Abstract
In the sharing economy, competitively important user data is often controlled by a single platform. This is due to concentration tendencies in the market as well as a platform’s gatekeeper function regarding user groups and user data. Presently, a platform only has to grant access to its user data if the refusal to deal is considered an abuse of market power under competition law. The legal requirements in the EU for such a claim are, however, very narrow. Since EU competition law further follows a strict case-by-case approach, it is ill-suited to provide a systematic solution for data access issues. With data becoming increasingly important, particularly in view of Big Data and AI, the question thus arises whether regulators need to step in and provide for a statutory data access regime. In this context, the impact on innovation and competition of mandatory access to data should be carefully assessed. This paper argues that since user data is, in principle, collected by sharing platforms as a by-product without additional substantial investments, a sector-specific statutory data access regime could promote competition as well as innovation.
Keywords
Introduction
Platforms operating in the sharing economy are no longer just facilitators and matchmakers for transactions between otherwise unknown and untrustworthy parties. With every new transaction, these platforms are collecting greater and greater data sets which they may utilize for a multitude of purposes. Besides using the collected data to promote and improve their services, these platforms can also monetise the data in ways that are beyond their role as matchmakers. This is illustrated in particular by the recent yet lesser-known business ventures of two big players in the field, Airbnb and Uber. In 2015, Airbnb launched ‘Airflow’, an open source platform for the processing of data (Miller, 2018). Airflow was originally built by Airbnb’s data engineers, data scientists and data analysts to process the fast-growing amounts of data the company has to handle each day and is now made available for free and marketed in particular to cities as a platform to store their data (Beauchemin, 2015). Two years later, Uber also launched a new tool called ‘Uber Movement’, which provides data and tools to city administrations for urban city planning (Eadicicco, 2017). Both ventures aim for the long-term utilization of urban data, which those sharing platforms generate daily in large quantities and technically as a by-product (Miller, 2018).
Although a great deal of research has been done on the legal and regulatory implications of the services of Airbnb, Uber and the like (Bond, 2015; Cohen & Sundararajan, 2015; Ranchordás, 2015) as well as on the economic and legal side of data markets and digital platforms (Acemoglu et al., 2019; Graef et al., 2015; Haucap, 2019; Richter & Slowinski, 2019; Schweitzer, 2019; Werden & Froeb, 2019), the academic discourse has yet to adequately address the fact that many sharing platforms already have vast collections of data at hand which might not only give these companies an immense competitive advantage, but also allows them to create business models that have little to do with the sharing that made the sharing economy big in the first place. This is particularly troublesome as the abundance of data possessed by many sharing platforms today was strongly enabled by the regulatory leniency that accompanied the sharing economy in its early days (Zale, 2018). Many questions arise from this, first and foremost relating to the legal classification of this development and how to address it from a regulatory perspective. This contribution aims to find answers to these questions. In particular, it has to be examined whether the possession of these large quantities of data constitutes a concern for effective competition that needs ex-ante regulatory intervention, possibly through obligatory data sharing. Such a regulatory solution would stand in contrast to the current sole option of ex-post antitrust sanctions that can only be imposed in the event of market power abuse (Czychowski & Siesmayer, 2018). In this context it has to be taken into account that the data collected through sharing activities varies greatly in terms of quality and its property as a mere by-product to the data of other prominent internet platforms outside the sharing economy for which this issue of obligatory data sharing has been in discussion for a while now (Monopolies Commission, 2015). Particularly the more recent use cases mentioned above, in which sharing platforms use user and transaction data not only for internal process optimisation and product/service enhancement but for completely new purposes of data analytics, demonstrate that the sharing economy is closely linked to the ever-growing data economy which is expected to reach 643 billion € in the European Union (‘EU’) by 2020, representing 3.17% of the overall EU GDP (EU Commission, 2017). And while we are still in the early stages of the data economy, this number clearly shows that any potential regulatory deficiencies should better be addressed sooner rather than later.
The remainder of this contribution is structured as follows: Section 2 outlines the main characteristics of multi-sided platforms in general and sharing platforms in particular as well as their respective effects on competition will be discussed on the basis of economic literature. Section 3 explores the current legal framework in the EU with regard to data access rights in the sharing economy. Section 4 discusses the different options for policy approaches when dealing with data access issues, whereas Section 5 outlines the legal limitations a statutory data access regime would have to adhere to. Section 6 critically appraises the findings. It will be argued that there are valid reasons for the implementation of a sector-specific data access regime for the sharing economy, but that such a regime should be closely monitored and tried out on a small-scale and specific (sub-)sector of the sharing economy first. Section 7 concludes.
Challenges for public governance and policy making
From a perspective of public governance and policy making, the data-heavy environment of the sharing economy presents many challenges. As will be shown in Section 3 below, means of competition law will not always suffice when addressing data access issues in the sharing economy that may eventually hinder effective competition and weaken the innovative strength of other market participants. As limited competition and innovation can lead to reduced consumer welfare which is borne by the public, public governance and policy researchers and practitioners should deal with these policy issues.
Background and motivation
In September 2018 the German government inaugurated the Commission of Experts on ‘Competition Law 4.0’ with the aim of establishing recommendations for a new competition framework for the digital economy (BMWi, 2019). In its final report, the commission found that mainly two economic factors of platform markets present challenges to the current competition framework: first, the strong trend towards concentration due to powerful positive network effects which make it difficult for competitors to challenge an already acquired position of power; and second, the so-called ‘gatekeeper’ function of platforms, which gives a platform control over the competition on the platform and potentially on adjacent markets as well. Taking both aspects into consideration, the costs of non-interference with anti-competitive behaviour are extremely high in platform markets (Kommission Wettbewerbsrecht 4.0, 2019). As the sharing economy is part of the overarching platform economy, the commission’s findings also hold true for data access issues in the sharing economy. It is therefore worth taking an in-depth look at the economic reality of platform markets. Moreover, any sound competition law analysis first needs to gain an understanding of how platforms work, because the specific characteristics of multi-sided platforms bring about certain implications that differ from other, especially single-sided markets (Graef et al., 2015).
