Local and Neighboring Effects of China’s New Energy Demonstration City Policy on Inclusive Green Growth

Abstract

Amid mounting global climate change, resource scarcity, and environmental pressures, regional economies are accelerating their transition towards green and inclusive growth models. This research examines how China’s New Energy Demonstration City (NEDC) policy influences inclusive green growth (IGG), including its underlying mechanisms. Harnessing policy interventions as quasi-natural experiments, we use 2006–2022 panel datasets of 284 Chinese cities to develop a spatial difference-in-differences (SDID) model for causal inference. The findings are as follows: (1) The NEDC policy significantly enhances IGG in pilot cities while generating positive spatial spillover effects on neighboring cities, exhibiting an inverted U-shaped pattern; (2) The policy effects demonstrate pronounced regional heterogeneity, with the strongest impact observed in western China; (3) Mechanism analysis confirms that green technology innovation serves as a critical pathway through which the NEDC policy drives IGG. These findings provide robust empirical evidence for designing scalable policy promotion mechanisms and refining innovation-driven governance frameworks.

1. Introduction

Inclusive green growth has emerged as a key strategic framework for nations to fulfill the United Nations’ 2030 Sustainable Development Agenda [1], emphasizing the synergy between economic prosperity, environmental sustainability, and social inclusion [2], and asserting that only green and inclusive economic growth is sustainable [3]. As a pivotal player in global sustainable development, China faces acute challenges, including environmental degradation, regional disparities, and frequent extreme weather events [4,5]. Clean technology innovation and energy transition represent vital pathways to reconcile ecological sustainability with economic growth while mitigating social inequities [6,7] objectives aligning closely with the “economy-ecology-society” tripartite synergy central to inclusive green growth (IGG) [1,2]. The NEDC policy exemplifies this approach, aiming to integrate clean energy deployment with urban development by advancing renewable industries, accelerating technology diffusion, and restructuring energy supply–demand dynamics. A thorough assessment of the NEDC policy’s impact is thus essential for achieving urban inclusive green growth.

IGG champions integrated advancement across economic prosperity, environmental sustainability, and social inclusion. Existing research has analyzed the NEDC policy’s impact from multiple dimensions. Economically, the NEDC policy will boost renewable energy adoption, thereby promoting economic growth [8,9]. Environmentally, some scholars hold the view that it can exert a positive impact on the ecological environment by modifying the energy consumption structure [10,11], reducing energy intensity [12], and lowering carbon emissions [13]. In addition, some scholars have verified its positive impact on regional green economic growth [5]. However, insufficient attention has been given to the impact on social inclusion. Social inclusion entails the equitable distribution of resources, benefits, and burdens in society [14,15], ensuring non-discriminatory access to basic resources [16,17]. Scholars have explored links between energy factors (e.g., consumption structure, efficiency, availability) and employment, income inequality, health disparities, and social welfare areas [7,18] where the NEDC policy likely exerts significant influence. Furthermore, energy policies in one city may influence neighboring cities through the diffusion of green technologies and the imitation effect of government competition [19]. Thus, taking into account the policy’s spatial spillover effects is crucial to prevent bias when assessing its impact. Previous studies have verified that the NEDC policy produces notable spatial spillover impacts on carbon emissions in adjacent cities [20].

Prior research has thoroughly investigated the influence of the NEDC policy on urban economic development and the ecological environment, offering critical theoretical underpinnings and empirical validation. However, three critical research gaps remain unaddressed. First, most of the literature focuses on single or dual-dimensional analyses, lacking direct assessments of social inclusion effects and systematic evaluations of integrated inclusive green growth (IGG) indicators. This oversight hinders a holistic understanding of policy impacts on sustainable development. Second, while spatial spillover effects are acknowledged, their attenuation patterns and spatial boundaries remain underexplored. This gap limits quantitative measurement of “neighboring effects” and impedes accurate identification of policy influence ranges. Third, insufficient research on the heterogeneity of policy effects and their transmission mechanisms weakens the precision and generalizability of policy recommendations, failing to address regional disparities in implementation.

