Impact Mechanisms and Empirical Analysis of Urban Network Position on the Synergy Between Pollution Reduction and Carbon Mitigation: A Case Study of China’s Three Major Urban Agglomerations

Abstract

Achieving the synergistic effect of pollution reduction and carbon mitigation (PRCM) is a core pathway for promoting green and low-carbon transition and realizing the “dual carbon” goals, as well as a crucial mechanism for coordinating ecological environment governance with climate action. Based on panel data from three major urban agglomerations (Beijing–Tianjin–Hebei, Yangtze River Delta, and Pearl River Delta) between 2008 and 2019, this study employs network centrality and structural holes to characterize urban network positions (UNP), and systematically investigates the impact mechanisms and spatial heterogeneity of urban network positions on PRCM synergy using a dual fixed-effects model. The findings reveal that (1) urban network positions exert significant inhibitory effects on the overall synergy of PRCM, meaning higher centrality and structural hole advantages hinder synergistic progress. This conclusion remains valid after robustness checks and endogeneity tests using instrumental variables. (2) Heterogeneity analysis shows the inhibitory effects are particularly pronounced in Type I large cities and southern urban agglomerations, attributable to environmental governance path dependence caused by complex industrial structures in metropolises and compounded pressures from export-oriented economies undertaking industrial transfers in southern regions. Northern cities demonstrate stronger environmental resilience due to first-mover advantages in heavy industry transformation. (3) Mechanism testing reveals that cities occupying advantageous network positions tend to reduce environmental regulation stringency and research and development investment levels. Conversely, cities at the network periphery demonstrate late-mover advantages by embedding environmental regulations and building stable technological cooperation partnerships. This study provides a theoretical foundation for optimizing urban network spatial configurations and implementing differentiated environmental governance policies. It emphasizes the necessity of holistically integrating network effects with ecological effects during new-type urbanization, advocating for the establishment of a multi-scale coordinated environmental governance system.

1. Introduction

Given that environmental pollution and greenhouse gas emissions share a common root, source, and process, the synergistic advancement of PRCM serves as a pivotal mechanism for fostering a comprehensive green and low-carbon transformation of economic and social development [1]. China’s 14th Five-Year Plan explicitly mandates that, through initiatives such as promoting coordinated environmental governance in key regions like the BTH region and the YRD, and strengthening innovation-driven technological development, the nation aims to achieve by 2025 an 18% reduction in CO₂ emissions per unit of GDP compared to 2020 levels and a more than 10% decrease in PM_2.5 concentration. These targets are designed to simultaneously drive continuous environmental quality improvement and high-quality economic development, thereby laying a solid foundation for building a Beautiful China and fulfilling long-term carbon neutrality commitments. Building upon this foundation, in 2022, seven ministries including the Ministry of Ecology and Environment (MEE) jointly issued the Implementation Plan for Synergizing Pollution Reduction and Carbon Mitigation, further emphasizing the strategic importance of coordinated PRCM efforts [2]. The 2024 Government Work Report unequivocally called for the integrated advancement of carbon reduction, pollution control, ecological expansion, and economic growth to comprehensively strengthen ecological civilization construction. Urban agglomerations, as the dominant form of China’s new urbanization, represent not only spatial units with concentrated carbon emissions but also serve as the pioneers and leaders in achieving early peak carbon [3]. Synergistically promoting PRCM within urban agglomerations constitutes a crucial breakthrough for alleviating the current structural, root-level, and trend-driven pressures on ecological and environmental protection. Consequently, investigating the synergistic effects of PRCM from the perspective of urban agglomerations and promoting balanced and sufficient regional development represent urgent and significant research imperatives.

The PRCM effect refers to the achievement of dual objectives—environmental quality improvement and climate change mitigation—through the integrated advancement of air pollutant reduction and greenhouse gas (GHG) control, thereby generating synergistic benefits characterized by “1 + 1 > 2”. This concept emphasizes moving beyond the singular “end-of-pipe treatment” mindset prevalent in traditional environmental governance towards synergistic source control, process optimization, and systematic management [4]. Existing literature often focuses on holistic assessments [5], driving factors [6], and spatiotemporal distribution characteristics [7] within single regions. These findings provide a solid theoretical foundation and practical support for China’s pollution control and carbon neutrality goals. Regarding holistic assessment, practical applications have been conducted across sectors such as agriculture [8], industry [9], and transportation [10], and at scales ranging from national [11], provincial [12], to key city [13] levels. Some scholars employ Computable General Equilibrium (CGE) models to quantitatively model the complex energy-economy-environment system for comprehensive PRCM effect assessment [14]. Due to the model’s intricate internal structure and strongly nonlinear interaction mechanisms, assessment results may exhibit sensitivity to parameter settings. Other scholars construct coupling coordination models to assess the synergistic potential between GHGs and air pollutants [15]. Such methods reduce the dimensionality of nonlinear interactions within complex systems, ensuring computational feasibility while effectively supporting multi-objective synergy and policy coordination assessment; however, their limited consideration of factors prevents the calculation of synergistic benefits. Regarding driving factors, Lin analyzed the evolution of CO₂ and PM₂.₅ emissions in the YRD urban agglomeration and their ten socio-economic drivers [16]; Yu developed a low-carbon development efficiency measurement method to assess efficiency levels across economic, social, and ecological dimensions within the YRD urban agglomeration and conducted low-carbon development zoning [17]; Yang used the Guanzhong Plain urban agglomeration in China to validate the role of spatial correlation network structures of carbon reduction capacity in fostering synergistic carbon reduction actions among cities [18]. Their research found a significantly uneven spatial distribution of carbon reduction capacity within the urban agglomeration, with differences in technological and economic levels influencing the formation of spatial correlations, while differences in urbanization level showed no significant impact. In spatiotemporal characteristic analysis, scholars employ spatial econometric models, spatial exploratory analysis, and other methods to analyze the spatial correlation and heterogeneity of PRCM synergy. For instance, Tang utilized a Geographically and Temporally Weighted Regression (GTWR) model to investigate the provincial spatiotemporal characteristics of the coupling coordination degree between carbon reduction and air pollutant control systems, finding an increase in the number of provinces within well-coordinated and excellently coordinated categories [19]. Building on this, Yin employed Kernel Density Estimation (KDE) to further study the dynamic evolution of China’s PRCM coupling coordination degree [20]. They observed a consistent yet diminishing increase in the coupling coordination degree over time. While inter-regional disparities widened, polarization phenomena gradually weakened. Assessments using different indicators and processes have not yielded entirely consistent conclusions, yet the overall pattern of the PRCM effect is broadly similar, exhibiting a provincial gradient disparity of “high in the east, low in the central and western regions” alongside significant agglomeration characteristics. Furthermore, scholars have explored spatial correlation relationships for synergistic pollution and carbon reduction from a complex network perspective, revealing that urban agglomeration PRCM synergy not only exhibits a complex network structure characterized by multiple clustered relationships but also that a node’s network position plays a decisive role in its ability to access heterogeneous information and scarce resources [21].

