Abstract: In the December 2020, the European Commission has presented its proposal for a Digital Market Act (DMA) aiming at promoting competition and preventing unfair practices on digital markets. The DMA creates a new category of platforms, “gatekeepers”, based on criteria relative to their turnover and number of users, that will be subject to specific rules and obligations. Analyzing the current proposal from a law and economics perspective and taking into account the particularities of the digital economy, this paper observes that such a regulation based on the size of platforms is likely to hinder competition, reduce innovation and harm consumers.

JEL classification: D43, K21, K24, L41, L44, L51, M38
Keywords: Digital platforms, Competition, Regulation, DMA, EU

To quote this paper: P. BENTATA, “Regulation ‘gatekeepers’: predictable ‘unintended consequences’ of the DMA for users’ welfare”, Competition Forum, 2022, n° 0031, https://competition-forum.com.

1. Introduction


The digital economy has witnessed the rise of platforms (Kenney 2016). Using the information and connectivity properties of the digital, they “provide a basis for delivering or aggregating services and content from service and content providers to end users.” (Nooren 2018, p. 2). In other words, they build bridges between independent parties that could mutually gain from interacting with each other.


These interactions can serve different purposes and involve various parties: easening the search for information (e.g. Google Search), communicating with other users (e.g. Skype, WhatsApp), sharing information or contents (Facebook, Vimeo, YouTube), facilitating transaction between the demand for and the supply of specific products and services (Amazon, Uber, Airbnb, booking.com, Deliveroo), but also creating an ecosystem of integrated and complementary services (Apple iOS and AppStore, Google Android and Google Play). 


Notwithstanding the differences between these platforms, they all share a common core feature: they aim at facilitating interactions between multiple groups that do not know each other in order to provide end-users with greater and more personalized consumption options. Hence, their ability to create value for end-users is correlated with the number and the diversity of the suppliers using the platforms. Consequently, platforms are more likely to be successful if they benefit from and develop network effects, especially indirect network effects, and economies of scope.


Indirect network effects appear when the value of a service for one group of users increases when new users of another group join the network (Rochet 2003, Caillaud 2003). And economies of scope appear when the marginal production cost of one service decreases when another related good is produced, help platforms providing complementary or integrated services that increase the overall value of the network for end-users. These characteristics, that are considered as key drivers of platforms’ success (Armstrong 2006, Parker 2005, Cennamo 2019, Panico 2015, McIntyre 2009) have raised concerns as they are deemed to provide competitive advantage to large platforms, favor the concentration of sectors and ultimately lead to winner-take-all situations (Lee 2006, Farell 1985, Katz 1985, Rohlfs 1974).


An abundant literature has focused on the potential adverse effects of the digital platforms’ economy. By channeling the relevant information across end-users and producers of information, platforms may filter the content and the treatment of information, reducing the diversity of information and trapping users into filter bubbles (Pariser 2011, Flaxman 2016) and echo chambers (Garett 2009, Sunstein 2001, Sunstein 2009, Sunstein 2019). Such a situation may reduce users’ welfare and lead to political instability as a restricted number of large platforms dominate the information sector (Goodman 2004, Holtz 2015). 


The economic literature has mostly focused on the potential adverse effects that large platforms may have on the competition. Indirect networks effects have been presented as barriers to entry, that could be used by the largest platforms for anticompetitive purposes (Khan 2017) and lead to monopolistic situations (Farell 1985, Katz 1985, Rohlfs 1974). Economies of scope have also been considered as potential market failures, due to the fact that large platforms providing diversified and interconnected products and services may become indispensable for both end-users and third-party firms and advertisers, getting de facto a monopolistic position (Nooren 2018). And network effects, when combined with lack if interoperability between platforms, could also reduce competition by creating harmful consumer lock-in (Martens 2016, Guibault 2013). Finally, large platforms may take advantage of their size to influence both the structure of the market, imposing their digital business model based on “free access” to any potential competitor and consumers’ choices (Bourreau 2020). 


