Governing the Green Transition: A Resource–Pressure Perspective on Urban Low-Carbon Sustainable Development in China

Abstract

Amid the accelerating global green transition, urban Low-Carbon Sustainable Development (LCSD) has emerged as a critical governance challenge. Despite a growing body of research on low-carbon initiatives, the role of local governments in shaping urban LCSD outcomes remains inadequately explored. To address this gap, this study develops a resource–pressure analytical framework that systematically examines how local governments’ resource endowments and pressures jointly condition LCSD. Drawing on panel data from 262 Chinese cities spanning the period 2011–2021, we construct city-level composite indicators of LCSD performance and investigate the underlying driving mechanisms through a combination of statistical analyses and geographically and temporally weighted regression. Our findings yield three principal insights: (1) although overall LCSD has progressed steadily, inter-regional disparities have widened, characterized by persistent structural misalignments and a discernible shift in spatial clustering from the northeast toward southeastern coastal regions; (2) supervisory pressure and economic resources consistently emerge as the most robust and influential determinants of LCSD; and (3) both resource-based and pressure-based drivers display significant spatiotemporal heterogeneity: economic and technological resources exert particularly strong effects in the northwest and central-west regions, respectively, while policy pressures generate differentiated impacts across cities. This research contributes to the theoretical refinement of low-carbon governance frameworks and furnishes robust empirical evidence to inform context-sensitive and regionally differentiated policy design.

1. Introduction

Urban low-carbon sustainable development has become a critical pathway for achieving the United Nations 2030 Sustainable Development Goals (SDGs), particularly Goal 11—“Make cities inclusive, safe, resilient and sustainable”—and Goal 13—“Take urgent action to combat climate change and its impacts.” With the rapid acceleration of global urbanization, cities are increasingly confronted with mounting resource constraints, environmental degradation, and climate-related risks, underscoring the urgent need for low-carbon transformation to attain green growth and sustainable prosperity. Low-Carbon Sustainable Development (LCSD) entails more than optimizing energy structures, enhancing energy efficiency, and advancing green technological innovation [1]. It also requires fundamental institutional reforms and governance innovations across sectors such as economic development and industrial restructuring. However, disparities in resource endowments, governance capacities, and development stages across cities have led to divergent transition pathways and uneven outcomes [2]. Consequently, a pressing challenge in contemporary urban governance is how to foster synergistic progress across economic, social, and environmental goals within the broader framework of the global sustainability agenda [3].

Nevertheless, local governments continue to face substantial implementation challenges. Amid escalating fiscal constraints, technological diffusion barriers, and underdeveloped administrative incentive structures, many cities exhibit insufficient investment in low-carbon initiatives and limited policy enforcement capacity [4]. These issues highlight the urgent need to address the practical dilemmas of urban low-carbon development and to explore effective institutional arrangements and policy tools to support locally grounded, yet globally coherent, sustainable urban transformation. These constraints are particularly salient in China. China has urbanized at an exceptional speed and scale, so city-level carbon governance has immediate policy relevance. At the same time, national climate goals are ambitious, but implementation is largely carried out by local governments. This creates a clear central–local governance tension. Chinese cities also differ widely in fiscal resources, industrial structure, and administrative capacity, making China a valuable setting for identifying how institutional and policy conditions shape low-carbon outcomes across cities.

In response to the pressing challenges of urban low-carbon transitions, a substantial body of literature has emerged both domestically and internationally, focusing on three major dimensions. First, scholars in environmental economics highlight the pivotal role of economic structure [5], energy efficiency [6], and fiscal instruments [7] in driving urban low-carbon transformation. A large number of empirical studies have demonstrated that green technological innovation and carbon emissions trading schemes significantly contribute to curbing urban carbon emissions [8,9].

Second, studies in political science and public administration examine how governmental behavior, governance capacity, and policy experimentation shape urban low-carbon development [10]. Particular attention has been paid to how local governments navigate the tension between central regulatory mandates and local development priorities [11]. These studies often analyze mechanisms such as promotion incentives for local officials, governmental agenda-setting, and government-business relations to reveal the deep influence of urban political dynamics on green governance strategies and policy implementation outcomes [12,13]. Third, research in urban and environmental geography emphasizes the role of urban spatial structure and the built environment in facilitating low-carbon sustainable development. This body of work often focuses on intra-urban inequality at smaller spatial scales, such as neighborhoods or urban grids. For example, scholars explore how the built environment contributes to carbon emission reduction and the mitigation of urban heat island effects [14], how urban expansion patterns relate to green growth [15], and how transportation networks affect the sustainability of urban development [16].

While existing research has provided important theoretical insights into LCSD, it remains limited in several respects [17,18]. Most studies focus on isolated mechanisms—such as policy instruments, governance logic, or the digital economy—without offering a systematic understanding of how local governments make decisions under the combined influence of resource constraints and external pressures [19,20,21]. As the central actors in the low-carbon transition, local governments must simultaneously navigate tightening fiscal budgets, growing technological demands, and increasing societal expectations [22]. However, few studies adopt a government-centered perspective to examine how these actors balance ambitious climate goals with real-world constraints.

To address the identified gaps, this study is guided by the following three research questions:

(1)

What are the spatiotemporal evolution patterns of LCSD across Chinese cities during 2011–2021?

(2)

Which factors significantly affect LCSD, particularly within the proposed resource–pressure analytical framework?

(3)

To what extent do the effects of these factors exhibit spatial and temporal heterogeneity across cities and periods?

Anchored in these questions, the study constructs a multidimensional LCSD evaluation framework and employs panel-based spatiotemporal analytical methods to identify both aggregate trends and heterogeneous driving mechanisms. The remainder of this paper is organized as follows. Section 2 introduces the theoretical framework. Section 3 outlines the research design, describes the data sources and details the operationalization of variables. Section 4 presents the empirical results. Section 5 discusses the findings, interpreting their theoretical implications, and the mechanisms through which local capacity shapes outcomes. Section 6 concludes by summarising the main contributions, offering practical recommendations, and suggesting directions for future research.