Network position theory posits that occupying advantageous positions within a network can confer superior competitive advantages [22]. However, existing research based on this theory reveals contradictory findings. Some scholars apply the core-periphery model from regional economic development to pollution and carbon reduction cooperation (PRCM), arguing that core cities leverage their locational advantages to concentrate superior resources and factors, becoming regional growth poles [21]. This collaborative dynamic amplifies their siphon effect, accelerating the flow of capital, labor, and other factors from peripheral cities towards the core. Conversely, other scholars present an entirely opposing view, suggesting that network positions can inhibit resource sharing and create resource redundancy, thereby hindering effective cooperation. For instance, Shi examined network position effects using centrality and structural holes as explanatory variables [23]. Their results indicate that while nodes occupying advantageous network positions gain access to heterogeneous information and resources, the dispersed and trust-deficient nature of their collaborators increases cooperation risks and reduces coordination efficiency. Similarly, Ma investigating from a network embedding perspective whether enhanced network positions improve urban capital allocation efficiency, found that nodes in more advantageous positions bear greater risks, consequently exerting adverse effects on outcomes [24]. These findings collectively suggest, to varying degrees, that the heterogeneous effects an individual node obtains within a network depend not only on the strength of inter-individual relationships but are also significantly influenced by the surrounding network structure. However, more empirical evidence is required to substantiate this within the specific context of PRCM networks. Furthermore, as research on regional collaborative governance deepens, scholars are increasingly focusing on PRCM synergistic effects at the city cluster scale [25]. In this context, social network analysis (SNA), as an effective tool for deconstructing cross-regional linkage mechanisms, is gradually being applied to this field. Nevertheless, existing literature on PRCM exhibits certain limitations: (1) Existing studies primarily rely on “attribute data” for exploring PRCM influences, overlooking the reality that city clusters have transcended geographical constraints. (2) Research predominantly focuses on national or provincial levels, with insufficient attention paid to the city cluster scale. (3) Current studies fail to systematically elucidate the conduction mechanisms through which network position heterogeneity leads to differential impacts on PRCM synergistic effects and lack a comprehensive theoretical framework.

Compared to existing research, the marginal contributions of this paper are as follows: (1) To reflect the evolving reality of spatial linkage network positions, this study adopts a unique relational perspective. It constructs “relational data” capable of characterizing cities’ network positions and empirically verifies their impact on PRCM. (2) This research conducts an impact analysis of urban network positions on PRCM at the city cluster level (e.g., BTH, YRD, PRD) using a dual fixed-effects model. (3) We systematically explain the conduction mechanisms underlying the differential impacts of network position heterogeneity on PRCM synergistic effects and construct a comprehensive theoretical framework to analyze the specific pathways and characteristic effects these differences may trigger.

Based on this framework, this study empirically examines the BTH, YRD, and PRD city clusters during 2008–2019. We employ network centrality and structural holes to characterize urban network positions, utilizing a dual fixed-effects model to systematically investigate the impact mechanisms of urban network positions on PRCM synergistic effects, their spatial heterogeneity patterns, and the mediating roles of Environmental Regulation Stringency and R&D Investment in the network position-PRCM relationship (research framework shown in Figure 1). This research aims to establish an empirical foundation for constructing collaborative PRCM governance systems in city clusters and advancing regional sustainable development.

2. Background and Theoretical Framework

Research Object

This study selects three major urban agglomerations as research areas to investigate the impact of urban network positions on the synergistic effects of PRCM (see Figure 2). The BTH, YRD, and PRD urban agglomerations represent regions with strong economic vitality, high openness levels, and significant development potential in China [26]. The BTH urban agglomeration, serving as China’s “Capital Economic Circle,” generated approximately 600 million tons of carbon emissions in 2020, accounting for 11% of the national total, which exceeded its proportion of national GDP [27]. The YRD urban agglomeration, as a crucial intersection point of the “Belt and Road Initiative” and the Yangtze River Economic Belt, centers on Shanghai and encompasses 26 cities across three provinces—Jiangsu, Zhejiang, and Anhui—holding strategic importance in China’s modernization and opening-up framework [28]. The PRD urban agglomeration constitutes the primary inland radiation zone of the Guangdong-Hong Kong-Macao Greater Bay Area and represents one of China’s most highly urbanized regions, with a permanent population of 134.44 million and 34.95 million registered vehicles in 2022 [29]. The high population density and advanced industrial and transportation systems inevitably result in high-intensity regional pollution and carbon emissions. Although the implementation of China’s environmental policies has decelerated carbon emission growth rates and significantly improved atmospheric conditions, the effectiveness of environmental governance investment and pollution prevention measures remains suboptimal. The three major urban agglomerations continue to face severe challenges regarding energy constraints and environmental issues.

3. Theoretical Analysis and Research Hypotheses

3.1. Urban Network Positions and Synergistic Effects of Pollution Reduction and Carbon Mitigation

Network position emerges from the relationships established between nodes, and the indicators measuring network position can be categorized into two types: those focusing on the structural forms composed of nodes and their adjacent points, and those examining the direct connections between nodes and all other points. Drawing on existing research [30], this study employs structural holes and centrality to measure network position.

Network Centrality and PRCM Synergistic Effects. Centrality, initially proposed by Freeman et al. and continuously refined, describes the direct connections of nodes within a network and can be used to measure the importance of a particular node in the network as well as the convenience of the node’s access to new resources. Existing research has found that regions with higher centrality possess high levels of openness and connectivity, effectively breaking down “barriers” between urban nodes, promoting inter-regional resource sharing and cooperation, accelerating the optimal reconfiguration of resources and collaborative innovation, reducing transaction costs, enhancing production efficiency, and thereby advancing the overall effectiveness of PRCM synergy [31]. However, the impact of network centrality on PRCM synergistic effects is not entirely positive. Cities with high network centrality leverage their network hub status to form resource control power, leading to a unipolar agglomeration pattern of innovation factors and creating a significant “resource siphoning” effect, characterized by the unidirectional concentration of capital, technology, talent, and other factors. Meanwhile, peripheral cities, suffering from factor outflow, become trapped in a predicament of weakened green transformation capacity, being forced to rely on high-pollution, high-energy-consumption industries to maintain economic growth, and compelled to make difficult trade-offs between “ecological protection” and “growth pursuit [32].” If they blindly undertake high-energy-consuming industries transferred from core cities, they may fall into a vicious cycle of “pollution transfer-ecological degradation”; if they strictly adhere to ecological red lines while suppressing industrialization processes, they may lack fiscal support for green technology application and low-carbon transformation due to weak economic foundations [33]. This spatial differentiation of “center polarization expansion-periphery passive lock-in” fragments the regional collaborative governance system, impeding the implementation of cross-regional mechanisms such as carbon emission trading and joint pollution prevention and control, ultimately forming a dual negative cycle of “core area overload-peripheral area lag,” weakening the overall network’s PRCM effectiveness.

Structural Holes and PRCM Synergistic Effects. Structural holes represent a classic sociological theory proposed by Burt [34]. When studying competitive relationships in social networks, they argue that “gaps” (i.e., structural holes) exist in social networks that are not directly connected, and individuals or organizations capable of bridging these gaps can gain competitive advantages through information benefits and control benefits [35]. The more structural holes a city possesses, the more it occupies an intermediary and bridging position in the network, connecting a greater number of otherwise unconnected cities. Therefore, cities occupying structural holes can control scarce resources in the network, concentrating resources regionally, thereby gaining advantageous positions in advancing PRCM synergy. However, when cities occupy multiple structural holes in the PRCM collaborative network, it also implies reduced cooperation and interaction between connected cities, embedding the city in a relatively dispersed network structure [36]. On one hand, dispersed network structures inevitably make information transmission within cities and between cooperative partners more difficult, with information sharing and collaboration potentially being constrained, leading to delays and incompleteness in information transmission or policy instrument conduction. On the other hand, dispersed network structures increase the intermediate nodes required for information transmission, further exacerbating risks of information distortion, delays, and misunderstandings, resulting in reduced information transparency and accuracy. Therefore, occupying structural hole positions may also be detrimental to information transmission and sharing, exacerbating resource monopolization and decision-making barriers among environmental governance actors, and inhibiting the overall PRCM collaborative process. Based on this analysis, the following hypothesis is proposed:

Hypothesis 1a: 

Ceteris paribus, the enhancement of urban network positions promotes pollution reduction and carbon mitigation synergistic effects.