Due to these specific adverse effects, it has been argued that large platforms should be subject to a specific set of regulations that would complement the current antitrust regulation (Evans 2015, Feld 2019, Newman 2019, Mazzucato 2020). Following these conclusions, the European Commission has recently published a draft law called the Digital Market Act (DMA) that aim at establishing “a level playing field to foster innovation, growth and competitiveness” (EC 2021). As part of a broader legislative initiative called “Europe fit for the digital age”, the DMA is designed to address perceived shortcomings in the European existing competition law instruments by imposing specific obligations to platforms identified as “gatekeepers”.


Yet, to define these “gatekeepers”, the DMA relies almost exclusively on quantitative size criteria: i) an annual turnover equal or above 6.5 billion euros in the last three years within the European economic area or an average market capitalization equal or above 65 billion euros, ii) more than 45 million monthly active end-users and more than 10,000 yearly active business users in the Union iii) during each of the last three years (EC 2020, p. 37). Given the peculiar characteristics of platforms, imposing a specific regulation on the largest platforms to promote competitiveness could have severe adverse effects and unintended consequences on the global level competitiveness of digital platforms that may harm end-users instead of improving their welfare.


The aim of this paper is to shed light on these adverse effects and show that regulators should not focus on the size of platforms if they want to implement rules that will promote innovation and protect end-users’ welfare. After describing the main features of the DMA proposal (Section 2), we provide theoretical economic arguments (Section 4) to explain that neither the spirit (Section 4) nor the text of this regulation (Section 5) are relevant if its main goal is to improve users’ welfare and innovation. Finally, section 6 concludes.


2. A glance at the DMA  

The first paragraph of the explanatory memorandum explicitly expresses the spirit of the DMA: digital platforms “increase consumer choice, improve efficiency and competitiveness of industry and can enhance civil participation in society. However, whereas 10 000 online platforms operate in Europe’s digital economy, most of which are SMEs, a small number of online platforms capture the biggest share of the overall value generated.” (EC 2020, p. 1)

Benefiting from networks effects, these large platforms represent a “key structuring elements” of the digital economy, and “act as gateways or gatekeepers between business users and end-users and enjoy entrenched and durable position, often as a result of conglomerate ecosystems around their core platform services, which reinforces entry barriers.” 


The aim of the DMA is thus clear: creating a specific regulation for large digital platforms that provide core platform services and could act as gatekeepers. According to the DMA, core platform services include a wide range of services, such as online intermediation, online search engines, social networking, video sharing, number-independent interpersonal electronic communication, operating systems, cloud services or advertising (Art. 2, §2). And to be qualified deemed to be a gatekeeper, a provider of core platform services should “i) have a significant impact on the internal market, ii) operate one or more important gateways to customers and iii) enjoy or are expected to enjoy entrenched and durable position in their operations.” (Art. 3, §1). 


However, as these definitions are broad, any platform providing a core service could be deemed to be a gatekeeper if its annual turnover in the European economic area is higher or equal to 6.5 billion euros in the last three years or if its market capitalization amounts to at least 65 billion euros, and if it welcomes more than 45 million monthly active users and more than 10,000 yearly active uses for the last three years. As a result, the DMA is designed to impose specific regulations to large platforms only. On this specific point, it is worth noticing that the Commission can adjust these thresholds to market and technological developments. (Art 3, §5). 


Designated “gatekeepers’’ should comply with a set of mandatory obligations, designed to prevent them from imposing unfair practices and raising entry barriers, unless the Commission decides to exempt them for overriding reasons of public interest (Art. 9). Among those obligations, “gatekeepers’’ will have to i) refrain from combining personal data sourced from different services they or third-party provide without end-users’ consent, ii) allow business users to offer the same products or services to end-users through third party online intermediation services at different prices and conditions, iii) allow business users to promote offers to and conclude contracts with end-users acquired via the core platform services regardless of whether they use the core platform services of the gatekeeper or not, iv) refrain from preventing or restricting business users from raising issues with any relevant public authority relating to any practice of the gatekeeper, v) refrain from requiring business users to use, offer or interoperate with an identification service of the gatekeeper, vi) refrain from requiring business users or end-users to subscribe to or register with any of the services proposed by the gatekeeper as a condition to access, sign up or register to any of its core platform services, vii) provide advertisers and publishers with information concerning the price paid by other advertisers and publishers and the amount or remuneration paid to publishers for the publishing of a given ad. 


