Baoqing Gan
Investor Sentiment Under the Microscope (job market paper)
Market-wide investor sentiment is known to exert influence on stock prices. Fewer studies have explored the impact of firm-specific investor sentiment on stock prices. Using the most granular intraday sentiment measures available, the minute-to-minute Thomson Reuters MarketPsych Indices (TRMI), we examine how the overnight build-up of investors' mood in news and social media affects opening stock returns. Our analysis reveals that sentiment formed during non-trading hours is a strong predictor of opening returns. Moreover, the sentiment generated by increasingly popular social media exerts a greater impact on opening prices than the sentiment found in traditional news media. Our results show that the impact of sentiment is asymmetric, with negative sentiment having a preeminent impact on opening returns. We find that the influence of sentiment quickly diminishes after the first minute of trading. These findings are consistent across a number of different models and specifications, providing further evidence against non-behavioural theories in this fast-paced digital era.
Do Emotions Trump Facts? Evidence from around the World (working paper)
The influence of investor sentiment on stock markets has been demonstrated previously, but the majority of the studies are either US-centred or focus on a single source of sentiment. In this study, we contrast the influence of social and news media to investigate how sentiment from these two sources impact markets in Australia, Brazil, Canada, the EU, France, Germany, Hong Kong, India, Japan, Singapore, Spain, Switzerland, the UK and the US. We find that the heightened social and news media sentiment during non-trading periods significantly affect the next-day opening returns even after accounting for previous day market activity. We discover that only the US stock market shows stronger reactions to social media sentiment compared to news, while other markets are more responsive to news media sentiment. Robustness tests demonstrate that the aggregation of sentiment up to three hours before the market open generates the most effective signals in predicting the index opening values. Overall, this study assists in our understanding of the price discovery process in international stock markets, with a novel dataset of high-frequency textual-based sentiment and an approach that helps to disentangle the return-and-sentiment feedback loop.
(Gan, B., Alexeev, V., Bird, R. and Yeung, D., 2020. Sensitivity to sentiment: News vs social media. International Review of Financial Analysis, 67, p.101390. )
We explore the rapidly changing social and news media landscape that is responsible for the dissemination of information vital to the efficient functioning of the financial markets. Using the sheer volume of social and news media activity, commonly known as buzz, we document three distinct regimes. We find that between 2011 and 2013 the news media coverage stimulates activity in social media. This is followed by a transition period of two-way causality. From 2016, however, changes in levels of social media activity seem to lead and generate news coverage volumes. We uncover similar evolution of lead-lag pattern between sentiment measures constructed from the tonality contained in textual data from social and news media posts. We discover that market variables exert stronger impact on investor sentiment than the other way around. We also find that return responses to social media sentiment almost doubled after the transition period, while return responses to news-based sentiment almost halved to its pre-transition level. The linkage between volatility and sentiment is much more persistent than that between returns and sentiment. Overall, our results suggest that social media is becoming the dominant media source.