Hypothesis 1b: 

Ceteris paribus, the enhancement of urban network positions inhibits the pollution reduction and carbon mitigation synergistic effects.

3.2. Urban Network Positions, Environmental Regulation Stringency, and Pollution Reduction and Carbon Mitigation Synergistic Effects

Based on social network theory, urban nodes positioned at the center of networks can more comprehensively access resources. As critical carriers of environmental governance, cities’ network positions profoundly influence the implementation effectiveness of environmental regulation stringency, thereby shaping the synergistic effects of PRCM. Cities with higher centrality maintain extensive connections and close relationships with other cities, enabling them to serve as coordinators in standard-setting processes by leveraging their abundant heterogeneous resources. This allows them to mitigate conflicts among network partners and enhance stakeholders’ compatibility and acceptance of technological standards. Concurrently, according to open innovation theory, environmental regulation stringency, as an external pressure, can stimulate urban nodes to participate in PRCM collaborative cooperation, facilitating effective information and technology exchange, joint research and development, and promotion of green technologies to meet environmental regulatory requirements. However, the impact of environmental regulation stringency on PRCM synergy is not uniformly positive. Xu indicates that environmental regulation stringency increases compliance costs for enterprises within cities, creating crowding-out effects on innovation investment and productive investment, which adversely affect firms’ research and development, management, and production activities, ultimately weakening cities’ PRCM synergistic capabilities [37]. Based on multi-level governance theory, Luo discovered that network-central cities bear greater environmental governance responsibilities and face multiple environmental regulatory requirements from different governmental levels, which tend to generate “beggar-thy-neighbor” environmental regulation implementation strategies [38]. Specifically, when urban nodes expand their ego-network size, they become more cautious in selecting cooperation partners and raise cooperation thresholds. Therefore, when confronted with environmental regulation stringency pressure, urban nodes may not necessarily identify more environmentally friendly and efficient technologies and solutions to adapt to new environmental requirements. City managers may also tend to adopt low-risk, conservative governance strategies, thereby inhibiting PRCM synergistic effects. Based on this analysis, the following hypothesis is proposed:

Hypothesis 2: 

Urban network position influences pollution reduction and carbon mitigation synergistic effects through the mediating effect of environmental regulation stringency.

3.3. Urban Network Positions, Research and Development Investment, and Pollution Reduction and Carbon Mitigation Synergistic Effects

Cities’ positions within networks determine their resource flow, knowledge diffusion, and policy radiation capabilities, thereby influencing PRCM potential [39]. Cities positioned at network cores can accumulate diversified resources, information channels, and a solid regional reputation, forming integration advantages in technological resources. These cities can effectively attract high-level talent and technical experts while rapidly absorbing domestic and international research and development spillover effects. These advantageous resources can powerfully promote the research, development, innovation, and application of PRCM synergistic technologies, accelerating the PRCM collaborative process. However, this process also faces realistic cost constraint challenges. High-level talent typically requires correspondingly high compensation packages, leading to substantial increases in talent recruitment and cultivation costs. Particularly, the comprehensive investment in talent recruitment (including housing subsidies, children’s education, medical insurance, and other supporting expenditures) may crowd out technological equipment funding, resulting in imbalanced Research and Development Investment structures [40]. Moreover, uncertainty exists regarding whether the generated benefits can effectively cover input costs. Therefore, when cities leverage network position advantages to obtain more development opportunities, decision-makers, based on realistic considerations of development costs and fiscal pressures, often tend to choose relatively low-cost development pathways while maintaining cautious attitudes toward large-scale recruitment of high-level talent or implementation of high-standard talent incentive policies. This cost constraint mechanism may restrict cities’ capabilities to acquire high-quality human capital, somewhat affecting the full realization of their innovation-driven development capabilities, thereby inhibiting urban PRCM synergistic development. Based on this analysis, the following hypothesis is proposed:

Hypothesis 3: 

Urban network position influences pollution reduction and carbon mitigation synergistic effects through the mediating effect of Research and Development Investment.

4. Research Design

4.1. Sample Selection and Data Sources

Based on data availability and consistency, this study selects relevant data from China’s three major urban agglomerations from 2008 to 2019 as the research sample. The city-level carbon emission data are sourced from the China Emission Accounts and Datasets (CEADs). This database represents a relatively complete, comprehensive, and high-precision carbon emission dataset, with a fitting accuracy of R² = 0.998 for carbon dioxide emissions, and has become an important data source for carbon emission research among domestic and international scholars. The PM_2.5 concentration data are based on satellite remote sensing inversion data of PM_2.5 concentrations published by the Atmospheric Composition Analysis Group at Dalhousie University, Canada. Using ArcGIS 10.6 software for data conversion, prediction, and adjustment, spatial raster data of PM_2.5 concentrations covering the entire territory of China from 2008 to 2019 were obtained (raster resolution: 0.01° × 0.01°), and weighted average processing was performed in combination with China’s prefecture-level administrative divisions.

The remaining raw data were sourced from various years of the China Statistical Yearbook, China City Statistical Yearbook, statistical bulletins, etc. Partially missing data were supplemented using the linear interpolation method.

4.2. Main Variable Definitions and Descriptions

4.2.1. Explained Variable

Synergistic Effects of PRCM. Atmospheric pollutants and carbon dioxide emissions exhibit co-sourced characteristics, with a close coupling relationship existing between CO₂ emission reduction and atmospheric pollution control. To characterize the overall efficacy of urban PRCM and reflect the synergistic development trends of different subsystems, this study employs a coupling coordination model to conduct dynamic coupling analysis of atmospheric pollutant emissions and carbon emissions, constructing a PRCM synergy index to measure synergistic effects. Considering the need for multi-regional comparative analysis, this study adopts carbon emission intensity to measure carbon reduction levels. Drawing on relevant research findings, the scope of “pollution reduction” is confined to the atmospheric pollution domain, with PM₂.₅ concentration selected as a proxy variable for atmospheric pollution [41]. Based on this foundation, this paper modifies the model by referencing the methodology of Wang [42], with the calculation formula for urban PRCM synergistic effects as follows:

$$C={\displaystyle \frac{2\sqrt{E_{P{M}_{2.5}}\times E_{C{O}_2}}}{E_{P{M}_{2.5}}+E_{C{O}_2}}}=\sqrt{[1-(E_{C{O}_2}-E_{P{M}_{2.5}})]{\displaystyle \frac{E_{P{M}_{2.5}}}{E_{C{O}_2}}}}$$

(1)

$$T=\alpha E_{P{M}_{2.5}}+\beta E_{C{O}_2}$$

(2)

$$S=\sqrt{C\times T}$$

(3)

Following the research of scholars, this study assigns equal weights to carbon emissions and air pollution to reflect the synergistic level of PRCM, with α = β = 0.5 [43,44,45]. The closer the S value approaches 1, the higher the coordination degree of PRCM, and the stronger the synergistic effect; conversely, the closer the S value approaches 0, the lower the coordination degree of PRCM, and the weaker the synergistic effect. The definitions of each variable within the model can be found in

Table 1. Symbols and meanings of variables in the PRCM synergistic index model.