To fill these gaps, this study makes three marginal contributions: (1) It pioneers the systematic evaluation of NEDC policy impacts on urban IGG, integrating economic, environmental, and social dimensions to overcome the limitations of one-sided or dual-sided analyses [2]. (2) By constructing multi-dimensional spatial weight matrices, it applies the spatial difference-in-differences (SDID) model to quantify the distance decay boundaries of policy effects, thereby providing the empirical evidence for measuring the thresholds of “neighboring effects.” (3) It clarifies the heterogeneous policy impacts across regions and, more importantly, conducts analysis of how NEDC policy affects IGG through green technology innovation as a key transmission mechanism. This not only enhances the targeted nature of policy recommendations but also specifically addresses the research gap in the existing literature.

The following sections are organized in the manner described below. Section 2 outlines the theoretical framework, Section 3 details the research methodology, Section 4 showcases the empirical findings, Section 5 engages in further discussion, and Section 6 concludes with policy recommendations.

2. Theoretical Analysis

2.1. Context of the New Energy Demonstration City Policy

Amid the growing severity of the global environmental and climate crisis, the contradiction between urban energy consumption, environmental governance, and economic growth has become increasingly acute. In response to this situation, China’s central government initiated the NEDC policy in 2012, entrusting the National Energy Administration with the responsibility of overseeing its implementation process. The policy centers on a renewable energy prioritization strategy. Guided by the “New Town, New Energy, New Life” concept, it integrates renewable technologies into urban sectors like power supply, heating, transportation, and construction. This boosts clean energy consumption and fosters sustainable urban development. This initiative aligns with global energy transition trends and provides a practical template for achieving urban IGG.

The NEDC policy implements a dual-mechanism of “top-down” and “bottom-up” application. The central government formulates the overarching policy framework, local governments submit applications on a voluntary basis, and the final demonstration list is established following the central government’s review and approval procedures. As stipulated in the relevant assessment criteria detailed within the Explanation for New Energy Demonstration Cities (Trial), applicant cities must satisfy dual thresholds for urban capacity and renewable energy utilization potential. In January 2014, the first batch included 81 cities (e.g., Beijing’s Changping District and Hebei’s Chengde City) and 8 industrial parks (e.g., the Sino-Singapore Tianjin Eco-City), spread across 22 provincial-level administrative regions, 4 autonomous regions, and 2 directly administered municipalities. Their practical achievements in energy structure optimization, industrial upgrading, and improving social inclusion. This offers a robust empirical context for examining the policy’s combined impacts on economic expansion, ecological conservation, and societal inclusion, holding substantial significance for enhancing both the theoretical framework and practical application of IGG.

2.2. Literature Review

• Literature on Inclusive Green Growth

The concept of inclusive green growth (IGG) first emerged at the 2012 “Rio+20” Summit and has since gained global attention. Many countries and regions have adopted IGG-centered development strategies [21].

Scholars have defined IGG from two perspectives. From a developmental economics perspective, it requires integrating improved resource efficiency, effective environmental control, and enhanced inclusiveness as a path to sustainable development [22]. From a welfare economics perspective, it focuses on equal opportunities for all, balancing economic growth, environmental protection, and social inclusion to raise welfare for current and future generations [21]. These perspectives are complementary, outlining IGG’s core: coordinated progress in economic growth, environmental protection, and social inclusion, along with a balance between short-term gains and long-term intergenerational welfare.

Scholars employ various methods to measure inclusive green growth (IGG), with most research focusing on the national level. Examples include Sustainability Window Analysis [23], index-based evaluations [24], and economic complexity approaches [25]. Notably, Data Envelopment Analysis (DEA) has been increasingly adopted for its strength in addressing IGG’s multidimensional complexity [6,26]. As a non-parametric method, DEA requires no predefined production functions and flexibly handles multiple inputs, desired outputs, and undesired outputs.

