The Spatio-Temporal Evolution and Influence Mechanisms of Intercity Cooperation Networks from the Perspective of Sustainable Regional Development: A Case Study of the Pearl River–Xijiang Economic Belt, China

Abstract

Intercity cooperation networks are critical for addressing regional imbalances and advancing sustainable regional development, yet existing studies typically focus on specific functional domains, rather than the overall intercity cooperation network. To bridge this gap, this study examines the intercity cooperation network in the Pearl River–Xijiang Economic Belt (21 cities, 2014–2023), analyzing its spatio-temporal evolution and influence mechanisms through Social Network Analysis (SNA) and partial least squares structural equation modeling (PLS-SEM). The results reveal the following: (1) the network has undergone three policy-driven development stages: initial–accelerated–steady; (2) a spatial pattern of “east—dominant, west—weak” has emerged, shaped by the radiating influence of core cities; and (3) institutional proximity and cooperation investment are key drivers of network formation, while geographical and organizational proximity exhibit negative impacts. These findings underscore the need for related regional development strategies to foster a more vital and open cooperation network. Overall, this study deepens the understanding of intercity cooperation by revealing its macro-level patterns and influence mechanisms, and provides practical implications for policymakers committed to promoting sustainable regional development.

1. Introduction

Globalization has significantly altered the roles and inter-relationships of nations, regions, and cities [1]. As this process reshapes social, economic, and political structures, city-regions have gained increasing importance [2]. To respond to this shift, many countries—guided by the New Public Management movement—have decentralized regulatory powers to the regional scale to enhance national competitiveness [3].

In contrast, China has implemented decentralization reforms to restructure central–local power relations since the late 20th century [4]. However, a top-down, government-led governance model still predominates at the regional scale, characterized by the coexistence of political centralization and economic decentralization [5,6]. This “Chinese-style federalism” has intensified horizontal competition among local governments, which, in turn, has resulted in fragmented and uncoordinated regional development [7]. Specific manifestations include growing cross-provincial economic disparities and a widening urban–rural gap [8], both of which undermine the goals of sustainable regional development. Moreover, as cross-boundary public issues increasingly emerge, individual local governments are unable to resolve them independently [9]. Similar governance challenges resulting from administrative fragmentation under decentralization have also been observed in other countries [10].

To address increasingly complex governance and development challenges, China introduced a coordinated regional development strategy in 2018, emphasizing the need to deepen regional cooperation mechanisms and strengthen exchange and cooperation between cities. Related studies also indicate that, in response to administrative fragmentation and regional development imbalances, China is pursuing a more cooperative regional governance mechanism [11,12].

In this context, intercity cooperation has emerged as a key approach to narrowing regional disparities and promoting sustainable development [13,14,15], attracting wide attention from both policymakers and scholars. Furthermore, in the face of governance challenges caused by administrative fragmentation, intercity cooperation offers a viable governance mechanism by integrating cross-boundary stakeholders, thereby facilitating more coordinated and sustainable social, economic, and environmental development [16,17,18,19].

Under this governance mechanism, intercity cooperation has become a widespread phenomenon, where political and administrative hierarchical structures are increasingly open to horizontal networks, both inter-sectoral and intra-sectoral [20]. By embedding local governments with citizen representatives, experts, public associations, and NGOs into collective governmental actions, a network governance mechanism for intercity cooperation is established [21]. A typical example of this network governance in action is the California Bay Delta in 1994, where over 25 federal and state agencies, along with more than 30 stakeholder and local groups, adopted a networked cooperative governance mechanism to address complex water resource management challenges [22].

In summary, analyzing intercity cooperation networks among local governments is essential for addressing widening regional disparities and development inequalities exacerbated by globalization. Such analysis facilitates more effective governance of cross-boundary public issues and contributes to sustainable regional development [23,24]. Although significant progress has been made in the theoretical and empirical study of cooperation networks, several important limitations remain: (1) Most existing studies focus on specific functional domains, such as the intercity cooperation network in environmental governance and public services [25,26,27], which limits the understanding of the overall structure and evolutionary characteristics of intercity cooperation networks. (2) Research attention has largely concentrated on coastal regions or well-developed city-regions, such as the Beijing–Tianjin–Hebei area, the Yangtze River Delta, and the Guangdong–Hong Kong–Macao Greater Bay Area [28,29,30], while the intercity cooperation networks in less developed and inland regions remain relatively underexamined. (3) Analyses of the influence mechanisms of intercity cooperation networks have largely focused on endogenous factors (e.g., cities’ economic level) [31,32], while exogenous factors have received comparatively less attention.

