中国研发人员性别结构分布演变、关联影响与差异来源——基于半参数核回归与Oaxaca分解的研究

激发女性研发人员创造活力，对于中国创新经济发展具有重要意义。然而，科研领域性别结构失衡是长期以来科学共同体内外所面临的问题。基于2010-2019年中国内地30个省份面板数据，围绕企业、研发机构、高校等创新主体，在对创新全要素生产率指数（以下简称创新TFP）进行测度的基础上，运用核密度估计对三大创新主体的性别结构、创新TFP分布的演进进行描绘。其次，采用半参数核回归法探析科研人员性别结构与创新投入、产出和创新TFP间的关联。最后，选用Oaxaca分解法探究STEM研究生性别差异对R&D人员性别差异的解释程度。结果表明：相比研发机构和高校，企业研发人员性别失衡更为严重，相比人口大省、经济发展落后地区，发达地区对研发人员性别结构的“兜底作用”更显著；创新投入、产出与性别结构之间存在非对称关系，在研发人员女性占比越高的地区，区域全要素生产率指数越高；STEM领域研究生数量上的性别差异可以解释约6成的研发人员性别差异，且解释程度呈逐渐增提升势。

The Evolution,Innovation Association and Formation of the Gender Structure of Chinese R&D Personnel：A Study Based on Semiparametric Kernel Regression and Oaxaca Decomposition

In recent years, although domestic and foreign governments, academia and media have called for more support for female scientists in various ways, and policy formulation has also changed from "system neutrality" to "being inclined and mainstreamed" in order to bridge academic and educational gender gaps. However, problems such as the missing of top female researchers in the scientific research field and the gender gap in the number of labor force in professional fields are still prominent. There is a big differences between female researchers and men in terms of the number of papers, patent achievements and academic influence. The disadvantaged status tends to be further aggravated. Most of the existing research focuses on the causes of the gap, the impact of gender factors on enterprise-level innovation output and performance improvement. In fact, there is also an "influence paradox" between gender differences and organizational innovation performance in scientific research. Therefore, it is necessary to explore the relationship between the gender structure of R&D personnel and innovation. Secondly, according to the order of time evolution, the sources of difference can be divided into theeducation system and the external labor market. Meanwhile there is also a lack of investigation into the difference in the numbers of male and female R&D personnel in China, based on the dichotomy of time between education and work.#br# This paper focuses on the innovation subjects such as enterprises, R&D institutions, and universities. Based on the panel data of 30 provinces in China from 2010 to 2019, it measures the innovation total factor productivity index (hereinafter referred to innovation TFP). The gender structure of major innovation entities and the evolution of innovation TFP distribution are described. Second, the semi-parametric kernel regression method is used to explore the relationship between the gender structure of scientific researchers and innovation input, output and innovation TFP. Finally, the Oaxaca decomposition method is used to explore the extent to which the gender difference of STEM graduate students explains the gender difference of R&D personnel.#br#It is found that firstly the gender structure adjustment of R&D personnel in universities and R&D institutions is better than that in enterprises. The underdeveloped areas are polarized, and the lower limit of developed areas is higher. Compared with enterprises, the proportion of female R&D personnel in universities and R&D institutions has risen more significantly. Meanwhile in terms of provincial distribution, regions with a high proportion of female R&D personnel include not only Xinjiang, Ningxia, and other relatively less developed provinces, but also economically developed regions such as Beijing. Secondly there is an asymmetric relationship between innovation input and output and the gender structure of R&D personnel. Regions with higher R&D investment intensity have a lower proportion of female R&D personnel, reflecting gender differences in funding from an inter-provincial perspective in China. In contrast, in regions with higher patent applications, the proportion of female R&D personnel is higher. The proportion of female R&D personnel is positively correlated with the innovation total factor productivity index, which also confirms that the positive mechanism of regional gender structure diversity is dominant. Thirdly the gender difference in the number of graduate students can explain more than 60% of the gender difference in the number of R&D personnel. For different innovation subjects, the master's index can explain more than 50% of the gender differences in R&D personnel in enterprises, while the doctoral index can explain more than 70% of the gender differences in R&D personnel in colleges and universities.This paper suggests that attention should be paid to the development of gender structure of R&D personnel in enterprises, and a guarantee mechanism should be provided for the balance of gender structure of scientific research in areas with large populations and relatively less developed economic development. Furthermore it is necessary to explore the reasons for the gender gap in research and development from the inside of the education system, and formulate differentiated gender structure adjustment policies for different innovation entities. While further stimulating the wisdom and vitality of female R&D personnel, this paper provides new potential development momentum for China's innovation.#br#