In IGG research, scholars have predominantly focused on the impacts of economic variables. Extensive studies have explored how factors such as foreign direct investment [27], digital economy agglomeration [24], green finance, technological innovation, and factor misallocation influence IGG. Additionally, specific policy impacts have been explored. The regional integration of the Yangtze River Economic Belt exerts a substantial influence on inclusive green growth [28]. Cities with Smart City pilots have significantly higher IGG growth than non-adopting ones, highlighting the critical role of Smart City initiatives in advancing IGG [29].

2.

Literature on New Energy Demonstration City Policy

Policy evaluations of NEDC policy primarily focus on impacts such as carbon emissions, energy efficiency, green technological innovation, and regional green economic growth. Compared to non-demonstration cities, the NEDC policy significantly improved energy efficiency in pilot cities by approximately 4.8%, as demonstrated by Zhou (2023) [30]. Urban carbon emission efficiency was measured, with findings that the NEDC policy enhances it by optimizing industrial structures, promoting green technological innovation, and reducing energy intensity [12]. The policy’s carbon reduction effects were also assessed by addressing the “carbon lock-in” dilemma [13]. Additionally, scholars have explored its impact on regional green economic growth. The policy’s positive effect on urban green economic growth was analyzed from a perspective of regional heterogeneity. Three potential channels were identified as technological innovation, industrial upgrading, and strengthened environmental regulations [6]. New energy development was confirmed to reduce carbon emissions and environmental pollution by improving energy efficiency and optimizing energy structures, ultimately generating economic and environmental benefits to achieve green development [31].

Social inclusion is increasingly a critical pillar for fostering sustainable development. Relatively few existing studies directly examine the NEDC policy’s impact on social inclusion. Including marginalized individuals in mainstream society has long been a core goal of social inclusion policies. These policies enhance human welfare and improve social conditions. They cover interventions in healthcare, education, social security, and housing. Many studies explore new and renewable energy use. They look at how it affects residents’ income, employment, health, and welfare. These areas are core to social inclusion policies. A positive correlation between improved energy efficiency and GDP, employment, and welfare in Canada was empirically confirmed [32]. Social inclusion also entails equitable access to resources. Increased consumption of new energy enhances energy availability, helping alleviate energy poverty [18]. Improved energy access significantly reduces income inequality and further impacts employment and health outcomes [7]. Exploring how the NEDC policy influences social inclusion holds evident and significant importance. In turn, its impact on inclusive green growth—which integrates economic growth, environmental protection—and social inclusion, merits research.

2.3. Research Hypotheses

China’s NEDC policy promotes IGG through economic, environmental, and social dimensions. Economically, the policy uses tax breaks and fiscal subsidies strategically. These drive green industrial shifts. Subsidizing companies draws in upstream and downstream businesses. This builds strong industrial clusters. This clustering effect generates a multiplier impact: it boosts local employment, attracts sustained investment, and enhances total factor productivity. As these elements interconnect, they create a self-reinforcing mechanism that drives urban economic growth. Environmentally, the policy strongly promotes renewable energy sources like wind and solar. This initiative introduces cleaner energy alternatives and promotes a transition to a more environmentally sustainable energy consumption structure. As a result, the reliance on fossil fuels such as coal, oil, and natural gas has been steadily declining. Clean energy adoption not only efficiently reduces greenhouse gas emissions but also substantially decreases the levels of particulate matter, sulfur oxides, and nitrogen oxides, thereby directly enhancing air quality. These changes together form a self-sustaining cycle, laying the foundation for long-term environmental sustainability in cities. Socially, distributed energy projects lower clean energy access costs, which refers to the economic and non-economic expenses incurred by residents and communities in obtaining such energy sources, particularly benefiting SMEs and low-income groups. This reduces energy distribution gaps and ensures equitable access to green public services, narrowing divides in sustainable energy adoption. By addressing cost barriers for vulnerable groups, these projects foster social inclusion through fair access to environmental benefits. Therefore, this study reasonably proposes Hypothesis 1:

Hypothesis 1. 

China’s NEDC Policy has brought about a notable improvement in the IGG level of pilot cities.

