时间：2024年3月21日 15点

地点：邵逸夫科学馆401

题目：Robo-Advisor Illusion

内容摘要：The rise of Robo-advisors in the financial market has garnered significant attention. This study aims to investigate whether investors exhibit a preference or aversion towards Robo-advisors and the underlying reasons behind such preferences. Specifically, we experimentally examine individuals’ preferences for investment advice from a Robo-advisor versus a Human advisor. We employ a unique design where the capabilities of the Robo-advisor and Human advisor in the investment task are objectively identical. We elicit individuals’ preferences for either the Robo-advisor or the Human advisor. We test two hypotheses to explain these preferences: 1. Preference for the Robo-advisor may stem from belief bias (illusion), wherein individuals believe that the Robo-advisor performs better in providing investment advice. 2. Preference for the Robo-advisor may arise from source preference, wherein investors lean towards the Robo-advisor while holding the belief that both types of advisors perform equally. Our findings provide supportive evidence for both hypotheses. Moreover, we find that belief bias and source preference are positively associated with individuals’ good prior experiences with advisors, but not familiarity or knowledge in Fintech. Overall, our results indicate that individuals possess distinct preferences for advice from Robo-advisors compared to Human advisors, and these preferences can be attributed to belief bias or source preference.

专家简介：李景景：深圳大学深圳南特金融科技学院副教授，香港科技大学经济学博士，德国Max Planck Institute of Economics博士后。研究兴趣为实验经济学、行为经济学和行为金融，他的工作重点是理解现有理论难以解释的选择和与社会重要问题相关的题目。

主要学术贡献：发表于2013年Experimental Economics 关于记忆偏见的论文 Li (2013) “Asymmetric Memory Recall of Positive and Negative Events in Social Interactions” 为以实验经济学研究记忆偏见的第一篇论文。Li (2017, Journal of Economic Psychology) 使用双语被试的实验研究发现语言（中文 vs. 英文）可以影响被试的社会偏好与决策，与使用英文相比，实验指引是中文时被试在信任游戏中更加信任他人与值得被信任， 这是实验经济学文献中最早研究语言如何影响行为的实验研究（工作论文版本于2010完成）。Li (2011, Journal of Risk and Uncertainty) 通过实验研究风险中决策决策者对不同控制方法的偏好, 发现了对于大多数表现出控制偏好的人来说，他们的偏好不是由于控制错觉（Ellen J. Langer，1975），而是源于来源偏好 (source preference), 是文献中最早研究随机偏好的实验研究。