A direct measure of investor sentiment

Abstract

A new measure of investor sentiment is introduced and tested in this study. Different from using search volume of certain macroeconomic terms to capture investor attention/sentiment, this new measure is crafted from analyzing and classifying the sentiment contents of textual comments of retail investors active in a major Chinese stock forum. Different from the traditional sentiment studies linking market-level sentiment to market-wide reactions, this uncomplicated measure is constructed for individual stocks and subsequently, reactions of the same stocks are tracked and examined, offering a more direct and precise correlation test. In our validation tests, we show a significantly positive correlation between investor sentiment and three stock market parameters that is, stock return, price volatility, and information efficiency. Specifically, a positive sentiment is associated with higher stock returns and a higher degree of information efficiency as well as higher price volatility. These associations seem to attenuate with improvement in the information environment, such as better investor protection.