摘要
上市公司违规不仅使财务报告的可信度急剧下降,而且严重损害了投资者的合法权益。以2007—2019年A股上市公司为研究样本,采用文本分析法识别年报非财务信息的“距离”“难度”“情感”等可视化信息,探讨上市公司在违规后对年报文本信息的操纵行为。研究发现:上市公司在发生违规后,公司年报文本信息的相似度显著降低、复杂度显著提高、语调显著乐观。同时,上述文本特征的系统性变化主要体现在违规之心较强、机会较多、需求较强烈以及被暴露概率较小组,表明公司利用年报文本信息精心掩饰违规是一种策略性行为。此外,上市公司精心掩饰违规行为减弱了短窗口市场反应、加剧了信息漂移现象、提高了股价崩盘风险。
Fraud of listed companies not only sharply reduce the credibility of financial reports,but also seriously damage the legitimate rights and interests of investors.Using 2007-2019 A-share listed companies as a research sample,taking text analysis to identify visual information such as“distance”,“difficulty”,“emotion”of the annual report non-financial information,and discussing the company’s response to the annual report text after a listed company violates regulations.This study find that after a listed company violated regulations,the similarity of the company’s annual report text information decreases,the complexity increases,and the tone becomes more optimistic.At the same time,the regular changes in the above-mentioned text characteristics are mainly reflected in the groups with strong violating intentions,more opportunities,strong demand,and low exposure probability,indicating that it is a strategic behavior for the company to use annual report text information to carefully conceal violations.In addition,the meticulous concealment of violations by listed companies has weakened the short-window market response,aggravated the phenomenon of post-earnings announcement drift,and increased the stock price crash risk.
作者
谭建华
王雄元
TAN Jianhua;WANG Xiongyuan(School of Government Audit,Nanjing Audit University,Nanjing 211815,China;School of Accounting,Zhongnan University of Economics and Law,Wuhan 430013,China)
出处
《中国软科学》
CSSCI
CSCD
北大核心
2022年第3期99-111,共13页
China Soft Science
基金
国家自然科学基金青年项目“基于多模态机器学习的公司财务舞弊研究:识别、侦测及风险指数构建”(72102110)
教育部人文社会科学研究一般项目“人工智能应用对企业成本管理的影响研究:基于成本粘性的视角”(21YJC630123)。
