本地知识基础对新兴产业创新集群形成的影响——以中国燃料电池产业为例

新兴产业的区位选择一直是经济地理学研究的重点问题,以中国燃料电池产业为例,本研究重点对“区域产业分叉”理论在燃料电池创新集群形成中的适用性进行探究。本文将知识基础区分为解析型知识、合成型知识、符号型知识,采集中国2000—2016年291个地级市燃料电池产业的专利信息、专利引证及被引信息、学术论文信息,通过ArcGIS、Stata软件进行空间和计量分析,得到如下结论：① 从格局上来看,燃料电池创新集群发展主要集聚于发达省会城市,并且在格局演化过程中路径依赖作用明显,291个地级市在2000—2005年、2011—2016年两个时期的创新集群指数相关系数高达0.904,呈现显著的强相关性。② 本地解析型、合成型、符号型知识库以及创新集群指数滞后一阶项的回归系数均显著为正,说明知识基础对新兴产业创新集群发展具有显著的积极影响,成为支持“区域产业分叉”的重要证据。③ 燃料电池产业创新集群形成主要以“区域产业分叉”方式进行,城市在知识基础上的优势对创新集群的形成具有积极作用,而创新集群的发展又使得知识基础进一步积累,实现良性循环,路径依赖现象显著;创新集群指数排名前20位城市中,直辖市和省会城市就占据了75%,一般城市的区位机会窗口较难打开。

The impact of local knowledge base on the formation of innovation clusters in emerging industries: The case study of China's fuel cell industry

The location choice of emerging industries has always been the focus of economic geography research. "Windows of Locational Opportunity" and "Regional Branching" have different understandings about the relationship between local knowledge base and location choice of emerging industries. This study distinguishes the concepts of innovation cluster and industrial cluster, and holds that innovation cluster is more conducive to focusing on the location selection of emerging industries. Then taking China's fuel cell industry as an example, this study focuses on the applicability of "regional industry bifurcation" theory in the formation of fuel cell innovation cluster. And this paper divides the knowledge base into analytical knowledge, synthetic knowledge and symbolic knowledge. This paper collects patent information, patent citation and cited information, academic papers information of 291 prefecture-level cities in China from 2000 to 2016. With the aid of ArcGIS and Stata software, this paper carries out a spatial pattern analysis and two-step panel system GMM regression analysis, and draws the following conclusions. (1) From the perspective of pattern, the development of fuel cell innovation cluster is mainly concentrated in developed provincial capital cities, and a few other major cities such as Dalian. In the process of innovation cluster development, path dependence is obvious. The correlation coefficient of innovation cluster index of 291 prefecture-level cities in 2000-2005 and 2011-2016 is as high as 0.904, showing significant strong correlation. (2) Local analytical, synthetic and symbolic repositories all have significant effects on the development of innovation clusters, and the lagged first-order term of innovation cluster index is significantly positive, which supports that the local knowledge base has a significant positive impact on the development of emerging industry innovation cluster, and becomes an important evidence to support "Regional Branching". (3) The formation of fuel cell industry innovation cluster is mainly carried out in the way of "Regional Branching". The advantages of developed cities on the knowledge base help them to occupy an advantage in the formation of fuel cell innovation cluster, which leads to the further accumulation of knowledge-based advantages. Therefore, the virtuous circle between knowledge base and innovation cluster is formed, and path dependence is significant. Among the top 20 cities in innovation cluster index, the number of municipalities (at provincial level) and provincial capitals account for 75%, almost all other cities are located in the eastern coastal areas, which shows that it is difficult to open the location opportunity window of ordinary cities.