Human-Robot Interactions in Investment Decisions

We study the introduction of robo-advising on a large set of employee saving plans. Different from many services that fully automate portfolio decisions, our robo-advisor proposes investment and rebalancing strategies, leaving investors free to follow or ignore them. The resulting human-robot interactions occur both at the time of the subscription and over time, as the robot sends alerts when the investor’s portfolio gets too far from the target allocation. We show that the robo-service is associated with an increase in investors’ attention and trading activities. Following the robot’s alerts, investors change their rebalancing behaviors so as to stay closer to their target allocation, which results in larger portfolio returns. Counterfactual returns induced by automatic rebalancing by the robot would be only slightly higher, suggesting that, on average, the financial cost of letting investors retain control is not large. This paper was accepted by Jean-Edouard Colliard, Special Issue on the Human-Algorithm Connection. Funding: This work was supported by Observatoire epargne europeenne, as well as LTI@Unito and Agence Nationale de la Recherche [Grant ANR-17-EURE-0010] to M. Bianchi. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.03886 .