In the past few decades, the new energy vehicle industry has formed a huge comprehensive knowledge base in the global innovation system through cooperation between automobile companies. As a typical region for innovation research in the new energy vehicle industry, what is the development stage of China's new energy vehicle industry? Have complex innovation networks been formed and driven by spillover effects from innovation networks as a whole? At present, there are relatively few studies on the above issues, and this paper attempts to focus on the analysis and discussion of the above issues.
The phenomenon of knowledge sharing, matching and spillover generated by innovation activities at different scales is the core content of current innovation research. Moreover, with the introduction of concepts such as flow space and network, network externalities, that is, the spillover effect of knowledge in the network structure and agglomeration externalities, have become the core concepts to characterize spatial spillover effects. However, due to the late development of new energy vehicles, there are relatively few relevant studies on the innovation network of the new energy vehicle industry in China, and only a small number of studies have analyzed the network structure, and the spillover effect in its development is rarely involved. At the same time, in the rapid development of China's new energy vehicle industry in the past 10 years, independent new energy vehicle brand enterprises represented by BYD, NIO, Xiaopeng, etc. have gradually occupied the new energy vehicle market, impacting the traditional automobile market dominated by state-owned automobile enterprises; Moreover, as an industry with a high technical threshold, the new energy vehicle industry may have an innovative development model different from the traditional automobile manufacturing industry.
Based on this, this study selects the innovation activities of China's new energy vehicle industry as the research object, and mainly analyzes two aspects: Based on the patent data of new energy vehicles, the social network analysis method is used to construct the urban cooperation matrix of China's new energy vehicle industry in the past 10 years from 2011 to 2020, and identify its innovation network structure and evolution process; Based on the city cooperative relationship in the innovation network, the network spillover effect is calculated by the spatial Dubin model, and compared with the spillover effect based on the city's adjacency relationship and distance relationship, in order to provide richer empirical evidence and development suggestions for the development model of China's new energy vehicle industry.
Analysis of the spatial structure of innovation network in the new energy vehicle industry
New energy vehicle industry innovation network data and methods
The data for this study is derived from the Patent Intelligence Service Platform of Patent Exchange. New energy vehicles refer to vehicles using unconventional power sources, so based on existing research, this study uses "pure electric", "new energy", "hybrid", "hydrogen power" and "fuel vehicle" as the search fields for patent search; At the same time, considering that the rapid development of the new energy vehicle industry in mainland China is mainly concentrated in the past 10 years, and the relevant statistical data used to calculate the spillover effect are only updated to 2020, this study screens the domestic valid patent data with the patent publication date from 2011 to 2020, and excludes the patent data that is completely filed by individuals, and obtains a total of 24 957 effective patent data for new energy vehicles, of which 2 328 cooperative patent data are completely completed by enterprises, universities and scientific research institutions. On this basis, the geographical location of each patent application unit is obtained one by one by using the "Aiqicha" enterprise information vertical search engine and display platform - when the geographical location of the patent applicant, university and scientific research institution is located in two different cities, there is 1 cooperative relationship between the two cities, and if there are three or more cities, then 1 cooperative relationship between the two is recorded, and then build an innovation network for China's new energy vehicle industry.
Based on the social network analysis method, this study calculates the overall and node characteristics of the innovation network of China's new energy vehicle industry by using a number of related indicators. In the calculation of the overall characteristics of innovation networks, the network connection ratio refers to the proportion of cooperative patents to all patents; Average refers to the average number of cooperative nodes of each node in the innovation network; Network density refers to the degree of direct cooperation between nodes in the innovation network; Average distance refers to the average distance between any two nodes in the innovation network that creates a cooperative link; The average clustering coefficient is based on the average distance, and the average probability of establishing a connection between two nodes connected to the same node in the network is measured. In the calculation of the specific node characteristics of the innovation network, node centrality refers to the number of nodes that have cooperative relationships with a node; Proximity centrality refers to the reciprocal of the average distance from a node to all other nodes with which it has partnerships; Mediation centrality refers to the number of shortest paths from a node to nodes with which it has a relationship.