To address these research gaps, this study takes the Pearl River–Xijiang Economic Belt as the study area. Connecting the highly developed Pearl River Delta with less developed inland regions, the Pearl River–Xijiang Economic Belt exemplifies the economic disparities, administrative fragmentation, and spatial coordination challenges that are characteristic of regional development in the globalization era. As a representative case, it offers significant research value for exploring ways to narrow regional disparities and promote sustainable regional development in less developed regions. In terms of the factors influencing the intercity cooperation network, this study adopts the theoretical perspective of multidimensional proximity [33], and draws upon analytical frameworks that integrate proximity theory with network theory [34]. Based on Boschma’s widely adopted five-dimensional proximity framework [35], which has been further developed in subsequent studies [36,37], this study focuses on three dimensions (geographic proximity, organizational proximity, and institutional proximity) by considering the measurability and data availability of cities in the Pearl River–Xijiang Economic Belt.

Therefore, the main aim of this study is to investigate the network characteristics and influence mechanisms of the overall intercity cooperation network in the Pearl River–Xijiang Economic Belt. This contributes to a deeper understanding of the evolution of city-regions, and provides valuable policy implications for policymakers in promoting regional coordination and achieving sustainable development. The structure of this paper is organized as follows. Section 2 introduces the study area, methods, and data collection process. Section 3 analyzes the spatio-temporal evolution of the intercity cooperation network in the Pearl River–Xijiang Economic Belt and examines the influence mechanism of multidimensional proximity. Section 4 discusses the research findings and provides relevant policy implications. Section 5 summarizes the key conclusions, highlights the main contributions, acknowledges limitations, and suggests directions for future research.

2. Materials and Methods

2.1. Study Area

The State Council’s approval of the Development Plan for the Pearl River–Xijiang Economic Belt in 2014 marked a strategic shift towards basin-wide cooperation as a regional development approach. The plan envisions the Economic Belt as a support zone for open development in the southwest and central-south regions, a pilot for the Eastern–Western Cooperation, an experimental area for ecological civilization construction, and a vanguard for the Maritime Silk Road. Given these diverse roles and the coexistence of regional economic disparities and developmental opportunities, the region offers a valuable case for examining how intercity cooperation can contribute to sustainable regional development.

This study takes the Pearl River–Xijiang Economic Belt as the study area. Based on the Development Plan, it includes 21 cities: Guangzhou, Foshan, Zhaoqing, Yunfu, Nanning, Liuzhou, Guilin, Wuzhou, Guigang, Yulin, Baise, Hezhou, Hechi, Laibin, Chongzuo, Qianxinan, Qiandongnan, Qiannan, Anshun, Wenshan, and Qujing. A cooperation network is constructed using these 21 cities as nodes, with intercity cooperation links serving as edges.

2.2. Method

2.2.1. Social Network Analysis

Given its focus on cooperative relationships between local governments, this study combines Social Network Analysis (SNA) with GIS spatial analysis to quantitatively analyze the intercity cooperation network. As a core concept in Social Network Analysis, the formula for calculating the degree of centrality is as follows:

$$C_D\left(n_i\right)=\sum _{j=1}^n{W}_{ij}$$

(1)

where $W_{ij}$ represents the weight of the edge between node i and node j (which, in this study, refers to the number of intercity cooperations between the two cities), and n represents the total number of nodes in the network.

In recent decades, scholars have extensively applied Social Network Analysis to examine network structures and their key characteristics [38,39,40]. To analyze the structural characteristics of the intercity cooperation network in the Pearl River–Xijiang Economic Belt, this study adopts the analytical framework proposed by Watt [41,42], focusing on three dimensions: transmission, clustering, and cohesive subgroups.

Specifically, this study measures the transmission and clustering of the cooperation network based on the average path length and clustering coefficient, respectively. The CONCOR algorithm in UCINET is then employed to identify cohesive subgroups, aiming to uncover the internal structure within the network, as well as the intricate interactional relationships between cities.

2.2.2. Structural Equation Modeling

To identify the factors influencing the intercity cooperation network, structural equation modeling (SEM) is employed. SEM accommodates simultaneous equations with multiple endogenous variables, allows for multiple indicators of latent constructs, and estimates reliability and validity. Unlike most econometric methods, SEM accounts for measurement error in both exogenous and endogenous variables. Moreover, SEM enables more general measurement models than traditional factor analytic structures, and enables researchers to specify structural relationships among latent variables [43]. Considering the limited number of sample parameters, this study applies partial least squares–SEM (PLS-SEM) using SmartPLS [44].

Specifically, SEM includes two models, the measurement model and the structural model.

