企业社会责任文本信息操纵与ESG评级分歧——基于自然语言处理技术的分析
    点此下载全文
引用本文:江媛1,2,张颖3,罗云磊1.企业社会责任文本信息操纵与ESG评级分歧——基于自然语言处理技术的分析[J].财经理论与实践,2025,(1):85-93
摘要点击次数: 83
全文下载次数: 34
作者单位
江媛1,2,张颖3,罗云磊1 (1.长沙理工大学 经济与管理学院湖南 长沙 4101142.湖南财政经济学院 会计学院湖南 长沙 4102053.中南大学 湘雅二医院财务部湖南 长沙 410083) 
中文摘要:以2011—2020年A股上市公司社会责任报告为样本,运用自然语言处理方法提取文本信息特征,考量企业社会责任文本信息操纵行为对ESG评级分歧的影响。结果显示:企业社会责任文本信息操纵程度越大,ESG评级分歧越大。机制检验发现,企业社会责任文本信息操纵会加剧信息不对称,加深第三方评级机构对私有信息的依赖,导致ESG评级分歧。鉴于此,建议制定统一的ESG信息披露细则,加强对企业社会责任信息披露的监管和验证。
中文关键词:社会责任报告  信息操纵  ESG评级分歧
 
Corporate Social Responsibility Text Information Manipulation and ESG Ratings Divergence—Analysis Based on Natural Language Processing Technology
Abstract:Taking the social responsibility reports of China’s A-share listed companies from 2011 to 2020 as a sample, the natural language processing method was used to extract textual information features and examine the impact of CSR textual information manipulation behavior on ESG rating divergence. The results show that the greater the degree of CSR textual information manipulation, the greater the ESG rating divergence. The mechanism test finds that CSR textual information manipulation exacerbates information asymmetry, deepens third-party rating agencies’ reliance on private information, and thus contributes to ESG rating divergence. In view of this, it is recommended to formulate unified ESG information disclosure rules to strengthen the regulation and verification of CSR information disclosure.
keywords:social responsibility report  information manipulation  ESG rating disagreement
查看全文   查看/发表评论   下载pdf阅读器