2. Literature Review and Theoretical Framework

Although existing studies have examined the drivers of LCSD from multiple perspectives—including the digital economy [23], government attention [24] and corporate emissions [25]—they still lack a systematic framework for understanding LCSD through the lens of local government behavior. While scholarly efforts to identify the determinants of urban low-carbon transitions vary in approach and emphasis, two fundamental and widely implicit hypotheses deserve closer attention. The first is the resources hypothesis [26], which posits that the resource endowments possessed by local governments are critical to sustainable development outcomes: cities with stronger foundational capacities are more likely to experience smoother transitions. The second is the pressure hypothesis [27], which contends that external and internal governance pressures serve as key drivers pushing local governments to accelerate low-carbon transformation. Building upon these theoretical foundations, this study constructs a “resources–pressure” analytical framework from the perspective of local government behavior [28], thereby offering a more integrated analysis of the forces shaping LCSD.

2.1. Resources Hypothesis

Local governments possess two broad categories of resources: institutional resources and self-acquired resources. While institutional resources are generally accessible to all localities [29], the key differentiating factor among local governments lies in the scope and effectiveness of their self-acquired resources. These refer to the resources proactively obtained by local governments through various strategic means, and can be further divided into internal resources and resources drawn from societal actors [30]. Our first conceptual innovation lies in integrating these resource categories into a unified framework that explicitly links local government capacity to low-carbon governance outcomes. By doing so, this study advances governance theory by highlighting the interplay between resource endowments and policy prioritization, offering a novel lens to understand how resource heterogeneity shapes environmental governance.

First, internal resources constitute a critical foundation for LCSD. Embedded within the broader “Beautiful China” strategy, LCSD must compete with diverse policy priorities for limited fiscal, planning, and implementation resources. Drawing on the principles of fiscal federalism [31], local governments operate within a multi-tiered fiscal system, in which their autonomy in budgetary allocation and reliance on intergovernmental transfers significantly influence the prioritization of environmental expenditures. In this framework, the green orientation of fiscal spending serves as a salient indicator of a local government’s commitment and capacity for low-carbon governance. Prioritizing environmental expenditure reflects an endogenous budgeting choice, signaling the extent to which green development objectives are institutionalized in local decision-making [32]. Simultaneously, such investment provides essential material support for green infrastructure, industrial upgrading, and long-term transition pathways, thereby enhancing the overall effectiveness of low-carbon policies [7].

Hypothesis H1a.

Local governments with higher fiscal autonomy are more likely to prioritize environmental expenditures, leading to better LCSD.

Beyond fiscal resources, institutional resources—defined as the completeness, coherence, and enforceability of local governance frameworks—represent a critical determinant of low-carbon performance. Drawing on institutional theory, these resources reflect the extent to which formal rules, procedures, and norms are established, internalized, and consistently applied within local governance systems [33]. They encompass not only the comprehensiveness and logical coherence of policy frameworks but also the standardization of enforcement mechanisms and the stability of institutional arrangements. According to Institutional Theory, well-structured and enforceable institutions shape the behavior of local actors by creating expectations, reducing uncertainty, and embedding policy routines. In the context of LCSD, a high-quality institutional system facilitates a favorable governance environment for low-carbon transition, enhances the efficiency and predictability of policy implementation, and promotes the development of policy feedback loops and institutional path dependencies—thereby supporting sustainable, long-term low-carbon urban development [34].

Hypothesis H1b.

Local governments with higher institutional quality, characterized by coherent and enforceable governance frameworks, will positively influence the effectiveness of LCSD.

Second, societal resources, defined as technological and human resources mobilized by local governments from enterprises and citizens, also play a pivotal role in supporting LCSD. In the current governance landscape, the relationships between governments, enterprises, and citizens have shifted from hierarchical control to collaborative, multi-actor partnerships.

Technological resources are indispensable for low-carbon transformation. Grounded in the innovation systems perspective, these resources encompass not only specific technological capabilities but also the networked interactions among enterprises, research institutions, and governmental actors that facilitate knowledge diffusion, learning, and innovation [35,36,37]. With the advancement of green technologies, enterprises have achieved notable progress in energy conservation, emissions reduction, industrial restructuring, and environmental innovation. These technological breakthroughs not only provide essential support for local governments’ policy implementation but also promote green industrial upgrading and sustainable economic growth. From the lens of Innovation Systems, corporate green innovation strengthens the technical capacity, adaptive capability, and resilience of local governance systems, thereby enhancing the efficiency, effectiveness, and systemic integration of low-carbon policy execution.

Hypothesis H1c.

Local governments that effectively mobilize technological resources will demonstrate higher efficiency and in achieving LCSD.

Additionally, human resources drawn from society constitute another key pillar of low-carbon governance. From a human capital perspective, the knowledge, skills, and expertise possessed by citizens, professional teams, and experts represent critical assets that enhance the capacity and effectiveness of local governments in implementing green policies [38,39]. Citizens’ environmental awareness, low-carbon behaviors, and civic engagement directly shape policy adoption and enforcement, while specialized professionals in environmental governance and green technologies provide essential technical guidance, consultancy, and operational support. By leveraging these human capital assets, local governments can improve policy precision, ensure procedural fairness, and facilitate the diffusion of low-carbon values throughout society, thereby fostering a more resilient and participatory governance ecosystem.

Hypothesis H1d.

Local governments that leverage human resources will achieve greater policy precision and resilience in LCSD.

2.2. Pressure Hypothesis

Governance pressure on local governments manifests in multiple dimensions, each grounded in well-established theoretical frameworks. Our second conceptual innovation is the synthesis of these multi-dimensional pressures into a cohesive framework that elucidates how external and internal forces jointly shape local low-carbon governance. This approach contributes to environmental political economy by unpacking the complex interplay of hierarchical, competitive, societal, and individual incentives in driving or hindering sustainable urban development.

Supervisory pressure, derived from higher-level authorities, aligns with bureaucracy theory, which emphasizes hierarchical control, rule-based governance, and institutionalized procedures [40]. Provincial governments impose regulatory mandates, performance evaluations, and institutional constraints on municipal administrations, shaping local decision-making and policy implementation. In environmental governance, these vertical pressures institutionalize low-carbon objectives, operationalized through inspections, formal regulations, and reporting systems. Such bureaucratic oversight creates path-dependent governance patterns, stabilizes local institutional arrangements, and ensures predictable, efficient, and enforceable execution of low-carbon policies, fostering sustained local capacity for environmental transformation [41].

Hypothesis H2a.

Higher levels of supervisory pressure from provincial governments lead to more effective LCSD outcomes at the local level.