The NEDC policy aims to significantly increase renewable energy use in urban consumption. This depends on improving green technology. Such technology helps boost renewable energy utilization. Green technological innovation significantly mediates the relationship between the NEDC policy and IGG. The NEDC encourages enterprises to boost investment in green technology R&D. It uses measures like tax breaks and research support. This lifts overall urban green innovation [33]. According to endogenous growth theory, technological progress can lead to increasing returns to scale, industrial upgrading, and productivity gains, thereby achieving urban economic growth. Upgrades in green technologies can effectively enhance energy utilization efficiency and refine the energy consumption pattern. This, in turn, reduces carbon emissions and improves environmental sustainability [34,35]. Simultaneously, green technology innovations reduce the cost of clean energy, improve energy availability, and promote social inclusion [18]. Thus, the NEDC policy has the potential to foster urban IGG through the promotion of green technology innovation. Drawing from the preceding analysis, we formulate the following hypothesis:

Hypothesis 2. 

Green technology innovation serves as an effective channel through which the NEDC policy impacts IGG.

Cities exhibit significant spatial dependence in their development [36]. The NEDC policy not only positively impacts the IGG of pilot cities but also generates positive spillovers for neighboring cities through “competitive imitation” and “technology diffusion”. After the new energy policy is implemented, pilot cities will advance new energy projects. They adjust industrial structures and refine management measures. Neighboring cities competitively emulate these strategies. This is driven by local governments’ “growth-oriented competition” [37]. Additionally, new energy projects in pilot cities are frequently accompanied by the adoption of relevant green technologies and innovations. Technologies and knowledge will naturally radiate and spread to the surrounding areas, thereby fostering their renewable sector development [38,39]. Furthermore, the policy also strengthens regional connections. Neighboring cities optimize resource allocation with pilot cities through energy collaboration and factor flow. This aids industrial restructuring and IGG promotion [40]. Based on the foregoing analysis, the following hypothesis is put forward.

Hypothesis 3. 

New Energy Demonstration City policies have a positive spatial spillover effect on inclusive green growth in surrounding cities.

3. Research Design

3.1. Data Sources

Considering that China’s Renewable Energy Law, implemented on 1 January 2006, may have an impact on policy evaluation. This study uses cities at the prefecture level and above in China from 2006 to 2022 as the analytical sample. Moreover, since the datasets of certain cities, including Lhasa, Kunyu, Sansha, and Danzhou, suffered from extensive data deficiencies, the research ultimately settled on 284 city samples. Among them, 58 were part of the pilot program, while the remaining 226 were non-pilot cities. The specific spatial distribution is shown in Figure 1. The NEDC list was obtained from the official website of the National Energy Administration (https://zfxxgk.nea.gov.cn/ accessed on 12 February 2025). Data were acquired from the China Urban Statistical Yearbook and the China Open Data Platform. To address data gaps, linear interpolation was employed for supplementation.

3.2. Variable Construction

Dependent Variable: IGG aims to harmonize economic growth, environmental governance, and social inclusion [42]. For measuring IGG levels in Chinese cities, we chose the undesirable output super-efficiency SBM model proposed by Tone (2002) [43]. This Data Enveloping Analysis tool accounts for input and output slack variables, with the objective of maximizing desirable outputs while reducing undesirable outputs to a minimum. By relaxing the constraints of the traditional SBM model, it differentiates decision-making units with an efficiency score of 1. Scores to exceed 1 are allowed. It lays the foundation for urban IGG measurement.

According to Sun et al. (2020), the measurement of China’s IGG level incorporates three indicator categories: input factors, desired outputs, and undesirable outputs [44]. Inputs (labor, capital, energy) reflect core production factors driving urban development, aligning with neoclassical growth theory and prior IGG studies. Desirable outputs capture economic and social gains. GDP is a commonly used and effective indicator of economic growth. In line with the core of social inclusion, we assess its achievement by examining outcomes of social inclusion policies in key areas such as healthcare, education, income, and consumption, among others. Therefore, based on theoretical analysis and data availability, we select the following indicators: per capita retail sales, health technicians per 10,000 persons, basic pension insurance participants, and student–teacher ratios to measure the desired outputs of social inclusion and the urban–rural income ratio as the undesired social output. Wastewater, soot, and sulfur dioxide emissions, typical industrial pollutants, act as undesired outputs in environmental governance.