Concentration tendencies on platform markets
Online platforms belong to a particular type of market known as ‘multi-sided’, which means they create value by bringing two groups of users together (Graef et al., 2015). The main characteristic of a multi-sided platform is the interdependence of the different sides of the platform (Evans, 2003; Graef et al., 2015). The more users that join one side of the platform, the more valuable the platform becomes for users on the other side of the platform which in turn attracts even more users on the first side and so on (Zale, 2018). In economic terms this phenomenon is called indirect or cross-platform network effects (OECD, 2018). A new user looking to offer his services/assets through a sharing platform will most likely join the platform that already has the biggest user base on the demand side and vice versa. The existence of indirect network effects therefore increases the likelihood that successful companies become dominant in a market (Graef et al., 2015). And since network effects are self-perpetuating, the first platform capable of scaling up, i.e. providing more transactions while reducing the transaction costs, has the highest chances of winning the race to the top. Despite the significance of first-mover advantages, the founder of LinkedIn, Reid Hofmann, pointedly stated: ‘First-scaler advantage beats first-mover advantage’ (Hoffman, 2015; Zale, 2018).
Besides enabling or strengthening a dominant position within a market, network effects can also lead to high entry barriers for new market entrants. A dominant market position, once acquired, can manifest itself strongly and in a short amount of time, so that (potential) competitors have only little chances of bringing it down. This so-called ‘tipping’ of a market, i.e. the tilting of a competitive market to a monopolistic/highly concentrated market, albeit not inevitable, presents a serious risk to platform markets in general and thus to the sharing economy in particular (Kommission Wettbewerbsrecht 4.0, 2019). Nevertheless, multi-sided markets are usually not ‘winner-takes-all’ markets since they evolve relatively slowly as it is difficult to attract customers on all sides of the platform. Also because of product differentiation, heterogeneous customer groups and the general possibility to use several platforms at the same time (referred to as ‘multi-homing’), several platforms are often able to coexist. Thus, these markets are better referred to as ‘winner-takes-most’ or ‘few-winners-take-all’ markets (Graef et al., 2015).
‘Gatekeeper Power’ of a platform
A platform with a strong intermediary position can cause either market side to be dependent on its intermediating services for enabling it to connect with the opposite market side. If no alternative platform has reached a critical mass of users due to strong network effects, the access seeker will be dependent on the market leader as it is the only viable connection to this user base. When focusing on the role of data, platforms can also act as gatekeepers of that user data as the access seeker would not be able to collect the data itself or access it via another platform in an economically viable way (Tamke, 2018). It is hence a common business strategy of platforms to extend their intermediating services to adjacent activities and offer multiple services as a bundle (Kommission Wettbewerbsrecht 4.0, 2019). Airbnb pursued this strategy when it started extending its services from just providing lodging for travellers to so-called ‘Airbnb experiences’ such as city tours and adventure packages in 2016 (Airbnb, 2019). If a platform has a dominant market position on the intermediation market this can result in the transfer of market power to adjacent markets (Kommission Wettbewerbsrecht 4.0, 2019).
Particularly if combined with a dominant market position on the platform market, the gatekeeper position bears the risk of distorting competition on the platform market and adjacent markets. In light of the strong effect of control a platform has over its users, the often fast pace of developments on digital markets and the significance of ‘first mover’ and ‘first scaler’ advantages, the costs of non-interference with anti-competitive behaviour are generally extremely high (Kommission Wettbewerbsrecht 4.0, 2019).
Data and market power
In addition to this very specific competition situation in platform markets, data has become an integral value-creating factor in the digital economy over all (Schweitzer, 2019). With the help of systematic data collection and analytics it is not only possible to optimize existing products or services, personalize them according to individual customer preferences or increase the efficiency of processes, but also to gain valuable insights that go well beyond the core business of the platform (Busch, 2019). The above mentioned tool ‘Uber Movement’, which gathers data from the traditional Uber transportation services and analyses it for purposes of urban city planning, is a case in point (Eadicicco, 2017). This is actually not very surprising since digital intermediation platforms find themselves in a particularly favourable position as they collect large quantities of data about transactions, user behaviour and customer preferences basically as a free by-product to their intermediation services. The privileged access to large data sets can already by itself establish a dominant position through ‘data power’ (Busch, 2019; see also Tamke, 2018).2
This is illustrated by the newly added Sec. 18(3a) No. 3 of the German Act Against Restraints of Competition (‘GWB’) which explicitly qualifies access to competitively sensitive data as one of the five factors in determining a company’s market power in a multi-sided market.
Data about the transaction behaviour of users also creates advantages for sharing platforms. The collection and analysis of information on the behaviour and preferences of users enables the platform to better predict what users will like based on their similarity with other users. In this way a platform can improve its recommendation system with every new transaction. A user is more likely to return to a platform that makes relevant suggestions for future transactions (Graef et al., 2015). In the new data economy, control over data can therefore lead to a key competitive edge. This is particularly true considering the growing significance of artificial intelligence (‘AI’) and its need of access to vast quantities of aggregated data (Schweitzer & Peitz, 2018). Because platforms are highly dependent on their user base, they have an interest in keeping their user data in a closed system (Graef et al., 2015). Smaller or new competitors could then be hindered in growth or even market entry, if they are not capable of collecting or at least buying relevant data in the required amounts someplace else. There is also the possibility that a well-established platform may have already collected such vast amounts of data that it is impossible for any competitor to catch up. This could be particularly relevant in cases where a platform has access to different data sources which allows for economies of scope (Tamke, 2018). Consequently, the question arises whether a dominant platform’s sole access to vast amounts of data alone could be considered an unfair advantage over competitors which would warrant intervention (König, 2018). This question is closely linked to the adjacent question whether such data can qualify as an essential facility and whether a refusal to grant access would constitute a breach of competition law (Graef et al., 2015). However, mere access to data is not the only relevant factor on the road to success. Just as important will be a competitor’s abilities to analyse, structure and, lastly, make use of the data to create an individual benefit for its users (Tamke, 2018).
