Next Article in Journal
Difficulties Encountered by SMMEs in uMhlathuze Municipality Due to COVID-19 Crisis
Previous Article in Journal
Toward Adjusting a Conceptual Model for Adaptive Reuse Practice in the Context of Sustainability: A Case Study of Historical Military Buildings in the North of Iraq
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 17
Issue 2
10.3390/su17020740
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Formation of Knowledge Flow Networks in the Yangtze River Delta, China: Knowledge Implicitness and Proximity Effect

by
Pengcheng Zhu
1,2,
Jianglong Chen
1,*,
Feng Yuan
1 and
Weichen Liu
1
1
Key Laboratory of Watershed Geographic Sciences, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 211135, China
2
University of Chinese Academy of Sciences, Beijing 100049, China
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(2), 740; https://doi.org/10.3390/su17020740
Submission received: 25 October 2024 / Revised: 5 January 2025 / Accepted: 16 January 2025 / Published: 18 January 2025
Browse Figures
Versions Notes

Abstract

:
Knowledge flow as the key to facilitating new technology production and diffusing innovation is crucial for achieving sustainable development. However, previous studies pay less attention to the type of knowledge in knowledge flow network construction, possibly leading to the deviation of conclusions. To fully show the panorama of knowledge flow, this study distinguishes between explicit and tacit knowledge based on the transfer of patent rights data and talent flow data, describes the spatial characteristics of flow networks and uses a multiple regression quadratic assignment procedure model to analyze the proximity mechanism of network formation in the Yangtze River Delta. We find that knowledge flow networks in the Yangtze River Delta cover a wide range but are extremely uneven, mainly concentrated along the Yangtze River and around Hangzhou Bay. In addition, the spatial structures of different types of knowledge flow networks vary. Different dimensions of proximity act in relatively consistent directions for both types of knowledge flows, with geographical and organizational proximity found to exert positive effects on facilitating knowledge flows while cognitive proximity has a negative impact. There is also a substitution effect between geographical proximity and organizational proximity, and a complementary effect with cognitive proximity. These findings provide significant implications for optimizing knowledge flow networks and promoting sustainable development.

1. Introduction

As sustainable development becomes a contemporary imperative, the demand for innovation is increasing [1]. However, due to the risk of innovation, technological complexity and environmental uncertainty, the contemporary innovation process has shifted from a traditional linear model directly linked by R&D, invention and innovation to an interactive model requiring local buzz and global pipelines [2]. Knowledge sharing and collective learning among innovation agents have become important supports for achieving innovation [3,4]. Obviously, static knowledge resources do not bring advantages in a fierce market, and the formation of competitive advantage mainly lies in the transmission, accumulation and creation of knowledge in the process of dynamic flow. Therefore, knowledge flow is not only a basic form of realizing the transformation of knowledge into innovation value but also an important channel for actors to expand their external networks and make up for a lack of resources, while innovation can be seen as a dynamic process of knowledge flow and resource activation [5]. In the context of the current global economic downturn, the intensification of Sino–US friction, and the complex and volatile international situation, China’s economic development also faces the risk of technological decoupling, for which the government accordingly proposes to build a new development pattern of “taking the domestic grand cycle as the main body and promoting both domestic and international cycles”. To give full play to the main role of the domestic circulation, production factors such as labor, capital and knowledge need to flow freely within the region. Therefore, it is urgent to reveal the spatial paths of regional knowledge flows and clarify the proximity influence mechanisms behind them, in order to promote scientific allocation of innovation resources and the enhancement of innovation capacity.
Knowledge refers to the integration of information by understanding or acquiring technology, and knowledge flow is the process of knowledge generation, transmission, integration, accumulation, absorption, value added, and recreation [6,7]. In this paper, the concept of knowledge flow refers to a narrow understanding, i.e., the spatial process of flow of knowledge factors from the perspective of geography. The existing literature mainly focuses on the following aspects: (a) Earlier theories of knowledge flow can be traced back to Adam Smith’s description in The Wealth of Nations and the industrial district theory proposed by Marshall, emphasizing the importance of experiential learning and knowledge diffusion from different perspectives [8]. Subsequently, theories such as knowledge-based theory, knowledge management theory and knowledge innovation theory have provided a rich basis for understanding knowledge flow and knowledge networks. Representative scholars include Amidon [9], Spender [10] and Nonaka [11]. (b) How to measure knowledge flow has been the focus of scholars worldwide, and a large number of empirical studies have been accumulated based on different knowledge carriers. Citation, patent transfer and movement trajectory of scientists are the most used data for characterizing knowledge flow [12,13,14]. Some scholars have also described the knowledge flow based on questionnaires [15]. For instance, Yu utilized collaborative publication and patent data to construct the knowledge and technological innovation networks within the Nanjing metropolitan area [16]. (c) Content mainly involves concept discrimination, forms, network, influencing factors and effect evaluation. Conceptually, a unified understanding of knowledge flow has not yet been formed, and its connotation is constantly enriched with the development of the times. Scholars have gradually realized that knowledge flow is a complex and dynamic process—it is not a simple transfer and exchange of knowledge between actors [17] but a process of knowledge reorganization, value-adding and reproduction—and have made useful additions to its connotation from different perspectives. The forms of knowledge flow include spillover, diffusion [18], transfer and sharing [19,20], while industry–university–research cooperation, industrial agglomeration, foreign direct investment, patents and informal networks are the main modes of knowledge flow [21]. The network is the main manifestation of the knowledge flow process. The related literature mainly focuses on exploring the topology and spatial pattern of networks, revealing the status and role of actors in the networks. Ma and Duan found that the knowledge flow network in China is shifting from a vertical-scale hierarchy to a horizontal network system, with a quadrilateral spatial pattern with Beijing, Shanghai, Guangzhou and Chengdu as the vertices gradually emerging [22,23]. Knowledge flow faces many obstacles, and these factors can be roughly divided into three categories: object factors, such as the characteristics of knowledge itself; subject factors, such as the willingness and ability of the source to share knowledge, the absorptive capacity of the recipient and the mutual relationship between subjects; and environmental factors, such as policies, culture, society, platforms and the convenience of transmission channels [7,12,24]. Object factors determine the occurrence of knowledge flow behavior, while the positive behavior and mutual relationship between subjects greatly drives knowledge flow, and environmental factors indirectly act on the object and subject through external forces. In addition, theoretical frameworks of multidimensional proximity have been widely used to explain cross-regional knowledge flows. Evolutionary economic geographers, represented by Boschma, have developed the concepts of geographic, cognitive, organizational, institutional and social proximity, and the role of proximity in facilitating inter-subjective knowledge flows has been increasingly emphasized and confirmed; excessive proximity may lead to a lock-in effect and thus inhibit the flow process [25,26]. Effect evaluation mainly focuses on the impact of knowledge flow on the regional innovation system and innovation performance and capability. Relevant studies show that the free flow of innovation factors can promote spatial knowledge spillover and lead to Pareto equilibrium results in regional economic development when the late-mover advantage is gradually eliminated [27]. (d) In terms of methods, the construction and visualization of knowledge flow are mostly based on the gravity model and network science methods [28], and influencing factor analysis adopts a combination of qualitative and quantitative methods. Zhang [29] used a negative binomial regression method to analyze the proximity mechanism of China’s technology transfer network.
The existing research has yielded many innovative results, but there is still some discrepancy in the perception of the spatial paths of knowledge flow and the effects of proximity. The reasons for this are that, on one hand, existing knowledge networks are often constructed on the basis of a single relationship, such as a citation, co-author or human mobility, with less attention paid to the characteristics of the chosen carriers. This treatment of knowledge as homogeneous may lead to exaggerated or skewed conclusions, as it reflects only part of the characteristics of knowledge flows. On the other hand, the knowledge network structure and the proximity mechanisms behind it may vary due to differences in economic levels and innovation capabilities between regions. Given these facts, we attempt to distinguish different knowledge types and construct a comprehensive multidimensional proximity analysis framework to fill the shortcomings in the existing research. Our work reveals the spatial flow characteristics of different knowledge types and the differences of various proximities in different types of knowledge flow networks, which provides a useful addition to the controversies that exist in knowledge flow and proximity studies, as well as practical insights into the construction of regional innovation integration and sustainable development in the Yangtze River Delta.
This paper aims to reveal the organization of knowledge networks and excavate the proximity mechanism of network formation in the Yangtze River Delta. We would like to address the differences in the formations of different kinds of knowledge networks and the role of proximity in knowledge flow. In the following sections, we describe the types of knowledge and propose an analytical framework of the multi-dimensional proximity and the impacts on knowledge flows. Then, we introduce the data and methodologies. The core empirical section presents the analysis of spatial characteristics of knowledge flow in the Yangtze River Delta and the results of the regression analysis. Finally, we offer the discussion and concluding remarks.