• Measurement Model

The measurement model delineates how latent variables are measured. The formula for the model is as follows:

$${\eta }_i=\sum _{i=1}{w}_{ij}{x}_{ij}+{\delta }_i$$

(2)

where ${\eta }_i$ signifies the latent variable. Within the scope of this paper, it pertains to network characteristics, geographical proximity, organizational proximity, institutional proximity, and cooperation investment. $x_{ij}$ denotes the observed variable that constitutes the latent variable. $w_{ij}$ represents the correlation between the observed variable and the latent variable. ${\delta }_i$ indicates the error term associated with the observed variable.

• Structural Model

The structural model elucidates the interconnections among latent variables. This model encompasses exogenous latent variables, which act exclusively as independent variables, and endogenous latent variables, which may function as dependent variables or serve in both capacities. The formula for the model is as follows:

$${\eta }_i=\sum _{i\ne j}{\beta }_{ij}{\eta }_j+\zeta$$

(3)

where ${\eta }_i$ signifies the endogenous latent variable. In this paper, endogenous latent variables refer to network characteristics. ${\eta }_j$ denotes the exogenous latent variables that influence endogenous latent variables, including geographic proximity, organizational proximity, institutional proximity, and cooperation investment. ${\beta }_{ij}$ represents the correlation between the endogenous latent variable and the exogenous latent variables. $\zeta$ indicates the error term associated with the relationship between the latent variables.

With reference to Kaihuang Zhang’s method [45], the observed variables for network characteristics include the degree of centrality and the average effective size of structural holes, while the observed variable for cooperation investment is derived from intercity cooperation data. In addition to these two variables, three proximity variables are introduced in the following section.

• Geographical Proximity

Geographical proximity can be measured in terms of both absolute and relative distance, such as the physical distance between two entities in geographical space [46]. This study quantifies geographical proximity using standardized shortest highway distances between cities, following existing studies [47]. The formula is as follows:

$${\mathrm{G}\mathrm{e}\mathrm{o}}_{ij}=1-\mathrm{l}\mathrm{n}(d_{ij}/\mathrm{m}\mathrm{a}\mathrm{x}{d}_{ij})$$

(4)

where $d_{ij}$ signifies the highway distance from city i to city j. $\max d_{ij}$ denotes the maximum road distance among all subjects in the study.

• Organizational Proximity

Organizational proximity refers to the closeness between organizational members arising from shared behavioral practices [46]. Drawing on Balland’s research [48], this study represents organizational proximity between different subjects using a dummy variable based on provincial boundary effects. Specifically, if cities i and j in cooperation are within the same province, then Org_ij = 1, otherwise Org_ij = 0.

• Institutional Proximity

Institutions refer to human-designed constraints, encompassing both informal institutions (e.g., traditions, customs) and formal institutions (e.g., laws) [49]. Institutional proximity arises from similarity in institutional frameworks between countries or regions, offering a foundation for cooperation.

In the context of China’s politico-economic system, intercity relations are shaped by hierarchical administrative structures, reflecting underlying power asymmetries and differentiated economic roles [50]. Cities with higher administrative levels typically have greater access to resources and preferential policies, enabling them to attract funding for large-scale cooperation projects prioritized by policymakers [51], and thereby gain an advantageous position in intercity cooperation. Therefore, this study quantifies institutional proximity based on the administrative-level gap between cities.

$${\mathrm{Ins}}_{ij}=\left\{\begin{array}{cc}\mathrm{City} i \mathrm{and} \mathrm{City} j \text{are both subprovincial or quasi-sub-provincial cities}\hfill & \hfill 3\\ \text{Only one of City} i \mathrm{and} \mathrm{City} j \text{is a sub-provincial city}\hfill & \hfill 2\\ \text{Only one of City} i \mathrm{and} \mathrm{City} j \text{is a quasi-sub-provincial city}\hfill & \hfill 1\\ \mathrm{City} i \mathrm{and} \mathrm{City} j \text{are both ordinary prefecture-level cities}\hfill & \hfill 0\end{array}\right.$$

2.3. Data Collection and Processing

This study collected intercity cooperation data from official government websites and policy documents, and organized the data collection and processing into three stages.

(1)

Initial Retrieval: This study first retrieved intercity cooperation news (2014–2023) from the official websites of 21 cities, using city names as keywords, and then extracted relevant policy documents from the 13th and 14th Five-Year Plans.

(2)

Manual Screening: This study conducted a meticulous manual review, scrutinizing both headlines and content to retain only records explicitly related to intercity cooperation. This process refined the dataset, enhancing its relevance and validity, and ensuring a solid foundation for subsequent analysis.