Similarly, access to data does not necessarily lead to market power and market entry barriers. This conclusion is prevalently true for relatively concentrated markets (Tamke, 2018). In particular, platform markets such as the sharing economy can show signs of market concentration due to the above mentioned network effects (Bamberger & Lobel, 2017). In view of the constantly increasing ‘data power’ of digital platforms, the question of access to data becomes more and more pressing when discussing the shape of the future regulatory framework for the platform economy (Busch, 2019). In fact, although data is commonly referred to as the ‘new oil’ of contemporary economies, the issue of trading and sharing data seems to be a complex one that calls for further research and appropriate legal framing which are both still in their infancy (Haucap, 2019; Kerber, 2017; Di Porto, 2016).
Market players need to have access to large and diverse data sets in order to take full advantage of the ever-increasing amounts of data generated by platform users (EU Commission, 2017). For one, access to user data is critical for the provision of intermediation services (Graef et al., 2015). Secondly and even more importantly, combining complementary data creates economic gains through data aggregation and constitutes an important source for innovation and competitiveness (Busch, 2018). With the help of data analytics, processes and decisions can be optimized, innovations can be accelerated and even future events can be predicted (EU Commission, 2017). Digitalisation has made it possible to collect and analyse extremely large quantities of data from different sources and in different formats at high speed. Herein lies enormous potential for the digital economy and beyond (Tamke, 2018).
However, this becomes difficult to achieve if the generators of the data keep it to themselves. The stakeholder dialogue initiated by the EU Commission as part of the public consultation on building the European data economy showed that stakeholders largely agree that more business-to-business data-sharing would be beneficial (EU Commission, 2018). Other studies indicate insufficient openness and reluctance to share data (Arnaut et al., 2018; Richter & Slowinski, 2019), or if access is granted, the imposition of unfair terms and conditions by the platform for access to its data (EU Commission, 2016b). This is often due to great differences in bargaining power which complicate the access seeker’s position further. In addition, universal standards or guidelines for the transfer of data as well as established, trusted trading platforms are not in place yet. In view of still under-developed data markets, data access claims have to be negotiated individually in each case and therefore create high and often prohibitive transaction costs (König, 2017). It can thus be concluded that although data shows a strong potential for innovation for society, economy, research and polity, sharing of data is still lacking behind. A long-term strategy on data access is needed so that these undesirable socio-economic developments cannot manifest.
Data in the sharing economy
The sharing economy brings about special characteristics that make it stand out within the platform economy, which is why data access issues in the sharing economy should be considered on their own merits.
Sharing platforms are traditionally aimed at consumers to provide the services or assets that are shared over the platform. This common narrative in the sharing economy, i.e. enabling mainly consumer-to-consumer transactions (although it has been repeatedly accused of being misleading), has led to regulatory leniency its early days in comparison to incumbent industries (Zale, 2018). Airbnb and Uber, for instance, both operated in regulatory grey areas in many jurisdictions in their early days and have in part been pushing the boundaries of existing regulation to their favour through what academia has termed ‘regulatory capture’ (Barry & Pollman, 2016; McKee, 2018). Although regulators are now trying to find adequate regulatory frameworks for many activities in the sharing sector, the early regulatory leniency has undeniably facilitated the sharing economy’s quick rise and thus the acquisition of these vast data sets which major sharing platforms such as Uber and Airbnb are now in possession of.
Moreover, the type and quality of the data in the sharing economy as well as its method of collection need to be considered. Uber and Airbnb, for instance, have large amounts of mobility respectively lodging data at their hands, which are of high value to many stakeholders such as city officials for planning purposes as well as private companies looking to develop new products or services (these could be competing or complementing products/services as well as products/services for unrelated markets). With their new ventures ‘Uber movement’ and ‘Airflow’, Uber and Airbnb have already shown there are ways of combining data generated through platform usage and mobile devices with algorithms and data analytics from the AI world to create something that can be used both for public purposes as well as private monetisation (Miller, 2018). In contrast to other prominent platforms which offer their services free of charge to at least one side of the market (e.g. Google, Facebook etc.), users on both sides of commercial sharing platforms are in most cases paying a fee or a commission to take part in the sharing activities. This is a major difference to the aforementioned platforms, whose business model is in fact aimed at the vast collection of data by providing free of charge services in exchange for user data (which can be turned into a profit through marketing to advertisers). The business model of sharing platforms on the other hand does not rely on the monetisation of collected data as the data is provided additionally to monetary contributions by its users. Thus, users in the sharing economy are normally not even aware that they are also paying with their data. Nevertheless, the data collected in the core business can freely be used for all kinds of purposes by the sharing platform. This includes, in particular, a platform’s attempts to expand its market power into adjacent or even completely unrelated markets, as the introduction of ‘Uber movement’ and ‘Airflow’ demonstrates. In contrast, competitors, public authorities and civil society associations have reported problems in accessing data from platforms, as a public consultation conducted by the EU Commission on the regulatory environment for platforms, online intermediaries and the collaborative economy shows (EU Commission, 2016a). These findings are in line with another study of the EU Commission on data sharing between companies in Europe which indicates that the majority of companies questioned was reluctant to share data, based on issues such as trade secrets, fear of misappropriation or considerations of commercial strategy (EU Commission & everis group, 2018). And while empirical evidence on data access and sharing is still limited, all the factors indicated above point towards a highly problematic competitive situation that will likely intensify over time.