Horizontal or peer pressure emerges from competition among similarly ranked jurisdictions, best explained by the yardstick competition theory [42]. Municipal governments continuously observe and benchmark the environmental strategies and outcomes of neighboring cities, emulating high-performing peers and avoiding relative underperformance. This lateral pressure promotes policy learning, encourages innovation in green governance, and accelerates the diffusion of sustainable practices across regions. By leveraging interjurisdictional comparison and reputational incentives, yardstick competition amplifies local governments’ commitment to effective low-carbon initiatives and strengthens regional policy convergence in pursuit of sustainable development [43].

Hypothesis H2b.

The presence of horizontal (peer) pressure among municipalities enhances the and effectiveness of LCSD.

Simultaneously, public pressure constitutes a bottom-up governance mechanism, grounded in public choice theory, which emphasizes citizen participation, social accountability, and collective monitoring of governmental behavior [44]. Citizens, civil society actors, and the media increasingly scrutinize environmental quality and policy outcomes, raising awareness of ecological issues and exerting social sanctions when local authorities underperform [45]. This participatory oversight fosters transparency, procedural fairness, and inclusive governance, channeling societal expectations into formal and informal policy processes. Public pressure complements hierarchical supervision and peer benchmarking, reinforcing a multi-layered governance environment conducive to sustainable low-carbon transitions.

Hypothesis H2c.

Increased public pressure positively correlates with the effectiveness of LCSD.

Finally, official-internal pressure reflects career-driven incentives among local officials, which are well captured by tournament theory [46,47]. In China’s cadre evaluation system, promotion prospects motivate local officials to prioritize short-term, high-visibility economic achievements over long-term low-carbon investments, despite increasing ecological indicators in performance assessments. This internalized pressure shapes policy priorities, influencing both the speed and scale of local interventions and creating a tension between immediate performance gains and sustainable development objectives.

Hypothesis H2d.

Official-internal pressure negatively affects the prioritization of long-term low-carbon investments by local governments, thereby hindering LCSD.

The resource–pressure framework integrates both the internal and societal drivers of LCSD with multi-dimensional governance pressures to provide a comprehensive explanation of local policy dynamics. As illustrated in Figure 1, this framework captures the interactive effects of resources and pressures on the effectiveness, efficiency, and sustainability of low-carbon governance.

3. Research Methodology

3.1. Analytical Strategy

Based on balanced panel data from 262 Chinese cities over the period 2011–2021, this study first employs the entropy method to objectively measure the level of LCSD across cities. Subsequently, it utilizes the Mann–Kendall trend test, spatial autocorrelation analysis, and the Dagum Gini coefficient to examine the spatiotemporal evolution of LCSD at the city level. Finally, the geographically and temporally weighted regression model is applied to investigate the spatiotemporal heterogeneity of LCSD drivers from the perspective of local government behavior.

3.1.1. Mann–Kendall Trend Test Method

To evaluate the temporal evolution of LCSD in Chinese cities from 2011 to 2021, this study adopts the Mann–Kendall (MK) trend test, a non-parametric method frequently used in urban sciences [48]. The MK test is particularly suitable for identifying consistent upward or downward trends in time series data without requiring the assumption of normality, making it highly applicable in the context of sustainability analysis. Specifically, the method examines the direction and consistency of change by comparing all possible pairs within the data series. The test procedure is formulated as follows:

$$S=\sum _{i=1}^{n-1}\sum _{j=i+1}^{n}sign\left(x_j-x_i\right)$$

(1)

The sign function indicates whether $x_j$ is greater than, less than, or equal to $x_i$, with $x_i$ and $x_j$ being the annual mean of LCSD. If the $S$ value is positive, it indicates an improvement in LCSD in certain cities over the study period and vice versa.

3.1.2. Spatial Autocorrelation

Spatial autocorrelation analysis is a key statistical approach for assessing the degree of similarity or spatial dependence among geographic units [49]. It plays a crucial role in evaluating the spatial distribution characteristics of variables across regions. In this study, we utilize spatial autocorrelation to investigate the spatial pattern of LCSD among Chinese cities. The global Moran’s I index is widely used to quantify overall spatial autocorrelation. It is computed based on both attribute similarity and spatial proximity, using the following expression:

$$I={\displaystyle \frac{n\sum _{i=1}^n \sum _{j=1}^n w_{ij}\left(x_i-\bar{x}\right)\left(x_j-\bar{x}\right)}{\sum _{i=1}^n \sum _{j=1}^n w_{ij}\sum _{i=1}^n {\left(x_i-\bar{x}\right)}^2}}$$

(2)

where $n$ is the total number of cities (262), $x_i$ and $x_j$ are the LCSD at cities $i$ and $j$, $\bar{x}$ is the mean of the LCSD, and $w_{ij}$ is the spatial weight between cities $i$ and $j$. The value of Moran’s I ranges between −1 and 1: positive values indicate positive spatial autocorrelation (clustering), negative values indicate negative spatial autocorrelation (dispersion), and values near zero suggest a random spatial distribution.

To further investigate the local spatial structure of LCSD, we apply the Getis-Ord Gi* statistic for hot spot analysis. This method identifies statistically significant spatial clusters of high (hot spots) and low (cold spots) values, thereby revealing the spatial polarization patterns within the study area [50].

3.1.3. Dagum Gini Coefficient

Compared with traditional inequality measures, the Dagum Gini coefficient allows for simultaneous analysis of overall disparities and their structural sources, making it well-suited for this study [51]. To assess the spatial inequality of LCSD across 262 Chinese cities, this study applies Dagum’s decomposition method: $G=G_W+G_{nb}+G_t$. Cities are grouped into three regions—east, central, and west. $G_W$ represents the within-region disparity, $G_{nb}$ represents the between-region disparity, and $G_t$ represents the super-variation density.

3.1.4. Geographically and Temporally Weighted Regression (GTWR)

GTWR model is an extension of traditional spatial regression that incorporates both spatial and temporal dimensions. It is particularly suited for analyzing datasets characterized by spatiotemporal heterogeneity, allowing for localized modeling across both space and time. GTWR has been widely applied in spatial analysis and regression modeling, especially when the underlying data-generating processes vary across geographic locations and time periods [52].