Symbols	Meanings
EPM2.5	Range-standardized value of annual average urban PM2.5 concentration (μg/m3)
ECO2	Standardized range of carbon emission intensity in cities (t/10,000 yuan)
α	Weighting coefficients for EPM2.5 and ECO2
β
C	Coupling degree of PRCM systems
T	Comprehensive coordination indicator
S	Coupling coordination degree of urban PRCM systems

.

4.2.2. Explanatory Variable

Network location. The spatial conduction channels of the PRCM synergistic effect in city clusters are identified in this research [46]. The gravity coefficient K is adjusted by dividing the S value of a particular city i by the total of the S values of that city and its associated city j, since the spatial correlation degree of the synergistic effect of PRCM in urban agglomerations is not equal. Equation (4) illustrates how the spatial correlation network of the synergistic effect of PRCM is constructed using the enhanced gravity model:

$$F_{ij}={\displaystyle \frac{S_i}{S_i+S_j}}\times {\displaystyle \frac{\sqrt[3]{P_i·G_i·S_i}\sqrt[3]{P_j·G_j·S_j}}{{d_{ij}}^2}},d_{ij}={\displaystyle \frac{di{s}_{ij}}{|g_i-g_j|}}$$

(4)

Transform the initial gravity matrix F into a binary matrix V. Using the row-wise mean of F as the threshold, elements greater than or equal to the threshold are assigned 1 (indicating spatial linkage of agricultural carbon emissions between corresponding regions), otherwise 0. This yields the PRCM Synergy Network Adjacency Matrix. The definitions of each variable within the model can be found in

Table 2. Correct the variable symbols and their meanings in the gravitational mode.

Symbols	Meanings
Fij	The Influence of City i on City j in Enhancing the Synergistic Effects of PRCM
P	Urban Population (100 million persons)
G	GDP (100 million yuan)
g	GDP per capita (10 thousand yuan/person)
dij	Inter-city economic distance
disij	Inter-city geographic distance
F	PRCM gravity matrix among cities within a city cluster

.

Common centrality metrics include Relative Centrality, Betweenness Centrality, and Closeness Centrality. Relative Centrality accounts for network scale effects and is thus adopted to measure regional centrality. The formal equation is:

$$Centralit{y}_i={\displaystyle \frac{k_i}{N-1}}={\displaystyle \frac{{\displaystyle \sum _{i\ne j}{\alpha }_{ij}}}{N-1}}$$

(5)

Structural holes, proposed and quantified by Burt [34], are measured via Constraint. Constraint assesses a node’s ability to utilize structural holes. Since constraint values may exceed 1, we measure structural holes by subtracting the constraint from 2. The formal equations are

$$Hol{e}_i=2-{\displaystyle \sum _j(p_{ij}+{\displaystyle \sum _{q,q\ne i,q\ne j}{p}_{iq}{p}_{qj}{)}^2}}$$

(6)

The definitions of each variable within the model can be found in

Table 3. Variable symbols and their meanings in network centrality and structural hole models.

Symbols	Meanings
Centralityi	Relative centrality of city i in the urban network
ki	Number of regions directly connected to city i
N	Total number of cities in the urban network
Holei	Structural holes of city i
piq	Proportion of city i’s relationships invested in city q to its total relationships

.

4.2.3. Mediating Variable

• Environmental Regulation Stringency: Following the approach of Guan [47], environmental regulation stringency intensity was measured using a weighted aggregation (via the entropy method) of three indicators: the wastewater treatment plant treatment rate, the municipal solid waste harmless treatment rate, and the comprehensive utilization rate of general industrial solid waste.
• Research and Development Investment: Following the approach of Wang [48], this was defined as the ratio of research and development expenditure to the total general public budget expenditure of the city.

4.2.4. Control Variable

To control for the impacts of environmental pollution control, resource utilization, and economic development on the synergistic effect of pollution and carbon reduction, this paper introduces the following control variables in the regression analysis: (1) per capita gross domestic product; (2) urban private and self-employed workers; (3) population density; (4) foreign direct investment; and (5) Share of Primary Sector. The definitions and specific measurement methods for each variable are detailed in

Table 4. Definition of variables.

Variable Type	Variable Name	Variable Symbol	Variable Definition
Explained variable	PRCM Synergy Index	ISEC	Coupling coordination analysis of PM2.5 concentration and carbon Emission intensity indicators
Explanatory variable	Network Position	SDC	Centrality, represented by out-closeness centrality as the centrality indicator
		SH	Structural holes, represented by the constraint index as the structural hole indicator
Mediating variable	Research and Development Investment	RD	Ratio of science and technology expenditure to city-level general public budget expenditure
	Environmental Regulation Stringency	ERS	Weighted Aggregation via Entropy Weight Method for Three Indicators: Centralized Treatment Rate of Municipal Sewage Treatment Plants, Harmless Disposal Rate of Domestic Waste, and Comprehensive Utilization Rate of General Industrial Solid Waste
Control Variable	per capita gross domestic product	PGDP	Ratio of GDP to Resident Population in the Same Period
	Urban Private and Self-Employed Workers	UPSEW	Logarithm of the sum of urban private sector employees and urban self-employed Individuals
	Share of Primary Sector	IND1rate	Ratio of value-added in the primary industry to gross domestic product
	Foreign Direct Investment	FDI	Actually utilized foreign direct investment amount in current year (Converted to RMB at current-year exchange rates)
	Population Density	POPDEN	Population density by registered residence at year-end relative to administrative land area

.

4.3. Model Specification

4.3.1. The Baseline Regression Model

This study illustrates the city network location features in terms of network centrality and structural holes, respectively, to examine the impact of city network location on the synergistic effect of PRCM. The random effect model is disproved by the Hausman test, and the following model is constructed by examining the prior hypotheses using the double-fixed-effect regression model:

$$ISE{C}_{it}={\alpha }_0+{\alpha }_{1}SD{C}_{it}+{\alpha }_{2}S{H}_{it}+\beta Contro{l}_{it}+{\mu }_i+{\nu }_t+{\epsilon }_{it}$$

(7)

The definitions of each variable within the model can be found in

Table 5. Variable symbols and their meanings in the double fixed effects model.

Symbols	Meanings
ISECit	Dependent variable, representing the synergistic effect of PRCM for city i in year t
SDCit	Core explanatory variable, representing the network centrality of city i in year t
α1	Impact of network centrality on PRCM effect
SHit	Core explanatory variable, representing the structural holes of city i in year t
α2	Impact of structural holes on PRCM effect.
Controlit	Control variables
μi	Entity fixed effects
vt	Time fixed effects
εit	Random disturbance term

.

