国内统一连续出版物号:CN 11-1384/F

国际标准连续出版物号:ISSN 1000-7636

当前位置: 首页  >>   最新刊发  >>   最新刊发
最新刊发

企业研发文本披露的同行溢出效应

企业研发文本披露的同行溢出效应

符号亮1 夏婷1 孙凤娥2

(1. 广东财经大学金融学院;2. 北京林业大学经济管理学院)

  内容提要:本文构建企业研发文本披露指标,分析同行企业研发文本是否对企业研发投入产生溢出效应。结果显示,中国上市企业研发文本披露存在溢出效应,同行企业披露的研发文本越多,目标企业的研发投入增加越多。机制分析结果表明,研发文本披露通过市场竞争机制和信息学习机制实现溢出效应:行业市场竞争会促使目标企业研发决策受到同行企业研发文本披露的影响;信息学习需求促使处于跟随地位的目标企业向行业领先企业学习。拓展研究结果显示,当企业融资约束程度低时,同行企业研发文本披露的溢出效应增强;当同行企业获得较多财政补贴时,研发文本披露的溢出效应增强;当同行企业信息可比度高时,研发文本披露的溢出效应增强。本文研究结论有助于相关部门完善企业研发文本披露规则和加快建立高效的企业创新生态。

  

  关键词:研发投入;信息披露;溢出效应;文本分析;信息可比度

  

  作者简介:符号亮,广东财经大学金融学院副教授,广州,510320;夏婷,广东财经大学金融学院讲师;孙凤娥,北京林业大学经济管理学院副教授,通信作者,北京,100083。


  基金项目:广东省哲学社会科学规划项目“绿色金融、数字赋能与经济高质量发展研究”(GD23XYJ15)

  引用格式:符号亮,夏婷,孙凤娥.企业研发文本披露的同行溢出效应[J].经济与管理研究,2023,44(11):125-144.DOI:10.13502/j.cnki.issn1000-7636.2023.11.007.

  

  


Peer Spillover Effects of Corporate R&D Text Disclosure

FU Haoliang1, XIA Ting1, SUN Feng'e2

(1. Guangdong University of Finance & Economics, Guangzhou 510320;

2. Beijing Forestry University, Beijing 100083)


  Abstract: Information disclosure is an effective way to improve information paradox. The current treatment of expensing and capitalization of R&D expenditures makes financial information unable to fully reflect the funds of R&D investment, thereby hindering the visibility of specific directions and projects of R&D. The disclosure of diverse non-financial textual information can fill this gap and become the main carrier for external decision-makers to understand a large quantity of R&D and innovation behaviors of enterprises.

  Based on the text mining of listed companies in China, this paper constructs R&D text disclosure indicators to investigate whether the R&D text of peer companies has a spillover effect on the R&D investment of target companies. It adopts the word vector technology of “seed vocabulary set+Word2Vec similar word expansion” in the field of machine learning to calculate the frequency of words that match the R&D dictionary in annual reports of listed companies on the WinGo Textual Analytics Database and conduct econometric regression tests. The results show a significant spillover effect of R&D text disclosure of listed companies in China. The target companies’ R&D investment increases proportionally with the disclosure of more R&D texts by peer companies. This spillover effect can be achieved through the market competition mechanism and the information learning mechanism. The former in the industry will enhance the significant impact of R&D text disclosure by peer companies on the R&D decision information of target companies; the demand for the latter drives target companies into a follower position to learn from industry leaders. The extended research reveals that the spillover effect is significantly enhanced when the target companies’ financing constraints are low, peer companies receive more financial subsidies, and the information of peer companies is highly comparable. Therefore, it is recommended to moderately increase the standardized regulation of non-financial information disclosure, formulate supporting policies for corporate innovation information disclosure, and strengthen the disclosure and comparability of information in the industry of science and innovation companies.

  The contributions of this paper lie in the following aspects. First, it analyzes the annual reports of Chinese listed companies through the WinGo Textual Analytics Database to construct more direct and comprehensive R&D information indicators. The relevant methods provide a reference for subsequent measurement of China’s R&D text disclosure. Second, it conducts an in-depth analysis of peer R&D information indicators, which is closer to the information needs of corporate innovation investment decisions. Third, it analyzes the mechanism and characteristics of the R&D information spillover effect among horizontal enterprises in the same industry, providing a qualitative information disclosure explanation for the horizontal information impact of enterprises.

  

  Keywords: R&D; information disclosure; spillover effect; text analysis; information comparability