Unlike conventional regression methods, GTWR accounts for spatial and temporal variations in the relationships between dependent and independent variables by allowing the regression coefficients to change over both space and time. The fundamental idea is to assign greater weight to observations that are closer in both geographic and temporal proximity. The general form of the GTWR model is expressed as:

$$y_i={\beta }_0\left(u_i,v_i,t_i\right)+\sum _{k=1}^p{\beta }_k\left(u_i,v_i,t_i\right)x_{ik}+{\epsilon }_i$$

(3)

In this equation, $y_i$ represents the observed value of LCSD for city $i$; $\left(u_i,v_i,t_i\right)$ denote the spatial coordinates and time index of city iii; ${\beta }_0\left(u_i,v_i,t_i\right)$ is the intercept term that varies across space and time; ${\beta }_k\left(u_i,v_i,t_i\right)$ is the spatiotemporally varying coefficient of the $k$ independent variable; $x_{ik}$ is the value of the $k$ explanatory variable for city $i$; ${\epsilon }_i$ is the random error term. The model thus captures the localized, non-stationary relationships between LCSD and its determinants. In this study, the GTWR model employs a fixed kernel, with a fixed bandwidth specification. The optimal bandwidth is determined by minimizing the AICc criterion through an automatic search within a predefined bandwidth range.

3.2. Variables and Data

3.2.1. Dependent Variable

Characterized by its multidimensional structure and spatial-temporal heterogeneity, LCSD poses significant challenges in terms of standardized measurement and cross-regional comparability. Addressing this empirical and methodological lacuna, this study draws upon established literature [53,54] to construct a comprehensive evaluation framework grounded in the dual imperatives of economic development and carbon reduction.

Specifically, the framework encompasses four core dimensions—low-carbon economy, energy, society, environment—and incorporates 15 representative, quantifiable indicators to systematically capture the diverse facets of LCSD (Figure 2). These dimensions collectively reflect a city’s capacity for industrial upgrading, energy system optimization, behavioral transformation, and low-carbon transportation restructuring, aligning with key theoretical and policy objectives in sustainability science.

The LCSD indicators were primarily constructed using authoritative national sources, including the China urban statistical yearbook (2012–2022) and the China environmental statistical yearbook (2012–2022). To improve coverage, missing city-level data were supplemented with information from municipal statistical reports and official bulletins. Where city-level data remained unavailable, provincial-level data were used, and any remaining gaps were filled through linear interpolation to ensure completeness and temporal consistency.

To ensure methodological rigor and transparency, we provide a step-by-step account of the index construction process. First, raw data were pre-processed through normalization to ensure comparability across heterogeneous units. For positive indicators, the standardized value $Z_{ijt}$ is calculated as follows (with the numerator inverted for negative indicators):

$$Z_{ijt}={\displaystyle \frac{x_{ijt}-\mathit{min}\left(x_j\right)}{\mathit{max}\left(x_j\right)-\mathit{min}\left(x_j\right)}}$$

(4)

Second, the Global Entropy Weighting Method was applied to derive indicator weights. Unlike traditional approaches, this method calculates information entropy based on the panel data of all cities across all years to derive objective weights ($w_j$). Finally, the LCSD index is computed through linear aggregation:

$${\mathit{LCSD}}_{it}=\sum _{j=1}^m{w}_j\times Z_{ijt}$$

(5)

where $Z_{ijt}$ represents the standardized value of indicator $j$ for city $i$ in year $t$, and $w_j$ is the global weight derived from the entropy method.

3.2.2. Independent Variables

Government economic resources. This study employs the proportion of local government expenditure on environmental protection relative to total fiscal expenditure to measure economic resources available for promoting low-carbon and sustainable development. Fiscal spending on environmental protection reflects the financial commitment and resource mobilization capacity of municipal governments in addressing environmental challenges and supporting green transformation [55]. The data are sourced from the China City Statistical Yearbook, which reports detailed public finance statistics at the city level. This indicator has been widely used in environmental economics literature as a proxy for the economic capability and environmental governance intensity of local governments [56,57].

Government institutional resources. Government institutional resources are measured by the intensity of low-carbon policies enacted by municipal authorities. Drawing on established methodologies in the literature [58], this study employs a text mining approach to systematically analyze official policy documents related to low-carbon development. Each policy is evaluated in terms of its content, regulatory strength, and implementation mechanisms, generating a quantitative score that reflects its potential impact. These scores are then aggregated at the city level to construct a composite index of policy intensity, capturing the institutional efforts and regulatory commitment of local governments in promoting low-carbon urban transitions. This approach provides a robust and replicable measure of institutional support for urban low-carbon sustainable development.

Government technological resources. This study uses the annual number of green invention patents granted in each city to measure the level of technological resources held by local governments. Green invention patents reflect the capacity for high-quality innovation in environmentally friendly technologies, which is critical for driving low-carbon and sustainable development [59,60]. Patent data are collected from the China National Intellectual Property Administration (CNIPA), and green patents are identified based on the WIPO’s International Patent Classification (IPC) Green Inventory. The application and grant of green patents are often influenced by public R&D investment and local policy support, making them a valid proxy for the technological resources mobilized or guided by governments in the context of low-carbon transitions.

Government human resources. This study uses the number of employees (in tens of thousands) engaged in the management of water conservancy, environmental protection, and public facilities to measure the human resources available to city governments for supporting low-carbon and sustainable development. These sectors represent key functional departments responsible for implementing environmental policies, managing ecological infrastructure, and providing essential public services. A larger workforce in these areas reflects stronger administrative capacity and operational readiness to carry out environmental governance tasks. The data are obtained from the China City Statistical Yearbook, which annually reports employment figures by industry at the city level.

Supervisory pressure from higher-level governments. Supervisory pressure from higher administrative levels is proxied by the environmental attention of provincial governments. Following established research practices [61], for each year we calculate the total number of words (excluding stop words) in provincial government work reports and the frequency of environment-related keywords (The keywords for this study encompass a range of environmental and sustainability topics, including haze and smog control, key air pollutants such as SO₂, CO₂, PM10, PM2.5, dust emissions, vehicle exhaust, industrial emissions, VOCs, and indicators of air quality such as blue sky.), then compute the ratio of keyword occurrences to total words as an annual indicator of environmental attention. This ratio reflects the relative salience of environmental issues in provincial-level political agendas and is interpreted as a form of top-down pressure exerted on municipal governments. The underlying rationale is that greater emphasis by higher-level governments on environmental concerns induces local governments to respond more actively in policy implementation and administrative action related to low-carbon development.