Hence, from a law and economics perspective, the DMA can be defined as an ex ante differentiated regulation that only targets the largest platforms, according to their turnover and their number of users. Although this particular feature is not intrinsically inefficient, it raises important concerns. First, the fact that the Commission could change the thresholds used to define a “gatekeeper’’ may be a source of uncertainty and ultimately of economic inefficiency, because firms need to know the obligations they will be subject to in order to adopt long-run welfare-enhancing strategies (Broulik2019). Second, a differentiated regulation may create “threshold effects’’ that reduce firms’ incentives to innovate and to compete with the most heavily regulated firms (Aghion 2021, Garcia 2019). Third, ex ante regulation can lead to inefficient results if it allows regulators to follow private goals that are not aligned with public interests (Stigler 1971, Laffont 1991) or if it prevents strategies which consequences are welfare-improving (Christiansen 2006).


3. An ill-spirited law: the size of platforms is not a relevant economic criteria to regulate them  

3.1. Competition is a dynamic process 

The first paragraph of the explanatory memorandum (EC 2020, p. 1) explicitly expresses the spirit of the DMA: digital platforms “increase consumer choice, improve efficiency and competitiveness of industry and can enhance civil participation in society. However, whereas 10 000 online platforms operate in Europe’s digital economy, most of which are SMEs, a small number of online platforms capture the biggest share of the overall value generated.’’ In other words, no matter their immediate impact on end-users, large platforms should be subject to a specific regulation because they are large. Yet, from an economic point of view, regulation should be driven by consumers’ welfare, not by market concentration (Lerner 1934, Shapiro 2018).


Market power and monopolistic position are not per socially harmful: to assess their impacts, it is first necessary to determine their origin. A monopoly may emerge because of legal barriers or economic barriers; the latter may result from the market structure – the presence of economies of scale (Pitelis 2017) – or from commercial strategies, such as product differentiation (Armentano1978). It is essential to review the causes of a monopoly to determine its positive or negative impact on consumers. If the barriers are legal or structural – like the existence of a non-duplicable technical infrastructure such as a telecommunications or energy distribution network – monopoly is said to be “natural’’. On the other hand, if it results from a correspondence between the services provided and consumers’ needs, then the monopoly is said to be “by merit’’, which means that consumers massively choose a particular service. In the first case, the monopoly appears despite the consumers; in the second, it is the consequence of the consumers’ decisions (Kirzner 1973).


Consequently, observing a market in a purely static way leads to interpreting a situation without knowing whether it is the cause of a market failure or the result of an efficient market process. For this reason, many economists criticize this approach, considering that a static analysis of the state of competition amounts to denying the reality of competitive processes. Indeed, the static analysis of competition implies that any strategy of price differentiation, advertising, marketing, packaging, any action aimed at attracting and satisfying consumers by offering a service that better meets their expectations is by definition “monopolistic” (Armentano 1972).


3.2. Entry barriers: obstacle to competition or result of a competitive process?

The same is true of entry barriers. The two main economic barriers are economies of scale and product differentiation. A recurring example among economists defending the static definition of monopoly is the car industry: car designs change every year to allow incumbent manufacturers to limit potential competitors’ entry to the market. Similarly, economies of scale provide market power to firms by reducing their unit price as their output increases, preventing new competitors’ entry.


This interpretation once again reverses the causal links (Mises 1963, p. 358 ff.): if car manufacturers can set higher prices by differentiating their products, it is because consumers pay more for new models, not the other way around. Product differentiation is a barrier if and only if consumers prefer this differentiation. If consumers preferred to buy less differentiated products, differentiation would not be profitable. The same argument applies to economies of scale: economies of scale can only act as a barrier if consumers want to pay less for mass-produced products. Otherwise, they would always have the means to differentiate themselves by paying more for a competitor’s product.


Moreover, it should be noted that these two barriers cannot exist simultaneously since differentiation implies charging more for a scarcer product. In contrast, economies of scale imply reducing competition by massively supplying a standardized product. The reversal of the causal link, which stems from a static interpretation of competition, leads to the imposition of rules aimed at strengthening competition to the detriment of consumers: to prevent differentiation, it is necessary to prohibit firms from differentiating themselves from competitors. To counteract economies of scale, it is sufficient to increase the firms’ costs benefiting from these economies through a tax. In both cases, public action aimed at restoring competition will limit consumer welfare.