2. Theoretical Background

2.1. Knowledge and Knowledge Flow

Knowledge can be divided into explicit knowledge and tacit knowledge regarding the difference described by Polanyi in The Tacit Dimension [30], i.e., what we know is more than what we can express. Explicit knowledge consists of formal skills and systematic language that can be encoded by means of texts or images and transmitted in the form of transfer, transactions or references, being easy to express and disseminate, while tacit knowledge is intended experience, which exists in people’s minds, is difficult to articulate and highly internalized and can only be acquired through informal learning behavior such as collaboration, experience sharing and people’s mobility and thus is more difficult in terms of flow [31]. Although the flows corresponding to these two types of knowledge differ significantly in terms of knowledge characteristics, storage forms, flow media, ease of transfer and spatial scope, they are not completely separated; they are interdependent and dialectically intertwined. The socialization, externalization, combination and internalization (SECI) knowledge creation model proposed by Nonaka and Takeuchi [11] well describes the spiral flow relationship between explicit and tacit knowledge.

2.2. Knowledge Flow and Its Proximity Mechanism

The existence and importance of knowledge flow and spillovers are widely acknowledged, and they are affected not only by the characteristics of knowledge itself, absorptive capacity, the external environment and the path, but also by the proximities between actors [32,33,34]. Actors are heterogeneous, have strong preferences for partners and are more likely to establish relationships with subjects similar to themselves [35]. The relationship between knowledge flow and proximity has also long been an important subject of economic geography, innovation geography and other disciplines.
Proximity was first proposed by the French School of Proximity in the 1990s, based on which scholars have entered the black box of externality to study the coordination mechanism in innovation geography. They have emphasized that in addition to geographical proximity, proximities in other dimensions play an important role in information interaction and innovation between organizations. Shaw and Gilly [36] divided proximities into geographical proximity, organizational proximity and institutional proximity. Evolutionary economic geographers such as Boschma proposed a five-dimensional framework of geographic, cognitive, organizational, institutional, and social proximity [33]. Since some proximities conceptually overlap, they can be broadly grouped into the three categories of geographical proximity, organizational proximity and cognitive proximity after referring to relevant studies [37,38].
Geographical proximity refers to the physical proximity between actors in geographical space. From the perspective of transaction costs, geographical proximity can reduce the costs of transportation and communication between actors, thereby improving the convenience of collecting and acquiring knowledge and technology resources, especially promoting the transfer of tacit knowledge [39]. Learning region theory points out that tacit knowledge cannot be clearly expressed and can be acquired only in practice, making it difficult to transfer remotely. Geographical proximity promotes the formation of face-to-face communication to overcome the obstacles caused by knowledge stickiness, and it is the key to the effective production and transfer of tacit knowledge [32,40]. In addition, innovation exists in an industrial atmosphere, and the formation of this industrial atmosphere depends on the geographical concentration of manufacturers and the mutual learning of workers in their daily face-to-face interactions. The concepts of geographical proximity and open learning are also important conditions for acquiring knowledge spillover. These theories all imply that the characteristics of knowledge flow attenuate with distance.
Organizational proximity refers to the closeness of organizational management, including similarity in the organizational structure, culture, system and social relations. It is the proximity formed by collective sense and similarity [37,41]. Embeddedness theory envisages that economic behavior is deeply embedded in the social network, in which personal relationships and trust are keys to effectively preventing various types of mutual destruction and fraud among partners, and rational economic activities are always carried out based on existing relationships [42]. Organizational proximity is conducive to the formation of trust mechanisms and internalized knowledge exchange; reducing transaction risks and costs in the process of knowledge spillover and providing a stable external environment for cooperation are also important institutional guarantees for knowledge spillover [43,44]. Brown and Pual [45] argued that the daily norms and practices shaped by organizations (or sub-groups within organizations) promote the creation and sharing of tacit knowledge.
Cognitive proximity refers to similarity in the knowledge base and development level. It is not only the basis of knowledge exchange but also the premise for the acquisition of knowledge spillover [46]. Resource-based theory holds that an organization is a collection of various resources, and unique heterogeneous resources are the source for it to maintain its competitive advantage [47]. When an organization’s existing knowledge base cannot meet its R&D innovation needs, it will seek new knowledge with heterogeneity and complementarity from the outside. However, it is not easy to incorporate external knowledge and technology into the organization’s own knowledge system. Having a similar knowledge system and economic capacity can reduce the stickiness caused by the professionalism and tacitness of knowledge [32], enable knowledge recipients to better understand, absorb and master the knowledge and technology offered by donors, and make it easier to realize inter-subjective knowledge exchange.
In addition, the impact of proximity on different knowledge flows differs due to the difference between explicit and tacit knowledge. Tacit knowledge is highly dependent on geographical proximity as it is difficult to encode. Howells [31] pointed out that tacit knowledge is more difficult to spread, and its own particularity makes the social background important. Hence, we propose the following hypotheses:
Hypothesis 1a.
Geographical proximity has a significantly positive impact on explicit knowledge flow.
Hypothesis 1b.
Organizational proximity has a significantly positive impact on explicit knowledge flow.
Hypothesis 1c.
Cognitive proximity has a significantly positive impact on explicit knowledge flow.
Hypothesis 2a.
Geographical proximity has a significantly positive impact on tacit knowledge flow.
Hypothesis 2b.
Organizational proximity has a significantly positive impact on tacit knowledge flow.
Hypothesis 2c.
Cognitive proximity has a significantly positive impact on tacit knowledge flow.
There is interaction between geographical and non-geographical proximity. On one hand, geographical proximity can play a catalytic role in the establishment of organizational proximity and cognitive proximity, providing both sides more opportunities to meet and communicate as well as the possibility of increased social relations. On the other hand, geographical proximity is not a necessary and sufficient condition for the occurrence of knowledge flow, and other forms of proximity can weaken the dependence of knowledge elements on physical distance [38]. The community of practice theory suggests that organizational or relational proximity and professional commonality are more conducive to the creation and flow of tacit knowledge than geographical proximity. Hence, we propose the following hypotheses (Figure 1):
Hypothesis 3a.
There is interaction between geographical proximity and organizational proximity.
Hypothesis 3b.
There is interaction between geographical proximity and cognitive proximity.

3. Materials and Methods

3.1. Study Area

This paper selects the Yangtze River Delta as study area, including Shanghai, Jiangsu province, Zhejiang province and Anhui province, with a total of 41 cities and an area of 359,000 km2. The Yangtze River Delta is located at the lower reaches of the Yangtze River in China (Figure 2), which is one of the most economically active and innovative regions in the country. By the end of 2022, the Yangtze River Delta region had a gross domestic product of CNY 29.02 trillion and a total of 472,606 invention patent applications, accounting for approximately 24.25% and 32.27% of the national total, respectively, making it an important engine driving China’s economic growth and an important strategic position in the construction of the country’s modernization system. With the integrated development of the Yangtze River Delta region elevated to a national strategy in 2018, the region is facing the challenges of transformation and upgrading. The Outline of the Yangtze River Delta Regional Integration Development Plan issued by the State Council clearly stated that efforts should be made to develop innovation integration, with the key to promoting the development of innovation integration being the realization of the cross-regional flow of innovation elements and the resulting knowledge spillover. Research on knowledge flow caters to the needs of regional high-quality development and has important strategic significance for the construction of cross-regional innovation networks in the Yangtze River Delta.