This raises the question if the current system of data appropriation in the sharing economy should be maintained, or alternatively, if a statutory data access regime should be implemented, particularly in in light of the delineated competitive restraints, market distortions, limits to innovation and the fact that the data in question was collected merely as a by-product to the primary intermediation service. Since it is the very purpose of economic regulation to guarantee functioning and competitive markets, policymakers should try to find adequate answers to these questions (Drexl, 2018).
Current legal situation in the EU
As a first step in finding answers to the questions set out above, the status quo of the current legal framework regarding rights to data access has to be assessed. This paper will focus on the legal situation in the EU. However, it will not try to find answers to the still legally undecided question who actually “owns” the data that is produced by platform users every day, which is the subject of an entirely own avenue of research (see Determann, 2018; Drexl et al., 2016; Kerber, 2016).
Under European law, the refusal to grant access to user and/or transaction data could qualify as an abuse of market power according to Art. 102 TFEU, which would in such a case lead to a compulsory antitrust license. The basic prerequisite for this legal consequence is an abuse of market power (Louven, 2018). It is to be noted that, in principle, competitive advantages and economic size alone do not give rise to competition concerns. It also does not present an issue under the current legal competition framework if a company grows big on its own efforts and, by doing so, acquires a leading position in the market (Louven, 2018). Even monopolies are legal if they are attained through internal growth (i.e. through commercial success on the market) (Körber, 2017). Instead, it (a.) has to be established that the data-holding platform also holds a dominant market position which in turn (b.) was unlawfully abused when refusing access to its data. Put simply: market dominance itself is not forbidden under competition law, only its abuse (Körber, 2017).
Market power
Proving market dominance according to legal standards already presents a challenge to any petitioner for data access in platform markets. First of all, in the case that market dominance is not already established by ordinary means, it is difficult to prove that control over certain data results in a relevant market dominance. Secondly, how to define the relevant market for data is still an open question, especially when the access claim is not aimed at individual data but rather large collections of data, for instance for data-mining purposes. Also it is still not fully resolved under which circumstances different data sets can be regarded as substitutable (Drexl et al., 2016). Moreover, market power can quickly shift to new players due to the fast pace and dynamics of digital markets (Di Porto, 2016). So far, the European Commission has to a great extent refrained from intervening in such markets by means of competition law, as evidenced in the Microsoft/Skype and Facebook/Whatsapp mergers (EU Commission, 2011, 2014; Drexl et al., 2016).
Abuse of market power/essential facilities doctrine
In the context of access rights, the criterion of market power abuse is further substantiated through the essential facilities doctrine. Originating from American antitrust law, this principle was initially applied to natural monopolies such as bridges or ports but has since been extended to legal monopolies such as patents and other intellectual property rights. In contrast, data access only presents a de facto monopoly due to de facto ownership of data (Louven, 2018). However, the European Court of Justice (ECJ) has applied the essential facilities doctrine accordingly when ruling on access to information (Magill – C-241/91 and C-242/91 [1995] ECR I-743, IMS Health C-418/01 [2004] ECR I-5039, Microsoft T-201/04 [2007] ECR II-3601).
Under the essential facilities doctrine, a dominant company only has to grant access to an essential, non-reproducible resource, if the use of this resource is necessary for a commercial activity on an adjacent market and if the refusal to grant access would fully exclude competition on this market without objective justifications (Czychowski & Siesmayer, 2018; Louven, 2018). In its case law, the ECJ has, however, formulated case-specific criteria for data access claims. In particular, the access seeker needs to prove that the data is essential for the development of a new product or service and that there is no other way to create or otherwise obtain it. As these judgments were issued under the assumption of intellectual property protection of the data in question, it is still unclear whether and how these findings can be applied to unprotected raw data (Drexl et al., 2016). In practice, many claims for data access will already fail on the first step of proving the necessity for the development of a new product or service (Louven, 2018). In particular, when dealing with ‘Big Data’, i.e. data that is large in volume and of unknown or unspecified content, companies usually do not know in advance precisely the product or service they might develop on the basis of such data before getting access to that data (Borgogno & Colangelo, 2018). This is even more true with regard to real-time data that could be supplied through technical interfaces such as application programming interfaces (‘APIs’) (Louven, 2018; Drexl et al., 2016). Also it will be difficult to prove that precisely the data held by one company is essential and cannot be substituted by other data that can be collected elsewhere or by one’s own means (Di Porto, 2016). Moreover, data protection obligations or the protection of trade secretes can constitute objective justifications for refusing access and regularly pose effective barriers to antitrust access claims (Louven, 2018). The essential facilities doctrine has thus been described as inherently ill-suited to tackle competition concerns in data-driven markets (Borgogno & Colangelo, 2018).
