Google’s dominant position in online search may distort services markets that depend on referrals from search engines. The EU Digital Markets Act seeks to tackle the root cause of platforms’ dominance: network effects. Network effects occur in search engines because more users generate more data that improve the quality of search and attract even more users to that search engine. The DMA seeks to break that data feedback loop by forcing very large “gatekeeper” search engines like Google to share user data with smaller search engines.
Asymmetric data sharing stimulates competition but also fragments the user dataset that search engines can access. Empirical evidence suggests that this reduces the quality of search, especially for rare keyword searches. Symmetric data sharing, irrespective of size, could reduce that tension between competition and consumer welfare objectives.
Recent advances in Generative AI, such as ChatGPT, open a new competition frontier in search. Search engines produce a list of webpages; users have to extract a reply from these pages. Chatbots produce a natural language reply to queries, based on the contents of relevant webpages. They are pre-trained on webpages and do not depend on network effects and user feedback. That facilitates competition. Chatbot replies are not always reliable and may require human verification, using relevant webpages. As such, chatbots and search engines complement each other. The DMA search data sharing obligation may still be useful because it can strengthen that interaction, provided it does not reduce the quality of search.
The Why Axis is a weekly newsletter distributed by Bruegel, bringing you the latest research on European economic policy.
B. Martens (2023) ‘What should be done about Google’s quasi-monopoly in search? Mandatory data sharing versus AI-driven technological competition’ Working Paper 10/2023, Bruegel, available here.