3.2. Data

Knowledge flow essentially is a kind of subjectively constructed “soft network” and an abstract relationship hidden in various direct and indirect links [48]. There are no objective data that can characterize it directly. Hence, how to construct the flow network to narrow the deviation from the objective reality is the basis of this study. Considering the close relationship between knowledge and the city carrier where it is located, it is feasible to construct a knowledge flow network through the cross-city behavior of knowledge carriers to some extent. Drawing on the dichotomy of knowledge mentioned above, the study divides knowledge carriers into explicit knowledge carriers attached to invention patents and tacit knowledge carriers embedded in human capital.
Patent transfer can be regarded as a tangible and traceable knowledge flow relationship. Nearly 90% of the global R&D output is included in patents. Patents condense high-value innovation elements and knowledge output, and they are the main indicators used to measure explicit knowledge. As one of the main routes of regional knowledge flow under the action of the market mechanism [49], patent transfer refers to the transfer of the patentee’s ownership or possession of a patent to the transferee, including the transfer of the patent application rights and the patent rights, of which the transfer of patent rights is the most direct manifestation of technology transactions, that is, the patent transfer referred to in this paper. Patent transfer data are derived from the patent information service platform (see http://search.cnipr.com/) under the State Intellectual Property Office, and a total of 18,624 valid samples were counted to construct the patent transfer database. By searching for the keyword “transfer” in the “legal status” column of the platform and using the Octopus Data Collector tool, a total of 94,458 transfer records in the Yangtze River Delta in 2018 were obtained. The records include detailed information such as the patent number, patent title, classification number, and name and address of the obligee before and after transfer. Then, we further screened the obtained records, and the specific process was as follows: (1) We eliminated some samples, including samples of transfer of patent application rights, samples in which the name and address of the obligee before and after transfer were missing, and duplicate samples. (2) Since this study discusses only the knowledge flow in the Yangtze River Delta, we eliminated samples in which the addresses of the obligee before and after transfer were outside the study area. (3) Related research shows that enterprises are more motivated to transform innovation into products than universities, research institutions and individual inventors, and enterprises are also the main actors who introduce innovations into the market. Considering the tacit knowledge carrier of this study, to select enterprise management talent, we divided the obligee into five categories based on the name—individuals, universities, research institutions, enterprises and others, including hospitals and government agencies—and selected only the patent transfer records between enterprises to ensure the matching of the analysis.
Talent flow is an implicit and highly personalized knowledge flow relationship. The process of talent entering a place is also the process of adding new tacit knowledge to the original knowledge base. The collision, integration and renewal of old and new knowledge helps to promote regional innovation and performance, which also establishes important knowledge links between the cities where talent lives. As a kind of talent, entrepreneurs use potential business opportunities related to their unique identity and background to play the role of boundary spanners and promote the exchange of market, product and technology information between clusters. Therefore, the inter-regional knowledge link is the result of entrepreneurs’ spatial migration [50]. Considering listed companies having a relatively complete organizational structure, we take directors, supervisors and senior management personnel in a company’s high-level structure as typical samples and obtain movement information by analyzing their resumes. Resume data were derived from the China Stock Market & Accounting Research Database (see https://www.gtarsc.com/), and a total of 11,586 samples were obtained, which were used as the basis for constructing the talent flow network. To avoid the repetition of annual statistics on personnel information, we selected the senior personnel samples of listed companies from November 2017 to November 2018 and combined the listed companies in the Shanghai Stock Exchange and Shenzhen Stock Exchange to eliminate samples that were not part of the study area and duplicate samples. Finally, a total of 11,586 samples were obtained. Then, we screened out samples containing job information in the resume, and 9493 valid samples were counted. On this basis, the geographical flow information of the samples above was extracted. We defined the first working place and the current working place of talent as the original working place and current working place, respectively, to determine the geographical flow path of enterprise management talent in prefecture-level cities (original working place → current working place), which was used as the basis for constructing the talent flow network.
Drawing on the principle of graph theory, taking cities as nodes, and taking the scale of patent transfer and talent flow between cities as edges, weighted asymmetric matrixes of patent transfer and talent flow in the Yangtze River Delta were constructed. These matrixes not only cover the direction of knowledge flow but also include the weight of flow intensity.

3.3. Analytical Methods

As the variables in this paper are all relational data in matrix form, they are not independent of each other and are structurally autocorrelated; thus, traditional methods such as ordinary least squares cannot be used for parameter estimation and statistical testing. The Multiple Regression Quadratic Assignment Procedure (MRQAP) is a non-parametric estimation method for relational data. The method regresses the explanatory and explained variables on the basis of multiple rank permutations of matrix data and performs random non-parametric tests along with the calculation of matrix correlations, which can effectively avoid problems such as endogeneity and spurious correlations [51]. Based on the analytical framework mentioned above, we use the MRQAP model to examine the impact of different dimensions of proximity (geographical proximity, organizational proximity, cognitive proximity) on different types of knowledge flow networks (patent transfer, talent flow) in the Yangtze River Delta. All variables are incorporated into the matrix representing the relationship between cities.
The selection and calculation of proximity variables are as follows: (a) Geographical proximity is represented by an adjacency matrix between cities. If the administrative boundary of two cities is bounded, the value is 1 and is 0 otherwise. (b) At the regional level, we mainly select institutional proximity and cultural proximity to measure organizational proximity. Institutional proximity mainly refers to the similarity of the formal institutional framework in terms of laws, regulations and planning systems in the region where the actors are located, and the study represents institutional proximity by the relationship between the administrative districts of cities. If two cities belong to one province, the value is 1 and is 0 otherwise. Cultural proximity refers to the similarity in informal systems, such as common standards, customs, values and daily norms, of which dialect is a prominent representative of regional culture [52]. Similar dialectal habits tend to share similar values, contributing to higher trust levels and reducing the risks and costs associated with communication barriers, thereby increasing the likelihood of knowledge spillovers [53]. We interpret the attribute of the urban dialect based on the Chinese Dialect Dictionary; if two cities belong to one dialect area, the value is 1 and is 0 otherwise. (c) Cognitive proximity is measured by technology proximity, industry proximity and economic proximity.
Technology proximity refers to the similarity between cities on the basis of technical experience and knowledge. Referring to the similarity coefficient of the technical structure proposed by Jaffe [54], the vector is constructed by the International Patent Classification (IPC) of invention patents and utility model patents authorized in the previous year.
T i j = ∑ k = 1 n f i k f j k / ∑ k = 1 n f i k 2 f j k 2
where fik and fjk are the total number of authorized patents of the previous year in class k of cities i and j, respectively, and the results are continuous variables ranging from 0 to 1. The larger the value is, the smaller the difference in the technical structure and knowledge level between cities. Data come from the platform of the State Intellectual Property Office (see https://www.cnpat.com.cn/).
Industrial proximity refers to the similarity in industrial structure between cities. We use the similarity coefficient of the industrial structure proposed by the United Nations Industrial Research Centre and calculate it based on the classification of industrial sub-sectors in the statistical yearbook.
S i j = ∑ k = 1 n X i k X j k / ∑ k = 1 n X i k 2 X j k 2
where Xik and Xjk are the proportion of the output value of industry k to the total output value of cities i and j of the previous year, respectively, and the calculation results are continuous variables ranging from 0 to 1. The larger the value is, the greater the degree of industrial homogeneity between cities.
Economic proximity refers to the similarity between cities in terms of the level of economic development; we use the difference in GDP of the previous year between cities for measurement. To unify the numerical logic with other indexes, we first standardize the range of the difference in GDP, then take its reciprocal, and finally replace it with the natural logarithm to obtain the economic proximity matrix. Higher values indicate that the economic strength of the two cities is closer. The industry output value and GDP data come from the statistical bureau of each province and city (see https://tjj.sh.gov.cn/, http://tj.jiangsu.gov.cn/, http://tjj.zj.gov.cn/, http://tjj.ah.gov.cn/).
Considering the MRQAP model is preferentially applied to process the binary adjacency matrix, the variables of technological proximity, industrial proximity and economic proximity are transformed into a binary relation matrix, with the average value as the cut-off value. The calculation results of each proximity are shown in Figure 3. On this basis, a regression with the knowledge flow relation matrix as the dependent variable and the proximity variable matrix as the independent variable is constructed.

