| Risk prevention in the stock market is not only a key factor in maintaining the stability of the financial system,but also an important condition for the stability and development of the national economy.In recent years,with the increasing integration of the global capital market and the frequent occurrence of uncertainty events,China’s stock market is full of "anomalies",such as mass selling and thousands of shares falling due to the spread of panic in the stock market crash;in addition,China’s stock market has the characteristics of large scale and irrational investors,and the traditional financial theory based on the rational man hypothesis cannot reasonably explain many "anomalies".The traditional financial theory based on rational man hypothesis cannot reasonably explain many "anomalies".In contrast,the "invisible" spillover mechanism associated with behavioral finance theory reveals the causes of irrational behavior in the financial market,thus opening up a new channel for explaining financial anomalies.With the development of the economy,investors’ access to information through the media has been increasing.Current research shows that media attention,media sentiment,investor attention,investor sentiment,and herding effect are the main behavioral factors that can be quantified and focused on,and their respective effects on the stock market and the spillover relationship between the "two-two" behavioral factors have been demonstrated.The study of the spillover relationship between the five behavioral factors and the stock market has not yet been studied systematically beyond the "two-two" spillover relationship.In addition,the analysis of information spillover mechanisms is an important way to identify and describe the sources and channels of risk transmission,and is one of the important research areas in financial market risk management theory,which deserves further in-depth study from new perspectives and approaches.Therefore,based on the framework of behavioral finance theory,this paper mainly takes Chinese stock market as the research object,applies the topological analysis method of information spillover network,incorporates five behavioral factors and stock market into a unified research system,and constructs the behavioral spillover network formed by the association among media attention,media sentiment,investor attention,investor sentiment,herding effect and stock market under smooth and extreme market conditions,respectively,for In order to understand the behavior spillover mechanism of China’s stock market scientifically and comprehensively,the article further analyzes the industry behavior spillover mechanism of stock market under different market states.Finally,on this basis,a stock market risk monitoring and early warning method based on the behavioral spillover network is established.The main research steps are as follows.First,under the smooth market state,given the existence of style differences in the stock market,the stock market is divided into large-cap stocks and small-cap stocks,and the information spillover effects related to investor attention,investor sentiment,media attention,media sentiment,and herding effect are analyzed separately,and the DY(Diebold and Yilmaz)model and BK(Barun(?)k and K(?)ehl(?)k)model are used to statically,dynamic perspectives to construct behavioral spillover networks in the time and frequency domains(short-,medium-,and long-term)for large-cap and small-cap stocks,respectively.In addition,the results are tested for robustness using TVP-VAR models,subsamples.The study shows that: firstly,the behavioral information spillover effect of small-cap stocks is higher than that of large-cap stocks,indicating that small-cap stocks are more susceptible to "invisible" mechanisms such as market sentiment and media opinion;secondly,the Chinese stock market is ranked by the spillover size of each behavioral factor: sentiment behavior >herd effect > attention behavior.Secondly,the world is currently in a period of turbulent change,with a significant increase in domestic and international uncertainty.Numerous studies have shown that the asymmetric distribution of stock market returns with "spikes and thick tails" corresponds to different risk characteristics.Therefore,constructing the information spillover network of investor attention,investor sentiment,media attention,media sentiment,herding effect behavior factors and stock market under extreme market conditions can provide new ideas for the identification and prevention of tail risk in stock market.The QVAR-based quantile spillover model is used to analyze the spillover effects of large and small-cap stocks in extreme conditions with quantile at 0.05(left tail)and 0.95(right tail)from static and dynamic perspectives,respectively.Robustness tests are conducted by transforming the model parameters.The study clarifies that the information connectivity between the Chinese stock market and media attention,media sentiment,investor attention,investor sentiment and herding effect is significantly enhanced in the extreme states.Among them,investor attention and investor sentiment have a stronger effect on large-cap stocks in the extreme down state;on the contrary,media sentiment and media attention have a stronger effect on small-cap stocks in the extreme up state.Then,in view of the fluctuating,staggered ups and downs in the stock market sectors and the "heterogeneity" feature,we further investigate the behavioral spillover effects of the materials,telecommunication services,industrial,utilities,financial,consumer discretionary,energy,daily consumption,information technology,health care and real estate sectors in our stock market.The behavioral spillovers of each industry in the stable and extreme market conditions are explored separately and tested for robustness.It is found that: first,there is a strong information spillover effect between each industry and the corresponding media attention,media sentiment,investor attention,investor sentiment and herding effect in China’s stock market;second,there is a strong information spillover effect between industrial,information technology and financial industries and multiple behaviors,while the behavioral spillover effect in utilities,telecommunication services and real estate industries is smaller compared with other industries;third,the behavioral spillover effect in each industry is stronger in the short run than in the other industries.Behavioral spillover effects in each industry are strongest in the short run,and investor behavior is in the leading position in the information spillover network of each industry,while the herding effect is more in the position of information receiver.Finally,based on the findings of stock market behavioral spillover research,a risk monitoring and early warning method based on the topology of multi-behavioral spillover network as a signal source is established to provide a new idea and method for the establishment of stock market risk monitoring and early warning system by making full use of the role of "invisible" mechanism.The article uses machine learning methods and logit models to evaluate risk factors,and the results prove that investor behavior and media behavior can be used as effective risk monitoring and early warning indicators for the stock market.Based on the study of behavioral spillover networks in China’s stock market under different market conditions,the main conclusions of this paper are summarized and countermeasures are suggested from the perspectives of regulators,media,enterprises and investors: First,there are close information spillover effects between China’s stock market and five behaviors: media attention,media sentiment,investor attention,investor sentiment and herding effect,and present style,frequency domain and time-varying differences.Thus,regulatory authorities should implement the information disclosure system,govern the development of disclosure blindness standardization and integration,and form a long-term mechanism for regular announcements.Second,in the extreme down(up)state,there is a large spillover effect between investor behavior(media behavior)and large-cap stocks(small-cap stocks).Therefore,investors should set their mindset to achieve rational investment;at the same time,the news media need to assume a number of functional responsibilities such as timely and accurate,orderly management and active cooperation to help promote the harmonious,stable and healthy development of the market.Third,the industry sector behavior spillover effect of the distinctive characteristics,so investors should fully examine the industry’s own development rules for decision-making,for example,risk averse to the advantages of the choice is the stability of the development of public utilities,telecommunications services and other industries;and for risk-averse,industrial,information technology and other factors due to the intense market environment,technology iteration is faster to grasp the wind,the choice to break through the beach.Fourth,the role of "invisible" mechanisms can be incorporated into the stock market to establish a system of risk monitoring and early warning methods.Therefore,listed companies should incorporate factors such as media opinion into the risk management structure system and establish a risk monitoring and early warning committee to ensure the implementation of risk prevention and control policies.This study responds to the call of the central government to focus on financial work in major party and state conferences in recent years,and provides theoretical evidence for the discussion of the mechanism of multiple behavioral interactions,as well as the significance of identifying,preventing and resolving financial risks. |