Overall characteristics of the innovation network of the new energy vehicle industry
The innovation capacity of China's new energy vehicle industry is constantly improving, but there is relatively little cooperative innovation. Table 1 shows that since 2011, the number of patents approved for China's new energy vehicle industry has increased year by year; Especially after 2015, China's new energy vehicle industry has entered a period of rapid development, and the number of patents approved each year has increased by more than 600 compared with the previous year, which shows that the innovation ability of China's new energy vehicle industry is improving year by year. On the contrary, although the proportion of network junctions reached a maximum of 22.57% in 2014, an increase compared with previous years, it still showed an overall trend of decreasing year by year, and by 2020, the network connection ratio was only 10.98%. This reflects that since 2011, although the number of cooperative patents approved in China's new energy vehicle industry has also increased as well as the total number of patents, more patents are applied for independently by enterprises, and the growth rate of the number of cooperative patents is much slower than the growth rate of the total number of patents.
The number of cities participating in China's new energy vehicle industry innovation network is increasing year by year, but the cooperation between cities is mainly concentrated in some cities. The average has increased year by year since 2011 and began to decrease after reaching the maximum in 2018. This reflects the gradual enrichment of cooperation between cities with the development of China's new energy vehicle industry and the increasing number of enterprises and cities involved in its production innovation. However, affected by the further decline of the domestic new energy vehicle subsidy policy, China's new energy vehicle production and sales declined for the first time in 2019, and new energy vehicle companies correspondingly reduced patent cooperation between enterprises, while in 2020, affected by the epidemic, the cost of inter-city cooperation increased, so the average degree after 2019 decreased significantly. Although more cities are participating in China's new energy vehicle industry innovation network, the newly participating cities are not closely cooperating with other cities, and only a few specific cities have carried out patent cooperation on new energy vehicles.
Trends in mean distance and mean clustering coefficient further explain the annual decline in network density since 2011. With the increasing number of participating cities in China's new energy vehicle industry innovation network, the average distance was stable at 2.2-2.4 after 2014, while the average clustering coefficient was only 5.089 in 2017, but the overall trend was upward. This reflects the gradual prominence of the small-world network characteristics of China's new energy vehicle innovation network, that is, new energy vehicle patent cooperation is often concentrated between two or a few cities, rather than random.
Node characteristics of new energy vehicle industry innovation network
The innovation activities of China's new energy vehicle industry are mainly concentrated in the three major urban agglomerations and other municipalities, provincial capitals and capital cities. As can be seen from Figure 1, the applicants for valid patents in China's new energy vehicle industry are mainly concentrated in Beijing and the Yangtze River Delta urban agglomeration, and there are also a small number of distribution in the Guangdong-Hong Kong-Macao Greater Bay Area, Chongqing, Chengdu, Wuhan, Zhengzhou and other regions. This reflects the geographical location of China's new energy vehicle R&D enterprises and scientific research institutions. For example, Beijing has Tsinghua University, Beijing Institute of Technology, various research institutes under the Chinese Academy of Sciences, as well as BAIC Group and State Grid Corporation of China; The Yangtze River Delta city cluster has new energy vehicle R&D units such as Shanghai Jiao Tong University, SAIC, Zhejiang University, Geely Automobile Group, Southeast University (Jiangsu), Nanjing University of Science and Technology, Guodian Nari Technology Co., Ltd., and Hefei University of Technology, Jianghuai Automobile Group and Chery Automobile Group (Anhui).
Beijing, Shanghai and other provincial capitals and capital cities are the core nodes of China's new energy vehicle industry innovation network, and have the ability to guide other cities to jointly carry out new energy vehicle patent cooperation. As can be seen from Table 2, Beijing's node centrality has always been the highest value among cities in China's new energy vehicle industry innovation network since 2011, and it is much higher than other cities. This reflects that Beijing not only has the largest number of new energy vehicle industry patents, but also carries out innovation cooperation with more other node cities in the innovation network than other cities. Beijing has also maintained a much higher value than other node cities in terms of intermediary centrality since 2011. This reflects that in the new energy vehicle industry innovation network, Beijing has a much higher ability than other cities to jointly carry out R&D activities in conjunction with two or more other cities. Shanghai's node centrality and intermediary centrality have increased significantly since 2011. At the same time, cities in other Yangtze River Delta urban agglomeration areas, such as Hangzhou and Nanjing, also have a high degree of node centrality and intermediary centrality. Combined with Figure 1, it can be seen that in the Yangtze River Delta urban agglomeration, Shanghai and other cities have formed a complex new energy vehicle industry innovation network, and also have the ability to jointly carry out R&D activities with two or more other cities. In addition, compared with 2011-2015, the top 10 cities in terms of node centrality and intermediary centrality from 2016 to 2020 have significantly increased in provincial capitals and capital cities, which is due to the fact that provincial capitals and capital cities usually have more policy preferences and richer enterprise and scientific research resources in the province and autonomous region. As more cities participate in China's NEV industry innovation network, provincial capitals and capital cities are better able to cooperate with cities that take the lead in developing NEVs, or independently carry out and guide other neighboring cities to participate in NEV R&D activities.