4. Results

4.1. Knowledge Flow Networks in the Yangtze River Delta

The study standardized the scale of patent transfer and talent flow between cities with the percentage of the maximum value, divided the links from high to low into four levels based on the system clustering tool, and then drew the spatial pattern of the knowledge flow network in the Yangtze River Delta with Gephi 9.2 (Figure 4). Node size corresponds to the centrality value. The thickness of links corresponds to the flow scale, and the arc clockwise direction is the flow direction of elements between cities. It should be noted that, centrality is used to describe the status and influence of target nodes in directed networks, and the centrality index selected in this study is derived from the Entropy Index [55]. Compared with traditional centrality indicators such as degree centrality, closeness centrality and betweenness centrality, this index can better reflect the directionality of flow in a directionally weighted network. The expression is E I i = − ∑ j = 1 j x j ln ⁡ x j ln ⁡ ( k i − 1 ) , where ki is the number of nodes associated with node I, and xj is the ratio of the flow from node j to node i to the flow from node j to all nodes. The higher the EIi value, the higher the importance of node i in the whole network, with a dominant position and priority.
There are differences in the spatial structure of different types of knowledge flow networks. In the patent transfer network, the flow scale of the first-level links (Shanghai → Suzhou, Shanghai → Nantong, Hefei → Bengbu) and second-level links (Shanghai → Jiaxing, Suzhou → Shanghai, Shanghai → Huzhou) accounts for 9.86% and 5.85% of the total, respectively, which spatially manifests as a radial structure with Shanghai as a divergent core and spreads towards neighboring cities; the trend of neighboring spreading is obvious. Hefei → Bengbu, as the only patent flow not related to Shanghai, is the key to the backbone of the patent transfer network in Anhui Province, and Hefei has also become a regional hub for patent output. In the talent flow network, the flow scale of the first-level links (Shanghai → Suzhou) and second-level links (Nanjing → Suzhou, Hangzhou → Taizhou, Shanghai → Hangzhou, Shanghai → Nantong, Hangzhou → Shaoxing, Shanghai → Shaoxing) accounts for 6.42% and 16.32% of the total, respectively, where the “Z”-shaped structure with Shanghai as the core along Shanghai–Nanjing, Shanghai–Hangzhou and Hangzhou–Taizhou has been spatially formed and has obvious hierarchical diffusion and adjacent diffusion characteristics. In terms of the flow scale of the first two-level links, 1.07% of the links in the patent transfer network account for 15.71% of the flow scale of the whole network, and 1.60% of links in the talent flow network account for 22.74% of the flow scale. Knowledge flow is mainly concentrated in a few core cities, and the “Matthew effect” is prominent.
The spatial distribution of knowledge flow networks is obviously unbalanced. There are 33 groups of three-level links in the patent transfer network, accounting for 5.90% of the overall links and 29.20% of the whole scale, while 31 groups of three-level links in the talent flow network account for 8.27% of the overall links and 30.38% of the whole scale, constituting the main link lines of knowledge flow networks. The joining of semi-periphery node cities such as Wuxi, Ningbo and Changzhou have significantly expanded the scale of the network, and the two-way flow pattern between core nodes is highlighted. Overall, the link-intensive areas are mainly concentrated along the banks of the Yangtze River and around Hangzhou Bay, and the intensity of knowledge flow is gradually reduced with the continuous expansion inland and on both sides. Moreover, the network structure in the northern plain area and the southern mountainous area of the Yangtze River Delta is relatively loose. Together, the two constitute the knowledge flow pattern in the Yangtze River Delta, showing the dual characteristics of agglomeration and dispersion.

4.2. The Effect of Proximities on Knowledge Flow

Before regression analysis, QAP correlation analysis is used to test for multicollinearity between the explanatory variables, and the correlation values are presented in Table 1. In general, correlation coefficients with absolute values below 0.4 are weakly correlated, above 0.6 are strongly correlated, and those in between are moderately correlated. The correlation coefficients between explanatory variables in this study are all below 0.4, and there is no significant co-linearity.
On this basis, we choose 5000 random permutations to perform regressions of the two types of knowledge flow matrices with each proximity variable matrix separately. Model 1 introduces all the proximity variables into the equation. Models 2–7 add the squared terms of the proximity variables in turn to model 1 to test whether the inverted “U” shape of the effect of proximity on knowledge flows holds. Models 8–12 add the product terms of geographical proximity and the remaining variables in turn to model 1 to test the interaction between proximities. In terms of the fitting degree of the model, the MRQAP coefficient based on the same data is generally lower because the pseudo-regression that may be brought by the OLS algorithm is avoided. The adjusted R2 values of all models range from 0.3 to 0.6, which are statistically significant at the 0.01 level, indicating that the regression results are ideal and have good explanatory power compared with previous experience [56].
Observing the regression results of model 1 in Table 2 and Table 3, the coefficients of geographical proximity, institutional proximity and economic proximity in the patent transfer network passed the significance test, and these proximities have a strong correlation with the formation and development of the patent transfer network, while cultural proximity, industrial proximity and technological proximity did not pass the test, indicating that the similarity of culture, technology and industrial structure among cities did not show a significant effect on patent transfer. In the talent flow network, all variables passed the significance test, showing an important influence on the occurrence and development of talent flow behavior. Specifically, the coefficients of the geographical proximity of the two networks are positive and significant in models 1–12, indicating that knowledge flow tends to decrease with an increase in distance. Geographical distance is still an important factor affecting knowledge flow, and Hypotheses 1a and 2a are verified. The coefficients of institutional proximity are all significantly positive in both networks, and the coefficient of cultural proximity is significantly positive in the talent flow network, indicating that organizational proximity has a positive role in promoting knowledge flows, and Hypotheses 1b and 2b are verified. Nevertheless, industry proximity and technology proximity are significantly negative in the talent flow network, and the economic proximity coefficient is also significantly negative in both networks. Results show that similarity in technology, the industrial structure and the economic level inhibits the inter-city flow of knowledge and technology in the Yangtze River Delta. Hypotheses 1c and 2c are not verified. Overall, the different dimensions of proximity work in the same direction for both types of knowledge flows.
In addition, from the results of models 8–12, the effect of geographical proximity on institutional proximity and economic proximity in the patent transfer network has a moderating effect, and the effect on institutional proximity, industrial proximity, technological proximity and economic proximity in the talent flow network also has a significant effect, confirming the moderating effect of geographical proximity on other dimensions of proximity; thus, Hypotheses 3a and 3b are verified. Specifically, observing the sign of the product term of geographical proximity and organizational proximity, it is found that there is a substitution effect, with spatial proximity effectively increasing the possibility of knowledge spillover between actors of different organizational backgrounds; however, cities with similar organizational frameworks and spatial proximity tend to form closed systems, resulting in excessive knowledge spillover and information redundancy, which inhibits the efficiency and willingness to exchange knowledge between regions. In terms of the sign of the product term of geographical proximity and cognitive proximity, a complementary effect is found, with geographical proximity reinforcing the negative effect of cognitive proximity on knowledge flows. Given the high level of cognitive proximity in the Yangtze River Delta, cities have a higher differentiated demand for knowledge and technology from neighboring cities.