Conclusions
These findings present a company demanding access to data with great difficulties, as the duty to grant access under EU competition law is contingent on narrow requirements that are hard to prove. In light of the above and considering the specific dynamics of platform markets it is to be expected that data access claims based on competition law will only be successful in very exceptional cases (Drexl et al., 2016). Accordingly, the predominant view in academic literature rejects the broad application of the essential facilities doctrine in contexts of data access claims. This view is primarily based on the concern that granting access to data constitutes a far-reaching interference with the legally protected entrepreneurial freedom to conduct a business (Art. 16 Charter of Fundamental Rights of the EU) (Czychowski & Siesmayer, 2018; Körber, 2016; Schweitzer & Peitz, 2018; Duch-Brown et al., 2017). While some scholars emphasize the importance of retaining sole power over one’s data as an incentive for future innovations and investments (Körber, 2017; Schweitzer & Peitz, 2018), others conclude that the current legal opportunities for data access are insufficient to meet the challenges of the digital economy, especially in light of ‘Big Data’ and the ‘Internet of Things’ (‘IoT’) (Czychowski & Siesmayer, 2018; Drexl et al., 2016)
There is good reason to think that competition law, in principle, is an inadequate tool to provide a systematic solution to data access problems because of its many inherent flaws: for one, competition law can only intervene in a dominant platform’s behaviour and thus fails to be a solution for access claims against non-dominant companies. On top of that, competition law is only enforced ex post and on a case-by-case basis which stands in the way of a uniform solution that would provide legal clarity upfront. Moreover, since a significant legal burden has to be met in order to force a dominant platform to grant competitors access to its data, actual enforcement will be limited (Drexl et al., 2016). Accordingly, it has been argued that the threshold for considering a refusal to deal as anti-competitive behaviour should be distinctively lower than when dealing with infrastructure or intellectual property rights and that national courts should thus apply the requirements of the ECJ more flexibly respectively modify them to the extent legally permissible. This is particularly true for cases in which the requested data is generated and collected as a side-product and without special investment (Schweitzer et al., 2018). Finally, the generally long duration of corresponding court proceedings creates an immense barrier for safeguarding potentially legitimate interests (Drexl et al., 2016).3
In the Magill case the proceedings lasted for 10 years and in the Microsoft case even for over 14 years (Drexl et al., 2016).
In summary, a high legal burden has to be overcome under Art. 102 TFEU as well as competition proceedings that are, in principle, too cumbersome and too time-consuming to provide the ‘quick and ready answers’ that are needed when dealing with data access issues (Körig, 2018).
When assessing whether regulatory measures should be adopted, the objectives of the regulatory approach should be clarified in advance. It is at the core of economic regulation to provide a fair level playing field for all contenders on a competitive market (Drexl, 2018). This is especially true with regard to the platform economy where competitive imbalances amongst the different platforms and other competitors are quicker to come up due to the special economic characteristics of such markets (see Section 2). With that in mind, and since competition law does not seem to be well-suited to deal with data access issues on platform markets, there might be a need for regulatory intervention to promote competition on those markets, possibly through imposing legally enforceable data access rights.
However, a similarly important objective of any regulatory action in the field of economics is the promotion of innovation and ensuring that necessary and sensible investments are made by market participants so that the economy as a whole can continue to grow. Introducing obligatory data sharing might on the one hand promote competition in certain parts of the market, but it might also deter innovation by incumbents that no longer want to invest in their data collections as such investments do not necessarily lead to a competitive advantage anymore. At the same time, obligatory data sharing might also deter innovation by competitors as they might refrain from investing in their own innovative solutions to collect the needed data themselves as they will have the right to access it from the incumbent (fear of so-called ‘free-riding’) (Tombal, 2019).
On the other hand, a free flow of information often enhances innovation, particularly in the field of technology. Follow-on innovations are more likely to occur if information relating to the basic technology has widely spread. From an innovation policy objective a well-regulated access to data is therefore in the public interest of an innovation-based society (Drexl, 2018).
Any regulation on access to data in the sharing economy should therefore be carefully assessed with respect to its multiple impacts on various policy objectives as it has to be placed between the two goals of restricting competitive constraints and promoting innovation. However, considering the systematic transfer of market power observed for Uber and Airbnb as well as the high entry barriers delineated in Section 2, there are valid reasons in favour of some sort of regulatory intervention.
Possible policy approaches
With that in mind, there are different routes for regulators to go about when trying to handle the challenges to competition and innovation presented by the data-heavy sharing economy.
Self-regulatory approach
Firstly, one could argue that there is no need for binding regulation just yet and good arguments support that view. The platform economy and with it the sharing economy is still evolving and new business models and uses of data are developed every day. Over-regulation, particularly if imposed too early, might stifle innovation so an alternative might be promoting forms of self-regulation while carefully monitoring the development of the market (Richter & Slowinski, 2019). Complementing hard regulation in the form of statutes, regulations or other forms of binding legislative measures, self-regulation such as codes of conduct or guidelines on when and how to share data can also shape and constrain behaviour. Self-regulation does not have to mean no regulation, but simply a shift of the regulatory responsibility to the market actors themselves. Cohen and Sundararajan, renowned scholars in the field of the sharing economy, have long been proponents of using platforms as partners in regulation rather than as subjects of hard regulation (Cohen & Sundararajan, 2015; Vitkovic, 2016).
Codes of conduct could be designed as suggestions, but they could also take a more rigorous form by including sanctions in the event of non-compliance. Essentially, it comes down to how much pressure is needed to persuade market participants to act as desired instead of not adhering to guidelines due to their non-binding character (Richter & Slowinski, 2019). In fact, self-regulation relating to innovation and competition has been in play for many years now in the form of standard essential patent licensing via FRAND declarations (FRAND stands for ‘fair, reasonable and non-discriminatory’ license terms). In the realm of standard essential patents (‘SEPs’), i.e. patents to a technology that has become an industry standard, FRAND principles have been successfully guiding licensing negotiations between different companies and even between competitors. In short, FRAND principles work as follows: if a company holding a SEP issues a FRAND declaration, i.e. agreeing to license to interested parties on fair, reasonable and non-discriminatory terms, the license seeker and the licensor need to follow specific procedures before either can take the case to court (Louven, 2018; Richter & Slowinski, 2019). Both the patent holder and the license seeker are incentivized to follow the procedure because the patent holder will be sheltered from antitrust remedies and the license seeker will be protected from the threat of injunctions (Borgogno & Colangelo, 2018).
