Regulating Recommender Systems? Effects of Data-Based Individualization (and its Limits) on Competition in the Digital World

ABSTRACT Data-based algorithmic recommender systems (DARSs) shape how users access, evaluate, and consume information and goods. These systems encompass search rankings tailored to estimated user preferences and direct recommendations. Their growing influence has prompted regulatory interest worldwide, with debates centering on their economic, social, and cultural implications. Drawing on attention economics and behavioral insights, the paper highlights the functional necessity of pre-selection mechanisms in information-overload environments. Personalized DARSs improve preference matching, expand the diversity of content receiving attention, and tend to intensify competition. However, DARSs also carry significant risks: they may reinforce biases through self-preferencing, amplify echo chambers, limit exposure to diverse viewpoints, and raise privacy concerns. Based on these challenges, this paper provides a comparative institutional analysis of regulatory options for DARS, evaluated through an economics-based framework. It examines regulatory effects across three dimensions: (i) preference fit, (ii) information transparency, and (iii) competition intensity. The paper evaluates a range of regulatory strategies and although each option addresses specific risks, the analysis shows that more interventionist regimes often come at the cost of reduced competition and diminished content diversity. Therefore, effective regulation should focus on addressing core pitfalls without eroding the systems’ welfare-enhancing functions.