(3)

Data Structuring: Following duplicate removal and manual curation, the remaining cooperation data were structured by cooperation date and city pairs, resulting in 3844 valid records. These records document intercity cooperation among the 21 cities in the Pearl River–Xijiang Economic Belt from 2014 to 2023, forming a comprehensive and reliable dataset.

3. Results

3.1. The Spatio-Temporal Evolution of the Intercity Cooperation Network

To investigate the spatio-temporal evolution of the intercity cooperation network in the Pearl River–Xijiang Economic Belt, this study reviewed changes in intercity cooperation data from 2014 to 2023 (Figure 1) and divided the evolution process into three distinct stages.

(1)

First Stage (2014–2015): The cooperation data increased only slightly from 143 in 2014 to 155 in 2015. Although the Development Plan of the Pearl River–Xijiang Economic Belt was officially approved in 2014, the supporting policies remained underdeveloped. During this stage, intercity cooperation within the Pearl River–Xijiang Economic Belt was characterized by low intensity, marking the beginning of regional cooperative efforts.

(2)

Second Stage (2016–2020): The cooperation data increased significantly from 289 in 2016 to 521 in 2020. This growth was supported by the opening of the Guiyang–Guangzhou and Nanning–Guangzhou high-speed railways and the enhancement of the Xijiang waterway, which together strengthened the region’s infrastructure foundation. In 2018, the integration of the Eastern–Western Cooperation into the poverty alleviation assessment further catalyzed cooperation, leading to the implementation of multiple paired assistance projects from Guangzhou to Qiannan. Driven by infrastructure improvements and policy incentives, intercity cooperation entered a stage of accelerated development.

(3)

Third Stage (2021–2023): The outbreak of the Omicron variant severely curtailed cross-provincial flows of people and goods, resulting in a sharp decline in cooperation data—from 586 cases in 2021 to 369 in 2022. However, with the implementation of Class B infectious disease management in 2023, cross-provincial cooperation activities, including study tours and investment promotions, recovered markedly. The successful hosting of the 20th China-ASEAN Exposition in Nanning in September further signaled the stabilization of intercity cooperation, with cooperation patterns becoming more institutionalized and normalized.

To analyze the spatial pattern of the intercity cooperation network, this study applied Jenks natural breaks classification to categorize the number of cooperations between cities from 2014 to 2023 into five levels (Figure 2). To better illustrate the dynamic evolution of this spatial structure, it focused on the top three levels with high-intensity linkages, excluding the lower two levels (Figure 3) [52].

Overall, this study identified the following notable features:

• Multi-core Radial Network Structure

Building upon the spatial classification results, the intercity cooperation network follows a clear radial structure characterized by a “one core–multiple sub-centers” pattern. Guangzhou, serving as the central hub, plays a pivotal role in facilitating and coordinating regional cooperation efforts. Sub-centers such as Nanning and Foshan jointly promote deeper regional cooperation in coordination with Guangzhou.

• Rapid Development of Cross-provincial Cooperation

The cross-provincial cooperation network between Guangdong and Guizhou has experienced rapid development, with the highest intensity of linkages between Guangzhou and Qiannan. This pattern reflects the close ties and cooperation needs between the two regions in economic, social, and policy coordination domains, primarily driven by the Eastern–Western Cooperation policy for poverty alleviation.

• Pattern of Regional Development Imbalance

The intercity cooperation network reveals a pattern of “east—dominant, west—weak”. This pattern is evident in the eastern region (e.g., Guangzhou, Foshan, Zhaoqing, Yunfu), which demonstrates more advanced intercity cooperation network development due to its robust economic base and well-developed infrastructure. In contrast, the western region (e.g., Qiannan, Qujing, Wenshan) remains relatively underdeveloped and geographically isolated, resulting in limited interaction with regional core nodes. This spatial disparity reflects regional development imbalance.

3.2. Analysis of Network Structure and Cohesive Subgroups

3.2.1. Analysis of Network Structure

The intercity cooperation network of the Pearl River–Xijiang Economic Belt exhibited a trend of gradual expansion and increasing closeness from 2014 to 2023, as reflected in the rising number of participating cities and the establishment of new cooperation links between different cities, alongside enhanced overall network cohesion, as evidenced by rising network density (

Table 1. Basic structural characteristics of the intercity cooperation network in the Pearl River–Xijiang Economic Belt.

Period	Number of Nodes	Number of Edges	Network Density
2014–2015	20	49	25.79%
2016–2020	21	126	60.00%
2021–2023	21	122	58.10%

). From the perspective of transmission and clustering, the network structure of the intercity cooperation network of the Pearl River–Xijiang Economic Belt shows distinct characteristics across different development stages (

Table 2. Analysis of the structural characteristics of the intercity cooperation network in the Pearl River–Xijiang Economic Belt from the perspective of transmission and clustering.

