An Empirical Study On The Impact Of Text Analysis-based Investor Sentiment Index On Stock Returns

With the continuous development of the Internet,we have entered the era of the information explosion.Every day,there is an endless stream of financial news,and the hottest news often has a great influence of the stock returns.It means investor sentiment can be reflected on the financial news,and the impact of investor sentiment on the stock market cannot be underestimated.With the development of behavioral finance,more and more people begin to pay attention to investor sentiment and start to do some research.The main research content of this article is the construction of investor sentiment indicators and the dynamic relationship between the investment indicators and stock market.We constructed three investor sentiment indicators.The first one is the financial news text sentiment indicator constructed using text analysis methods.The second one is the market sentiment indicator constructed using PCA(principal component analysis)method.The last one is comprehensive sentiment indicator,which also constructed by PCA.We used correlation analysis,trend analysis and regression analysis,which are qualitative and quantitative methods,to find out which sentiment is most effective for studying the relationship between investor sentiment and stock return.Then we found that the market sentiment index is invalid among the three sentiment indexes.The financial news text sentiment index is effective,but its effectiveness is not as good as the comprehensive sentiment index which is constructed by market and financial new text sentiment indicators.The reasons are as follow.Firstly,the text analysis method is conducive to improving the accuracy of the indicator,so that it can reflect more investor sentiment.Meanwhile,the comprehensive sentiment index covers not only the sentiment of institutional investors but also the individual investors,so the wider the coverage is,the more accurate investor sentiment will be reflected.Based on VAR model,we used Granger causality test,impulse response function and Variance decomposition to discuss the dynamic relationship between the comprehensive investor sentiment index and the CSI 300 index return.The results are as follow.Firstly,the comprehensive investor sentiment index and the CSI 300 index return rate are Granger reasons for each other,indicating that the two are mutually influential.Secondly,the impact of investor sentiment on the stock market return rate is positive and dramatic,but the duration is short,which is less than 2 days.The results of these studies have an important and positive influence to securities institutions and financial regulators,which can help them understand the investor sentiment in a better way.This means,they can do better in guiding investors in making rational investment decision and the effective development of the Chinese stock market will be improved.There are two main innovations in this article.First of all,this article has made innovations in the selection of sentiment indicators.When selecting the proxy indicator of sentiment indicators,we not only consider the characteristics of the Chinese stock market but also combine the financial news sentiment indicators.As result,we finally select three index,which is margin buying,A.D.line and diversified financials industry.These indicators are more suitable for ultra-short term research.Second,this article studies the relationship between investor sentiment indicators and stock market returns on a daily basis,and innovatively derives the trend of the impact between the investor sentiment and the stock return within a week,which has certain reference value for other researchers and financial institutions.