3.3. That a monopoly exists does not legitimize regulation if it is contestable

Based on a dynamic approach to competition, it appears that a monopoly is only problematic if it harms consumers. However, this is only possible if demand is inelastic – consumers must be captive and with no alternative to the service provided (Mises 1963). This does not imply that there is only one company at a given moment, but that the resources needed to produce a similar service are no longer available, for legal reasons – prohibition of exploitation, prohibitive taxation – or economic reasons – non-renewable resources whose stock has been depleted or are too distant to be used at an acceptable cost. 

In other words, only the inability to enter a market can define a monopolistic structure (Baumol 1982). It is, therefore, the inability to “contest’’ the situation of a market that matters. However, the market’s contestability depends on one thing only: the entry costs, i.e., the unrecoverable costs that a competitor will have to incur to enter the market and provide an attractive service to consumers.


4. “Gatekeepers” inevitably compete with each other 

4.1. Networks effects and economies of scope drive digital competition

Regarding platforms’ economics, the ability to provide a service that competes with existing platforms primarily depends on creating a platform, i.e., to benefit from the direct and indirect network effects enjoyed by existing platforms. These effects depend mainly on the ability to capitalize on consumer data and define a financing model that simultaneously satisfies the different parties involved (Rochet 2003): producers and consumers in the case of a two-sided markets, and different groups of producers, consumers, and third-party financers in the case of a multi-sided markets. On this point, the static analysis of the markets on which platforms operate often portrays them as uncontestable markets since they are characterized by a high degree of concentration. Due to strong network effects, a small number of operators would attract the majority of consumers and benefit from a dominant position (Brousseau 2007).

This “monopolistic” structure is an inevitable consequence of the way the digital economy operates. As the amount of personal data collected and processed increases, the platform can develop better and more personalized services; but in so doing, it attracts more consumers, which makes the platform more attractive to businesses and advertisers. And the increase in the number of advertising companies and businesses on the platform provides it with more means to develop its algorithms and refine its data collection and processing methods, thus creating a virtuous circle that translates into a mechanical increase in the size of the platform and increasingly intense consumer captivity (Haucap2014). For this reason, platform markets should be regulated to restore consumer sovereignty by allowing competitors to enter the market.

Specifically, proponents of such an approach consider that platforms are in a monopoly position and pose a danger to consumers if they benefit from economies of scale, economies of scope, are able to control or lock down access to personal data and can merge with other platforms to increase the amount of data under their control (Rubinfeld 2017).

Despite its apparent logic, such an analysis does not correspond to the reality of the digital economy. Competition in traditional markets is defined by the number of firms offering similar products and targeting similar customers (Chen 1996). Hence, competition concerns the products within a specific market, and market shares and firms’ dominance is relatively easy to observe. Platform competition is different. As the number of products and services sold through the platform may vary and may come from different traditional markets, platform market boundaries are expanding and cannot be easily defined or taken as fixed and given (Panico 2015). The key point to define a market is not the number of firms providing substitutable goods to a given number of potential consumers, but the potential competitive interactions that may arise between platforms as they expand their consumption options to provide consumers with interconnected and complementary products and services (Cennamo 2019).

This specificity is due to the fact that on digital markets, what is valued by end users is not the separate use of each product or service available but the diversity of functionalities that their joint use makes possible (Baldwin 2009, Cennamo 2019b). Therefore, competition exists, but instead of concerning individual products, it occurs both within platforms – e.g. between apps available on the same app store – and between platforms – to provide the best integrated ecosystem of complementary services (Eisenach2017). Consequently, “because of their generative and connectivity nature, digital markets are unbounded in that the value of the digital product/service transcends its own boundary, and spills into new, and often contested domains” (Cennamo 2019).