Period	Average Path Length	Clustering Coefficient
2014–2015	2.026	0.617
2016–2020	1.405	0.754
2021–2023	1.429	0.731

).

• Initial Development Stage (2014–2015)

In 2014, the National Development and Reform Commission of China officially issued a notice on the implementation of the Development Plan of the Pearl River–Xijiang Economic Belt. At the same time, the Guiyang–Guangzhou and Nanning–Guangzhou high-speed railways began operations. Together, these initiatives provided both a policy framework and a physical infrastructure foundation for intercity cooperation in the region.

During this stage, the intercity cooperation network was at a relatively low development level. The average path length was 2.026, and the clustering coefficient was 0.617. The relatively high average path length indicates that the transmission efficiency of intercity exchanges and cooperation was low, leading to relatively high diffusion costs for factors like capital, population, technology, and information. Meanwhile, the moderate clustering coefficient suggests that while some local clusters existed within the network, they were not prominent, and the overall clustering effect was not significant.

• Accelerated Development Stage (2016–2020)

As part of the government’s poverty alleviation strategy, Guangzhou replaced Shenzhen as the designated partner for Qiannan in 2013, and further strengthened its support by dispatching a dedicated work team in 2016, which, together, intensified intercity cooperation between the two cities. In parallel, infrastructure development further supported intercity cooperation. In 2017, the 3000-ton capacity expansion project for the Xijiang waterway was completed, achieving full 3000-ton navigability along the Guangdong section of the river. In 2018, construction began on the 3000-ton navigation channel from Guigang to Wuzhou, further enhancing regional inland waterway connectivity.

During this stage, the average path length of the intercity cooperation network declined from 2.026 to 1.405, indicating optimization of transmission and accessibility between cities. Meanwhile, the network clustering coefficient increased from 0.617 to 0.754, indicating that as the network grew, cities tended to select partners with certain similar characteristics for cooperation. This led to a “local area network” phenomenon within the intercity cooperation network.

• Steady Development Stage (2021–2023)

In 2021, China adjusted the new round of the Eastern–Western Cooperation pairing relationships, and issued policy documents such as the Implementation Opinions on Establishing Closer Paired Assistance Relations Between Guangdong and Guizhou Provinces to Forge an Exemplary Model of Eastern–Western Cooperation and the Agreement on Guangdong–Guizhou Eastern–Western Regional Cooperation During the 14th Five-Year Plan Period. Against this policy background, in addition to the original Guangzhou–Qiannan partnership, Guangzhou–Anshun and Foshan–Qiandongnan also forged new intercity partnerships. However, due to the impact of the COVID-19 pandemic, local governments experienced intensified financial pressure, which, in turn, constrained investment in intercity cooperation.

During this stage, both the average path length and clustering coefficient of the intercity cooperation network remained largely unchanged compared to the previous stage, indicating that the intercity cooperation network temporarily entered a relatively steady and balanced development stage.

3.2.2. Analysis of Cohesive Subgroups

By applying the CONCOR algorithm in UCINET, this study identifies four cohesive subgroups among the 21 cities in the Pearl River–Xijiang Economic Belt (Figure 4). Furthermore, to better highlight the specific features of intercity cooperation within each subgroup, this study retains only cities participating within each group to draw the internal intercity cooperation network diagram (Figure 5) [53].

• Subgroup 1: Policy-Driven Paired Assistance Network

Subgroup 1 consists of four cities from Guangdong Province (Guangzhou, Foshan, Zhaoqing, and Yunfu) and three from Guizhou Province (Qiandongnan, Qiannan, and Anshun), which receive paired assistance from Guangdong—Guangzhou supports Qiannan and Anshun, while Foshan assists Qiandongnan. The most frequent intercity cooperation is observed between Guangzhou and Qiannan (966) and between Foshan and Yunfu (251), reflecting the radiating influence of core cities in promoting coordinated regional development.

This subgroup, centered around Guangzhou and Foshan, engages in Eastern–Western Cooperation through fiscal transfers and industrial relocation with Guizhou cities. This policy-driven cooperation reflects the central government’s efforts to narrow regional disparities through resource redistribution.

• Subgroup 2: Boundary Bridge Interwoven with History and Geography

Subgroup 2 consists of Wuzhou and Hezhou, which are located at the boundary between Subgroup 1 and Subgroup 3. Formerly administered under Wuzhou, Hezhou became a separate city only in 2002. The two cities have shared infrastructure, cultural identity, and governance experience for a long time, reflecting the enduring impact of geographical proximity and historical connections on intercity cooperation.