5. Discussion

5.1. Proximity Mechanisms of Knowledge Flow

Although we have confirmed the impact of proximity on patent transfer and talent flow in the Yangtze River Delta, different scholars have different views on the direction of the impact of different dimensions of proximity on knowledge flow.
Lv [5] pointed out that explicit knowledge can overcome geographical friction and spread faster and cheaper due to its coding properties. However, our research suggests that geographical proximity still can facilitate the patent transfer. Patent transfer is a kind of market trading behavior and still requires offline negotiation and contracts to ensure the smooth realization of the transaction process. Geographical proximity can effectively reduce the travel costs in offline interaction, which is also conducive to the exchange of tacit knowledge to better promote the understanding and absorption of new technologies by demanders. Therefore, geographical proximity is of great significance for explicit knowledge flow [57]. In addition, we find that the role of physical distance varies in the talent flow network at different scales. From the perspective of urban internal mobility, the average talent stickiness rate of cities in the Yangtze River Delta is as high as 81.0% compared with patents (58.7%), indicating that the solidification phenomenon is serious. Among them, the stickiness rate (Ri) reflects a region’s ability to stick to knowledge elements. The expression is Ri = Li/Ti, where Li is the number of patents or talent left in city i, and Ti is the number of original patents or talent in city i. Most talent rarely changes their work place once they settle in the right city. Physical distance has a significant blocking effect on talent flow; from the perspective of inter-city mobility, the constraint of physical distance on talent flow is not prominent. Human capital theory points out that the higher the human capital is, the smaller the flow obstacles are [58]. Therefore, the talent flow in the Yangtze River Delta has the dual characteristics of neighborhood diffusion and long-distance hierarchical diffusion.
Institutional proximity and cultural proximity, as two variables at the regional level of organizational proximity, have positive effects on patent transfer and talent flow in the Yangtze River Delta. The institutional environment of different administrative regions will be different and will create barriers to communication and affect the formation of knowledge networks [52]. Policies such as local protection and high-tech control may also hinder the cross-regional flow of knowledge to some extent. Therefore, similar institutional frameworks can effectively reduce the obstacles caused by institutional differences. Having a similar linguistic and cultural foundation can lead to a higher sense of cultural identity, less cultural friction and better communication between actors, which in turn facilitate the flow of knowledge. However, the regression results of Model 3 show that the squared term of institutional proximity is significantly negatively correlated, indicating that the curve of the effect of institutional proximity on knowledge flows is a parabola with a downward opening, i.e., above a certain threshold, institutional proximity will have a negative effect on knowledge flows. In addition, we also found significant differences in the extent to which institutional proximity acts on different types of knowledge flow. The average talent stickiness rate of each province reaches 51.7% compared with that of patents (32.7%), indicating that administrative boundary has stronger constraints on talent flow. We argue that for talent, inter-provincial mobility not only raises the quantifiable direct costs of transport, communication and information searches but also increases the invisible costs incurred by giving up their original social network ties, thus discouraging their willingness to move across provinces. In contrast, the standardization of and improvement in patent information centers and transfer trading platforms reduce the cost of information searches and the risk of transactions to some extent, making the resistance of provincial boundaries to patent transfer less prominent.
Regarding the explanation for the negative effect of cognitive proximity on knowledge flow in the Yangtze River Delta, we think that, on one hand, it is due to the inhibitory effect caused by excessive cognitive proximity. Observing the results of models 2–7, the squared term of economic proximity is significantly negatively correlated, indicating that the effect of cognitive proximity on knowledge flows exhibits an inverted “U” shape. In addition, the average values of the industrial and technological structure similarity coefficient and similarity level of economic development are 0.883, 0.948 and 2.55, respectively. For the optimal cognitive proximity value (0.5) proposed by Grossman [59], the study suggests that the current degree of homogenization of industrial, technological and economic development levels among cities in the Yangtze River Delta region is too high. The convergence of perceptions has led to cities having similar needs in terms of knowledge and technology, and local protectionism has developed among actors for their own development, thus limiting the cross-regional flow of knowledge and technology resources. Romero [60] also points out that too little cognitive distance may reduce the heterogeneity of the knowledge base and the possibility of cooperation and sharing between actors. On the other hand, an appropriate knowledge gap is also the basis for realizing knowledge diffusion. Traditional technology gap theory argues that regional technology gaps and economic gaps are the motivation for technology transfer, and the existence of gaps creates conditions for knowledge flow, enabling knowledge diffusion from high potential energy subjects to low potential energy subjects [12]. The knowledge flow network in the Yangtze River Delta presents the characteristics of core cities as the center and follows the gradient law of knowledge transfer, confirming the conclusions above. However, this does not mean that cognitive proximity is insignificant. When the cognitive proximity between actors is lower due to the enormous differences in professional fields, there will be obstacles in communication [61]. Successful cooperation still requires cognitive proximity.
Furthermore, we find that there is a moderating effect between geographic proximity and non-geographic proximity, with a substitution effect between geographical proximity and organizational proximity, and a complementary effect with cognitive proximity. Specifically, organizational proximity can weaken the dependence of knowledge elements on physical distance to some extent, i.e., the transfer of knowledge and technology can be achieved through organizational proximity even if actors are geographically distant, and the lack of organizational proximity can also be compensated for by geographic proximity. Geographical proximity can catalyze and reinforce the effects of cognitive proximity, allowing actors with similar cognitive levels to deepen their ties through frequent face-to-face exchanges. But this also means that the moderating effect of geographical proximity can only be activated if actors have a relatively similar technical and economic base; if there are large knowledge gaps and technical barriers between them, it is difficult to generate sparks even if face-to-face exchanges are offered.

5.2. Policy Implications

The role of knowledge flow, combination and reproduction in the innovation process has become increasingly prominent, and the findings of this paper provide the following implications for optimizing regional knowledge flow pattern and promoting the sustainable development in the Yangtze River Delta. Firstly, most cities have an obvious spatial tendency in establishing knowledge connections. That is, they choose cities that are geographically adjacent or within the same administrative region. Thus, breaking regional and administrative barriers is an effective way to promote knowledge exchange. On the one hand, it is necessary to speed up the construction of “hard” infrastructure to improve regional accessibility and reduce the time and economic costs of cooperation. On the other hand, it is also necessary for governments to narrow the differences in the institutional design of industrial development, transportation strategies and human resources to accelerate the construction of “soft” infrastructure, such as institutional guarantees. Secondly, excessive cognitive proximity inhibiting knowledge flow provides new insights for regional development. That is, the differentiated development path is the internal demand of high-quality integrated development for the Yangtze River Delta. Local governments should clarify the positioning of cities in regional development, take guiding and encouraging measures, give full play to local comparative advantages, strengthen the regional division of labor and cooperation and avoid homogeneous competition in industry and technology, to obtain regional sustainable competitive advantages. Thirdly, there are obvious institutional obstacles in the talent flow in the Yangtze River Delta. Thus, it is necessary to build formal or informal platforms such as regional innovation alliances to promote the exchange of talent and enhance the transmission, sharing and spillover of tacit knowledge. In addition, the establishment of a cross-regional linkage mechanism for talent training, introduction and sharing is an effective measure to achieve the complementary advantages of human resources. Fourthly, the core of knowledge flow lies in the exchange and sharing of innovation resources. Thus, it is necessary to avoid the obstacles caused by an excessive polarization of knowledge elements in the pure market economy while revitalizing innovation elements.