4.3.2. Mechanism Variable

Drawing on Jiang Ting’s methodology for mediating effect analysis, we empirically examine the mediating role of environmental regulation and technological investment, with the model constructed as follows [49]:

$$Z_{it} ={\gamma }_0 +{\gamma }_1 SD{C}_{it} +{\gamma }_2 S{H}_{it }+\beta Contro{l}_{it} +{\mu }_i+{\nu }_t +{\epsilon }_{it}$$

(8)

The definitions of each variable within the model can be found in

Table 6. Variable symbols and their meanings in the mediating effect model.

Symbols	Meanings
Zit	Mediating variable, representing the environmental regulation and Research and Development Investment of city i in year t
γ1, γ2	Effect of the PRCM synergistic effect on the mediating variables

.

5. Empirical Results Analysis

5.1. Descriptive Statistics

Figure 3 illustrates the evolution of the PRCM Synergy Index across China’s Three Major City Clusters during 2008–2019. As depicted in Figure 3, the PRCM Synergy Index for all three megaregions exhibited a fluctuating upward trajectory throughout the 2000–2019 period. This trend signifies progressively enhanced co-benefits in pollution control and carbon mitigation across urban agglomerations, indicating accelerated manifestation of synergistic outcomes between urban economic development and eco-environmental conservation efforts.

Simultaneously, based on the measurement results, this study uses 2019 as an example to illustrate the spatial distribution and spatial correlation network of PRCM synergy. As shown in Figure 4, the BTH urban agglomeration exhibits a distinct dual-core dominant structure. Beijing and Tianjin form the strongest core spatial linkage. However, peripheral cities such as Shijiazhuang and Tangshan demonstrate relatively low PRCM synergy indices, resulting in a pronounced core-periphery structure. Concurrently, the inter-city linkage network is relatively sparse, indicating uneven development characteristics. In contrast, the YRD urban agglomeration demonstrates characteristics of polycentric coordinated development. It possesses the highest overall level of PRCM synergy index, which is also evenly distributed, reflecting a relatively mature integrated pattern. The development levels among its cities are relatively balanced, with minimal intra-regional disparities. The PRD urban agglomeration has the smallest geographical scope, yet exhibits generally high PRCM synergy indices with a relatively uniform spatial distribution. The cities exhibit tight inter-city linkages, forming a high-density urban network. Guangzhou and Jiangmen share the strongest spatial linkage and occupy the core positions, while Foshan and Dongguan hold secondary core positions.

Table 7. Descriptive statistics.

Variables	Unit	Obs	Mean	SD	Min	Max
ISEC	--	600	0.759	0.125	0.463	0.930
SDC	Nodes	600	20.115	7.699	10.788	40.313
SH	--	600	0.182	0.069	0.098	0.364
RD	Percentage	600	0.033	0.022	0.004	0.117
ERS	--	600	0.208	0.112	0.031	0.712
PGDP	10,000 yuan per capita	600	10.992	0.598	9.58	12.136
UPSEW	People	600	5.891	0.436	4.899	6.860
IND1rate	Percentage	600	0.069	0.054	0.001	0.203
FDI	10,000 yuan	600	0.031	0.021	0.003	0.104
POPDEN	persons per km2	600	0.067	0.036	0.010	0.228

presents the descriptive statistics of the variables and the correlation coefficients between them. As shown in Table 7, the PRCM synergy index ranges from a minimum of 0.463 to a maximum of 0.930, with a mean of 0.757 and a standard deviation of 0.125. These results indicate substantial variation in the PRCM synergy index across different cities. Analysis of the descriptive statistics for the network position metrics further reveals significant disparities in PRCM synergy indices associated with different network positions. Furthermore, the Variance Inflation Factor (VIF) was calculated for all variables. The maximum VIF value was 4, and the mean VIF was 2.07, suggesting a low probability of multicollinearity in this regression model.

5.2. Baseline Regression Analysis

Table 8. Benchmark regression results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC
SDC	−0.006 ***		−0.006 ***		−0.006 ***		−0.002 ***		−0.001 **
	(−0.001)		(−0.001)		(−0.001)		(−0.001)		(0.000)
SH		−0.618 ***		−0.635 ***		−0.680 ***		−0.230 ***		−0.116 **
		(−0.070)		(−0.069)		(−0.068)		(−0.074)		(−0.053)
PGDP			0.044 ***	0.044 ***	0.026	0.026	0.013 *	0.013 *	−0.004	−0.004
			(−0.013)	(−0.013)	(−0.016)	(−0.016)	(−0.007)	(−0.007)	(−0.007)	(−0.007)
UPSEW			−0.030 **	−0.030 **	−0.044 ***	−0.044 ***	0.070 ***	0.070 ***	−0.004	−0.004
			(−0.015)	(−0.015)	(−0.014)	(−0.014)	(−0.010)	(−0.010)	(−0.008)	(−0.008)
POPDEN			−0.101	−0.100	−0.154	−0.153	−0.419 ***	−0.419 ***	−0.098	−0.098
			(−0.161)	(−0.161)	(−0.172)	(−0.172)	(−0.094)	(−0.094)	(−0.077)	(−0.077)
FDI			−0.519 **	−0.520 **	−0.446 *	−0.447 *	−0.124	−0.124	0.057	0.057
			(−0.227)	(−0.227)	(−0.228)	(−0.228)	(−0.116)	(−0.116)	(−0.083)	(−0.083)
IND1rate			−1.262 ***	−1.262 ***	−1.111 ***	−1.110 ***	0.793 ***	0.793 ***	0.169	0.169
			(−0.161)	(−0.161)	(−0.159)	(−0.159)	(−0.228)	(−0.228)	(−0.159)	(−0.159)
id	NO	NO	NO	NO	NO	NO	YES	YES	YES	YES
year	NO	NO	NO	NO	YES	YES	NO	NO	YES	YES
_cons	0.870 ***	0.871 ***	0.671 ***	0.673 ***	0.954 ***	0.957 ***	0.226 ***	0.226 ***	0.839 ***	0.839 ***
	(−0.013)	(−0.014)	(−0.166)	(−0.166)	(−0.201)	(−0.201)	(−0.076)	(−0.076)	(−0.091	(−0.091)
N	600,000	600,000	600,000	600,000	600,000	600,000	600,000	600,000	600,000	600,000
R2	0.114	0.115	0.231	0.231	0.282	0.283	0.918	0.918	0.962	0.962

Note: “***”, “**”, and “*” express significance levels at 1%, 5%, and 10%. p-values are shown in brackets. id = region fixed effects; year = year fixed effects; YES = included; NO = not included.

presents the regression results examining the impact of urban network position on PRCM synergistic effects. The findings reveal that the enhancement of urban network status exerts a significantly negative influence on PRCM synergy. In column (7), the coefficient of network centrality is negative at the 1% significance level, indicating that the more prominent the core position occupied by urban agglomerations within the network, the weaker their PRCM synergistic effects. This phenomenon may be attributed to the fact that core cities, due to economic agglomeration effects, attract energy-intensive industries (such as manufacturing and heavy chemical industries), resulting in concentrated energy consumption and pollutant (PM_2.5) emissions, while carbon emission intensity remains persistently high. In column (8), the regression coefficient of structural holes is −0.109 (p < 0.05), demonstrating that the more significant the structural hole positions occupied by urban agglomerations, the weaker their PRCM synergistic effects. Cities occupying structural holes serve as “intermediaries” connecting different groups. When cities occupy multiple structural holes in the PRCM synergy network, network density correspondingly decreases, forming a relatively dispersed network structure. This network configuration impedes the effective transmission of environmental governance information and technology both within cities and among collaborative partners, causing information-sharing delays and inefficient coordination mechanisms, ultimately undermining inter-regional environmental governance cooperation effectiveness. The aforementioned analysis demonstrates that the enhancement of urban network status has a significantly negative impact on PRCM synergy. Therefore, Hypothesis 1a is not supported, while Hypothesis 1b is validated.

