Research On Investor Behavior In Multi-level Stock Market From The Perspective Of Social Media

Under the dual direction and guidance of national policy and market,China’s stock market has achieved great development since its establishment,which has played a significant role in optimizing the financing and resource allocation of participating enterprises.However,China’s stock market is still an emerging market compared with European and American capital markets,and the investors in the stock market are mainly composed of medium and small investors.Medium and small investors often have a certain degree of overconfidence and cognitive deviation and have a strong speculative psychology,their professional skills and investment ideas are not mature enough.So that their irrational investment behavior has a significant impact on the stock market,to a certain extent,restricting the optimized allocation function of China’s stock market.On the other hand,with the development of China’s stock market,the popularity of the Internet in China also continues to deepen.Various financial media and instant messaging tools are changing the way in which financial information is shared,published,communicated and judged based on their enhanced social features.Financial social media not only broadens the channels for investors to obtain financial information,but also facilitates the interaction between investors and greatly promotes the learning behavior among investors,which is likely to magnify the impact of investors’ irrational factors on the stock market.In order to investigate the influencing factors and mechanisms of investors’ irrational behavior on the stock market in the context of social media,this dissertation adopts the method of combining theoretical and empirical research and follows the research thinking and technical route of "literature review--theoretical research--empirical research--conclusion and suggestion",takes the "social media horizon,multi-level stock market investor behavior research" as the theme,centers on the two key issues of "what behavior mode do investors interact with in social media" and "what impact does social media investor behavior information have on the stock market",and constructs a multi-level and all-round systematic study on the stock market investor behavior from the perspective of social media.This dissertation firstly sorts out the research background and current situation of investor behavior in the era of social media big data.Then,the dynamic model of social media investor behavior is established under the theoretical framework of evolutionary game,and the theoretical mechanism of investor behavior on the stock market is explored.Furthermore,through the relevant indicators of investor behavior extracted and constructed from the user interaction data of investors in financial social media,multiple scenarios are selected from the three levels of micro individual stocks,medium industry and macro market,to comprehensively investigate the impact of investors’ irrational behavior on the stock market.Finally,on the basis of the above research,it provides decision support for the stock market regulators,the managers of listed companies and the majority of investors in China.Specifically,the research work of this dissertation is mainly divided into theoretical research and empirical research.The theoretical research part puts forward a social media investor behavior evolution game model.This model considers the public opinion events heat and investor psychological deviation on its information learning strategies,can vividly show investors in social media to study the dynamic process of interaction and decision-making information.Through the evolutionary stability strategy analysis of the evolutionary game model of investor behavior,it is found that the key factor to determine the stable state of the system is whether the net return of investors who actively analyze communication in social media interaction can compensate for their dissatisfaction with investors who blindly follow negative imitation.In addition,the results of the simulation experiment also show that both the popularity of public opinion events and the satisfaction degree of investors’ psychology will have an impact on the decision-making evolution process of social media investors.The empirical research part studies the impact of social media investor behavior on the market at three different market levels: micro individual stock,medium industry and macro market.At the level of micro individual stocks,this thesis adds social media investor behavior factor into the Fama-French five-factor pricing model to study the impact of investor behavior on asset pricing at the level of individual stocks,so as to verify the effectiveness of social media investor behavior information.The results show that social media investor behavior factors not only contain factors that can significantly affect the excess returns of individual stocks--changes in investor sentiment and investor attention,but also can enhance the explanatory power of market factors on asset pricing.The subsequent application of stock price trend prediction based on fusion information also achieved good results,which also confirmed the effectiveness of social media investor behavior information.At the medium industry level,this thesis constructs an industry panel data including social media investor behavior,and uses the fixed effect model to test the impact of investor behavior on industry volatility.The results show that the improvement of investor sentiment and attention increases the volatility of the stock market,while the change of investor sentiment and attention and the divergence of opinions among investors restrain the volatility of the stock market.Among them,investor sentiment is the most profound behavioral factor affecting the fluctuation of the stock market.Then,the existence of industry emotion spillover effects is further investigated by the DAG-SVAR model,and some industry emotion spillover paths are described.At the macro market level,this thesis selects the CSAD model based on the absolute dispersion of return rate to investigate the existence of herd behavior in China’s stock market,and extends it by using social media investor behavior indicators to test the impact of investor behavior on market herd behavior.Results show that the overall stock market in China showed significant herding effect,mainly concentrated in low and high yields of discrete degree of quantiles,investor sentiment and investor attention and changes that can reduce the degree of market herding effect,and investors in the process of social media interaction,the higher the degree of disagreement,the easier to induce the generation of herd behavior.Subsequently,the samples were further decomposed from different dimensions to investigate the differences in the impact of different market situations,subsectors and investor behaviors before and after the epidemic on herding,which enriched relevant research results and provided more detailed evidence.Finally,this dissertation analyzes and summarizes the theoretical and empirical research results and conclusions,and puts forward corresponding decision-making suggestions for the three participants of the stock market by combining with the real environment.Market regulators should pay more attention to the behavior of social media investors,improve the management system of social media,and maintain the healthy development of the stock market.For the managers of listed enterprises,they should standardize their own information release process,fully respect the diversity and objective facts of investor groups,and avoid releasing information with emotional tendency to disturb the online public opinion environment to cater to some investor groups.For the vast number of investors,they should make efforts to improve their professional knowledge and skills,to prevent the understanding of their own lack of expertise and the market as the irrational trading behavior,also should pay attention to the overall investor behavior,and take the official information and social media platform into the common learning and decision-making process to reduce the investment risk.In general,the innovation of this dissertation is mainly reflected in the following aspects:First,a comprehensive research system of multi-level stock market investor behavior is constructed based on the perspective of social media research;Secondly,the interaction behavior patterns of investors on social media platforms are mined through the evolutionary game research framework;Thirdly,this study optimizes the quantitative method of investor behavior indicators and promotes the integration of interdisciplinary research.Finally,this dissertation proposes the spillover effect of investor sentiment,which expands and enriches the research results related to investor behavior.