Table 1. Indicators of urban IGG measurement.

Indicators	Categories	Specific Indicators
Inputs	Labor input	Year-end employment figures (10,000 persons)
	Capital input	Fixed asset investment stock (10,000 CNY)
	Energy input	Electricity consumption of the whole society (in thousands of kilowatt-hours)
		Water consumption of the whole society (10,000 cubic meters)
Desirable Outputs	Economic output	Gross domestic product (ten thousand yuan)
	Social output	Per capita retail sales (CNY)
		Health technicians per 10,000 persons (persons)
		Basic pension insurance participants (persons)
		Ratio of students to teachers in primary and secondary schools (%)
Undesirable Outputs	Social output	Urban-rural residents per capita disposable income ratio (%)
	Environmental output	Total wastewater discharge (tons)
		Total soot (dust) emissions (tons)
		Total sulfur dioxide emissions (tons)

presents the specific indicators.

Core explanatory variable: The binary variable DID denotes NEDC policy adoption by city i in year t, with 1 indicating implementation and 0 non-implementation. After excluding cities with significant data deficiencies, the final sample consisted of 58 treatment cities and 226 control cities.

Mechanism variable: GTI is measured by the logarithmic value of total green patent applications. Due to the lag and instability of green patent authorization caused by factors such as administrative approval, some patents may be used in the patent application process. Thus, using logarithmic scaling for total green patent applications as a measure is more justifiable than the total volume of green patent authorizations [45]. Green patent applications, a common GTI proxy, have shorter lags than authorizations but still face time lags and quality limitations. They are retained due to scarce urban green commercialization data and consistency with prior studies, with future research advised to integrate multiple proxies.

Control variables: Following Zhang Tao et al. [46], we control for: (1) Education Development (ED), which is measured by the ratio of local government education expenditure to general budget expenditure.; (2) Population Density (PD), which is calculated by taking the logarithm of population distribution per square kilometer. (3) Financial Development (FD), which is quantified through the ratio of the city’s financial institutions’ total loan balances as of year-end to its gross domestic product (GDP); (4) Science and Technology Development (STD), which is measured by the ratio of fiscal investment in science and technology to local governments’ overall budgetary spending.

3.3. Model Specification

Given that the New Energy Demonstration City policy constitutes a quasi-natural experiment, naturally forming treatment and control groups with a time-varying discontinuity, the differences-in-differences (DID) model is the most appropriate method to isolate the policy effects. The DID can largely solve the endogenous problem. However, this model ignores the spatial correlation among geographical units. The Spatial Durbin DID model (SDM-DID) addresses this limitation [47]. Our specification is as follows:

$${\mathrm{I}\mathrm{G}\mathrm{G}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+\mathsf{\rho }\mathrm{W}{\mathrm{I}\mathrm{G}\mathrm{G}}_{\mathrm{i}\mathrm{t}}+\mathsf{\beta }{\mathrm{D}\mathrm{I}\mathrm{D}}_{\mathrm{i}\mathrm{t}}+\mathsf{\theta }\mathrm{W}{\mathrm{D}\mathrm{I}\mathrm{D}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_{\mathrm{i}}{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\theta }}_{\mathrm{i}}\mathrm{W}{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(1)