6. Conclusions

6.1. Main Conclusions

Based on patent transfer data and talent flow data, this paper distinguishes between explicit and implicit knowledge, describes the spatial characteristics of flow networks in the Yangtze River Delta, and uses the MRQAP model to analyze proximity mechanism. The main findings can be summarized as follows: (1) Knowledge flow networks in the Yangtze River Delta cover a wide range but are extremely uneven. The patent transfer network spatially forms a radial structure, with Shanghai as the divergent core and with spread towards neighboring cities. The talent flow network presents a Z-shaped structure, with Shanghai as the core along Shanghai–Nanjing, Shanghai–Hangzhou and Hangzhou–Taizhou in space. (2) Different dimensions of proximity have different effects on knowledge flow. Geographical and organizational proximity have significant positive effects, while cognitive proximity has a negative impact. The effect of institutional proximity on different types of knowledge flow is significantly different, and the administrative boundary is more restrictive for talent flow. Moreover, there is a substitution effect between geographical proximity and organizational proximity, and a complementary effect with cognitive proximity.

6.2. Limitations and Future Work

Although this paper analyzes the spatial characteristics of knowledge flow and provides an analytical framework for exploring its proximity mechanism, some limitations of this study should be acknowledged. First, there are various storage forms and flow media for knowledge, including language, text, pictures, data, the human brain, etc. However, the paper only selects patents and talents to characterize explicit and tacit knowledge, which only reflects a part of the actual situation of regional innovation. Empirical research integrating multidimensional indicators and data is also needed to present a more comprehensive and systematic panorama of knowledge flow. Second, knowledge flow is a dynamic process, and the influence of different dimensions of proximity on knowledge flow may change with the development of the network. In future research, the evolution of knowledge flow networks and changes in proximity mechanisms can be studied based on continuous time data. Moreover, due to the applicability and limitations of the methodology, this study only explores the impact of proximity factors on knowledge flow. Future research needs to adopt other appropriate techniques to explore the impact of the local innovation base and other external environments. Research on such issues may obtain interesting findings.