Competition pressure among local governments. Competition pressure among local governments is measured by the LCSD levels of other cities within the same province. Drawing on competitive dynamics frameworks in regional governance, this study uses the LCSD index of neighboring cities to capture the competitive pressures faced by municipal governments [62,63]. The assumption is that cities are in constant competition for resources, investment, and political legitimacy, leading them to closely monitor and emulate the low-carbon practices and policies of other cities in the same province. A higher LCSD performance by neighboring cities signals greater pressure for a city to adopt and strengthen its own low-carbon initiatives, fostering a competitive environment for sustainable development.

Citizen attention pressure. In this study, citizen attention pressure is proxied by public environmental concern, measured using the Baidu Search Index. The index is available by search channel—total search index, PC-based, and mobile-based—with the total search index calculated as the weighted sum of the PC and mobile indices. Following prior studies on public engagement and environmental governance [64,65], we selected a set of environmental keywords, including terms such as “smog,” “environmental pollution,” and “carbon emissions.” The daily search indices for these keywords are collected and aggregated to obtain annual averages for each city, capturing both temporal trends and cross-city variations in public attention. Higher index values indicate stronger public concern, generating societal pressure on local governments to implement policies for low-carbon and sustainable development.

Promotion pressure of local officials. Promotion pressure is measured using the tenure length of municipal Party secretaries, drawing on official government websites and publicly available search records. Based on findings from political economy and environmental governance research [66,67], longer tenure is associated with stronger promotion incentives for local officials. When tenure increases, local leaders are more likely to prioritize measurable economic growth over long-term environmental goals in order to signal political performance to higher authorities. Therefore, tenure length serves as a proxy for upward mobility pressure, which may crowd out attention to sustainable development and low-carbon initiatives.

In summary, this study constructs a balanced panel dataset of 262 Chinese cities from 2011 to 2021. Missing values for certain variables are addressed through linear interpolation. The final descriptive statistics are presented in

Table 1. Descriptive statistics of variables.

Variables	N	Mean	Sd	Median	Min	Max
LCSD	2882	0.33	0.23	0.26	0.00	0.99
economic resources	2882	0.01	0.00	0.01	0.00	0.03
institutional resources	2882	2.35	3.88	0.00	0.00	46.75
technological resources	2882	124.75	481.69	16.00	0.00	10,010.00
human resources	2882	228.85	251.94	145.39	0.00	2460.70
supervisory pressure	2882	0.01	0.00	0.01	0.00	0.02
competition pressure	2882	0.33	0.12	0.33	0.05	0.82
attention pressure	2882	24.10	26.93	13.74	0.00	164.51
promotion pressure	2882	2.60	1.64	2.00	1.00	10.00

, and detailed definitions and measurements of all variables are provided in the Table S1.

4. Results and Analysis

4.1. Spatiotemporal Distribution of Low-Carbon Sustainability Development

As shown in Figure 3, we present the overall characteristics of LCSD in urban areas of China. Due to limitations in the available data, the regions of Tibet, Hong Kong, Macau, and Taiwan have been excluded from this study. According to the MK test results shown in Figure 3a, the LCSD in China exhibits an initial period of fluctuating increase, followed by a continuous upward trend after 2017. However, the MK test did not indicate a statistically significant trend, suggesting that the overall rise is more a result of fluctuations rather than a sustained increase. This trend is likely driven by national policies such as carbon emissions trading and low-carbon city initiatives, reflecting an increasing nationwide emphasis on LCSD during this period [68,69].

Based on the analysis presented in Figure 3b, this study further examines the inter-annual variations in LCSD across the eastern, central, and western regions of China. Overall, the LCSD trends in all three regions exhibit fluctuations with a general upward trajectory, maintaining consistency. However, the eastern region consistently holds the highest LCSD values, indicating that, as a leading region in LCSD, it has made substantial progress and achieved significant outcomes. In contrast, the central and western regions show relatively similar inter-annual variations, with some degree of overlap between their trends. Nevertheless, both regions lag behind the eastern region and this gap appears to be widening, potentially signaling an increasing regional inequality in LCSD.

Figure 3c illustrates the spatial and temporal patterns of LCSD across Chinese provinces from 2011 to 2021. Overall, municipalities such as Beijing, Shanghai, and Tianjin consistently maintained high LCSD levels, highlighting their leading roles in the low-carbon transition. Beijing reached the highest value in 2019 (0.953), reflecting strong performance in sustainable development, while Qinghai recorded the lowest in 2017 (0.0196), indicating notable regional disparities. At the city level, a similar pattern emerges: top-ranking cities, including Fushun, Fuxin, and Ziyang, consistently scored above 0.97, whereas cities such as Linfen, Yuncheng, and Jincheng remained below 0.03, emphasizing pronounced heterogeneity within provinces. In terms of improvement, Chongqing showed the greatest increase, with its LCSD rising from 0.175 in 2011 to 0.859 in 2021, suggesting steady progress in its low-carbon development. In contrast, provinces like Gansu and Guangdong experienced slight declines, indicating ongoing challenges in certain regions.

Following the spatiotemporal mapping of LCSD in Chinese cities from 2011 to 2021, we further apply the Dagum Gini coefficient to examine regional disparities and their evolution. As shown in

Table 2. The Dagum Gini coefficient of the LCSD in China’s cities from 2011–2021.

Year	Overall G	Intra-Group Differences			Inter-Group Differences			Contribution Rate (%)
		East	Central	West	C-E	W-E	W-C	${\mathit{G}}_{\mathit{w}}$	${\mathit{G}}_{\mathit{n}\mathit{b}}$	${\mathit{G}}_{\mathit{t}}$
2011	0.34	0.30	0.35	0.37	0.34	0.34	0.36	13.9	33.3	52.8
2012	0.35	0.31	0.37	0.38	0.35	0.35	0.37	13.1	33.4	53.6
2013	0.31	0.28	0.34	0.32	0.32	0.30	0.33	11.2	33.7	55.1
2014	0.34	0.28	0.38	0.34	0.34	0.31	0.37	1.5	33.7	64.7
2015	0.35	0.28	0.42	0.33	0.36	0.31	0.38	7.8	33.6	58.6
2016	0.34	0.28	0.40	0.33	0.34	0.31	0.37	8.1	33.8	58.1
2017	0.36	0.30	0.40	0.36	0.37	0.35	0.38	16.0	33.2	50.8
2018	0.37	0.32	0.40	0.37	0.37	0.36	0.39	17.1	33.1	49.8
2019	0.37	0.33	0.38	0.39	0.38	0.37	0.39	18.2	33.0	48.9
2020	0.38	0.35	0.39	0.39	0.38	0.38	0.39	10.1	33.8	56.0
2021	0.39	0.34	0.39	0.42	0.38	0.39	0.41	12.4	33.4	54.2

, the overall Gini coefficient rose from 0.34 in 2011 to 0.39 in 2021, indicating an intensification of spatial inequality in LCSD—consistent with the spatial pattern shown in Figure 3, where eastern cities consistently lead, and central-western regions lag.

