美联储言辞沟通与中国投资者情绪
Federal Reserve Communications and Chinese Investor Sentiment
姚登宝 1葛新杰2
作者信息
- 1. 安徽大学经济学院(合肥,230601)
- 2. 北京大学经济学院(北京,100871)
- 折叠
摘要
本文运用词典法和大语言模型测度美联储言辞沟通和中国投资者情绪,分析美联储言辞沟通对中国投资者情绪的影响效应.结果表明:美联储言辞沟通与中国投资者情绪保持显著的正向关联,消极词汇和强弱语气词汇对投资者情绪的影响程度相对更强;仅有消极词汇对投资者情绪的影响存在显著的欠反应偏差,大部分词汇均对投资者情绪具备预期效应且具有显著的异质性;美联储言辞沟通对中国投资者情绪的影响效应具有非对称性,正向变动的边际效应强于负向变动;消极词汇在短期能提高投资者情绪的日内波动,而长期影响逐渐消失.
Abstract
This paper employs dictionary methods and large language models to measure the Federal Reserve's commu-nications and the sentiment of Chinese investors,and analyzes the impact of the Federal Reserve's communications on Chinese investor sentiment.The results show that there is a significant positive correlation between the Federal Reserve's communications and the sentiment of Chinese investors,with negative and strong/weak words having a relatively stronger impact on investor sentiment.Only negative words exhibit a significant underreaction bias in their impact on investor sen-timent,while most words have an expected effect on investor sentiment with significant heterogeneity.The impact of the Federal Reserve's communications on Chinese investor sentiment is asymmetrical,with the marginal effect of positive changes being stronger than that of negative changes.Negative words can increase the intraday volatility of investor senti-ment in the short term,but their long-term impact gradually disappears.
关键词
美联储言辞沟通/投资者情绪/文本分析/关联模式/影响效应Key words
Federal Reserve communication/investor sentiment/text analysis/correlation pattern/influential effect引用本文复制引用
基金项目
国家自然科学基金青年基金(71803002)
安徽省哲学社会科学规划项目(AHSKY2021D135)
出版年
2024
NETL
NSTL
万方数据