Research On The Measurement And Impact Of Investor Sentiment On Stock Market Under The Background Of COVID-19

With the continuous development of financial markets and the emergence of various financial anomalies,the modern financial theory represented by the efficient market hypothesis has been challenged.Scholars began to pay attention to the psychological characteristics and behavior patterns of micro-individuals.On this basis,behavioral finance came into being.Since then,as an important factor to quantify investors’ psychological characteristics,investor sentiment has gradually become an important branch of behavioral finance,and it has attracted widespread attention in academic research and investment activities.At the end of 2019,the COVID-19 broke out and spread rapidly around the world,which had a significant impact on people’s life and social development,besides,it also caused investor panic and capital market turmoil.The impact of this epidemic is different from previous emergencies: first,it has a wide range of impacts;second,it lasts for a long time;third,it severely restricts the flow of factors;fourth,there has been a black swan event in which the circuit breaker has been tripped for four time.Therefore,the performance of financial markets and investors under this epidemic is an issue worth exploring.After sorting out relevant literature and theories,the main directions of the research are clarified: the first one is the construction of high-frequency investor sentiment indicators based on the background of big data;the second is to analyze the impact of investor sentiment on the capital market from multiple perspectives.Specifically,more than 34 million Twitter texts in the range of 2019.11.01-2021.12.31 are obtained through data mining,and the deep learning method — CNN and LSTM are used to analysis the sentiment of tweets,the results are aggregated at a frequency of half an hour to construct a high-frequency investor sentiment indicator.Then,the return and volatility of the S&P 500 index are used to measure the market performance,and the quantile regression model are selected as the main measurement model,empirical analysis is carried out from various perspectives,including different quantiles,different stages of epidemic development,considering epidemic information,intraday and overnight.The results show that:(1)The impact of investor sentiment on stock market returns is different at different quantile.(2)The role of investor sentiment is strongest during the outbreak period.(3)Epidemic information strengthens the influence of investor sentiment on the market.(4)The influence of intraday investor sentiment on the market is mostly expressed through volatility.(5)Overnight investor sentiment has a significant impact on market within the first two hours.