing a “open platform” model for gatekeepers to create greater standardization across platforms, whereby free entry and complete fungibility of products and services across platforms become standard. Would such an approach create more competition among existing platforms or allow newcomers to challenge existing players? Research suggests probably not. Once the sources of differentiation among platforms are removed so they become like standardized infrastructures, size of the platform network becomes the most powerful driver of value, and users will tend to choose the bigger platform.
Keeping platform markets contestable does not imply breaking up a specific platform to parcel one market among many similar contenders. It should rather aspire to preserve opportunities for platforms to differentiate themselves. The disruptor that will successfully challenge incumbents will not be providing more of the same but something different, and there are clear benefits to that.
Airbnb is a case in point. Despite strong offline gatekeepers (real estate operators), and online gatekeepers (booking platforms such as Booking.com), Airbnb has managed to create an alternative ecosystem of hospitality services by leveraging distinct, peer-to-peer interaction dynamics. This has expanded the option of hospitality services for consumers (e.g., alternative and cheaper accommodation options), with the list still expanding in additional complementary services such as local attractions or “experiences.” While Airbnb may be contested on other grounds (e.g., raising housing prices), the initial idea of utilizing idle spaces had clear benefits to consumers.
Finally, regulators should hold gatekeepers accountable, but not tell them what to do. The traditional permission-based regulatory approach prescribing specific permissible conduct (dos and don’ts) is largely inadequate to police a domain that has no clear boundaries, is hard to monitor, and evolves rapidly as firms continue to innovate their business models and data practices. Regulation generally performs poorly in dynamic markets with rapid technological change. As experience with traditional public utilities indicates, regulation setting fixed prescriptions can easily fail to meet its objectives as technology evolves, changing the structural conditions of the economic environment.
What can be done to this end? One solution might be to move to a decentralized, data-driven accountability regulatory system, whereby gatekeepers would provide data APIs for public auditing, perhaps to ad-hoc digital agencies. This mechanism would not force platforms to disclose their algorithms, and so would therefore protect the value of their innovation. At the same time, it will allow oversight over the system, which can act also as deterrent of the potential abuse of orchestration power. Such a system might balance the various interests at stake; it can enhance transparency and higher standards of behavior while not discouraging innovation by preserving a firm’s latitude of action.
Without effective calibration of the DMA to reach only anticompetitive behavior, the DMA threatens to distort platform industries across the economy and hurt small businesses and consumers who rely upon these platforms, while the proposed rules have doubtful efficacy as for promoting competition. We believe a different approach is needed to design an appropriate regulatory framework for the digital economy; one that builds general principles that should guide actions while leaving to antitrust (specialized) units to derive criteria to implement those principles to specific issues on a case-by-case. This can guarantee that regulatory institutions can learn and adapt to the dynamic digital environment. The principles we proposed here are but a starting point in this direction.