This subgroup serves as a key hub linking the Guangdong–Hong Kong–Macao Greater Bay Area with the southwestern hinterland. In 2014, the Guangdong–Guangxi Cooperative Special Pilot Zone was established along the Wuzhou–Zhaoqing border to explore new models of cross-regional cooperation, highlighting the strategic value of geographical proximity in shaping intercity relations.

• Subgroup 3: Intra-provincial Cooperation Network Led by Administrative Center

Subgroup 3 consists of Nanning, Guilin, Hechi, Laibin, Chongzuo, Liuzhou, Guigang, Yulin, and Baise, all of which are located in Guangxi Province and situated in the central part of the Pearl River–Xijiang Economic Belt. Within Subgroup 3, the most frequent intercity cooperation occurs between Nanning and Liuzhou (88), followed by Nanning–Baise (82), Nanning–Yulin (77), and Nanning–Chongzuo (75), illustrating a single-core radial network structure with Nanning as the central hub.

In 2019, Guangxi issued the Implementation Opinions on Advancing the Strategy to Strengthen Provincial Capitals, which emphasized strengthening Nanning’s primacy to become a growth pole and enhancing its capacity to stimulate regional development. This reflects the agglomeration effect of regional central cities in the intercity cooperation network, and their role in driving the development of surrounding cities.

• Subgroup 4: Weakly Connected Clusters at Marginal Area

Subgroup 4 consists of Qianxinan, Wenshan, and Qujing, located in the western part of the Pearl River–Xijiang Economic Belt. These cities are situated more than 500 kilometers from both Guangzhou and Nanning, and are geographically distant from the main Xijiang River channel.

This geographical remoteness constrains intercity cooperation and leads to weaker linkages with the core region, positioning this subgroup as a peripheral area within the network.

Despite their remoteness, the cities in Subgroup 4 possess considerable mineral potential, particularly in coal and metal resources. To unlock this potential, enhanced policy support and targeted intercity cooperation mechanisms are required. This highlights the need to optimize development strategies for peripheral areas within the network.

3.3. The Proximity Mechanism of Cooperation Network Evolution

This study rigorously tested the model’s reliability and validity using SmartPLS to ensure that it complied with established academic standards. The results (

Table 3. Tests for reliability and validity.

Variable	Cronbach’s Alpha	Composite Reliability	Average Variance Extracted
Geographical proximity	1	1	1
Organizational proximity	1	1	1
Institutional proximity	1	1	1
Cooperation investment	1	1	1
Network characteristic	0.746	0.886	0.886

) show that the Cronbach’s alpha values for all variables were greater than 0.7, the composite reliability values were greater than 0.8, and the AVE (Average Variance Extracted) values were greater than 0.5, indicating that the model passed the test.

The results of the structural equation modeling test (

Table 4. Tests for structural equation modeling.

	Path Coefficients	R2	SRMR Value
Institutional proximity -> Network characteristic	0.510	0.684	0.053
Cooperation investment -> Network characteristic	0.539
Geographical proximity -> Network characteristic	−0.171
Organizational proximity -> Network characteristic	−0.254

) indicate that, at the 0.05 significance level, geographical proximity, organizational proximity, institutional proximity, and cooperation investment all exerted significant direct effects on network characteristics. Among them, cooperation investment had the largest path coefficient (0.539), followed by institutional proximity (0.510), while geographical and organizational proximity both exhibited negative effects. The model’s R² value was 0.684, which was greater than the standard threshold of 0.5. In terms of model fit indices, the SRMR value was 0.053, which was less than 0.08, indicating that the model was generally acceptable.

Building upon the empirical results of the structural equation modeling, which highlight the varying effects of different dimensions of proximity, this study further explores the underlying influence mechanisms that drive the evolution of the intercity cooperation network.

• A Trend of Preferential Connections and the Matthew Effect

Among the various factors shaping the evolution of the intercity cooperation network in the Pearl River–Xijiang Economic Belt, institutional proximity plays a particularly central role. High institutional proximity fosters shared norms and values, enhancing trust and reducing cooperation barriers [54]. This dynamic is further reinforced by cooperation investment, which reflects a city’s capability and willingness to strengthen intercity cooperation. In China’s politico-economic system, cities with higher administrative levels often possess more abundant resources and can leverage comparative advantages in funding, technology, and information to accumulate network capital and achieve self-reinforcement, thereby increasing their cooperation investment. While institutional proximity generally facilitates cooperation, it may also lead to institutional inertia, impeding innovation and the reconfiguration of existing cooperation frameworks [55].