Applying FRAND principles to cases on access to data could prove to be useful. Started by the European Union (EU Commission, 2017), several authors have continued the discussion on how to apply the concept of FRAND in the context of data sharing (Borgogno & Colangelo, 2018; Drexl, 2017a; Früh, 2018; Louven, 2018; Richter & Slowinski, 2019). The main difference between SEPs and data, however, lies in the fact that SEPs, just like other patents, are protected by intellectual property rights which grants the patent holder exclusivity with regard to the patented technology. Data, in contrast, is not subject to such a legal protection – it is only protected by factual circumstances as it is saved on computer systems that are controlled by a certain party. In other aspects SEPs and data are very alike: both can be essential for certain market activities and are non-rivalrous in nature (Richter & Slowinski, 2019). However, courts have not been uniform in their decisions on the enforceability of a FRAND declaration: some courts view FRAND declarations only as a statement of intent with the licensing obligation (and its prerequisites) still resulting from competition law whereas others regard it as a binding offer that constitutes an independent legal basis for a license (Richter & Slowinski, 2019). If applied to the context of data sharing the same dispute would arise. It would not be until data sharing according to FRAND principles was the subject of extensive case law before ex ante legal clarity for market participants could be achieved (Richter & Slowinski, 2019).4
In this regard it is notable that the lack of sanctions and clear enforceability of FRAND rules resulted in extensive global litigation that was only put to an end by the intervention of competition authorities and courts after years of seemingly unlimited court proceedings (Richter & Slowinski, 2019).
However, even regardless of competition policy, providing official guidance on data sharing, promoting technical standardization as well as adopting standard terms as currently contemplated by the EU Commission in an attempt to reduce sharing transaction costs is a wise approach in any case (EU Commission, 2017; König, 2018).
A second option would take into consideration the need for further research with regard to the competitive effects of a regulatory solution as well the overall framework of an effective regime of access. In light of this, it could seem preferable to refrain from regulatory intervention for now and closely monitor market developments in the sharing economy instead so that potential market failure could be identified early on. There are good reasons for this approach as well: in view of the diverse factual constellations in different market sectors of the sharing economy, the complex effects on competition and the high error costs in case of misjudgements, any new data access regime needs to be thoroughly vetted (Kommission Wettbewerbsrecht 4.0, 2019).
Statutory data access regime for the sharing economy
Another option would be to introduce a new data access regime for companies in the sharing economy by way of statutory provision. Such statutory data access rights are not completely new within the European legal framework. For one, there is the EU Regulation on Type Approval of Motor Vehicles (Directive 2007/46/EC) which grants independent repairers and the like a right to unrestricted and standardized access to vehicle repair and maintenance data. Besides implementing the access right itself, the regulation also provides for certain modalities of the data sharing, in particular with regard to technical requirements and the appropriateness of the access fee (König, 2018). The regulation is aimed at preventing car manufacturers from abusing their market power to foreclose competition on the aftermarket for car repairs and maintenance (Directive 2007/46/EC). It is thus an example for regulatory intervention via a new data sharing regime in response to competition concerns in a distinct market.
Similarly, the EU Directive on Payment Services 2 (‘PSD 2’, Directive (EU) 2015/2366) implemented a right for payment service providers to access their customer’s bank account data from the respective banks. The Directive thereby effectively brought down artificial market entry barriers for providers of innovative online payment services that were created and upheld by incumbent banks trying to prevent intensified competition on the payment market (Drexl, 2018).
Even the EU General Data Protection Regulation (‘GDPR’, (Regulation (EU) 2016/679) could serve as a reference for future data access regimes. In Article 20 GDPR the first ever, cross-sectoral data portability right with regard to personal data was implemented. The legislators of the GDPR explicitly viewed the portability right not only as a tool to strengthen the data autonomy of the private individual, but also to promote the free flow of personal data in the EU and to foster competition between data holders (Drexl, 2018). Hence, with regard to competition policy as well as from a dogmatic legal standpoint, a new data access right would resemble Article 20 GDPR in many ways (Drexl, 2017b). Since its introduction, an active discussion has emanated in jurisprudential literature on the question whether the data portability right of the GDPR could serve as a starting point for a general data access regime (Drexl et al., 2017; Richter & Slowinski, 2019; Tombal, 2019).
However, it is advisable to restrict the applicability of any access regime to a specific sector as well as adapt its individual provisions to the needs of the sector in question. This way a more targeted and therefore more effective regime can be put in place while minimizing the competitive risks identified in Subsection 3.4, i.e. reduced innovative strength and investments (Drexl, 2018). Since the sharing economy is not a single market but constitutes of multiple, very heterogenous markets, such access regulation should preferably address a specific sector or sub-sector of the sharing economy or, as the delineation of sectors is often problematic, any other clearly defined market area. The decision in favour or against access regulation should reasonably be based on a careful economic assessment of the competitive respectively innovation-related impacts such an access regime would have on the respective sector and/or sub-sector and the sharing economy as a whole. Introducing a data access regime in the strictly limited frame of a specific and carefully selected sub-sector of the sharing economy to gain experiences and test solutions could pose a conciliatory solution to the ‘wait-and-see’ approach above (Kommission Wettbewerbsrecht 4.0, 2019).
Legal limitations of any statutory data access regime
In any case, there are certain legal limitations a statutory data access regime would have to respect, of which the most pressing ones will be discussed in the following chapter.
Data protection law
Insofar as the access request is directed at (amongst others) personal data, data sharing is limited by data protection law. Data protection law can thus strengthen the de facto position of control over data (Schweitzer et al., 2018). This can be avoided, however, if the data request only relates to anonymized personal data, i.e. data rendered anonymous in such a way that the data subject is not or no longer identifiable, which places the data request outside of the scope of the GDPR (Regulation (EU) 2016/679). True anonymization sets an extremely high bar which can often only be achieved by way of data aggregation, i.e. merging individual data to a set of aggregated data in such a way, that it is impossible to identify its individual components (Ernst, 2018). Consequently, data protection can even be considered a driver of innovation as the digital economy can only strive if it is accepted by consumers due to their privacy concerns being respected (Drexl, 2018). The implementation of a data access regime could even pose an opportunity to resolve the existing uncertainties on the legal requirements for proper anonymisation.