5.3. Endogenous Test

Instrumental Variable Method

The location of city networks might encourage the process of pollution and carbon reduction synergies, according to the results calculated using the double fixed-effects model. However, endogeneity issues may skew the regression results. The occurrence of reverse causality and the bias brought on by the exclusion of variables may be the primary causes of the endogeneity issue in this investigation. On the one hand, the urban network’s location is influenced by the synergistic impact of PRCM; on the other hand, it is unavoidable to leave out significant variables because the synergistic effect of PRCM is a comprehensive outcome of various elements, some of which are hard to measure. This paper further uses the instrumental variable method for estimation in an attempt to lessen the negative effects of the endogeneity problem. The first-order lag terms of network centrality and structural holes are used as instrumental variables, respectively, and further tests are conducted using two-stage least-squares estimation. The results are displayed in

Table 9. Endogeneity test results.

Variables	(1)	(2)
	ISEC	ISEC
SDC	−0.003 ***
	(−0.001)
SH		−0.386 ***
		(−0.105)
PGDP	−0.004	−0.004
	(−0.006)	(−0.006)
UPSEW	−0.004	−0.004
	(−0.008)	(−0.008)
POPDEN	−0.086	−0.086
	(−0.069)	(−0.069)
FDI	0.105	0.105
	(−0.083)	(−0.083)
IND1rate	0.159	0.159
PGDP	(−0.175)	(−0.175)
id	YES	YES
year	YES	YES
_cons	0.804 ***	0.804 ***
	(−0.099)	(−0.099)
Kleibergen-Paap Wald rk F	154.47	154.59
	(13.27)	(13.27)
Kleibergen-Paap rk LM	7.88 **	7.87 **
N	550	550
R2	0.965	0.965

Note: “***” and “**” express significance levels at 1% and 5%.

. Table 9 shows that city network placement has a considerable negative impact on network centrality and structural gaps, which is in line with the benchmark regression. Both the instrumental variable under-identification test and the weak instrumental variable test were passed by the instrumental variables.

5.4. Robustness Test

5.4.1. Reduction in Sample Period

The 2008 global financial crisis had a major impact on economic and environmental policy, which could lead to abnormal data (such as a sudden decline in industrial activity or temporary reductions in pollutants). The 2008–2009 data can be excluded to exclude the impact of the post-crisis recovery period and guarantee that the sample accurately represents the impact of pollution and carbon reduction in the norm. Columns (1) and (2) of

Table 10. Robustness test results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC
	Reduction in Sample Period		Standard Error Clustering Hierarchy		Shrinking of Sample Data
SDC	−0.001 **		−0.001 **		−0.001 **
	(−0.001)		(0.000)		(0.000)
SH		−0.127 **		−0.116 **		−0.116 **
		(−0.059)		(−0.047)		(−0.053)
PGDP	−0.005	−0.005	−0.004	−0.004	−0.004	−0.004
	(−0.007)	(−0.007)	(−0.006)	(−0.006)	(−0.007)	(−0.007)
UPSEW	−0.006	−0.006	−0.004	−0.004	−0.004	−0.004
	(−0.009)	(−0.009)	(−0.012)	(−0.012)	(−0.008)	(−0.008)
POPDEN	−0.097	−0.097	−0.098	−0.098	−0.098	−0.098
	(−0.082)	(−0.082)	(−0.093)	(−0.093)	(−0.077)	(−0.077)
FDI	0.103	0.103	0.057	0.057	0.057	0.057
	(−0.094)	(−0.094)	(−0.119)	(−0.119)	(−0.084)	(−0.084)
IND1rate	0.139	0.139	0.170	0.170	0.170	0.170
	(−0.176)	(−0.176)	(−0.217)	(−0.217)	(−0.159)	(−0.159)
id	YES	YES	YES	YES	YES	YES
year	YES	YES	YES	YES	YES	YES
_cons	0.864 ***	0.864 ***	0.838 ***	0.839 ***	0.838 ***	0.838 ***
	(−0.097)	(−0.097)	(−0.102)	(−0.102)	(−0.092)	(−0.092)
N	550	550	600	600	600	600
R2	0.966	0.966	0.967	0.967	0.967	0.967

Note: “***” and “**” express significance levels at 1% and 5%.

display the regression results. The coefficients of SDC and SH are still significantly negative, indicating that the conclusions remain sound even after the sample time interval is substituted.

5.4.2. Standard Error Clustering Hierarchy

The coefficients of network centrality and structural holes are considerably negative in columns (3) and (4) of Table 10, and the standard errors of clustering at the city cluster level are used for the regression in this paper in consideration of potential heteroskedasticity and autocorrelation issues.

5.4.3. Shrinking of Sample Data

As indicated in columns (5) and (6) of Table 10, this study uses a two-sided shrinkage treatment at the 1% quantile for all continuous variables (including core explanatory variables and control variables) to account for the geographic differences of the study sample, which spans different regions of the nation, and to prevent the interference of extreme outliers that may be caused by geographic distribution heterogeneity. The results of the robustness tests are consistent with the previous findings and remain largely consistent.

6. Further Analysis

6.1. Heterogeneity Analysis

6.1.1. Heterogeneity of Urban Size

The study categorizes cities into megacities (e.g., Beijing, Shanghai), metropolises (e.g., Tianjin, Hangzhou), and Type I cities (e.g., Shijiazhuang, Ningbo) to analyze how network location impacts PRCM synergy.

Table 11. Results of heterogeneity test.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC	ISEC
	Megacity		Metropolis		Type I Large Cities		the Southern Part of the Country		the Northern Part of the Country
SDC	−0.003		−0.004		−0.001 **		−0.001 ***		0.005
	(−0.005)		(−0.004)		(0.000)		(0.000)		(−0.005)
SH		−0.390		−0.426		−0.113 **		−0.145 ***		0.516
		(−0.523)		(−0.459)		(−0.052)		(−0.042)		(−0.542)
PGDP	0.171 *	0.170 *	−0.157 ***	−0.157 ***	−0.001	−0.001	−0.002	−0.002	0.030 *	0.030 *
	(−0.098)	(−0.098)	(−0.039)	(−0.039)	(−0.007)	(−0.007)	(−0.006)	(−0.006)	(−0.016)	(−0.016)
UPSEW	−0.104	−0.104	−0.035	−0.035	0.008	0.008	−0.001	−0.001	−0.014	−0.014
	(−0.089)	(−0.089)	(−0.028)	(−0.028)	(−0.009)	(−0.009)	(−0.009)	(−0.009)	(−0.016)	(−0.016)
POPDEN	5.752	5.754	−2.696	−2.696	−0.118	−0.118	0.025	0.025	0.203	0.203
	(−6.151)	(−6.15)	(−1.872)	(−1.872)	(−0.078)	(−0.078)	(−0.081)	(−0.081)	(−0.159)	(−0.159)
FDI	1.761	1.762	−0.343	−0.343	0.111	0.111	0.081	0.081	−0.076	−0.076
	(−1.182)	(−1.182)	(−0.304)	(−0.304)	(−0.085)	(−0.085)	(−0.077)	(−0.077)	(−0.221)	(−0.221)
IND1rate	−0.654	−0.655	−3.98	−3.98	0.853 ***	0.853 ***	0.118	0.118	3.948 ***	3.947 ***
	(−0.508	(−0.508)	(−3.837)	(−3.837)	(−0.293)	(−0.293)	(−0.128)	(−0.128)	(−0.981)	(−0.981)
id	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
year	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
_cons	−0.491	−0.489	3.141 ***	3.142 ***	0.706 ***	0.706 ***	0.836 ***	0.837 ***	0.017	0.017
	(−1.459)	(−1.458)	(−0.386)	(−0.386)	(−0.094)	(−0.094)	(−0.088)	(−0.088)	(−0.253)	(−0.253)
N	48	48	36	36	516	516	432	432	168	168
R2	0.957	0.957	0.995	0.995	0.969	0.969	0.957	0.957	0.968	0.968

