Financial bots and retail trade

Abstract

In this study, I examine the pervasive presence of financial bots, which are unverified automated accounts on social media platforms involved in online stock discussions, and their impact on retail trade. I find that financial bots generate 70% of stock-related tweets, and that financial bots’ human followers generate half of human tweets. Individual financial bots are 2-3 orders of magnitude more active than humans but interact less with other users. Most financial bots show strong affiliation with external websites. Financial bots are more active on days with major firm events, and their coverage is positively associated with firm size, analyst following, recent stock returns, and shareholder count, but negatively associated with institutional ownership, suggesting that financial bots cater to the information demand of retail investors by covering highly visible stocks that attract retail investors. By comparing the retail trading activity during half-hour windows around the post of bot tweets, I find that retail trading volume significantly increases upon and after bot tweets, consistent with financial bots triggering retail trade. Furthermore, the ability of retail order imbalance to predict subsequent stock returns decreases upon bot tweets, consistent with financial bots induce attention-based trade rather than information-based trade.

Date of Award	2022
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Amy Yunzhi ZANG (Supervisor)