Author Contributions

Conceptualization, P.Z. and J.C.; methodology, W.L.; software, W.L.; validation, P.Z., F.Y. and W.L.; formal analysis, P.Z.; investigation, P.Z.; resources, P.Z.; data curation, P.Z. and W.L.; writing—original draft preparation, P.Z. and F.Y.; writing—review and editing, P.Z., J.C., F.Y. and W.L.; visualization, P.Z. and W.L.; supervision, J.C.; project administration, J.C. and P.Z.; funding acquisition, J.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China (No. 42101177) and the Science and Technology Service Network Initiative (STS) Project of the Chinese Academy of Sciences (KFJ-STS-QYZD-202).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are contained within the article.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Sun, Y.; Wang, T.; Gu, X. A Sustainable Development Perspective on Cooperative Culture, Knowledge Flow, and Innovation Network Governance Performance. Sustainability 2019, 11, 6126. [Google Scholar] [CrossRef]
  2. Bathelt, H.; Malmberg, A.; Maskell, P. Clusters and knowledge: Local buzz, global pipelines and the process of knowledge creation. Prog. Hum. Geogr. 2004, 28, 31–56. [Google Scholar] [CrossRef]
  3. Sun, Y.; Du, D. Determinants of industrial innovation in China: Evidence from its recent economic census. Technovation 2010, 30, 540–550. [Google Scholar] [CrossRef]
  4. Asheim, B. The Changing Role of Learning Regions in the Globalizing Knowledge Economy: A Theoretical Re-examination. Reg. Stud. 2012, 46, 993–1004. [Google Scholar] [CrossRef]
  5. Lv, L.C. The Geography of Innovation; Science Press: Beijing, China, 2017. [Google Scholar]
  6. Rowley, J. The wisdom hierarchy: Representations of the DIKW hierarchy. J. Inf. Sci. 2007, 33, 163–180. [Google Scholar] [CrossRef]
  7. Dong, K.; Xu, H.Y.; Cui, B. A review of knowledge flow research. J. China Soc. Sci. Tech. Inf. 2020, 39, 1120–1132. [Google Scholar] [CrossRef]
  8. Argote, L.; Mcevily, B.; Reagans, R. Managing knowledge in organizations: An integrative framework and review of emerging themes. Manag. Sci. 2003, 49, 571–582. [Google Scholar] [CrossRef]
  9. Amidon, D.M. Dialogue with Customers: Secret to Innovation Strategy. Int. J. Innov. Manag. 1997, 1, 73–87. [Google Scholar] [CrossRef]
  10. Spender, J.C. Making knowledge the basis of a dynamic theory of the firm. Strat. Manag. J. 1996, 17, 45–62. [Google Scholar] [CrossRef]
  11. Nonaka, I.; Takeuchi, H. The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation; Oxford University Press: New York, NY, USA, 1995. [Google Scholar]
  12. Liu, C.L.; Niu, C.C. Spatial evolution and factors of interurban technology transfer network in Northeast China from national to local perspectives. Acta Geogr. Sin. 2019, 74, 2092–2107. [Google Scholar] [CrossRef]
  13. Ibrahim, S.E.; Fallah, M.H.; Reilly, R.R. Localized sources of knowledge and the effect of knowledge spillovers: An empirical study of inventors in the telecommunications industry. J. Econ. Geogr. 2008, 9, 405–431. [Google Scholar] [CrossRef]
  14. Sun, Y.; Latora, V. The evolution of knowledge within and across fields in modern physics. Sci. Rep. 2020, 10, 12097. [Google Scholar] [CrossRef] [PubMed]
  15. Behrend, F.D.; Erwee, R. Mapping knowledge flows in virtual teams with SNA. Prog. Hum. Geogr. 2009, 13, 99–114. [Google Scholar] [CrossRef]
  16. Shi, Y.; Zhai, W.; Yan, Y.; Wang, X. The Influence of Proximity on the Evolution of Urban Innovation Networks in Nanjing Metropolitan Area, China: A Comparative Analysis of Knowledge and Technological Innovations. ISPRS Int. J. Geo-Inf. 2024, 13, 273. [Google Scholar] [CrossRef]
  17. Szulanski, G. Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strat. Manag. J. 1996, 17, 27–43. [Google Scholar] [CrossRef]
  18. Kang, X.; Liu, D. Transmission mechanism of knowledge diffusion path to knowledge evolution in agricultural enterprises: Based on intermediary role of knowledge sharing and regulating role of knowledge. Sci. Technol. Manag. Res. 2018, 38, 184–190. [Google Scholar] [CrossRef]
  19. Teece, D.J. Technology transfer by multinational firms: The resource cost of transferring technological know-how. Econ. J. 1977, 87, 242–261. [Google Scholar] [CrossRef]
  20. Bartol, K.M.; Srivastava, A. Encouraging knowledge sharing: The role of organizational reward systems. J. Leadersh. Organ. Stud. 2002, 9, 64–76. [Google Scholar] [CrossRef]
  21. Zhu, Y.W.; Zeng, G.; Cao, X.Z.; Si, Y.F. Research progress of innovation network and knowledge flow. World Reg. Stud. 2017, 26, 117–125. [Google Scholar] [CrossRef]
  22. Ma, H.; Fang, C.; Pang, B.; Wang, S. Structure of Chinese city network as driven by technological knowledge flows. Chin. Geogr. Sci. 2015, 25, 498–510. [Google Scholar] [CrossRef]
  23. Duan, D.; Du, D.; Chen, Y.; Zhai, Q. Spatial-temporal complexity and growth mechanism of city innovation network in China. Sci. Geogr. Sin. 2018, 38, 1759–1768. [Google Scholar] [CrossRef]
  24. Marques, C.S.; Leal, C.; Marques, C.P.; Cardoso, A.R. Strategic knowledge management, innovation and performance: A qualitative study of the footwear industry. J. Knowl. Econ. 2016, 7, 659–675. [Google Scholar] [CrossRef]
  25. Hu, Z.; Cao, Z.; Du, A.; Tu, Q. Where does it matter? Revisiting the role of proximity in knowledge spillovers. Am. J. Econ. Sociol. 2024, 1–26. [Google Scholar] [CrossRef]
  26. Capello, R.; Caragliu, A. Proximities and the intensity of scientific relations: Synergies and nonlinearities. Int. Reg. Sci. Rev. 2016, 41, 7–44. [Google Scholar] [CrossRef]
  27. Malerba, F.; Mancusi, M.L.; Montobbio, F. Innovation, international R&D spillovers and the sectoral heterogeneity of knowledge flows. Rev. World Econ. 2013, 149, 697–722. [Google Scholar] [CrossRef]
  28. Zhi, Z.; Liu, J.; Liu, J.; Ai, L. Geospatial Structure and Evolution Analysis of National Terrestrial Adjacency Network Based on Complex Network. J. Geovis. Spat. Anal. 2024, 8, 12. [Google Scholar] [CrossRef]
  29. Zhang, Y.O.; Gu, R.X.; Ma, S. Spatial characteristics and proximity mechanism of technology transfer among cities in China. Prog. Geogr. 2019, 38, 370–382. [Google Scholar] [CrossRef]
  30. Polanyi, M. The Tacit Dimension; The University of Chicago Press: Chicago, IL, USA, 1966. [Google Scholar]
  31. Howells, J.R.L. Tacit knowledge, innovation and economic geography. Urban. Stud. 2002, 39, 871–884. [Google Scholar] [CrossRef]
  32. Gertler, M.S. Tacit knowledge and the economic geography of context, or the undefinable tacitness of being (there). J. Econ. Geogr. 2003, 3, 75–99. [Google Scholar] [CrossRef]
  33. Boschma, R. Proximity and innovation: A critical assessment. Reg. Stud. 2005, 39, 61–74. [Google Scholar] [CrossRef]
  34. Capello, R. Spatial spillovers and regional growth: A cognitive approach. Eur. Plan. Stud. 2009, 17, 639–658. [Google Scholar] [CrossRef]
  35. Crescenzi, R.; Nathan, M.; Rodríguez-Pose, A. Do inventors talk to strangers? On proximity and collaborative knowledge creation. Res. Policy 2016, 45, 177–194. [Google Scholar] [CrossRef]
  36. Shaw, A.T.; Gilly, J. On the analytical dimension of proximity dynamics. Reg. Stud. 2000, 34, 169–180. [Google Scholar] [CrossRef]
  37. Oerlemans, L.; Meeus, M. Do organizational and spatial proximity impact on firm performance? Reg. Stud. 2005, 39, 89–104. [Google Scholar] [CrossRef]
  38. Torre, A.; Rallet, A. Proximity and localization. Reg. Stud. 2005, 39, 47–59. [Google Scholar] [CrossRef]
  39. Breschi, S.; Lissoni, F. Mobility of skilled workers and co-invention networks: An anatomy of localized knowledge flows. J. Econ. Geogr. 2009, 9, 439–468. [Google Scholar] [CrossRef]
  40. Tödtling, F.; Trippl, M. One size fits all? Towards a differentiated regional innovation policy approach. Res. Policy 2005, 34, 1203–1219. [Google Scholar] [CrossRef]
  41. Caragliu, A.; Nijkamp, P. Space and knowledge spillovers in European regions: The impact of different forms of proximity on spatial knowledge diffusion. J. Econ. Geogr. 2015, 16, 749–774. [Google Scholar] [CrossRef]
  42. Granovetter, M. Economic Action and Social Structure: The Problem of Embeddedness. Am. J. Sociol. 1985, 91, 481–510. [Google Scholar] [CrossRef]
  43. Miörner, J.; Zukauskaite, E.; Trippl, M.; Moodysson, J. Creating institutional preconditions for knowledge flows in cross-border regions. Environ. Plan. C Polit. Space 2017, 36, 201–218. [Google Scholar] [CrossRef]
  44. Vasin, S.M.; Gamidullaeva, L.A.; Wise, N.; Korolev, K.Y. Knowledge Exchange and the Trust Institution: A New Look at the Problem. J. Knowl. Econ. 2020, 11, 1026–1042. [Google Scholar] [CrossRef]
  45. Brown, J.S.; Paul, D. Organizational learning and communities-of-practice: Toward a unified view of working, learning, and innovation. Organ. Sci. 1991, 2, 40–57. [Google Scholar] [CrossRef]
