Essays on investor herding behavior and stock market

Abstract

The dissertation consists of two essays on the investor herding behavior in the stock market. The first essay examines how investor herding behavior affects a firm’s equity price and return volatility. This study assumes agent in the economy displays a herding behavior as empirically verified whereas prior models assumes agents agree to disagree: adhere to own opinions even after they notice the difference in opinions from each other. Agents become to have similar precision in forecasting economic variable that is closely related to stock price due to herding behavior. Consequently, their change in supply and demand have almost equal power on the response of stock price Therefore, observed disagreement increase the sensitivity of stock price with respect to an external economic shock, and stock return volatility. In addition, we expect the return predictability for stock with an excessive volatility. Simulation study finds that stock return volatility and predictability increases in agents’ disagreement with the existence of investor herding behavior. Empirical study also supports this argument. Stocks with high analyst forecast dispersion and higher herding behavior displays higher stock return volatilities, and positive future stock returns.

The second essay studies the locality of behavior contagion and spillover over investor cognitive network constructed through their participation in an online social media for investment. This study examines whether cognitive network captures investors’ attention distribution onto stock market and serves as a channel for that multiple behaviors (e.g., sentiment, disagreement, and overconfidence) can pass through. Empirical studies confirm that attention-grabbing stocks takes investors’ limited cognitive resources away from cognitively close stocks and behaviors are contagious only to cognitively close stocks. Moreover, behavioral aspects not only transmits, but also affects its market variables related to uncertainty. These findings imply that social influence among investors are so crucial that must be reflected into asset pricing.