Stock-level investor sentiment and attention during the coronavirus pandemic.

Krystyniak, Karolina | $24,835

Ontario Ontario Tech University (University of Ontario Institute of Technology) 2021 SSHRC

In current digitalized world factual information often competes with easily disseminated personal preferences and emotions. Investors tend to be overly pessimistic or optimistic, as evidenced by past sharp declines and increases in stock prices, such as bubble in the 1990s, financial crisis in the 2008 and coronavirus pandemic in 2020. The sentiment dominating stock market can drive the asset prices away from the fundamental values, and the unsentimental investors may fail to correct the mispricing due to limits to arbitrage. Especially in times of crisis, when societal anxiety is high, it is likely that investors may turn to sentiment-based trading decisions.

Another factor that drives stocks away from their fundamental values is investor distraction. Investors have limited attention; they focus their trading on attention-grabbing stocks and are slower to react to important firm events if distracted by macroeconomic announcement.

The purpose of this project is to improve the understanding of stock-level investor sentiment and attention and analyze how they impact asset prices before and during the Covid pandemic. Our research will consist of two subprojects.

The first subproject will focus on investor sentiment. First, we will identify a stock-level sentiment measure and analyze its determinants and characteristics. We will use the sentiment score from Bloomberg that aggregates Twitter and Stock Twits data and decompose it to remove the influence of fundamental information. Second, we will analyze stock-level and aggregate market sentiment before and during the pandemic. Third, we will test the sentiment measure’s predictive power on stock returns. The expectation is that negative sentiment would drive undervaluation and subsequent positive returns, while positive sentiment would be followed by negative returns. We expect the mispricing to become even more extreme during Covid pandemic due to uncertainty. We will also perform cross-sectional analysis to see if investor sentiment has stronger price effect on hard-to-value stocks.

The second subproject will examine the fluctuations in attention and its impact on stock prices around the Covid pandemic. On the one hand, the investors, especially retail ones, may be more distracted by Covid-related news and pay less attention to firm announcements and events. On the other hand, frequent lockdowns and limited leisure activities may result in investors having more time to process trading information. First, we will analyze trends in stock-level attention before and during the pandemic as measured by daily Google searches of a firm’s stock ticker symbol. Second, we will examine how the fluctuations in attention impact stock price reactions to important events, like earnings announcements. Third, we will examine how stock prices and volume responses to firm news were affected on days with highly distracting news.

This project will contribute to the growing literature on the importance of sentiment and attention in the asset pricing, by adding an important stock-level dimension. It will add to our understanding about how emotions, political opinions, unverified gossip, and cognitive limitations to investor attention affect decision making and, as a result, stock markets and economy as a whole. The findings will be useful to policy makers whose goal is to reduce the uncertainty and stabilize economy in times of crisis. The results will also serve investors, both retail and institutional, in understanding the drivers of asset prices and making appropriate trading decisions accordingly. Finally, the findings will hold educational value to the students, furthering their understanding of behavioural finance and improving their financial literacy.

With funding from the Government of Canada

Please complete this short survey to help us understand our impact. Thank you!

Give Feedback