Data Concentration and Competition in Digital Platforms

Policy discussions in major countries regarding access to data as a means of promoting competition in the digital platform market are taking place within the framework of existing legislation. The existing legal framework determines whether access to data is necessary based on whether the data is essential for competition in the relevant market and whether it can only be obtained from the dominant platform. In competition law, discussions about whether other companies can access a dominant firm’s resources have traditionally taken place under the so-called essential facilities doctrine. Whether data is essential for competition in a given market cannot be determined uniformly. However, existing literature suggests that narrowly defined essential data does not exist in many markets, and it is also difficult to find such data in markets related to large digital platform companies like Google or Amazon. This is because businesses can be established and operated in search engines or online marketplaces without having extensive user data. From this perspective, the role of competition law in ensuring access to data under the existing legal framework is inherently limited. These limitations arise not only in relation to the essential facilities doctrine but also in the process of applying competition policy. Competition policy typically takes a long time to implement, while the rapidly changing business environment of the digital platform market requires companies to quickly identify business opportunities in order to enter the market. Even if it is proven that a dominant company has used its data in an anticompetitive manner, it is extremely difficult to impose appropriate remedies that could restore the market to its previous competitive state. Additionally, designing complex remedies and continuously monitoring compliance with competition policy is not easy, which is why there are ongoing discussions about the need for ex ante regulation. However, ex ante regulation also has its own costs and drawbacks, so the practical benefits of implementing such regulations must be carefully evaluated. One potential measure to limit the scale of user data held by market incumbents is to restrict the retention period of raw data, such as search queries entered by users or location data. According to Chiou and Tucker (2017), even with a shortened retention period, the learning capacity of platforms inferred from data did not significantly diminish. By applying incremental learning techniques, where new models are trained based on existing models and continuously refined with new data, the platform's learning capacity could remain largely unaffected, even with a reduced data retention period. However, if a significant amount of raw data is quickly deleted, there may be some impact on the flexibility and efficiency of the learning models. This is suggested by the fact that, despite Chiou and Tucker's (2017) findings, many analyzed platforms extended their data retention periods after initially shortening them, citing the need to enhance personalized services. On the other hand, if data retention periods are shortened, market incumbents with large user bases may still gather significant amounts of data within the limited retention period, while market entrants with smaller user bases may struggle to collect enough data for meaningful analysis, potentially reinforcing the competitive advantage of incumbents. Additionally, a shorter data retention period could limit the amount of user data shared by incumbents with new market entrants. This would likely restrict the ability of entrants to narrow the gap in insights derived from data compared to incumbents. (the rest omitted)