Intra-regional disparities are most pronounced in the western region, suggesting a dual developmental dilemma for some cities. Inter-regionally, Gini coefficients between East–Central, West–East, and West–Central regions remain high, with relatively small differences between the central and western regions, reflecting shared developmental constraints.

In terms of inequality sources, intra-regional contributions remain low (mostly <20%), while inter-regional and transvariation contributions are higher—around 33% and above 50%, respectively, with the latter peaking at 54.2% in 2021. This indicates that inequality arises not only from mean differences but also from overlapping distributions and entrenched hierarchies—where underperforming cities exist in leading regions and vice versa.

Overall, while LCSD levels improved during 2011–2021, regional disparities widened. This trend aligns with earlier analyses, highlighting the increasing developmental divide and internal heterogeneity, particularly within central and western regions.

4.2. Spatial Autocorrelation Analysis of Low-Carbon Sustainability Development

To investigate whether the LCSD levels in Chinese cities exhibited spatial clustering, we calculated the global Moran’s I for the period 2011–2021. As shown in

Table 3. Global Moran’s I for LCSD during 2011–2021.

Year	Moran’s I	E(I)	sd(I)	Z Score	p-Value
2011	0.054	−0.004	0.000	5.538	0.000
2012	0.041	−0.004	0.000	4.309	0.000
2013	0.043	−0.004	0.000	4.483	0.000
2014	0.074	−0.004	0.000	7.380	0.000
2015	0.094	−0.004	0.000	9.302	0.000
2016	0.093	−0.004	0.000	9.246	0.000
2017	0.077	−0.004	0.000	7.688	0.000
2018	0.065	−0.004	0.000	6.543	0.000
2019	0.055	−0.004	0.000	5.581	0.000
2020	0.059	−0.004	0.000	5.940	0.000
2021	0.050	−0.004	0.000	5.090	0.000

, all values of Moran’s I are positive and statistically significant $(p < 0.01)$, indicating a clear spatial autocorrelation in LCSD—suggesting that cities with similar LCSD performance tend to cluster geographically. The Moran’s I peaked in 2015 at 0.094 and declined slightly thereafter, yet remained statistically significant throughout. This pattern suggests that spatial dependence strengthened in the mid-period and slightly weakened in later years, but overall, the spatial distribution of LCSD did not become random, and spatial interconnectedness among cities remained relatively strong.

To further explore the spatial heterogeneity of LCSD across Chinese cities, the Getis-Ord Gi* statistic was applied to identify spatial clusters of significantly high or low values, as shown in Figure 4. In the first half of the study period (2011–2015), the northeastern region experienced a notable transition from non-significant areas to strong hot spots, indicating relatively high LCSD levels during this phase. In contrast, the central region, particularly the Beijing-Tianjin-Hebei area, exhibited persistent cold spots with expanding spatial extent, reflecting lagging LCSD performance.

During the second half of the period (2017–2021), the northeastern hot spots gradually weakened, eventually turning into non-significant or even cold spot areas. At the same time, cold spot areas in the central region began to contract, suggesting marginal improvements. Meanwhile, southeastern coastal cities evolved into emerging LCSD hot spots, indicating substantial progress in low-carbon development and sustainability achievements.

In sum, the hotspot analysis based on the Gi* statistic enriches our understanding of the spatial and temporal heterogeneity of LCSD. It captures a clear regional shift—from early-stage advantages in the northeast to later-stage dominance in the southeast—revealing both the dynamics and disparities of low-carbon development across urban China.

4.3. The Role of Local Government Behavior in Driving Low-Carbon Sustainability Development

Building upon the identified spatial clustering characteristics of LCSD, this study further investigates its driving mechanisms from the perspective of local government behavior. To mitigate potential multicollinearity among explanatory variables, Variance Inflation Factor (VIF) tests were conducted, confirming the absence of serious collinearity. Given the spatiotemporal non-stationarity in the relationship between LCSD and its influencing factors, the GTWR model was employed. Compared with OLS and GWR models (

Table 4. Comparison of model performance.

Model	${\mathit{R}}^2$	$\mathbf{Adjusted} {\mathit{R}}^2$	AICc
OLS	0.182	0.182	−961.668
GWR	0.308	0.306	−1305.490
GTWR	0.350	0.348	−1366.750

), the GTWR model exhibits superior explanatory power and lower AIC values, validating its suitability for capturing the dynamic effects of local government behavior on LCSD.

We then explore the temporal evolution of governmental behavioral drivers underlying LCSD. We standardize the coefficients of the eight independent variables derived from the GTWR model and examine changes in the relative importance of driving factors within the resource–pressure framework, both overall and across individual years. As illustrated in Figure 5, the distribution of variable importance exhibits a clear structural pattern at the aggregate level. Supervisory pressure ranks first with an average relative weight of 52.00%, indicating its overwhelming dominance within the mechanism of organizational behavior. This result suggests that institutional constraints, superior oversight, and accountability mechanisms play a substantially greater role in shaping LCSD outcomes than other types of resources or pressures [70]. Economic resources follow with a weight of 26.20%, reflecting the foundational importance of fiscal capacity, investment intensity, and industrial structure. These economic fundamentals constitute the practical basis upon which local governments implement low-carbon policies. The strength of a locality’s economic resource base directly shapes both the efficiency and effectiveness of LCSD, providing critical support for environmental protection efforts [71]. In contrast, the remaining variables contribute relatively little to the overall structure, each accounting for approximately 2.7%. Although these factors are theoretically relevant, they fail to function as systematic drivers or constraints in the practical governance of urban low-carbon transition.