In summary, with the development of China's new energy vehicle industry, its innovation network has begun to take shape and has a trend of rapid growth. However, on the whole, the cooperation between cities is not close, and as the number of cities participating in the innovation network increases, some municipalities, provincial capitals and capital cities rely on their economic and policy advantages to make the small-world network characteristics of the innovation network more prominent. Among them, Beijing is the core node of China's new energy vehicle innovation network, and it also has a strong ability to promote new energy vehicle R&D cooperation in other cities.
Innovation network spillover effect of new energy vehicle industry
Characteristics of spillover effect in the new energy vehicle industry
Spillover effects reflect externalities generated in economic activities, where agglomeration externalities and network externalities explain spillovers in innovation activities from two perspectives. The former emphasizes the distance cost of knowledge dissemination - more adjacent locations are often more conducive to knowledge sharing, matching and learning between enterprises; Especially in environments with relatively poor traffic conditions, the impact of distance on knowledge spillover in space is more significant. The latter believes that with the continuous expansion of production networks, the cross-regional cooperation of industrial clusters becomes more and more obvious - the innovation power of different regions depends not only on their own endogenous factors, but also on their division of labor and cooperation in production networks at multiple scales. With the promotion of the integration of production, education and research, universities and scientific research institutions outside the enterprise have also begun to participate in innovation activities. Therefore, the spillover effect of innovation activities is not only limited to enterprises in the vertical industrial chain, but also exists in the knowledge chain between horizontal enterprises and universities and scientific research institutions.
As an emerging high-tech threshold industry, the industrial chain of the new energy vehicle industry involves R&D and cooperation between many different types of enterprises, between universities and scientific research institutions, and between enterprises and universities and scientific research institutions. From the perspective of cooperative relations in the new energy vehicle industry innovation network: (1) Part of the cooperation in the new energy vehicle industry is the cooperation between new energy vehicle enterprises and relevant universities and scientific research institutions. In the early stage of the development of China's new energy vehicle industry, the research and development capabilities of new energy vehicle technology were mainly concentrated in universities and scientific research institutions, and emerging new energy vehicle enterprises must rely on government policy support and cooperate with scientific research institutions to carry out new energy vehicle technology innovation. However, new energy vehicle enterprises and universities and scientific research institutions with new energy vehicle research and development capabilities are not completely coupled in spatial distribution, so it is easy to form cross-regional urban cooperation. (2) Cooperation in the new energy vehicle industry is also a collaboration between new energy vehicle companies. With the development of the new energy vehicle industry, more large automobile companies have participated in research and development activities in this field, and such automobile companies have sufficient scientific research funds to spontaneously carry out the research and development of new energy vehicle technology. In order to reduce production costs, enterprises divide the production of different products and the corresponding innovation departments from within the enterprise to other cities for production and R&D through internal division of labor, or divide the R&D departments separately into cities with better innovation environment, so as to form a cross-city cooperative relationship. Knowledge also creates a network spillover effect in two ways.
Spatial autocorrelation analysis
There is a certain spatial agglomeration phenomenon in the effective patents of China's new energy vehicle industry between cities, and the degree of agglomeration is on the rise. It can be seen from Table 3 that the global Moran index of China's new energy vehicle patents from 2011 to 2015 and 2016-2020 passed the significance test of 5%, and the Moran index was positive, but low overall, which indicates that China's new energy vehicle patents have weak spatial agglomeration between cities. Compared with 2011-2015, the Moran Index from 2016 to 2020 has improved, which also reflects that as more cities participate in the research and development of new energy vehicles, the spatial agglomeration phenomenon of patent applications is gradually prominent.