Note: “***”, “**”, and “*” express significance levels at 1%, 5%, and 10%.

reveals that network centrality and structural holes exhibit insignificant coefficients in megacities/metropolises but significantly negative coefficients in Type I cities. This suggests that larger cities’ administrative advantages—strict environmental policies, robust infrastructure, and cross-sector coordination—mitigate the adverse environmental effects of network structures. Conversely, Type I cities, constrained by fragmented governance, policy delays, and weaker centrality, face intensified inter-regional resource competition. This triggers a “race to the bottom,” where loosened environmental regulations and industrial chain manipulation redirect pollution to peripheral areas, undermining PRCM benefits. Additionally, limited incentives for industrial upgrading in Type I cities reinforce path dependency, locking network advantages into inefficient capacity expansion and worsening environmental externalities.

6.1.2. Heterogeneity of Regional Economic Organization

To examine the heterogeneous impacts of urban agglomeration network positions on PRCM synergistic effects, drawing upon relevant research, this study categorizes urban agglomerations into northern and southern groups based on geographical location for comparative analysis, thereby identifying the differentiated characteristics of network position effects under distinct regional economic organizational features. The northern urban agglomeration includes the BTH urban agglomeration, while the southern urban agglomerations comprise the YRD urban agglomeration and the PRD urban agglomeration.

Regarding regional economic organizational heterogeneity results, as shown in columns (7)–(10) of Table 11, the regression coefficients of network centrality and structural holes are not significant in northern urban agglomerations, whereas in southern urban agglomerations, the regression coefficients of both network centrality and structural holes are significantly negative at the 1% level. This indicates that when southern urban agglomerations occupy central positions or structural hole advantages within the network, their PRCM synergistic effects are conversely weakened, while northern regions do not exhibit this characteristic. The underlying reason may stem from differences in regional economic organizational models between the north and south: southern urban agglomerations demonstrate higher levels of openness and marketization, resulting in intense resource competition within urban agglomeration networks. Under these circumstances, central nodes and structural hole occupants may overly focus on economic efficiency while neglecting environmental collaborative governance, or even exacerbate environmental burdens by transferring high-pollution industries to network peripheral areas. Taking the YRD region as an example, central cities such as Shanghai and Suzhou, leveraging their superior locational advantages, educational and technological resources, and advanced manufacturing strengths, have taken the lead in completing industrial structure optimization and upgrading. Meanwhile, coastal areas in Jiangsu Province (relatively peripheral regions such as Lianyungang and Yancheng), due to relatively lenient PRCM targets, have objectively enhanced their attractiveness to high-energy-consumption and high-emission industries, gradually becoming concentrated zones for undertaking pollution- and carbon-intensive industrial transfers, substantively bearing the environmental costs externalized by central cities. In contrast, northern urban agglomerations are characterized by government-led economies with strong rigidity in environmental policy implementation, where resource allocation advantages brought by network positions are weakened by administrative intervention, resulting in decoupling between economic and environmental effects of network structure. Furthermore, industrial homogenization competition in southern urban agglomerations may undermine network collaboration efficiency, while the inherent pollution rigidity of northern heavy industrial clusters may mask the marginal impacts of network structure.

6.2. Mechanism of Action Analysis

(1)

Mediating effect test of environmental regulation stringency

The mediating role of urban environmental regulation is presented in

Table 12. Mechanism test analysis results.

Variables	(1)	(2)	(3)	(4)
	IER	IER	RD	RD
SDC	−0.000 ***		−0.000 ***
	(0.000)		(0.000)
SH		−0.007 ***		−0.053 ***
		(0.002)		(0.010)
xc2	−0.003 ***	−0.003 ***	0.012 ***	0.012 ***
	(0.000)	(0.000)	(0.002)	(0.002)
xc3	−0.001 ***	−0.001 **	0.003 *	0.003 *
	(0.000)	(0.000)	(0.002)	(0.002)
xc4	−0.023 ***	−0.023 ***	−0.101 ***	−0.101 ***
	(0.005)	(0.005)	(0.022)	(0.022)
xc5	0.032 ***	0.032 ***	0.066 **	0.066 **
	(0.007)	(0.007)	(0.032)	(0.032)
xc6	3.075 ***	3.075 ***	0.050 **	0.050 **
	(0.005)	(0.005)	(0.022)	(0.022)
id	YES	YES	YES	YES
Year	YES	YES	YES	YES
_cons	0.048 ***	0.048 ***	−0.103 ***	−0.103 ***
	(0.005)	(0.005)	(0.023)	(0.023)
N	600.000	600.000	600.000	600.000
R2	0.999	0.999	0.520	0.520

Note: “***”, “**”, and “*” express significance levels at 1%, 5%, and 10%.

. Columns (1) and (2) show significantly negative regression coefficients for both SDC and SH, indicating that higher network centrality or occupying abundant structural holes suppresses Environmental Regulation Stringency. The underlying mechanism may be explained as follows: Cities with advantageous network positions prioritize economic growth (e.g., investment attraction, industrial expansion) by leveraging their resource allocation and information control advantages. They adopt strategic avoidance behaviors in environmental governance, passively benefiting from other cities’ pollution control achievements through resource siphoning effects while reducing their own regulatory investments. Conversely, cities at the network periphery, constrained by weak resource control capabilities, cannot secure development space through economic negotiations. Instead, they embed environmental regulations into political promotion games, utilizing policy labels (e.g., “Civilized City” campaigns) to position PRCM as an “admission ticket” for resource allocation. In summary, urban network positions influence PRCM synergistic effects through the mediating role of Environmental Regulation Stringency, validating Hypothesis 2.

(2)

Mediating effect test of research and development investment

The mediating role of R&D Investment is presented in Table 7. Columns (3) and (4) show significantly negative regression coefficients for both SDC and SH, indicating that elevated urban network status substantially suppresses Research and Development Investment. Higher network centrality or occupation of abundant structural holes reflects city embedding within open yet sparse network structures, characterized by low network density and unstable relational ties. This structural configuration increases uncertainty costs for technological cooperation, compelling enterprises to face higher risk premiums when investing in PRCM technologies. Consequently, it undermines their absorptive and transformative capacities for technological innovation. In contrast, although cities at the network periphery have limited connections, they typically establish stable and deep cooperative relationships with select key nodes. This facilitates persistent technological learning mechanisms and dedicated innovation investment, aligning with the late-mover advantage theory. Unconstrained by incumbent technological trajectories, these cities demonstrate greater potential for technological leapfrogging. Thus, urban network positions influence PRCM synergistic effects through the mediating role of Research and Development Investment, validating Hypothesis 3 as empirically verified.

