Pengfei Sui, Chinese University of Hong Kong-Shenzhen
Abstract: Using novel data on social interactions and individual trading records in the Bitcoin market, we document evidence of social learning leading to sentiment contagion. Investors significantly update their beliefs about Bitcoin in the same direction as the average peer sentiment, although it is not informative about future prices. Our findings indicate inefficiency in social learning, consistent with the echo chamber effect and selective interpretation of signals. Moreover, social learning affects both individuals’ trading decisions and aggregate market outcomes. We construct a novel measure for the intensity of sentiment contagion resulting from social learning, which significantly predicts Bitcoin volatility, volume, and crash.
Discussant: William Mullins, University of California-San Diego
Abstract: We investigate the effect of social media adoption on stock market participation in the United States. Using plausibly exogenous variation in the early adoption of Twitter across counties, we show that a 10% increase in social media usage is associated with a 2.5% higher rate of stock ownership and an overall increase in stock market wealth. Consistent with the idea that social media can lower the cost of accessing information, we find that Twitter adoption is associated with a decline in the number of financial advisors and has larger effects on stock ownership in counties with lower levels of pre-existing stock market knowledge. Twitter adoption also fuels interest in “meme stocks”, which tend to be more volatile and owned by retail investors. Overall, our results suggest a distinct impact of social media platforms on household portfolio choices that differs from that of other modern information technologies.
Discussant: J Cookson, University of Colorado-Boulder
Abstract: We provide causal evidence of the peer effect on equity investment in a large-scale natural experiment in Taiwan. We show that retail investors respond to the investment decisions of their military peers who were randomly assigned in compulsory military drafts: retail investors participate more in the stock market, invest more in stocks that peers hold, and obtain more dividend gains and capital gains. Our investigation indicates that retail investors learn valuable information from their peers to make profitable investment decisions. These effects are more pronounced among peers who are more sophisticated and among stocks entailing less behavioral bias. Stocks with more peer clientele outperform stocks with less clientele.
Discussant: Byoung-Hyoun Hwang, Nanyang Technological University
Abstract: Fraud indicators in the Paycheck Protection Program (PPP) COVID relief program are highly geographically concentrated. Areas with high PPP fraud also have heightened indicators of suspicious Economic Injury Disaster Loan (EIDL) advances and unemployment insurance claims. Zip codes and counties with high rates of suspicious PPP loans exhibit strong social connections to one another with evidence of fraud spreading over time through social connections. Additionally, individuals in suspicious social media groups have higher rates of PPP fraud, and socially connected zip codes frequently use the same specific FinTech lenders and EIDL agents, consistent with social connections influencing detailed loan decisions.
Discussant: Karsten Müller, National University of Singapore