Figure 2 further illustrates the specific spatial distribution characteristics of the spatial agglomeration phenomenon of patent applications. (1) Tianjin, Baoding and Zhangjiakou, which are adjacent to Beijing, are all "high-high" clustering areas, but there are also "low-high" clustering areas such as Tangshan and Langfang. This reflects that Beijing, as the core city of new energy vehicle research and development, has a certain knowledge spillover effect on its neighboring cities, but its spillover has a more obvious direction - although some cities are geographically close to Beijing, Beijing knowledge spillover does not have a more obvious impact on it. (2) In the Yangtze River Delta urban agglomeration, there is a significant spatial agglomeration phenomenon of new energy vehicle patents, and cities represented by Shanghai, Nanjing, Hangzhou and Ningbo have formed a "high-high" cluster. At the same time, comparing the spatial pattern of the local Moran index from 2011 to 2015 and 2016-2020, it can be found that from 2011 to 2015, there were a large number of cities belonging to the "low-high" cluster on the periphery of the cities belonging to the "high-high" cluster, mainly distributed in the southern part of Zhejiang Province, the northern part of Jiangsu Province and the eastern part of Anhui Province. With the development of the new energy vehicle industry, these cities have gradually changed from "low-high" clusters to "high-high" clusters, and the "low-high" clustering areas have further extended to the periphery, which shows a significant knowledge spillover effect based on geographical proximity. (3) There are contiguous "low-low" clustering areas in western Guangdong Province, Guangxi Zhuang Autonomous Region and Yunnan Province. These three provinces all have a certain automobile manufacturing base, but relatively few companies are involved in new energy vehicles. Although Guangdong Province has important new energy vehicle companies such as BYD, its knowledge spillover is limited to the Guangdong-Hong Kong-Macao Greater Bay Area, and it is difficult to produce knowledge spillover in more obvious geographical proximity.
Significance analysis of spillover effects
According to the results of the Moran index, China's new energy vehicle industry may have certain spillover effects. Therefore, this study further uses the spatial Durman model to measure the spatial spillover effect of China's new energy vehicle industry. Existing research points out that explicit knowledge and tacit knowledge have different modes of transmission: tacit knowledge is usually transmitted informally, and its transmission distance is more restrictive, while explicit knowledge is easier to spread across regions in multiple ways, and the two complement each other. In addition, multidimensional proximity also emphasizes that there are many other proximities in addition to geographic proximity. Therefore, when constructing the model, the number of patents in each city was used as the explanatory variable, and 10 indicators such as the per capita gross domestic product (GDP) of the city, the R&D capital investment of industrial enterprises above designated size, and the expenditure on science and technology were used as the explanatory variables. On the basis of setting the network adjacency matrix, the urban spatial distance matrix and the spatial adjacency matrix were set as controls to measure the impact of spatial proximity, spatial distance and network proximity between cities on the spillover effect of the new energy vehicle industry.
The results of spatial spillover effects of China's new energy vehicle industry based on different spatial matrices show that: (1) The innovation and development of China's new energy vehicle industry has not formed a spillover effect to neighboring cities as a whole. Among the spillover effects based on the distance relationship and adjacency relationship of cities, the spatial autoregressive coefficient (RHO) was not significant, which indicates that the innovation activities of China's new energy vehicle industry from 2011 to 2020 did not form a significant spillover from cities to neighboring cities on a national scale. (2) There is a certain network spillover effect in the innovation and development of China's new energy vehicle industry. The RHO of the network spillover effect is significant and positive at the 1% level, indicating that the innovation activities of China's new energy vehicle industry from 2011 to 2020 have a certain positive spillover in the urban network. (3) The R&D investment of enterprises has played an important role in promoting the innovation and development of the new energy vehicle industry. In the main effect, the internal expenditure of R&D funds of industrial enterprises above designated size was significantly positively correlated with the number of new energy vehicle patents, while in further effect decomposition, the internal expenditure of R&D funds of industrial enterprises above designated size in the city was still significantly positively correlated with the number of patents in the new energy vehicle industry in the direct effect, but the indirect effect was not correlated. This means that although the company's local R&D investment has a certain role in promoting local new energy vehicle innovation activities, it has not produced a significant spillover effect.
Therefore, from the above analysis, it can be seen that new energy vehicle enterprises are the main body of new energy vehicle innovation activities in China. However, on the one hand, the development of China's new energy vehicle industry is still in its early stage, and a relatively sound industrial system has not been formed; On the other hand, the research and development of new energy vehicles has a high technical threshold, and cooperation between enterprises is difficult to carry out. The network connection ratio of China's new energy vehicle industry innovation network and its downward trend also indicate that most of the research and development of new energy vehicles is basically completed by enterprises independently. As a result, China's new energy vehicles did not have a significant spillover effect from 2011 to 2020, and only had a certain concentration of innovative activities in the Yangtze River Delta urban agglomeration.
Conclusions and recommendations
conclusion
This study takes the innovation activities of China's new energy vehicle industry as the research object, uses the cooperative patent data of the new energy vehicle industry to construct the urban cooperation matrix of China's new energy vehicle industry, and identifies the structure and evolution process of its innovation network. On this basis, the spatial Dubin model is used to calculate the network spillover effect of innovation activities in China's new energy vehicle industry, and compared with the spillover effect based on city distance relationship and proximity relationship, three specific conclusions are obtained.
(1) From 2011 to 2020, the innovation activities of China's new energy vehicle industry have begun to take shape, and a certain scale of innovation network has been formed; However, compared with the overall innovation activities of the new energy vehicle industry, there is relatively little innovation cooperation between cities and enterprises, the cooperative relationship is not close, and the innovation network is still in the early stage of development.
(2) In the process of participating in the new energy vehicle industry innovation network, the scale and level advantages of cities are significant. Larger provincial capitals and capital cities have more enterprise and policy advantages to attract other enterprises or scientific research institutions to carry out R&D activities of new energy vehicles, and promote neighboring cities to participate in relevant R&D activities.
(3) The R&D investment of enterprises is the core driving force to promote the innovation and development of the local new energy vehicle industry, but its role scale is mainly manifested in the city at the current stage. Although the innovative development of the new energy vehicle industry has a certain spatial agglomeration phenomenon in the urban agglomeration area of the Yangtze River Delta, on the whole, there is no significant spillover effect in the geographical proximity and network proximity of the city.
suggestion
(1) Continue to implement subsidy policies for innovation and research and development in the new energy vehicle industry, and promote cooperation between enterprises and enterprises, enterprises and universities and scientific research institutions within the new energy vehicle industry. When formulating preferential policies for new energy vehicle enterprises, in addition to reducing and exempting relevant taxes, the government also needs to implement relevant policies that are conducive to enterprises to carry out technology research and development, so as to promote the integration of production, education and research in the local new energy vehicle industry, and encourage local enterprises to cooperate with universities and scientific research institutions and other enterprises with complementary technologies. At the same time, relevant policies conducive to R&D cooperation between cities will be formulated at different spatial levels (such as within provinces, metropolitan areas, urban agglomerations, etc.), reduce the cooperation barriers caused by administrative boundaries in the process of enterprise cooperation in different cities, and promote the development of new energy vehicle industry innovation networks.
(2) According to the development stage, innovation resource endowment and location conditions of the new energy vehicle industry in different regions, improve the construction of the new energy vehicle industry chain and knowledge chain in a targeted manner, and give full play to the driving role of the core node cities in the innovation network to promote the network development of new energy vehicle industry innovation in the region. As far as specific cities and regions are concerned: (1) Continue to give full play to Beijing's advantages in scientific research resources, increase R&D funding for universities with new energy vehicle R&D strength such as Tsinghua University and Beijing Institute of Technology, as well as scientific research institutions such as the Institute of Microelectronics of the Chinese Academy of Sciences and the Institute of Electrical Engineering of the Chinese Academy of Sciences, and encourage scientific research institutions to carry out R&D cooperation with local and neighboring cities of new energy vehicle enterprises. (2) Give full play to the advantages of relatively complete industrial chain and close cooperation among enterprises in the Yangtze River Delta city agglomeration, appropriately reduce the policy restrictions on enterprise cooperation between different cities and provinces, and promote the development of the new energy vehicle industry innovation network in the Yangtze River Delta city agglomeration. (3) The Guangdong-Hong Kong-Macao Greater Bay Area has many new energy vehicle brands such as BYD, Xpeng, and GAC AION, and the production, R&D of key components upstream of the industrial chain are distributed in various cities in the region, but there is a lack of cooperation between enterprises in the industrial chain. In the future development of the new energy vehicle industry, upstream and downstream enterprises between different cities should be encouraged to cooperate to achieve an effective combination of upstream and downstream industries. (4) On the one hand, the central and western regions should make use of local land rent, manpower and other resource advantages to attract investment from new energy vehicle enterprises in the eastern region, and jointly carry out relevant R&D activities with other new energy vehicle enterprises to participate in the new energy vehicle industry innovation network; On the other hand, we should give full play to the resource advantages of new energy vehicle brands such as Changan in the central and western regions, and gradually drive the new energy vehicle industry in the central and western regions to innovate and develop.
(Authors: Xiong Zhifei, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences; Wenzhong Zhang, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Key Laboratory of Regional Sustainable Development Analysis and Simulation, Chinese Academy of Sciences, School of Resources and Environment, University of Chinese Academy of Sciences. Contribution to Proceedings of the Chinese Academy of Sciences))