The year-by-year dynamic analysis further reveals the structural stability and phase-specific adjustments within the dual resource–pressure framework. Since 2011, supervisory pressure has consistently dominated the overall configuration, maintaining a relative weight around 50%, demonstrating the persistent role of top-down enforcement in carbon governance. The trajectory of economic resources follows a “decline-then-rise” pattern: its weight fell from 34.44% in 2011 to 20.61% in 2014 amid intensifying economic downturn pressures and fiscal strain at the local level. However, its importance gradually rebounded to 30.18% by 2021, driven by the rise of green finance tools and enhanced central government fiscal support for green transition during the 14th Five-Year Plan [72]. This trend highlights that economic resources are not only critical enablers of local low-carbon transition but are also closely tied to fiscal capacity, national policy orientation, and external macroeconomic conditions. Meanwhile, competition pressure shows a clear downward trend, decreasing from 7.44% in 2011 to just 1.59% in 2021. This decline reflects the weakening of inter-local green competition. On the one hand, the motivation for local governments to “catch up” in green development has diminished due to increasing uniformity in performance assessments and emission targets [73]. On the other hand, as green development becomes normalized, the governance model is shifting from competitive dynamics toward cooperation and standardized regulation [74].

The remaining variables exert relatively limited influence on LCSD promotion. Technological resources increased marginally from 2.46% in 2011 to 2.67% in 2021. Although technology constitutes a critical foundation for green transformation, local governments’ capacity in terms of innovation and technical reserves remains constrained. Institutional resources have remained relatively stable at around 2.67%, indicating a consistent but limited accumulation in local policy tools and governance capacities. Promotion pressure declined slightly from 2.13% in 2011 to 1.99% in 2021, underscoring the diminishing role of traditional GDP-centered promotion incentives amid the institutional shift toward green performance evaluation and accountability mechanisms [75]. Human resources, with a steady weight around 2.62%, suggest that human capital and professional capacity building have not yet emerged as central drivers of green transformation. Finally, attention pressure also remained stable at around 2.62% throughout the period. Although its weight is comparatively low, it reflects the potential influence of public opinion and civil society on the implementation of low-carbon policies by local governments.

Under the resource–pressure framework, this study further uses the natural breaks classification to divide the coefficients into five classes, thereby revealing the spatial heterogeneity and intensity of how both resource and pressure factors influence LCSD. As illustrated in Figure 6, different variables exert markedly varied spatial effects.

Regarding resource factors, technological resources show significant positive impacts in central and western China, indicating that green technological advancement in less-developed regions can effectively enhance LCSD. Economic resources are particularly influential in the northeast and northwest, with Urumqi reaching a maximum coefficient of 12.10, underscoring the critical role of fiscal capacity. In contrast, human resources generally exhibit weak effects, underscoring the pivotal role of financial capacity in facilitating environmental governance. Institutional resources display strong spatial divergence—positive in the northeast (e.g., Heihe, 0.007), negligible in central China, and negative in some parts of the west—highlighting variations in institutional adaptability and governance efficiency.

For pressure factors, supervisory pressure presents a polarized pattern: it negatively affects LCSD in places like Hulunbuir (−24.71), but positively in Baotou (13.22). This divergence supports the notion that the outcome of external oversight is contingent on local capacity. In regions with weak governance capabilities, intense top-down supervision may trigger “institutional overload,” where local governments lack the administrative and financial resources to genuinely comply. Competition pressure is broadly positive, though its negative effect in cities likely reflects maladaptive competition under structural constraints, where a “race-to-the-bottom” on costs or over-investment in specific sectors harms holistic sustainability. Citizen attention pressure, representing public environmental concerns, has limited spatial variance but notably promotes LCSD in key political regions such as Beijing–Tianjin–Hebei. Lastly, promotion pressure manifests negative effects in some southern cities (e.g., Sanya, −0.022), suggesting a potential misalignment between bureaucratic career incentives and environmental outcomes, which may foster strategic decoupling or performance gaming by local officials.

Overall, the GTWR results reinforce the core argument that LCSD is jointly determined by the interaction between resource endowments and multi-level pressures. However, as the direction and magnitude of the coefficients vary significantly across space—with some even showing unexpected negative effects—Hypotheses 1 and 2 are only partially supported. This heterogeneity underscores that the green transition is not driven by a singular policy logic but emerges through context-specific interactions of resources and constraints.

5. Discussion

5.1. Spatiotemporal Inequality of Low-Carbon Sustainability Development

In China, low-carbon sustainable development is not only a necessary response to global climate change but also an important goal for the nation’s socio-economic development. However, from a spatiotemporal perspective, the distribution of LCSD exhibits significant inequality. Our results show that there is a considerable imbalance in the level of LCSD between different regions, with the highest levels concentrated in the economically developed eastern regions, while the central and western regions exhibit lower levels of low-carbon development. This finding aligns with similar observations in other studies [76,77].

In terms of spatiotemporal distribution, LCSD demonstrates a marked spatial clustering pattern. Eastern coastal cities, supported by stronger economies, technological advantages, and governance capacity, achieve higher levels of low-carbon development and readily form clusters under policy incentives. In contrast, western and inland regions face fiscal constraints, weaker industrial bases, and slower technological adoption, which hinder their transition [78]. This uneven distribution not only reflects economic disparities but also institutional and infrastructural differences, and further reveals a path dependence in which leading regions accumulate advantages while lagging ones risk a vicious cycle of weak capacity and low performance [79]. Therefore, narrowing regional gaps requires not only resource transfers and supportive policies but also the enhancement of local innovation capacity, institutional reform, and cross-regional cooperation to achieve equitable and sustainable low-carbon development.

Second, the results of spatial autocorrelation analysis further support this conclusion. Through both global and local spatial autocorrelation analysis, this study finds that there is a strong spatial correlation in LCSD levels, meaning that regions with higher levels of low-carbon development are often closely connected with neighboring regions. This spatial clustering phenomenon reveals the systemic spatial inequality in low-carbon sustainable development, which is consistent with the spatial heterogeneity [80]. Additionally, local spatial autocorrelation analysis also identifies some regions that are relatively lagging in low-carbon development, which are geographically isolated and lack endogenous drivers for low-carbon transition.