where ${\mathrm{I}\mathrm{G}\mathrm{G}}_{\mathrm{i}\mathrm{t}}$ denotes the urban inclusive green growth (IGG) level. In the model, i and t, respectively, indicate the city and year; ρ acts as the spatial auto-regressive coefficient, which gauges the magnitude of spatial auto-correlation; $\mathrm{W}{\mathrm{I}\mathrm{G}\mathrm{G}}_{\mathrm{i}\mathrm{t}}$ refers to the spatial lag term of the dependent variable; The spatial weight matrix W is constructed using Queen contiguity criteria to capture regional spatial interactions via adjacency and connectivity strength; ${DID}_{it}$ denotes the implementation status of NEDC policies, and β is its spatial regression coefficient; $W{DID}_{it}$ represents the policy variable’s spatial lag effect, with θ measuring the elasticity coefficient.; $X_{it}$ represent multiple control variables, βi are their spatial regression coefficients; $W{X}_{it}$ refers to control variables’ spatial lag specification, θi represents the elastic coefficients. For modeling spatial-temporal dynamics, ${\mu }_i$ and ${\delta }_t$ account for spatial fixed effects and temporal fixed effects, respectively, while ${\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$ represents the random perturbation term.

4. Empirical Analysis

4.1. Analysis of Spatial Correlation in Urban IGG

Spatial auto-correlation analysis is a critical approach for examining the spatial distribution characteristics of socioeconomic activities, capable of revealing their agglomeration and dispersion patterns. This research utilizes ArcGIS10.8.1 software alongside Global Moran’s I statistics to conduct a comprehensive analysis of spatial auto-correlation in urban IGG. As illustrated in Figure 2, the significance test results of the global Moran’s I indices for urban IGG in 2006 (a), 2012 (b), 2017 (c), and 2022 (d) show significant positive spatial correlation (p < 0.05). Specifically, the spatial polarization pattern—where high-value areas cluster with high-value areas (HH agglomeration) and low-value areas cluster with low-value areas (LL agglomeration)—persists. This phenomenon highlights significant spatial dependence in urban IGG. Although the intensity of spatial agglomeration fluctuates over time, the overall spatial structure remains stable.

4.2. Spatial-Temporal Evolution of Urban IGG

Using natural breaks classification in GIS, IGG levels are categorized into three tiers. Spatial distribution patterns of urban IGG in 2006, 2012, 2017, and 2022 are shown in Figure 3. It shows IGG at the municipal level is generally on the rise. Specifically, 135 cities were in the [0.226062, 0.455098] range in 2006, accounting for 47.5% of the total sample size; 4 cities were in the range of [0.692766, 1.106256], accounting for 1.4%. By 2022, cities in the lowest tier decreased to 112, accounting for 39.4 percent, and those in the highest tier increased to 33, accounting for 11.2%.

From a spatial perspective, over time, the eastern coastal regions, including the Beijing–Tianjin–Hebei region, the Yangtze River Delta, and Pearl River Delta, have consistently remained the high-value core zones for IGG, and the synergy within the region has been continuously enhanced, evolving from “core radiation” to “multi-pole linkage”. The gradient difference in the central region has gradually narrowed, and the provincial capital cities represented by Wuhan, Zhengzhou, and Changsha have formed a “mid-high level core pole”, driving the transition of the surrounding areas to the medium level. Western China exhibits growth poles in Chengdu, Chongqing and Xi’an, but remote areas (such as Ningxia and Qinghai) remain at a low level for a long time. Northeast China demonstrates deteriorating spatial patterns, with core cities such as Shenyang and Dalian declining, and resource-based cities such as Daqing and Hegang falling into the low tier. It is notable that low IGG persists across much of western China, despite progress in some central/western cities.

4.3. Baseline Regression Results and Robustness Tests

4.3.1. Model Testing

To identify the specific form of the spatial econometric model, corresponding model tests were performed, as presented in

Table 2. Model test results.

Statistics	Hausman Test	LR_Test	Wald_Test
Chi2	186.75 ***
SAR_chi2		192.02 ***	196.68 ***
SEM_chi2		208.85 ***	211.00 ***

Robust t-statistics in parentheses: *** p < 0.01.