In this context, digital platforms compete with each others, by providing integrated, interconnected and complementary services, in order to maximize end-users’ experience. As a result, they may even incite platforms to purposefully restrict access to end-users and to third-party suppliers. Indeed, maximizing end-users’ experience may imply to impose screening mechanism and specific standards of quality that restrict access to the platform (Wareham 2014). These restrictions are of primary importance when users value specific functionalities and a specific organization of the platform, i.e. the platform’s identity (Jacobides 2018) and when their satisfaction is deeply rooted in the interconnectedness and integration of the products and services available on the platform. In this case, platforms face a trade-off between increasing their size and differentiating themselves from other platforms, either through a specific technology (Zhu 2012), specific contents (Cennamo2013, Seamans2014) or specific market positioning (Bresnahan 2014).

Consequently, platform size is the most important determinant of value creation only when consumers and complements have the same impact on the indirect network effects, any extra user do not reduce the value of the platform for other users, and users’ utility is positively correlated with the amount of content or complement available (Cennamo 2019, Panico 2015, McIntyre 2009). Hence, the size is determinant only for pure multi-sided transaction platforms. Information and innovation platforms do not meet these conditions, as on these platforms, users value quality and integration more than quantity of products and services (Zhu & Iansiti 2012, Binken & Stremersch 2009). As a result, on these markets, platforms of different size may coexist, compete, and survive (Cennamo 2019, Cennamo 2013, Anderson 2014), as it is the case for search engines, browsers softwares, e-commerce platforms, cloud services, social networks, online video platforms, or even operating systems.


4.2. Some evidence of the intensity of competition between platforms

Due to the specific features of digital markets, large platforms operating in one domain face fierce competition from other large platforms originally operating in another domain. And this prevents any possibility to reach a winner-take-all situation. In fact, as soon as the technological architecture or the organization of the ecosystem differ between platforms – functionalities, degree of openness – none will be able capture the entire market, even if they share the same type of end-users. 

For example, Apple and Google compete on the mobile OS market but have very different approaches: Apple aims at providing an “unmatched user interface experience and an optimal software-hardware-service integration”, whereas Google aims at providing “unmatched capabilities and database in the search categorization domain” (Cennamo 2019). These strategies imply different architectures: Apple has an integrated architecture, imposing great control over external developers whereas Google is open, both for apps developers and for the developers of the OS platform. Consequently, it would be extremely complex and costly for each platform to imitate the other platform’s strategy in order to attract users of both platforms, as attracting users of the other platform would imply massive investment (to change the OS design, adapt the model of development of the OS, change the business model and the ecosystem) and could reduce the experience and the satisfaction of the current users. 

As the value created by platforms depends on their ability to provide end-users with integrated, interconnected and complementary services, what is supposed to strengthen monopolistic situation in traditional markets favors competition in digital markets. As explained earlier, because they develop through the use and enhancement of data, platforms benefit from economies of scale and scope that allow them to improve their services as they grow and diversify the services they offer. The more they develop, the more they can access new markets with similar operations and business models. Therefore, entry costs into a new market are partially offset for a platform that has already developed in its initial market.

From a theoretical perspective, this specificity means that markets are all the more contestable where there are economies of scale and scope. For example, Facebook, originally a social network, offers e-commerce services, competing directly with Amazon. Similarly, Amazon can create a video distribution platform and compete with Netflix and YouTube; Twitter becomes an alternative to sending text messages on iMessage or emails via Yahoo mail or Google mail); a job search can be done on Facebook, LinkedIn or Le Bon Coin (in partnership with Pôle Emploi, the French public employment service), and Amazon or Facebook can be used as search engines. Because these platforms have already partially assumed the development costs and attracted enough end-users that are looking for a complex bundle of integrated services, they can easily penetrate markets where other platforms are present. In other words, if the platforms benefit from economies of scale, allowing them to have a multi-market reach, they inevitably compete with each other.

Based on the Form 10-K filled by Amazon, Apple, Facebook, Google, Microsoft and Netflix over the last decade, a recent study observes that among these platforms, only Facebook does not rank competition as a major risk for its business and development, meaning that for the majority of the largest platforms, competition exists and is considered as an existential threat (Petit 2020, p. 43). It also appears that platforms perceive competition as a more important risk than other large companies operating in other sectors, such as airline, telecommunications, pharmaceuticals, medias, credit reporting or drug wholesaling.