This development trend is specifically reflected in the spatio-temporal evolution characteristics of the intercity cooperation network. Guangzhou and Nanning, both cities with higher administrative levels, serve as core cities within the intercity cooperation network, playing a significant radiating and driving role and emerging as regional growth poles. Accordingly, the intercity cooperation network in the Pearl River–Xijiang Economic Belt exhibits a pattern of agglomeration growth, characterized by “preferential connections and the Matthew effect”. Peripheral cities often rely on core cities as intermediaries to gain access to the network, which also explains the emergence of a “local area network” phenomenon in the structure of the intercity cooperation network.

• Negative Lock-In under the “Proximity Paradox”

Geographical and organizational proximity are conventionally seen as enablers of intercity cooperation by promoting frequent exchange of information and knowledge, and reducing uncertainty and opportunism risks [56,57]. However, excessive proximity may produce counterproductive effects. High geographical proximity can lead to competition for the same resources, resulting in homogenization and reducing willingness to cooperate [58]. Similarly, excessive organizational proximity may restrict mutual learning and limit the organizational flexibility, ultimately hindering further development [59]. This “proximity paradox” highlights the potential for diminishing or adverse effects when proximity becomes excessive.

The empirical results from the Pearl River–Xijiang Economic Belt validate the “proximity paradox”, indicating that both geographical and organizational proximity have a negative impact on the intercity cooperation network. This may stem from excessive similarities in industrial structure and cognition, which are themselves a consequence of high geographical and organizational proximity, thereby reducing complementarity and diminishing the benefits of cooperation. In contrast, backed by substantial national policy support (e.g., the Eastern–Western Cooperation), many cities are actively overcoming geographical constraints by engaging in cross-provincial cooperation with economically or technologically complementary partners. This outward-oriented cooperation has become the dominant mode of network expansion, surpassing the slower pace of intra-provincial cooperation. However, such cooperation remains largely policy-driven. For example, following the formal designation of the Guangzhou–Anshun paired-up assistance in 2021, their intercity cooperation data increased significantly, highlighting the decisive role of policy intervention in shaping intercity cooperation patterns in China.

4. Discussion

This study adopts quantitative methods to examine the spatio-temporal evolution and influence mechanisms of the intercity cooperation network in the Pearl River–Xijiang Economic Belt. Based on the analysis, the main findings are summarized as follows:

• Policy-Driven Staged Development of the Intercity Cooperation Network

Largely shaped by national and regional policy interventions (e.g., the Development Plan of the Pearl River–Xijiang Economic Belt and the Eastern–Western Cooperation), the development of the intercity cooperation network has progressed through three distinct stages—initial, accelerated, and steady. This observation aligns with existing studies that underscore the ongoing dominance of government-led regional governance in China [60], wherein regional development remains heavily reliant on top-down directives. By quantitatively aligning network structural indicators with policy implementation timelines, this study provides robust empirical evidence for a causal relationship between administrative actions and network transformation, thereby adding methodological rigor to earlier qualitative assessments. Furthermore, the transition from the accelerated to the steady stage may reflect a plateauing effect of short-term policy incentives. This highlights the need to strengthen the self-organizing capacity of local actors and to institutionalize more enduring mechanisms to sustain intercity cooperation beyond the initial stimulus provided by policy interventions.

2.

Imbalance in Regional Development

The intercity cooperation network exhibits pronounced spatial asymmetries, reflecting persistent development imbalances between the eastern and western subregions. Cities such as Guangzhou and Nanning, functioning as regional cores, leverage their comparative institutional and economic advantages (e.g., higher administrative levels) to attract resources and exert radiating influence on surrounding cities. While prior studies have emphasized the economic benefits and spatial spillovers associated with such growth poles, they have also acknowledged the risks of overconcentration and increasing core–periphery divergence [61]. Unlike these studies, which primarily focus on economic disparity, this study demonstrates how spatial inequality is embedded within the intercity cooperation network. Specifically, it reveals how core cities dominate intercity cooperation, while peripheral cities such as Qianxinan and Wenshan remain relatively marginalized. This phenomenon undermines the inclusiveness of regional cooperation and calls for institutional mechanisms to rebalance participation and ensure more equitable access to cooperative opportunities across the region.

3.

Paradoxical Effects of Multidimensional Proximity

The findings further reveal that while institutional proximity significantly enhances intercity cooperation, both geographical and organizational proximity are associated with counterintuitive negative effects—an outcome indicative of the “proximity paradox”. Existing studies have primarily examined this phenomenon at the micro-level (e.g., partnerships involving universities or industries) [58,59,62], while this study extends the concept to a new empirical setting: government-led, policy-driven intercity cooperation networks in China. The results suggest that excessive geographical and organizational proximity may lead to intensified homogeneous competition and overlapping resource demands, thus impeding cooperation effectiveness. In this context, institutional proximity acts as a compensatory force that can overcome geographical and organizational barriers, facilitating more effective intercity cooperation. This mechanism may partly explain the predominance of cross-provincial and outward-oriented partnerships in the expansion of the intercity cooperation network within the Pearl River–Xijiang Economic Belt.

