科研管理

科研管理 ›› 2021, Vol. 42 ›› Issue (8): 150-159.

• 论文 • 上一篇    下一篇

基于S-SEM的区域创新系统影响因素和路径研究

薛永刚   

  1. 广东药科大学 医药商学院,广东 广州510006
  • 收稿日期:2019-01-21 修回日期:2019-12-29 出版日期:2021-08-20 发布日期:2021-08-13
  • 通讯作者: 薛永刚
  • 基金资助:
    全国教育科学规划2019年度教育部重点课题：“基于空间结构方程模型的高校协同创新系统关键因素、影响机制和制度创新研究”（DFA190319，2019—2021）。

A research on the influence factor and path of regional innovation based on the spatial structure equation model

Xue Yonggang   

  1. Medical Business School, Guangdong Pharmaceutical University, Guangzhou 510006, Guangdong, China
  • Received:2019-01-21 Revised:2019-12-29 Online:2021-08-20 Published:2021-08-13

摘要：     创新是驱动经济增长和提高区域经济竞争力的一个关键因素，是保持经济持续增长的一个主要驱动力，在经济转型时期有必要研究创新影响因素及创新策略，增强区域科技创新能力。现有文献主要从创新影响因素、区域创新空间特征、区域创新效率、创新对企业的影响等方面展开研究，取得了相关研究结论和成果。区域创新系统是一个多因素交互作用复杂社会系统，并且存在空间溢出效应，因此研究区域创新系统有必要考虑不同要素之间的交互作用效应和空间溢出效应对区域创新的影响。基于相关理论分析提出九个假设，建立了广东省区域创新系统的空间结构方程模型，从创新主体、创新要素、创新政策、内部经济环境、外部经济环境、专利申请、新产品销售交互作用视角下实证分析广东省区域创新系统影响因素和路径。以广东省21个市域创新数据为样本，基于ESDA方法分析了广东省区域创新系统空间特征和空间相关性，基于空间结构方程模型研究了广东省区域创新系统影响因素和影响机制。通过研究得到以下结论：1）创新政策、创新要素和外部经济环境对广东省区域创新系统具有正向影响，创新政策影响最大；2）创新主体对广东省区域创新具有不利影响；3）内部经济环境对区域创新系统具有正向直接影响，但是考虑到各因素交互作用后对新产品销售具有不利影响；4）区域创新空间相关性对广东省区域创新系统具有正向影响。根据研究提出以下建议： 1）适当管控外资引入规模，限制有技术壁垒的外资企业，提高外资利用质量；2）改进现有高校和科研机构创新评价和激励制度，实施多元化和市场化的创新评价体系；3）建立政府主导的广东省创新系统公共平台，提供创新系统所需的信息交互、多方合作、成果交易和需求发布等功能；4）加大教育科技投入力度，营造创新氛围，增强创新人才的培养。

