基于文本的创新测度及对企业绩效的影响研究  

Text-Based innovation measurement and its impact on enterprise operating performance

作　　者：郭冬梅[1] 王继彬 胡瀚清 尹志锋[1] GUO Dongmei;WANG Jibin;HU Hanqing;YIN Zhifeng(School of Economics,Central University of Finance and Economics,Beijing 100081,China;Bank of China,Beijing 100818,China)

机构地区：[1]中央财经大学经济学院,北京100081 [2]中国银行总行,北京100818

出　　处：《系统工程理论与实践》2024年第6期1896-1912,共17页Systems Engineering-Theory & Practice

基　　金：国家社会科学基金重大项目(21&ZD085);国家自然科学基金面上项目(72274231);中央财经大学重大研究支持计划(2022)。

摘　　要：本文基于中国A股上市企业的研报文本数据,使用LDA主题模型、情感分析等方法构建了文本创新指标,并通过考察其与企业绩效的相关性,验证该指标的合理性.研究发现:第一,文本创新指标具有全面性,不仅与研发投入、专利等传统创新指标具有高度的相关性,还能在一定程度上测度企业的非技术创新;第二,文本创新指标具有稳健性,相较于传统创新指标,文本创新指标与企业经营绩效的正相关性更稳健;第三,文本创新指标具有精准性,能部分减缓传统创新指标因测量误差、研发操纵等导致的估计偏误.本研究有益于拓展创新度量指标并据此准确估计创新对于企业经营绩效的作用效果,同时还为政府制定相关创新政策提供度量依据.Based on the text of analyst research reports of A-share listed companies,this paper uses LDA model and sentiment analysis to construct a text-based innovation indicator.We verify the rationality of this indicator by examining the correlation between it and enterprise performance.The results show that:First,this indicator is comprehensive.It is not only highly correlated with traditional innovation indices such as R&D and patent,but also measures non-technological innovation activities to a certain extent.Second,this indicator is robust.Compared with traditional innovation indicators,it has a more robust positive correlation with enterprise performance.Third,this indicator is more accurate.It can partially reduce the estimation errors of traditional innovation indicators,which are caused by measurement errors and R&D manipulation.Overall,this study expands innovation measures and estimates the effect of innovation on enterprise performance accurately.At the same time,it also provides a measurement basis for the government to formulate relevant innovation policies.

关 键 词：文本创新 研发投入 专利 非技术创新 企业绩效 

分 类 号：F273.1[经济管理—企业管理] F276.4[经济管理—国民经济]