管理层讨论与分析的信息含量及其市场反应 -国际经济与贸易(本)专业

摘要

随着信息披露制度的不断完善，管理层讨论与分析（MD&A）作为上市公司定期报告中的重要组成部分，日益成为投资者解读企业经营状况、评估未来前景的关键信息来源。然而，由于其内容具有较强的主观性和非结构化特征，MD&A的信息含量及其对资本市场的实际影响仍存在较大争议。本文旨在系统评估MD&A的信息价值，并深入探讨其在资本市场中的反应机制。通过构建文本分析模型，我们从语义情感、信息量和战略导向三个维度对2010年至2022年间A股上市公司的MD&A文本进行量化处理，并结合事件研究法与多元回归模型，分析其与股价波动、交易量变化及分析师预测修正之间的关系。实证结果显示，MD&A中积极语调和前瞻性信息披露程度越高，短期内市场反应越显著，且分析师预测偏差随之降低，表明管理层话语确实蕴含着增量信息。进一步研究表明，高质量的MD&A有助于缓解信息不对称，提升公司透明度，尤其在治理水平较低和分析师关注度较少的企业中，这种效应更为突出。本文的创新之处在于将自然语言处理技术引入中国资本市场的信息披露研究，构建了适用于中文文本的量化指标体系，并揭示了MD&A在不同制度环境下的异质性作用机制。本研究不仅丰富了信息披露与市场反应领域的理论成果，也为监管机构优化披露规则、提升信息有效性提供了政策参考。

关键词：管理层讨论与分析；文本分析模型；信息不对称；资本市场反应；自然语言处理

Abstract

With the continuous improvement of information disclosure regulations, Management Discussion and Analysis (MD) has increasingly become a crucial component of listed companies' periodic reports and a key source of information for investors to interpret corporate performance and assess future prospects. However, due to its subjective and unstructured nature, the informational content of MD and its actual impact on capital markets remain subject to considerable debate. This paper aims to systematically evaluate the informational value of MD and explore in depth its reaction mechanism within the capital market. By constructing a text analysis model, we quantitatively process the MD narratives of A-share listed companies from 2010 to 2022 along three dimensions—semantic sentiment, information richness, and strategic orientation—and employ event study methodology and multiple regression models to examine the relationships between MD characteristics and stock price fluctuations, trading volume changes, and analyst forecast revisions. Empirical results indicate that higher levels of positive tone and forward-looking information disclosure in MD are associated with more pronounced short-term market reactions, while also leading to reduced analyst forecast errors, suggesting that managerial discourse indeed contains incremental information. Further analysis reveals that high-quality MD helps alleviate information asymmetry and enhance firm transparency, particularly in firms with lower governance quality and less analyst coverage, where this effect is more pronounced. The innovation of this study lies in introducing natural language processing techniques into disclosure research within the context of the Chinese capital market, establishing a quantitative indicator system tailored to Chinese textual data, and uncovering the heterogeneous mechanisms through which MD operates under different institutional environments. This research not only enriches theoretical contributions in the field of information disclosure and market response but also provides policy implications for regulators seeking to optimize disclosure rules and improve the effectiveness of information dissemination..

Key Words：Management Discussion And Analysis ；Text Analysis Model；Information Asymmetry；Capital Market Reaction；Natural Language Processing

摘要 I

Abstract II

第1章绪论 1

1.1 研究背景和意义 1

1.2 研究现状 2

1.3 研究方法 3

第2章管理层讨论与分析信息的基本特征与披露机制 4

2.1 管理层讨论与分析的制度框架与发展沿革 4

2.2 管理层讨论与分析内容的构成要素解析 4

2.3 不同行业背景下MD&A信息披露的差异性 5

2.4 我国上市公司MD&A披露质量的现状评估 6