Investor Information Behavior And Market Influence

The internet not only provides investors a flood of information,but also records the information behavior of investors,such as post and interaction.The information and recording bring new research problems for academic research.Based on the behavior and content of investor post and interaction in social media,this dissertation extracts several characteristics of investor information behavior,such as investor post,investor information interaction and the investor posting mood,and studies the relationship between these social media records and market returns,stock price efficiency,and individual buy-sell decisions.This dissertation has three parts:The first part examines whether investor post in social media,like traditional mass media coverage,can describe the cross-sectional risk of stocks and thus influence asset prices.In the Internet era,social media has become a major information source for investors.Compared to mass media,social media can affect investors' investment decisions and hence asset prices in the same way,while social media has some heterogeneous features,such as a larger amount of information and a more diverse sources.As a result,it is an important question whether social media has the same or different risk effects on the stock market as mass media.From the perspective of asset pricing,this dissertation proves that stocks with low social media post,equivalent to mass media coverage,can bring a higher expected return,and the risk premium can't be explained by known factor model and mass media coverage factor.Introducing the pricing factor constructed from this effect into the existing factor model can improve the pricing power of the model.The second part studies how different investor information interaction modes in social media affect the stock price efficiency.The information interaction between investors affects their information set of investment decision-making,however,it is difficult to observe interaction process in traditional way.Social media provides us with a great opportunity to observe and measure the modes of investor interaction.Based on the ”post-comment-reply” behavior of investors in social media,this dissertation uses complex network model to express their information interaction into binary interaction,single interaction and null interaction,and thus studies their influence the stock price efficiency separately.The results show that single interaction can improve the market price efficiency,while binary interaction and null interaction can't.The results also show that this effect can be explained by private information and investor recognition.The third part verifies the mood channel of exogenous variables(weather etc.)influencing investor buy-sell decisions.Studies in behavioral finance suggest that exogenous variables,which are not directly related to market performance,can affect investor mood,then influence their buy-sell decisions,and finally impact on the financial market.Nevertheless,because investor mood is hard to observe,it is impossible to prove these variables actually play a role in the transmission.However,investor mood can be expressed by posting behavior in social media,thus can be directly measured by the content.This dissertation analyzes the content of investor post by NLP method and match investor mood and their buy-sell decision at city level.The results confirm that haze weather can induce investor negative mood,which leads to more selling behavior of investors.Further empirical research finds that mood will affect investor information attention bias and risk preference,which will ultimately affect investment decisions.In summary,this dissertation provides new viewpoints and discoveries for the related theory of investor behavior under the background of Internet,and puts forward a new understanding on the formation mechanism of stock price and individual investor behavior.It also provides the theoretical and policy decision basis for China's financial management practice in the Internet era.