More importantly, observing how seventeen market research firms, specialized in business analysis, market research and competitive analysis, assess the level of competition between these platforms, the same study observes that they all compete with each others. For example, Google competes fiercely with Amazon, Apple, Facebook and Microsoft; Facebook faces some strong competition from Google and Microsoft but also to a lower extent from Amazon and Netflix; Apple competes with Google and Amazon but its main competitor is Microsoft, and Amazon, though its services of cloud data storage and streaming, directly competes with Mcrosoft, Apple, Netflix but also with Google (Petit 2020, p. 54).


Direct observation of their strategies confirms the intensity of competition on digital markets. Indeed, their research and development spending is continually increasing, demonstrating that they are continually seeking to improve their services’ quality. According to the latest PwC innovation report, Amazon and Alphabet are the world’s two largest investors in R&D, spending $22.6 billion and $16.2 billion respectively in 2018. Microsoft, Apple, and Facebook rank sixth, seventh, and fourteenth, respectively. In comparison, Volkswagen, Roche, and Novartis, the largest European investors, spent €15.8 billion, €10.8 billion, and €8.5 billion, respectively, in 2018. In the digital sector, SAP and Ubisoft, the two largest investors, spent €4 billion and €0.9 billion, respectively. Putting these investments into perspective, digital platforms’ investments reflect their need to innovate, renew, and improve the services they offer to maintain their appeal. Indeed, for companies that are considered as global giants, these expenditures are significant: Amazon, for example, spent the equivalent of 38% of France’s domestic R&D expenditures on their own R&D in 2018.


5. Ill-formulated rules: DMA could limit competition, slow down innovation and harm users

5.1. Some well-known unintended consequences of ex-ante differentiated regulation

To prevent any unfair or anti-competitive practice, the Commission has opted for an ex ante regulation mechanism that focuses only on large platforms designated as “gatekeepers”. This type of regulation is not new and has been extensively discussed in the law and economics literature, making it possible to define its relative costs and benefits compared to an ex-post liability mechanism.

The main benefit of this kind of regulation lies in the regulator’s ability to ensure that each player’s behavior regulated in this manner can be verified, thereby preventing any harmful action for the company. On the other hand, it has a significant disadvantage: the cost of implementation. By definition, ex-ante regulation implies controlling all companies’ behavior subject to such regulation, regardless of the consequences of such action. Compared to a liability mechanism that would apply only in the event of harmful or undesirable behavior, ex-ante regulation is very costly since it involves mobilizing supervisory agents and efficient administration upstream of any activity (Dewees 1996, Hiriart 2004, Boyer 2004). 

The best illustration of the costs and benefits of such an approach is undoubtedly found in the field of technological and environmental risk prevention: as soon as they are likely to cause severe health or environmental damage, companies are subject to specific, more stringent regulations requiring them to adopt special prevention measures and to communicate information to the regulator concerning their operations and the degree of risk that their activity represents. \footnote This ex-ante regulation is justified because, in the event of a nuclear accident or an accident resulting in widespread water or soil pollution, the damage would be far greater than the costs of prevention imposed by the regulations.

However, this regulation is accompanied by several additional costs and shortcomings. In the French case, its implementation was insufficient because of the drastic costs that each company’s strict monitoring would impose (Bentata 2014). Moreover, ex-ante regulation has proven to be incapable of genuinely regulating behavior since it is virtually impossible to monitor companies’ day-to-day activities, and level of care (Bentata 2013). And finally, the individuals in charge of designing the rules and enforcing them have sometimes been “captured” by the companies whose activities they were supposed to regulate (Hylton2002, Boyer 2001, Hiriart 2010).This example reminds us that even in areas where it appears to be perfectly adapted, ex-ante regulation is only partially effective. The costs of implementation and enforcement remain very high in relation to the expected benefits (Bentata 2012b, Faure 2007).
For this reason, ex-ante regulation is preferable to an ex-post liability mechanism only on the condition that the potential damage in case of corporate misconduct is irreversible and exceeds the financial capacity of the companies. Otherwise, and particularly if companies have heterogeneous operations, face heterogeneous costs, or operate in different markets, ex-post liability regulation is preferable (Pacces 2012). Also, with regard to platforms, ex-ante regulation does not seem to be effective, as the potential damage they could cause is not irreversible, and their operations are heterogeneous. 