EU Database Directive
The EU Database Directive (Directive 96/191/EC) could potentially pose the biggest hurdle for extensive data sharing. Irrespective of the question whether data should be subject to a new data ownership right for the manufacturer of the data, the company collecting the data could be entitled to an intellectual property right regarding the database as such. Article 7 of the EU Database Directive stipulates a sui generis intellectual property right for the benefit of the database manufacturer if substantial investments have been made for the procurement, presentation or validation of database contents. This sui generis right prohibits the extraction or re-utilization of database contents in whole or in substantial parts unless the database manufactures expressly consents.
As any regulation would have to be in accordance with existing legal provisions, particularly with higher-ranking EU legislation, even a newly imposed statutory data access right could not curtail the database manufacturer’s sui generis right pursuant to the EU Database Directive, as far as applicable. Although concerns about overprotection and restrictive effects on competition in aftermarkets have been voiced from the very start, the ECJ has shown a tendency to interpret the directive in favour of the data manufacturer in its jurisprudence over recent years (Hugenholtz, 2017; König, 2017 with further references to the ECJ case law). Even when the database in question was not the result of a deliberate investment decision but rather a by-product of a company’s primary commercial activity (so-called ‘spin-off database’) that would have been generated either way with or without legal protection, the ECJ awarded the manufacturer exclusivity through an intellectual property right. In light of the directive’s goals of providing incentives for innovation and protecting investments it is fair to say that the ECJ’s case law has led to an unnecessary overprotection in those cases where the incentive function was clearly not very relevant to the procurement of the database due to the specific circumstances of the case (König, 2017; Tombal, 2019).
It is thus a legitimate concern that the enforcement of a potential data access right would be severely restricted if a data holder effectively invokes its database protection right. A possible solution that takes the interests of both sides into account would be to adopt a stronger purposive interpretation of the “significant investment” criteria by the courts (König, 2017). Then the incentive function can achieve its full effect without overly restricting competition.
Critical appraisal and final assessment
As elaborated in Section 2, there are certain competition concerns in the sharing economy when one or only a few companies have exclusive access to particular data sets and restrict others from accessing it without legitimate business justification. This kind of behaviour can limit innovation and eventually hurt consumers, so when this kind of anti-competitive behaviour can be found in a market, there are good reasons for policymakers to intervene (Castro & Steinberg, 2017). At the same time, innovation could also be affected negatively if former drivers of innovation are discouraged from continuing to invest in the collection of raw data because the competitive upside is no longer worth the corresponding investment. Particularly the trade-off between long-term and short-term effects on competition has to be thoroughly assessed when considering introducing new means of data access. Obliging a dominant firm to grant access to one’s data may stimulate competition in the short-term, but it might simultaneously reduce innovation in the long run (Graef & Husovec, 2017). The crucial question therefore is to what extent data-related innovation will be promoted or slowed down if policy makers change the current legal system of data appropriation in the sharing economy (Graef & Husovec, 2017).
Irrespective of the above, a regulatory solution would have certain advantages over data access proceedings solely based on competition law. For one, a regulatory framework can be more specific and tailored to the particular needs of the (sub-)market in question, for instance with regard to technical aspects of sharing data or determining appropriate compensation for granting access, thereby avoiding some of the practical issues inherent in competition law enforcements (König, 2018). The scope of antitrust measures is also limited by its case-by-case approach. Regulatory intervention would thus be more suited to deal with core problems in competition that go beyond an individual case (Borgogno & Colangelo, 2018). A regulatory initiative would also provide more legal certainty as the prerequisites for access claims are clearly specified in advance so that market participants are less dependent on the judgment of individual courts. Moreover, psychological barriers to request or even sue for access are taken down if the access seeker can rely on official regulation instead of case law that might be altered in the future, therefore the number of data access requests – and eventually actual data sharing – would likely go up.
The EU Commission has, for now, opted to give priority to the principle of freedom of contract and to only give some guidance by proposing key principles when engaging in data sharing agreements (EU Commission, 2018a; Tombal, 2019). Accordingly, the EU draft directive on online intermediary platforms only proposes new standards for more transparency and fairness without establishing a specific data access right (EU Commission, 2018b). Although a step in the right direction, the draft directive does not pose a solution for strong discrepancies in bargaining power which can still lead to market failures and hence to reduced innovation and public welfare. The access seeker would have to resort to antitrust law with its limited scope of application (see Subsection 3.2) and might thus be left without legal remedy, for instance in cases where the access request is aimed at a non-dominant platform (Tombal, 2019).
From an innovation point of view, and particularly with regard to the sharing economy, there are reasons to believe that the EU draft directive with its focus on transparency and fairness will not be enough. For personal data, the GDPR provides for a novel data access regime that promotes the free flow of personal data and competition. For non-personal data, an equivalent to the GDPR is still absent, even though it can be assumed that a regulatory data access regime should have positive effects on innovation and competition if the data in question was generated as a mere by-product without substantial investments flowing into its collection. At least insofar as sharing platforms are aimed at the sharing of goods respectively the provision of services in the non-digital world, such as for instance in the case of the sharing economy’s most prominent contenders Uber and Airbnb, the focus of the platform’s business model does not lie in the procurement of user and transaction data beyond the scope of service provision and optimisation. This constitutes a fundamental difference to other platforms such as social media platforms or search engines, whose business models are primarily directed at the analysis and exploitation of user data. It can thus be argued that user data does not legitimately belong in the sole hands of a sharing platform as it was obtained from its users on top of monetary considerations. In that sense, aggregated user data rather ‘belongs’ to the users collectively as they collectively invested in that data through the execution and payment of transactions over the sharing platform. The data should, therefore, be made available to the users, and in extension of this argument, to society as the individual user generally lacks use for the data in question. In this regard, certain user data from sharing platforms resembles public sector information, which society indirectly invested in through tax money, and which has to be made publicly available based on the Directive 2003/98/EC on public sector information (Custers & Bachlechner, 2017).