Building on the above findings, this study proposes several policy implications to further optimize the intercity cooperation network in the Pearl River–Xijiang Economic Belt and to promote sustainable regional development.

(1)

Enhance Local Autonomy: To build a more stable and enduring intercity cooperation network, it is crucial to reform the current government-led regional governance mechanism. The key lies in enhancing the self-organizing capacity of local actors (e.g., local governments, enterprises, and social organizations). Within a cooperative governance framework, future intercity cooperation should prioritize local autonomous cooperation mechanisms [63], rather than relying solely on top-down political mobilization or short-term policy interventions. This transformation can enhance the vitality, adaptability, and openness of the intercity cooperation network.

(2)

Promote Inclusive Cooperation: To narrow regional disparities, targeted resource redistribution mechanisms could be adopted to expand cooperation opportunities for peripheral cities. As ecological barriers, upstream cities such as Qianxinan and Wenshan can engage in cross-regional cooperation with downstream cities through ecological compensation mechanisms [64]. This approach would promote mutual benefits by aligning environmental preservation with regional economic interests, thus offering a synergistic pathway toward sustainable regional development.

(3)

Encourage Differentiated Development: To address the “proximity paradox” highlighted in this study, policymakers should prioritize differentiated development strategies that reduce homogeneous competition and promote functional complementarity among cities. This includes aligning upstream cities’ mineral and hydropower resources with downstream industrial capabilities to enhance the integrity of regional industrial chains. Simultaneously, regulatory mechanisms should be strengthened to prevent redundant construction and promote efficient resource allocation.

(4)

Strengthen Institutional Mechanisms: To address the lack of stable mechanisms in current intercity cooperation—manifested in weak self-organization, unbalanced cooperation opportunities, and homogenized competition—it is crucial to move toward institutionalized regional governance. This involves embedding cooperative mechanisms into robust legal and regulatory frameworks, clarifying the roles of key actors and enhancing regional policy coordination. Only by replacing fragmented unilateral actions with coherent and comprehensive planning can cross-boundary public issues be effectively addressed [65], thereby providing a resilient foundation for sustainable and inclusive regional development.

5. Conclusions

This study focuses on the Pearl River–Xijiang Economic Belt and analyzes intercity cooperation data from 2014 to 2023 using Social Network Analysis and PLS-SEM. The findings reveal the following: (1) a three-stage development process driven by policy interventions; (2) a significant spatial disparity in cooperation, characterized by an “east—dominant, west—weak” pattern; and (3) differentiated effects of proximity dimensions, whereby institutional proximity promotes network development, while geographical and organizational proximity exert negative influences.

The main contributions of this study are as follows:

(1)

Theoretical Contribution: It proposes a holistic analytical framework for intercity cooperation networks by integrating multidimensional proximity theory with Social Network Analysis. This approach deepens the understanding of their structural evolution and influence mechanisms, shifting the focus from specific functional domains to a broader perspective on overall intercity cooperation networks.

(2)

Empirical Contribution: By selecting the Pearl River–Xijiang Economic Belt—a region connecting economically developed and less developed cities—as a case study, the study provides valuable policy implications for narrowing regional disparities and promoting sustainable development in China and other emerging economies.

Despite its contributions, this study has certain limitations. Firstly, the exclusive reliance on publicly available government data may fail to capture informal cooperation, potentially leading to an underestimation of the overall intensity and diversity of intercity cooperation. Secondly, the selection of influencing factors is constrained by data availability, potentially omitting other relevant variables that affect the intercity cooperation network. Lastly, the study primarily adopts quantitative approaches (e.g., Social Network Analysis and PLS-SEM), which, while effective in revealing structural patterns, may fall short in capturing the underlying institutional logic, contextual dynamics, and policymaking processes.

To address these limitations, future research should consider incorporating emerging data sources (e.g., big data, social media) to refine the measurement of intercity cooperation and its influencing factors. Additionally, integrating qualitative methods—such as interviews, case studies, or discourse analysis of policy documents—could complement quantitative models by offering deeper insights into the mechanisms of intercity cooperation formation, performance, and institutional interplay. These enhancements would not only improve the granularity and timeliness of cooperation data, but also contribute to more nuanced theoretical understanding and provide stronger empirical support for the development of targeted and inclusive policy strategies aimed at sustainable regional development.