  46. Kiiski, S.; Pohjola, M. Cross-country diffusion of the Internet. Inf. Econ. Policy 2002, 14, 297–310. [Google Scholar] [CrossRef]
  47. Mancusi, M.L. International spillovers and absorptive capacity: A cross-country cross-sector analysis based on patents and citations. J. Int. Econ. 2008, 76, 155–165. [Google Scholar] [CrossRef]
  48. Castells, M. The Information City; Oxford: Basil Blackwell, UK, 1989. [Google Scholar]
  49. Maggioni, M.A.; Uberti, T.E.; Nosvelli, M. Does intentional mean hierarchical? Knowledge flows and innovative performance of European regions. Ann. Reg. Sci. 2014, 53, 453–485. [Google Scholar] [CrossRef]
  50. Saxenian, A. The New Argonauts: Regional Advantage in a Global Economy Cambridge; Harvard University Press: Cambridge, MA, USA, 2006. [Google Scholar]
  51. Dekker, D.; Krackhardt, D.; Snijders, T.A.B. Sensitivity of MRQAP tests to collinearity and autocorrelation conditions. Psychometrika 2007, 72, 563–581. [Google Scholar] [CrossRef] [PubMed]
  52. Si, Y.F.; Sun, K.; Zhu, Y.W.; Cao, X.Z. Spatial structure and influencing factors of knowledge network of highly cited Chinese scientists. Geogr. Res. 2020, 39, 2731–2742. [Google Scholar] [CrossRef]
  53. Hegde, D.; Tumlinson, J. Does Social Proximity Enhance Business Partnerships? Theory and Evidence from Ethnicity’s Role in U.S. Venture Capital. Manag. Sci. 2014, 60, 2355–2380. [Google Scholar] [CrossRef]
  54. Jaffe, A.B. Technological opportunity and spillovers of R&D: Evidence from firms’ patents, profits, and market value. Am. Econ. Rev. 1986, 76, 984–1001. [Google Scholar]
  55. Limtanakool, N.; Dijst, M.; Schwanen, T. A theoretical framework and methodology for characterising national urban systems on the basis of flows of people: Empirical evidence for France and Germany. Urban Stud. 2007, 44, 2123–2145. [Google Scholar] [CrossRef]
  56. Qin, Q.; Wu, L.; Li, F.; Cheng, S.K.; Zhang, D.; Cheng, X.P. A social-network-based study on geo-relations in Southeast Asia. Acta Geogr. Sinica. 2018, 73, 2014–2030. [Google Scholar] [CrossRef]
  57. Lyu, K.; Cai, D.; Hao, M. Dynamic Innovation Collaboration Based on Complex Network Analysis: Evidence from the “Belt and Road” Initiative. J. Knowl. Econ. 2024. [Google Scholar] [CrossRef]
  58. Faggian, A.; McCann, P. Human capital, graduate migration and innovation in British regions. Cambr. J. Econ. 2008, 33, 317–333. [Google Scholar] [CrossRef]
  59. Grossman, E.B.; Yli-Renko, H.; Janakiraman, R. Resource search, interpersonal similarity, and network tie valuation in nascent entrepreneurs’ emerging networks. J. Manag. 2012, 38, 1760–1787. [Google Scholar] [CrossRef]
  60. Romero, C.C. Personal and business networks within Chilean biotech. Ind. Innov. 2018, 25, 841–873. [Google Scholar] [CrossRef]
  61. Balland, P.A. Proximity and the evolution of collaboration networks: Evidence from research and development projects within the global navigation satellite system (GNSS) industry. Reg. Stud. 2012, 46, 741–756. [Google Scholar] [CrossRef]
Sustainability 17 00740 g001
Figure 1. Impact of proximity on knowledge flow: An analytical framework.
Figure 1. Impact of proximity on knowledge flow: An analytical framework.
Sustainability 17 00740 g001
Sustainability 17 00740 g002
Figure 2. Study area: (a) Location of the Yangtze River Delta in China; (b) Administrative boundaries of the Yangtze River Delta.
Figure 2. Study area: (a) Location of the Yangtze River Delta in China; (b) Administrative boundaries of the Yangtze River Delta.
Sustainability 17 00740 g002
Sustainability 17 00740 g003
Figure 3. Different dimensions of proximity in the Yangtze River Delta: (a) geographical proximity; (b) institutional proximity; (c) cultural proximity; (d) technological proximity; (e) industrial proximity; (f) economic proximity.
Figure 3. Different dimensions of proximity in the Yangtze River Delta: (a) geographical proximity; (b) institutional proximity; (c) cultural proximity; (d) technological proximity; (e) industrial proximity; (f) economic proximity.
Sustainability 17 00740 g003
Sustainability 17 00740 g004
Figure 4. Knowledge flow network in the Yangtze River Delta: (a) patent transfer; (b) talent flow.
Figure 4. Knowledge flow network in the Yangtze River Delta: (a) patent transfer; (b) talent flow.
Sustainability 17 00740 g004
Table 1. Results of correlation analysis.
Table 1. Results of correlation analysis.
VariablesGeographical
Proximity		Institutional
Proximity		Cultural
Proximity		Industrial
Proximity		Technological
Proximity		Economic
Proximity
Geographical
proximity		1.000 ***
Institutional
proximity		0.357 ***1.000 ***
Cultural
proximity		0.232 ***0.256 ***1.000 ***
Industrial
proximity		0.095 ***0.171 ***0.220 ***1.000 ***
Technological
proximity		−0.0330.044−0.0020.277 ***1.000 ***
Economic
proximity		0.090 ***0.256 ***0.037 ***0.245 ***0.096 *1.000 ***
Note: *** Significant at the 1% level; * significant at the 10% level.
Table 2. Results of regression (patent transfer network).
Table 2. Results of regression (patent transfer network).
VariablesModel 1Model 2Model 3Model 4Model 5Model 6Model 7Model 8Model 9Model 10Model 11Model 12
Geo_Prox0.083 ***0.083 ***0.084 ***0.086 ***0.083 ***0.083 ***0.083 ***0.083 ***0.083 ***0.083 ***0.083 ***0.083 ***
Inst_Prox0.069 **0.068 **0.074 **0.069 **0.069 **0.070 **0.066 **0.075 **0.068 **0.068 **0.068 **0.072 **
Cul_Prox0.0560.0560.0570.0380.0560.0550.0520.0580.0580.0580.0560.058
Ind_Prox0.0740.0750.0740.0740.0740.0720.0820.0710.0730.0780.0730.060
Tech_Prox0.0310.0310.0300.0370.0310.0240.0390.0310.0310.0340.0350.045
Eco_Prox−0.265 ***−0.265 ***−0.263 ***−0.263 ***−0.265 ***−0.266 ***−0.203 ***−0.261 ***−0.264 ***−0.262 ***−0.263 ***−0.215 ***
Geo_Prox squared 0.014
Inst_Prox squared −0.044 *
Cul_Prox squared 0.060
Ind_Prox squared 0.000
Tech_Prox squared −0.023
Eco_Prox squared −0.191 ***
Geo_Prox × Inst_Prox −0.049 *
Geo_Prox × Cul_Prox −0.008
Geo_Prox × Ind_Prox −0.029
Geo_Prox × Tech_Prox −0.015
Geo_Prox × Eco_Prox −0.161 ***
Sample size164016401640164016401640164016401640164016401640
Adjusted R−square0.329 ***0.328 ***0.335 ***0.340 ***0.329 ***0.328 ***0.417 ***0.335 ***0.327 ***0.334 ***0.328 ***0.393 ***
Note: *** Significant at the 1% level; ** significant at the 5% level; * significant at the 10% level.
Table 3. Results of regression (talent flow network).
Table 3. Results of regression (talent flow network).
VariablesModel 1Model 2Model 3Model 4Model 5Model 6Model 7Model 8Model 9Model 10Model 11Model 12
Geo_Prox0.070 **0.071 **0.071 **0.073 **0.071 **0.071 **0.070 **0.069 **0.068 **0.070 **0.070 **0.070 **
Inst_Prox0.220 ***0.221 ***0.227 ***0.220 ***0.219 ***0.218 ***0.217 ***0.232 ***0.218 ***0.216 ***0.216 ***0.223 ***
Cul_Prox0.119 **0.120 **0.120 **0.105 **0.121 **0.120 **0.115 ***0.123 **0.133 ***0.123 **0.118 **0.120 **
Ind_Prox−0.096 *−0.097 *−0.095 *−0.095 *−0.089 *−0.094 *−0.088 *−0.101 *−0.101 *−0.084 *−0.100 *−0.111 **
Tech_Prox−0.097 *−0.098 *−0.099 *−0.092 *−0.097 *−0.090 *−0.090 *−0.096 *−0.098 *−0.090 **−0.074−0.081 *
Eco_Prox−0.196 ***−0.195 ***−0.194 ***−0.195 ***−0.195 ***−0.194 ***−0.140 ***−0.190 ***−0.189 ***−0.188 ***−0.188 ***−0.141 ***
Geo_Prox squared −0.024
Inst_Prox squared −0.063 **
Cul_Prox squared 0.046
Ind_Prox squared −0.027
Tech_Prox squared −0.024
Eco_Prox squared −0.173 ***
Geo_Prox × Inst_Prox −0.097 ***
Geo_Prox × Cul_Prox −0.058
Geo_Prox × Ind_Prox −0.081 **
Geo_Prox × Tech_Prox −0.078 **
Geo_Prox × Eco_Prox −0.179 ***
Sample size164016401640164016401640164016401640164016401640
Adjusted R−square0.413 ***0.413 ***0.430 ***0.431 ***0.412 ***0.413 ***0.519 ***0.440 ***0.425 ***0.437 ***0.430 ***0.518 ***
Note: *** Significant at the 1% level; ** significant at the 5% level; * significant at the 10% level.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhu, P.; Chen, J.; Yuan, F.; Liu, W. The Formation of Knowledge Flow Networks in the Yangtze River Delta, China: Knowledge Implicitness and Proximity Effect. Sustainability 2025, 17, 740. https://doi.org/10.3390/su17020740

AMA Style

Zhu P, Chen J, Yuan F, Liu W. The Formation of Knowledge Flow Networks in the Yangtze River Delta, China: Knowledge Implicitness and Proximity Effect. Sustainability. 2025; 17(2):740. https://doi.org/10.3390/su17020740

Chicago/Turabian Style

Zhu, Pengcheng, Jianglong Chen, Feng Yuan, and Weichen Liu. 2025. "The Formation of Knowledge Flow Networks in the Yangtze River Delta, China: Knowledge Implicitness and Proximity Effect" Sustainability 17, no. 2: 740. https://doi.org/10.3390/su17020740

APA Style

Zhu, P., Chen, J., Yuan, F., & Liu, W. (2025). The Formation of Knowledge Flow Networks in the Yangtze River Delta, China: Knowledge Implicitness and Proximity Effect. Sustainability, 17(2), 740. https://doi.org/10.3390/su17020740

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop