Research On The Mechanism Of Media Reports And Investor Attention On Stock Returns

Information is the key factor of capital asset pricing and the core of capital market operation.With the continuous development and progress of information technology,society has entered an era of informatization and digitalization.Media has also been quantified into digital,and new media such as network and mobile phone have gradually become the leading role.It can be seen that Internet news has become a high-frequency basic network application,which can meet users' information needs to obtain news and information,and it is an important channel for people to collect information and express opinions.In the stock market,media play an increasingly important role along with the extension of digitalization,and it is one of the most important sources for investors to obtain information.Meanwhile,every transaction of investors is accompanied by information collection,and search has become the main channel for investors to screen information.Existing studies have some confusion about the role of media reports and investor attention,and have not found the difference and connection between the influence of these two factors on stock returns.Therefore,there is no consensus on the deep reasons and internal mechanism of the influence of media reports and investor attention on stock returns.At present,there are three main explanations for the impact of stock returns: risk,mispricing and data mining.Therefore,this paper speculates that the impact of media reports and investor attention on stock returns may come from the following sources: Firstly,media reports and investor attention change the risk of stock;Secondly,because investors have partial expectations for stocks,stock returns are corrected when the amount of information increases and the limited attention is alleviated;Thirdly,the impact of media coverage and investor attention on stock returns may be the result of mispricing.In the end,the impact of media coverage and investor attention on stock returns is only due to the emergence of accidental correlations between large amounts of data.Based on this,the paper aims to further explore the relationship and difference between media reports and investor concerns affecting stock returns,and explore the mechanism of their impact on stock returns.Specifically,this paper proceeds as follows: Firstly,it empirically analyzes the separate and interactive impact of media reports and investors' attention on stock returns,and reveals the differences and connections between media reports and investors' attention in the process of affecting stock returns;Secondly,through the impact of non trading day media reports and investors' attention on the opening of the next period,verify the possibility of reverse causality of media reports and investors' attention on the performance of the stock market;Finally,on the basis of clarifying the impact of media reports and investor concerns on stock returns,this paper further analyzes the impact mechanism,and tests the sources of the impact of media reports and investor concerns on stock returns from the perspectives of risk,mispricing and data mining.The main conclusions of this paper are as follows:(1)Baidu media index and search index are reasonable as proxy variables of media reports and investors' attention.According to the univariate regression results,the media reports and investor concerns represented by Baidu index are significantly correlated with the stock market returns.According to the AR model with the addition of the lagging Baidu index item,the historical media reports and investor attention have a significant impact on stock returns,highlighting the unique value of the Baidu index as a proxy variable of media reports and investor attention that is different from the existing market indicators.(2)Media coverage and investor attention have a positive impact on stock returns,and there is a substitution effect between media attention and investor attention.A panel regression model with the daily excess return of stocks as the explained variable is constructed.After controlling the market return rate,the size of listed companies,the price-to-book ratio and other factors,the influence of media reports and investors' attention on stock returns is significantly positive,and the interaction term between media reports and investors' attention is significantly negative.At the same time,the data of media reports and investor attention on non-trading days are used to verify that the media reports and investor attention cause the movements of the stock market,rather than the performance of the stock market.(3)Risk cannot explain the effect of investor attention and media coverage on stock returns.First,through a dynamic model that only allows for time-varying risk premium,comparing the same stock returns on the same day,the higher the stock that is covered by the media and paid attention by investors still has higher returns.Second,by looking at what can be changed,by adding market returns to the model,we find that high investor attention and news coverage do not lead to an increase in ? value.Finally,media reports and investor attention are not the risk factors that explain the volatility of stock returns,and the "risk pricing hypothesis" has not been supported by the data of China's capital markets.(4)Investors' attention to the impact on stock returns is due to the correction of biased expectations,while the media coverage of the impact on stock returns is mainly due to arbitrage restrictions.Firstly,the residual income valuation model is used to preliminarily determine that there is mispricing in China's stock market.With the increase of the stock intrinsic value to market ratio of the sample companies,the annual return rate of the portfolio generally increases.Secondly,from the perspective of biased expectations,taking analysts' prediction error as the explained variable,it is found that investors' attention has a significant negative impact on analysts' prediction error,and the positive impact of investors' attention on stock returns is the correction result of low expectations,while the results reported by the lag media do not pass the significance test.Finally,taking the stock closing price,the proportion of institutional investors and the standard deviation of analysts' forecasts as the research objects,we find that the more stringent the market arbitrage restriction is,the stronger the negative effect of media reports and subsequent stock returns is,while investors' attention does not show this feature.(5)Data mining alone cannot explain the positive impact of media coverage and investor attention on stock returns.Due to the impact of stock returns results may only be a large amount of data to multiple test "deviation",therefore,this article also from the perspective of data mining is verified,the results show that data mining can not explain stock returns.In addition,to the median value of the sample enterprises are classified into large sampled stocks and small sampled stocks,according to the sample regression estimation results.It is found that the abnormal return difference between stocks of large and small enterprises is more significant under the influence of media reporting days and investor attention.Data mining alone cannot explain the difference between stocks of large and small enterprises.Compared with the existing research,the innovation of this paper is mainly reflected in the following three aspects:(1)This paper takes media reports and investor concerns into an analytical framework to analyze the relationship and difference between them in the process of affecting stock returns.In addition,since the data during the trading day and after the market break cannot be distinguished well,the relationship between the current variables cannot indicate whether the change in market performance is caused by media reports or investor attention,or whether the market performance is caused by media reports or investor attention.Therefore,this paper also studies the relationship between non-trading day media reports and investors' attention on the jump in stock opening returns and tests the reverse causality.(2)This paper will now explain the three main factors affecting stock returns:Risk,the mispricing and data mining into the scope of inspection at the same time,the use of China's listed companies,and baidu index data,news and investors pay close attention to the mechanism of action of affect stock returns,investors pay close attention to the impact on the stock returns is correction of biased expectation,and media influence on stock returns stems mainly from arbitrage restrictions.It is not only beneficial to provide test evidence for the validity of foreign research conclusions in emerging markets like China,but also can provide a supplement for the research on the influence mechanism of investors' attention and media reports on stock returns.(3)The innovation of variable indicators.In order to investigate the dynamic risk,eliminate the "multiple test bias" caused by the amount of data,and test the mechanism of the influence of media reports and investors' attention on stock returns,21 variables related to the influence of stock returns are summarized from the core journals of finance,accounting and economics that are widely recognized in the domestic academic circle.The abnormal variables in this paper are constructed to eliminate the accidental correlation caused only by a large amount of data between variables.In addition,since the Baidu index does not provide downloads in any format and uses traditional image recognition methods,the presence of the Baidu Watermark is around 5% off.Therefore,this article writes a Python 2.7-based invocation program to retrieve encrypted data from the Baidu Index,and then to decrypt the program to get the accurate Baidu index original data.