A balanced access regime could, for instance, make it a prerequisite for justified data access claims that the generated data was produced as a by-product without substantial investments having been made into it in order to prevent free-riding tactics. In this case, the EU database directive would no longer be applicable and thus not pose a problem for effective data sharing. Also, congruity with the legislative intent behind intellectual property rights (IPRs) would be achieved. The ratio legis of IPRs is to incite innovations and protect the necessary investments by providing an opportunity to capitalize on them (Custers & Bachlechner, 2017). As has been pointed out above, ownership of data acts as a de facto right over the data whereas IPRs provide de jure positions of power over the protected intangible assets (Louven, 2018). This raises the question why a solely de facto position of power should provide a stronger protection than the carefully balanced legal protection under an intellectual property rights regime. Intellectual property rights regimes principally show a thorough consideration of the interests of the rightsholder (opportunity to capitalize on investments) and the interests of society (free access and use of intellectual property). This is reflected in legal barriers to protection such as in the stipulation of pre-conditions (for instance, a certain level of originality for copyright protection), registration requirements, terms of protection and, most importantly, permitted uses (in particular pursuant to the fair use doctrine). However, a solely de facto data ownership currently provides for a stronger protection of data than it would have under intellectual property protection as no explicit statutory exceptions and permitted uses exist outside of general competition law claims with its inherent flaws outlined above. This is also true if the data in question was obtained without substantial investments having flown into it, so that a justification for such high degree of protection cannot be found even from this angle. A data access regime could realign the imbalances in the degree and justification of data protection based solely on ownership compared to the protection of other intangible assets essentially resting upon intellectual property rights.
Moreover, in light of the fact that platform markets such as the sharing economy are particularly prone to tipping, a data access regime could introduce a counteracting force in the market. Particularly in the event that a sharing platform has gained its prominent market position at least in part through operating in regulatory grey areas in its early days, it seems justified to introduce a counterforce for other market participants that were not able to enjoy the benefits of the somewhat questionable start-up support.
When it comes to data and platform markets, a forward-looking solution is needed which does not solely focus on the presence, but also bears future developments in mind. Competition economics are generally well suited to detect problems with market power and competition early on. The high level of threat to innovation and competitiveness with regard to data in the sharing economy as illustrated in Section 2 justifies early intervention. If regulators hesitate until market failures manifest, they might have not only missed the chance to steer innovation processes in a more favourable direction. It will also be much harder to effectively resolve the market power and competition problems that have consolidated by then.
However, as we are dealing with a part of the economy that is still in its infancy which makes competitive effects difficult to predict, any new data access regime should first be tested on a small scale, e.g. in a strictly limited and carefully selected sub-sector, before broader application over multiple sectors of the sharing economy. In a test environment, the dynamics of the market in question can be closely monitored and valuable experiences can be gained which can form the basis for further evidence-based policies.
Conclusion
The analysis has shown that the sharing economy presents certain challenges when it comes to questions of data access and data sharing with respect to ensuring effective competition and promotion of innovation. As part of the bigger platform economy, markets in the sharing economy tend to show similar key economic features such as concentration tendencies due to strong network effects. Another distinct characteristic of platform markets is the gatekeeper function of the platform which is based on its intermediary position between two market sides. The gatekeeper function is not only exercised in relation to the user group on the opposite side of the market but also to the user data collected by the platform. In light of Big Data, artificial intelligence and data analytics, access to large amounts of data is becoming increasingly crucial for competitiveness and innovation in the digital economy. Market participants can therefore depend on gaining access to a platform’s user data in order to develop competing or complementary products or services. If a contractual agreement on data access cannot be obtained, for instance because of differences in bargaining power between the platform and the firm seeking access, current EU law only provides the latter with a legal claim to data access only if the refusal to deal is considered an abuse of market power pursuant to Art. 102 TFEU. However, due to its narrow requirements and often long proceedings, competition law is a particularly inadequate tool to provide an effective solution for data access issues. Moreover, since competition law strictly follows a case-by-case approach, it is not suited to regulate the specifics of access provision on a bigger scale. Also, it does not provide the legal clarity for multiple stakeholders to ensure that the preconditions to facilitate innovation and competition in the sharing economy are met. It can thus be argued that the question of data access should better be answered through a statutory data access regime that not only addresses all the corresponding questions of data sharing (under what requirements should data access be made mandatory? Which data shall be subject to the data access? How should data access be technically implemented and against what remuneration? etc.) but also adopts the provisions to the needs of the specific (sub-)sector. Considering the fact that user data in the sharing economy is, in principle, collected as a by-product without additional substantial investments having been made for its collection, there are good reasons to believe that a sector-specific data access regime for the sharing economy will have a stimulating effect on competition and innovation even beyond the regulated sector, provided data protection regulation and the protection of company and trade secrets are ensured, for instance through data anonymization by way of data aggregation. However, given the heterogenous nature of different markets within the sharing economy as well as its comparatively short existence, a sector-specific data access regime for the sharing economy should be thoroughly tested within a restricted framework to avoid false positives in regulation.
The right to public sector information pursuant to the EU Open Data and Public Sector Information Directive (Directive (EU) 2019/1024) presented a first step on the way to a comprehensive European data culture that promotes not only the responsible use of data but also an open culture of innovation. Through cloud infrastructures the advantages of data sharing can easily be made available to a variety of stakeholders. Besides market participants this also includes the civil society as well as public bodies for the enhancement of public services such as city planning. A carefully designed sector-specific data access regime to data in the sharing economy could prove to be an important element in advancing data-driven innovations in the EU and could become a cornerstone in the European vision for the new digital era: guaranteeing a justified access to and a better usage of data for the benefit of the entire society.