In this case, a differentiated ex ante regulation would not only be ineffective but potentially detrimental for end-users. Indeed, differentiating regulatory obligations according to the size of platforms, would involve defining a group of platforms considered to have abnormally dominant positions in their market based on their size, market share, economies of scale and scope, and ability to control and restrict consumer data access. As explained earlier, from a purely economic standpoint, these criteria do not appear to be relevant in judging competitive intensity in the case of platforms and could even harm end-users who are looking for a bundle of integrated and interconnected services. Moreover, this approach seems particularly risky for two reasons highlighted by the economic analysis of the law.

First, the introduction of differentiated regulation automatically leads to market distortions. Once defined as “gatekeepers” platforms will be subject to more stringent regulation, reducing their incentives to develop beyond the level at which the new regulation is imposed. Therefore, differentiated regulation is inevitably accompanied by a reduction in competitive intensity, thereby freezing market strategies (Aghion 2021). This is an unavoidable consequence that results in the maintenance of the status quo in a given market.
While this may be legitimate in the case of particularly risky activities – for example, with regard to the protection from environmental and health risks, which differentiates regulations according to how dangerous companies are – the costs and benefits of such regulation for society as a whole should always be compared: tighter regulation improves the protection of individuals from certain risks but to the detriment of market dynamics. 

For platforms, differentiated regulation will be ineffective since the assumed risk is that of imperfect competition. In other words, imposing differentiated regulation to strengthen competitive processes implies reducing competition in the name of the competition itself.
Second, to define efficient criteria for differentiating platforms, the regulator must identify the platforms’ costs and operations, which means developing close interactions with the companies it will have to regulate. This is another inevitable constraint since the information needed to introduce efficient regulatory standards is by its very nature confidential, known only to the companies. Furthermore, to obtain this information, it is necessary to establish close ties with the companies that will be subject to the new regulations, which increases the risk of “capturing” the regulator (Laffont & Tirole, 1991, 1993).

This risk, widely documented in economic literature, translates into a two-fold problem: firstly, the regulator may be led, consciously or unconsciously, to implement ineffective regulations or to misapply regulations; secondly, and more seriously, by imposing specific regulations, the regulator becomes, in the eyes of consumers and society in general, responsible for the development of the regulated sector, thus providing an opportunity for firms to shift the burden of responsibility away from them.

5.2. Predicable adverse effects of Article 5

The choice of an ex ante differentiated regulation may be ill-suited to promote competition and protect end-users’ welfare. Yet, the extent to which it will harm end-users and reduce incentives to compete and to innovate depends on the precise formulation of the rules. This section details the predictable adverse effects that may arise for each of the seven general rules defined by the Article 5.


i) Refrain from combining personal data sourced from different services they or third-party provide without end-users’ consent, and v) refrain from requiring business users to use, offer or interoperate with an identification service of the gatekeeper.

When platforms combine data from different services, end-users benefit from an integrated bundle of complementary services with zero transaction costs. They do not need to sign-in every time they use another service from the same ecosystem and they have the guarantee that any service available within the ecosystem is compatible with the others. And these compatibility and interconnectedness are the core value of platforms for end-users. Therefore, preventing them from combining data from different sources will reduce end-users welfare.
Moreover, as these rules prevent “gatekeepers” from favoring data interoperability between its services and third-party services available on its ecosystem, complementary services from third-party business may lack compatibility, which in turns may reduce their value for end-users who will prefer using services that are directly under the control of the “gatekeeper”. In this case, these rules will provide an advantage to the largest platforms instead of favoring competition.


ii) Allow business users to offer the same products or services to end-users through third party online intermediation services at different prices and conditions.

This obligation, that aims at giving more business opportunities to business users, may reduce the overall value of platforms at the expense of end-users. This is particularly true for multi-sided transaction platforms, such as e-commerce platforms. For end-users, the value of these platforms depends on their ability the provide ever-expanding consumption options at the best price. Under this rule, end-users will then have to compare prices from different sources and may have to create multiple accounts on different websites or platforms. As a result, they will face higher transaction costs to find out the best consumption opportunities.