7. Research Conclusions and Policy Recommendations

7.1. Research Conclusions

This study examines the impact of city network location on the synergistic effect of PRCM by building a double fixed-effects model using panel data from the three main metropolitan agglomerations of the BTH, the YRD, and the PRD from 2008 to 2019. By fusing the diverse features of various regions with the mediating effect of ecological protection, this research also examines how network location affects the synergistic effect of PRCM. The findings of the study indicate the following:

(1)

Several robustness tests and the instrumental variables approach to endogeneity have not changed the conclusion that the location of urban networks hinders the overall synergistic process of pollution and carbon reduction. This might be because of the network’s central location, which intensifies economic activity and makes it challenging to decouple pollution and carbon emissions in the short term. Additionally, the network’s high concentration of factor resources may crowd out ecological management inputs, leading to “diseconomies of scale.”

(2)

In particular, the location of the city network in I-type cities and southern urban agglomerations presents a significant inhibitory effect on the synergistic effect of PRCM. The results of the heterogeneity analysis demonstrate that the influence of city network location on the synergistic effect of PRCM presents heterogeneity in different city sizes and regional economic organizations. This could be because large cities have more complicated industrial structures, pollution control, and carbon emission reduction have a path dependence, and the scale impact raises the additional marginal cost of environmental control. While northern cities have had a better base for environmental governance and have seen the change of heavy industry earlier, the South has a greater degree of economic outward orientation and bears the environmental pressures brought on by industrial transfer.

(3)

Mechanism Effect Testing reveals that urban network positions primarily influence the synergistic effects of PRCM through two pathways: environmental regulation stringency and research and development investment. Cities in advantageous network positions adopt conservative strategies toward environmental governance and technological investment due to their network structures and environmental compliance costs, manifested as significant reductions in environmental regulation intensity and R&D investment levels. Consequently, these reductions inhibit the promotion, application, and co-development of PRCM technologies. Thus, urban network positions exert suppressive effects on PRCM synergies by diminishing environmental regulation stringency and research and development investment.

7.2. Policy Recommendations

Based on China’s current development stage and the preceding empirical findings, this study proposes the following strategic recommendations to expedite and enhance the efficiency of achieving the “Dual Carbon” goals:

(1)

Optimize Urban Agglomeration Network Structure Hierarchically and Establish Interest-Compatible Mechanisms. Implement a “pilot-first, gradient advancement” strategy. Commence trials in sub-regions with strong existing cooperation foundations, such as Hangzhou Bay and Shenzhen-Dongguan-Huizhou. Establish an “Urban Agglomeration Green Transition Fund” to provide fiscal transfers to core cities bearing environmental governance costs and offer special green development funds to constrained peripheral cities. Concurrently, cultivate secondary node cities (e.g., Xiongan New Area, Jiaxing, Zhuhai) to disperse environmental pressures from core cities through enhanced transportation connectivity and industrial gradient relocation.

(2)

Implement Differentiated Governance Strategies Tailored to Local Conditions. BTH Urban Agglomeration: Establish a “BTH Coordinated Development Commission” led by the National Development and Reform Commission (NDRC) to coordinate environmental governance resource allocation. Implement an “Environmental Carrying Capacity—Economic Development Rights” linkage mechanism, binding Hebei’s industrial undertaking scale to carbon intensity reduction targets. Establish a Green Technology Trading Center in Tianjin and Hebei. YRD Urban Agglomeration: Leverage the Shanghai-Hangzhou-Hefei Science and Technology Innovation Corridor to overcome new energy technology bottlenecks. Pilot cross-provincial mutual recognition of environmental standards and joint law enforcement. Establish a YRD Green Development Bank to provide financial support for cross-regional environmental projects. PRD Urban Agglomeration: Establish a carbon footprint traceability system for export products and implement carbon border adjustment mechanisms (CBAM). Deepen cooperation on green finance within the agglomeration. Establish environmental cooperation mechanisms with Hong Kong and Macao, utilizing external pressures to catalyze the refinement of internal coordination mechanisms.

(3)

Strengthen Supporting Safeguard Systems. Promote the enactment of the Urban Agglomeration Coordinated Development Law to provide a legal foundation for cross-regional environmental governance. Incorporate cross-regional environmental synergy performance into the performance appraisal system for leading officials, establishing differentiated evaluation mechanisms (e.g., increasing the weight of ecological restoration assessment to 30% in BTH). Construct a multi-source urban agglomeration database to support policy simulation and evaluation, and establish a real-time monitoring system for environmental governance effectiveness. Implement a “Personal Carbon Account” system, incentivizing low-carbon public behavior through carbon credits. Establish cross-regional talent exchange mechanisms for environmental governance to enhance professional capacity building.

(4)

Overall Implementation Pathway and Risk Mitigation. Phased Implementation Strategy: Phase I (2025–2027): Focus on establishing institutional frameworks in pilot areas, initiating interest compensation mechanisms, completing the legislative process for the Urban Agglomeration Synergistic Development Promotion Law, and building foundational data platforms and monitoring systems. Phase II (2028–2030): Gradually expand implementation scope based on pilot successes, refine cross-regional coordination mechanisms, deepen market-oriented reforms, and strengthen international cooperation. Phase III (Post-2031): Comprehensively scale up successful experiences, establish institutionalized and normalized urban agglomeration collaborative governance mechanisms, and achieve long-term synergistic development in Pollution and Carbon Reduction Coordination (PRCM). Risk Mitigation System: Policy Implementation Risk Early Warning: Establish a multi-stakeholder coordination and communication platform to promptly identify and resolve interest conflicts and coordination obstacles during implementation. Supervision Safeguard Mechanism: Establish an independent Policy Implementation Oversight Committee to conduct regular effectiveness evaluations, ensuring the tangible implementation of all measures. Emergency Response Mechanism: Develop contingency plans for unexpected situations during policy implementation, activating emergency response and coordination protocols promptly for major environmental incidents or significant policy implementation blockages.

7.3. Research Limitations and Future Directions

The following are the study’s shortcomings that still need to be addressed:

(1)

This study focuses solely on Environmental Regulation Stringency and Research and Development Investment as mediating variables for empirical investigation at the mechanism level. However, the transmission paths between urban network positions and the synergistic effects of PRCM are multifaceted. Future research should explore other theoretically significant pathways for analysis, such as industrial structure and the level of economic agglomeration.

(2)

This paper chooses the data from the three main metropolitan agglomerations for the study to guarantee the study’s relevance when choosing research samples and objects. To conduct more universal and focused research and to more thoroughly examine the impact of network location on the synergistic effect of PRCM, the study could be further refined to the enterprise level in the future or its scope could be extended to the entire nation.

(3)

The construction of network metrics faces limitations regarding endogeneity treatment. Current research employing centrality metrics assumes a relatively stable network structure during the measurement period; however, real-world urban networks exhibit high dynamism, which may lead to an underestimation of the impact of network evolution on synergistic effects. Additionally, PRCM synergies may conversely influence the evolution of cities’ network positions, creating bidirectional causality that existing models struggle to fully exclude.