Understanding the impact of robo-advisors on individuals' stock investment : evidence from a quasi-experiment

Abstract

Financial institutions are increasingly adopting robo-advisor services due to their advantages over human advisors, such as cost-efficiency and accessibility. However, it remains unclear to what extent robo-advisors influence individuals’ investment decisions, particularly in high-risk assets, and the mechanisms that drive their effectiveness. To address this gap, we utilize proprietary stock-transactional data from Sinolink Securities, encompassing over 100,000 clients across diverse wealth brackets. In our study, we conduct a quasi-experiment comparing free and paid users of robo-advising services with investors who do not use these services. By observing the signal of robo-advising with a one-period latency for free users, we precisely estimate the direct impact of robo-advising on mimicking stock investment decisions and its spillover effect on investments within the same industry. Additionally, we examine three key channels that enhance the usage of robo-advising: information disclosure, performance, and salience. Our findings reveal that information scarcity, the relative outperformance of robo-advisers, and the salience of recommended stocks significantly contribute to the positive effect of robo-advisors on users’ subsequent investment decisions. We also confirm that mimicking robo-advisors’ recommended stocks yields positive raw and adjusted returns, particularly for paid investors. Further analysis highlights that demographic and financial factors, such as gender, experience, and financial wealth, shape investor responsiveness to robo-advisor recommendations. In summary, our study provides insights into the impact and mechanisms underlying the growing adoption of robo-advisor services. These findings have important implications for financial institutions and investors seeking to optimize the use of robo-advisory tools.