【主讲人】江嘉骏 教授
【主持人】高明 长聘副教授
【报告题目】Behavioral Personalization and Algorithmic Explainability: Evidence from a Field Experiment on Robo-Advising

【报告摘要】Algorithmic financial advice has the potential to improve investment performance but often suffers from low adoption. We conduct a large-scale field experiment in partnership with a major commercial bank in China to provide the first causal evidence on how behaviorally informed personalization and algorithmic explainability affect investor take-up of robo-advice. We find that a simple rule-based algorithm that tailors recommendations based on elicited behavioral traits including risk and loss aversion significantly increases the likelihood of investing in the recommended asset category by 8.11% relative to a standard Sharpe ratio–maximizing baseline. Providing investors with feedback on their elicited preferences, along with an explanation of how these inform the recommendation, also increases uptake by 14.24%. Similar significant increases are observed in both the amount held in the recommended asset type and its share in the overall portfolio. These effects are shown to be persistent over time. Our findings highlight the value of incorporating behavioral preferences and explainability in promoting more inclusive adoption of robot advice.

【主讲人介绍】复旦大学经济学院副教授。北京大学光华管理学院金融学博士，数学与应用数学学士。主要研究领域为投资者行为偏误、智能投顾、机器学习与资产定价等。主持国家自然科学基金、上海市哲学社会科学规划课题等研究项目，研究成果在《经济研究》《金融研究》《管理科学学报》和Journal of Economic Behavior & Organization、Journal of Money, Credit and Banking等国内外权威期刊发表。