The impacts of investors' sentiments on stock returns using fintech approaches

Abstract

This study uses fintech approaches, including web crawler technology with distributed architecture to select internet news messages largely and efficiently and a dictionary-based linguistic text mining to create sentiment variables, to explore the respective impacts of investors' optimism and pessimism on stock returns. The construction of sentiment variables in network- and dictionary-based messages is more precise and variable than that in traditional-based messages. Our results show that firms with investors' optimistic sentiments have significantly higher stock returns in the current month, whereas those with pessimistic sentiments have significantly opposite effects. The effect of both investors' optimism and pessimism on stock returns subsequently reverses. Then, the negative impacts of investors' largely pessimistic sentiments on stock returns are larger than the positive impacts of their largely optimistic sentiments within a quarter. Next, investors' optimistic sentiments significantly raise stock return volatility by approximately a quarter, but their pessimistic sentiments have the opposite effects. Furthermore, investors' high optimism more significantly and persistently raises stock return volatility than their general optimism, but the negative effects of their high pessimism on volatility become smaller and the persistence is shorter than general pessimism. In addition to the advantage of our methodology in creating sentiment variables, the simultaneous consideration of investors' optimism and pessimism to analyze the effects on the returns and the volatility of individual stocks in this study is more complete than previous related studies.