Abstract: Based on the Sci-Tech innovation board listed stocks, this paper focuses on the effect of investor sentiment on the market return before and during the COVID-19. First, we classify the sentiment tendency using daily stock comments in the Eastmoney bar from December of 2019 to March of 2020 using Bi-LSTM deep learning technology and then establish the investor sentiment index. By constructing the two-way fixed effects simultaneous equations model, this paper adopts the 2SLS estimator to investigate the impact of investor sentiment on the return of the Sci-Tech innovation board. Then we verify the difference of this impact between normal market state period and the COVID-19 period. The empirical results and robustness check both reveal that investor sentiment has a significant positive effect on the return, and this positive effect is significantly transmitted through the trading volume at the 1% significance level. The influence of investor sentiment on the market return remains robust in both the normal period and the COVID-19 period. Furthermore, investor sentiment plays a more important role during the outbreak of COVID-19. The results are of important significance not only for the securities market regulators in China to improve the trading mechanism of Sci-Tech innovation board, but also for the small and medium-sized investors to optimize their investment strategy.

Key words: investor sentiment, Sci-Tech innovation board, text mining, simultaneous equations model, COVID-19

摘要： 本文以科创板市场为主要研究对象,基于文本数据挖掘方法探究了新冠疫情发生前和疫情期间投资者情绪对市场收益率的影响及其作用机制。利用东方财富股吧2019年7月至2020年3月的日度科创板股票评论数据,基于Bi-LSTM深度学习技术对文本数据情感倾向进行分类,建立投资者情绪指数。通过构建双向固定效应的联立方程模型,采用2SLS方法估计投资者情绪对科创板市场收益率的作用,并检验在经济平稳运行和受新冠疫情冲击期间该作用的差异性。实证分析及稳健性检验的结果均表明,投资者情绪通过影响交易量进而影响科创板股票市场收益率,这种正向作用在1%的置信水平下显著。此外,投资者情绪对科创板收益率的影响在经济平稳运行和受新冠疫情冲击期间均保持稳健,且在新冠疫情期间作用更强。本研究成果对于新冠疫情期间我国证券市场监管层完善科创板交易机制,以及对中小投资者优化投资战略具有重要意义。

关键词: 投资者情绪, 科创板, 文本挖掘, 联立方程模型, COVID-19