These inequalities are not unique to China but also reflect a broader global pattern. Worldwide, low-carbon sustainable development is marked by pronounced unevenness: developed economies generally possess stronger financial resources, technological capacity, and institutional support, which enable them to advance more rapidly toward low-carbon transitions [3]. In contrast, many developing countries face the dual pressure of limited resources and urgent socio-economic development needs, making it difficult to prioritize low-carbon pathways. Moreover, such disparities are not confined to the global North–South divide. Even within countries such as the United States [81], members of the European Union [82], or major emerging economies, progress in low-carbon development is far from uniform [83]. Some regions have achieved notable advances in renewable energy deployment and green infrastructure, while others remain heavily dependent on carbon-intensive industries. This pattern underscores the structural complexity and multi-scalar unevenness that characterize global low-carbon sustainable development.

5.2. Local Government Behavior and the Resource–Pressure Framework

Grounded in the “resource–pressure” framework, this study further reveals the critical role of local government behavior in advancing LCSD. The findings demonstrate that low-carbon governance practices at the local level are jointly shaped by a range of factors, including top-down supervisory pressure, horizontal competitive pressure, economic resources, and institutional resources, with pronounced spatial and temporal heterogeneity. These nuances explain why the theoretical hypotheses regarding resources and pressures are only partially supported: their effects are contingent rather than universal.

First, from an institutionalist perspective, traditional new institutional economics emphasizes how formal and informal institutions shape government behavior [84]. Our findings extend this theory within China’s multi-level governance system by showing that institutional pressures are not automatically effective but are mediated by local implementation capacity. Results show that top-down supervisory pressure plays a dominant role, underscoring the central role of institutional constraints within China’s local governance system. This aligns with the hierarchical accountability logic of the bureaucratic system: local governments primarily respond to coercive pressure from higher-level authorities rather than to market signals or societal feedback [85]. However, where local resources are insufficient, this pressure may lead to symbolic implementation or institutional overload rather than substantive change. This institutional arrangement reinforces the role of superior-level supervision in steering and constraining local governments’ carbon governance performance, forming an implementation mechanism centered on top-down directives. At the same time, promotion pressure faced by officials, as a form of informal institution, also influences local governments’ engagement with LCSD. The interplay of these dual institutional logics reinforces the prioritization of LCSD by local governments, highlighting the institutionalized nature of China’s green transition.

Second, from the evolutionary perspective of the environmental Kuznets curve (EKC), this study uncovers a non-linear, multi-phase trajectory in the influence of fiscal resources on low-carbon transitions [86]. Our results reflect a multi-phase institutional dynamic. Specifically, the importance of fiscal resources rose after 2010, reflecting the growing capacity of local governments to invest in green infrastructure and technologies alongside economic growth and strengthened public finances. However, the decline in fiscal influence around 2014 may be attributed to macroeconomic downturns, fiscal decentralization tensions, and deleveraging policies that constrained fiscal support for green development. Entering the 14th Five-Year Plan period, national-level efforts to boost green fiscal investment and intergovernmental redistribution have revived the role of fiscal capacity in supporting LCSD. This suggests that the curve is driven less by automatic market mechanisms and more by shifting policy regimes and fiscal institutions. This multi-stage evolution not only extends the understanding of the EKC but also emphasizes that green transition outcomes are dynamically shaped by the interplay of economic structure, fiscal mechanisms, and broader policy environments.

Third, the study finds a marked decline in “competitive pressure” among local governments, suggesting that performance evaluations in environmental governance have become increasingly standardized and institutionalized, with diminishing interjurisdictional competition. This pattern is highly consistent with the notion of “normative isomorphism” described in institutional theory: as national environmental policies and evaluation systems become more unified, local governments have converged in their carbon governance approaches, with competitive incentives giving way to rule-based regulation [87]. In the context of multi-level governance, this indicates a shift from a “tournament model” to a “compliance model” driven by central standardization.

Finally, the GTWR model reveals strong spatial heterogeneity and directional reversals in factor effects, highlighting the importance of adaptive governance in local low-carbon transitions. While resources and pressures represent common external drivers of green development, their actual efficacy depends on how local governments strategically respond to and translate these drivers within their institutional and socio-economic contexts. This underscores the need to avoid a one-size-fits-all approach in promoting low-carbon transitions. Instead, cities should design context-sensitive, flexible, and responsive policy frameworks tailored to their specific conditions [88].

5.3. Limitations

Despite its contributions, this study has several limitations that should be acknowledged. First, the temporal scope of the analysis is confined to 2011–2021. Future research could extend the dataset to include more recent years to capture the latest dynamics of low-carbon sustainable development. Second, our analysis is conducted at the city level; investigations at finer spatial scales (e.g., districts or communities) may provide richer evidence on the micro-mechanisms underlying spatial heterogeneity. Third, while our indicator system and empirical models are relatively comprehensive, they can be further refined.

Importantly, our empirical design does not fully resolve potential endogeneity concerns between government resources/pressures and LCSD outcomes. Future research could strengthen causal identification by adopting quasi-experimental designs such as difference-in-differences (DID) using exogenous policy shocks, or employing instrumental-variable strategies and related causal inference tools.

6. Conclusions

6.1. Key Findings

This study contributes theoretically by showing that LCSD is jointly shaped by the interaction between resource endowments and multi-level institutional pressures, rather than by a single policy driver.

Using panel data from 262 Chinese cities (2011–2021), this study shows that LCSD has improved overall, but regional inequality has widened, with the east maintaining an advantage and the west showing deeper internal imbalance. We also find a clear spatial agglomeration pattern, with hotspots shifting toward the southeastern coast and fewer cold-spot cities over time. The driver structure is stable but heterogeneous: supervisory pressure remains dominant, economic resources are important but phase-sensitive, and the effects of resource- and pressure-based factors vary markedly across regions. These findings contribute new evidence on how uneven local capacity shapes low-carbon transition outcomes in emerging economies and support differentiated policy design.

6.2. Policy Implications

To promote balanced low-carbon development, policy interventions must be precise and differentiated. First, central fiscal policies should include dedicated transfer payments for green transitions. Developed regions should financially compensate less developed areas for their carbon sink efforts and conservation work. Second, performance evaluation systems need to move beyond uniform targets. Local governments should be grouped based on their resource and pressure conditions; for areas with limited resources, assessment metrics should reduce the weight of rigid economic goals and increase incentives for technology adoption and capacity building. Third, regional cooperation mechanisms should be strengthened to address spatial differences. This involves creating shared investment funds for low-carbon projects and facilitating technology sharing platforms, particularly to support cities in central and western China. Finally, reducing excessive inter-city competition requires coordinating industrial planning to avoid duplicate investments, fostering a collaborative approach to achieving national carbon goals.