关键词: 区域创新系统, 空间结构方程模型, 影响因素, 影响机制

Abstract:      Innovation is considered a key driver to economic growth and competitiveness. Traditional industrial economies are now transformed into knowledge economies where innovation is considered one of the mains drivers for sustained economic growth. Therefore, most of the advanced industrialized nations are keen to build their regional innovation systems. It is necessary to strengthen the scientific and technological innovation and to research regional science and technology innovation strategies. It is particularly important to apply scientific methods to study regional innovation now. This paper attempts to research regional innovation system of Guangdong province based on the exploratory spatial data analysis and the spatial structure equation model.
    We find that most of the domestic and foreign literature focuses on the following four aspects: (1) research on the influencing factors of regional innovation; (2) research on the spatial characteristics of regional innovation; (3) research on the efficiency of regional innovation; (4) research on the influence of regional innovation on enterprises. The existing literature finds that technological innovation plays an important role in determining the foreign direct investment-economic growth relationship. The level of collaboration with different partners can enhance firms′ innovation capabilities only if the focal firms′ managers have developed the capacity to scan and acquire external knowledge. There are positive effects of internal R&D and externally sourced innovation practices, as well as a positive influence of organizational innovation on the realization of technological innovations. The efficiency of innovation between regions has obvious spatial linkage and dependence. There is a significant positive spatial spillover effect. Unrelated variety hinders regional innovation efficiency and spatial spillover effects are not significant.
    Regional innovation is a complex system with the interaction of multiple factors, and there exist spatial spillover effects in the regional innovation system. Therefore, it is necessary to study the regional innovation according to the interaction of multiple factors & spatial spillover effects. Our study differs from the existing studies in two aspects: (1) according to the interactive effects of multiple factors including the innovation subjects, elements, policy and economic environment, and (2) expanding the structure equation model to the spatial structural equation model based on the exploratory spatial data analysis.
    Firstly, this paper analyzes the relation of the innovation elements, innovation subjects, innovation policy, the external economic environment and the internal economic environment. Then nine hypotheses are proposed as following: Hypothesis 1: The innovation subjects have a positive effect on the regional innovation; Hypothesis 2: The innovation elements have a positive effect on the regional innovation; Hypothesis 3: The innovation policy have a positive effect on the regional innovation; Hypothesis 4: The innovation policy have a positive effect on the innovation elements; Hypothesis 5: The innovation policy have a positive effect on the innovation subjects; Hypothesis 6: The internal economic environment has a positive effect on the regional innovation; Hypothesis 7: The internal economic environment has a positive effect on the innovation elements; Hypothesis 8: The internal economic environment has a positive effect on the innovation subjects; Hypothesis 9: The external economic environment has an inconclusive effect on the regional innovation. The Structure Equation Model of the regional innovation system is established based on the theoretical analysis and above 9 hypotheses.
    The exploratory spatial data analysis and structural equation model are adopted in the paper. Spatial analysis is statistically important because it can enhance inference accuracy, and at the same time reduces estimate bias by considering spatial proximity and dependence. The exploratory spatial data analysis should be considered as a descriptive step to explain the spatial patterns understudy and before estimating and testing more sophisticated regression models. The exploratory spatial data analysis can reveal complex spatial phenomena. Global autocorrelation is assessed by global Moran′s I statistic. A positive Moran′s I value indicates a general pattern of clustering in space of similar values. Structural equation modeling has roots in two different techniques developed in two different fields. Path analysis with its graphical representations of effects and effect decomposition comes from genetics research. Structural equation modeling is a confirmatory technique, which means that a model is formulated based on the theory, and it is judged whether this model should be rejected by fitting the model to data. If multivariate normality of the data holds, the variance-covariance matrix of the interest and the sample size is sufficient to fit models to the data. In order to test hypotheses, these hypotheses have to be translated in the a statistical model. The statistical model can be formulated in different ways. The structural equation model is defined by two main parts: the first part-the Structural Model- refers to the relationships among the latent variables, while the second part-the Measurement Model- represents the relationships between manifest and latent, endogenous and exogenous variables, respectively.
   All the data in this paper comes from the Guangdong Statistical Yearbook, Guangdong Intellectual Property Yearbook and China City Statistical Yearbook. The data time range is from 2005-2016. The space character and relativity were analyzed based on the exploratory spatial data analysis, the influence factor and mechanism were researched based on the spatial structure equation model with the example of 21 prefecture-level cities in Guangdong. We can gain the following conclusions: (1) The innovation policy, innovation elements and the external economic environment have a positive effect on the innovation, the innovation policy has more effect than the others; (2) The innovation subjects have a negative effect on the innovation; (3) The internal economic environment has a positive effect on the innovation and shows negative effect according to interaction; (4) The space relation has a positive effect on innovation. 
    The following suggestions are recommended: (1) It should restrict the scale of foreign investment and improve the quality of foreign capital; (2) The innovation evaluation and incentive system in colleges and universities should make better and implement innovation evaluation system of diversification and marketization; (3) The public platform of innovation system should be built by the government in Guangdong and provide the function of information exchange, multi-party cooperation, achievement trading and demand release; (4) The investment in educational science and technology must be enhanced and to create the scope of innovation and enhance the training of innovative talents.

Key words:  regional innovation system, spatial structure equation model, influence factor, influence mechanism