But this rule could have another adverse effect. When buying goods and services online, both business users and end-users tend to rely more on well-established large platforms because they expect them to be safer (Goldberg 2021, Johnson 2021, Geradin 2020). Hence, if end-users do not change their behaviors and continue relying on large e-commerce platforms, the rule will have no positive impact on smaller platforms and business users and may result in an increase of prices on the “gatekeeper’s” platform.

iii) Allow business users to promote offers to and conclude contracts with end-users acquired via the core platform services regardless of whether they use the core platform services of the gatekeeper or not, and vi) refrain from requiring business users or end-users to subscribe to or register with any of the services proposed by the gatekeeper as a condition to access, sign up or register to any of its core platform services.

These rules aim at preventing bundling and tying, which is a source of value for end-users and the reason why some services can be provided without any payments for end-users. For example, device manufacturers can use Android OS for free because the counterpart is that Google’s search engine is set by default, and this service is the principal source of revenues for Google. These rules will enable device manufacturers to keep using Android OS without this default setting. But such an opportunistic behavior could force Google to change its business strategy, as visits and revenues from smartphones’ users will decrease, and make manufacturers pay for its OS. Hence, end-users may end up paying for services they currently enjoy for free and be charged higher prices for devices that support platforms’ ecosystems. 

Moreover, should these rules be applied with great severity – preventing for instance “gatekeepers” to impose prices on currently free services – less efficient competitors will be able to free-ride on gatekeepers’ efforts to innovate, which may ultimately deter innovation. Indeed, smaller competitors will prefer competing within the current ecosystem rather than innovating to create their owns. As a result, smaller competitors will have no incentives to look for disruptive innovation, which will in fact protect the status quo on digital markets (Jamison 2020).

iv) Refrain from preventing or restricting business users from raising issues with any relevant public authority relating to any practice of the gatekeeper, and vii) Provide advertisers and publishers with information concerning the price paid by other advertisers and publishers and the amount or remuneration paid to publishers for the publishing of a given ad.

These two rules aim at empowering third-party business when interacting with “gatekeepers”. Although they will not have any adverse effect on end-users’ welfare, they may be socially costly, as the first rule may increase administrative costs and the second one may reduce competition between platforms strongly relying on advertising to finance their business.

The first rule provides third-party businesses with incentives to fill complaints or to sue large platforms whenever they perceive an opportunity to do so, reducing the probability of out-of-courts settlements and the benefits of designing proper contracts with large platforms. Thus, number of strategic behaviors could emerge which consequences will depend on administrative and legal decisions. For instance, if these decisions tend to favor third-party businesses, the number of complaints and claim will increase, raising the administrative costs and incentivizing large platforms to either integrate services currently provided by third-party business or impose less financially attractive conditions to third-party businesses in order to cover their administrative costs. Hence, the diversity of services may be reduced and the digital market may end up being more concentrated.

The second rule aims at promoting price transparency. Although this may provide better information to advertisers and publishers, it might also lead to collusive practices, as “gatekeepers” knowing each other prices and conditions, may end up adopting similar strategies and prices, which could reduce competition among those “gatekeepers”.



Regulating large platforms according to their size may have important drawbacks. Due to the particularities of the digital economy and especially the way value is created therein, what is considered as market failures in traditional markets, especially network effects and economies of scope, enhance competition in digital markets.
In this context, concentration should not be assumed to be socially detrimental and market power should not be observed based on firms’ turnover or any other criteria focusing on the size of firms. Such an approach has led the European Commission to define a specific category of large platforms as “gatekeepers” and to target them with specific obligations that may both hinder the competitive process on digital market and reduce end-users’ welfare.

Showing the potential adverse effects of such a regulation, the present paper contributes to the current debate on platforms’ regulation by demonstrating that no regulation could be efficient without adopting an economic approach that takes into account the value creation process platforms and aim at only protecting end-users.

In other words, the DMA is likely to miss is goal because it aims at preventing platforms’ expansion instead of looking at consumers’ welfare.





Submit a Comment

Your email address will not be published. Required fields are marked *