. The Hausman test strongly favored the spatial fixed-effects Spatial Durbin Model (SDM) over the random-effects model at the 1% significance level, prompting the use of estimation results from the fixed-effects model. Moreover, by benchmarking against the spatial auto-regressive model (SAR) and Spatial Error Model (SEM), the specification was subjected to diagnostic checks using LR and Wald tests. At the 1% significance threshold, the test statistics were strongly rejected, providing evidence that spatial spillover effects exist among at least certain regions and variables. This outcome underscores the inappropriateness of simplifying the model to either the SAR or SEM specification. Collectively, these findings establish the fixed-effect SDM as the most suitable choice for this research.

4.3.2. Benchmark Regression Results

Using Equation (1),

Table 3. Baseline regression results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect
DID	0.0437 ***	0.184 ***	0.227 ***	0.0144 **	0.0405 **	0.0549 ***
	(0.00708)	(0.0178)	(0.0197)	(0.00730)	(0.0185)	(0.0213)
ED				0.178 ***	−0.489 ***	−0.311 **
				(0.0688)	(0.142)	(0.151)
PD				0.0224	−0.685 ***	−0.662 ***
				(0.0373)	(0.0829)	(0.0835)
FD				0.0125 ***	0.0672 ***	0.0797 ***
				(0.00304)	(0.00525)	(0.00514)
STD				0.719 ***	0.893 ***	1.612 ***
				(0.166)	(0.342)	(0.358)
time	Yes	Yes	Yes	Yes	Yes	Yes
individual	Yes	Yes	Yes	Yes	Yes	Yes
Observations	4828	4828	4828	4828	4828	4828

Robust t-statistics in parentheses: *** p < 0.01, ** p < 0.05.

presents the SDID model estimates. Columns (1)–(3) display effect decomposition in the absence of control variables, while columns (4)–(6) feature empirical results incorporating such variables alongside fixed effects. The regression results indicate a direct effect coefficient of 0.0144 for the NEDC policy in column (4), significant at the 5% threshold. This finding confirms Hypothesis 1: the policy significantly elevates IGG levels in pilot cities. Column (5) reveals an indirect effect coefficient of 0.0405, also significant at the 5% level, supporting Hypothesis 3 that pilot cities’ status generates positive spillovers on surrounding cities’ IGG.

4.3.3. Robustness Test

• Parallel Trend Test

The difference-in-differences (DID) method hinges on the parallel trend requirement, stipulating that treatment and control groups must show no significant pre-policy divergences. An event study framework validated this assumption empirically, complemented by an analysis of the policy’s dynamic effects [48]. The expression is as follows:

$${IGG}_{it}={\alpha }_0+\rho W{IGG}_{it}+{{\sum }_{\mathit{t}\ge -7,\mathit{t}\ne -1}^9{\beta }_{t}DID}_{it}+{\sum }_{\mathit{t}\ge -7,\mathit{t}\ne -1}^9{\theta }_t{WDID}_{it}+{\beta }_i{X}_{it}+{\theta }_{i}W{X}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}$$

(2)

This study examines the coefficient ${\mathsf{\beta }}_{\mathrm{t}}$ quantifying the IGG gap between pilot and non-pilot cities in the t-th year of NEDC implementation. Figure 4 shows no major pre-policy differences between the groups. This satisfies the parallel trend assumption. It also validates the DID model for assessing the NEDC’s post-implementation effects.

2.

Altered Timeframe

The time span is shortened from 2006–2022 to 2010–2022. Then, the regression analysis is rerun. Results in Column 1 of

Table 4. Results of robustness test versus mechanism test.

	(1)	(2)	(3)
			GTI
Direct effect	0.0284 ***	−0.00376	0.252 ***
	(0.00863)	(0.00786)	(0.0466)
Indirect effect	0.0630 ***	0.0126	1.612 ***
	(0.0209)	(0.0191)	(0.161)
Control variables	Yes	Yes	Yes
Time	Yes	Yes	Yes
Individual	Yes	Yes	Yes
Observations	4828	4828	4828

Robust t-statistics in parentheses: *** p < 0.01.

show that the core variable DID had effects—a direct component of 0.0284 and an indirect component of 0.0630. Both are statistically significant at the 1% level. This indicates the policy’s direct and indirect effects remain significant with a shorter time span, demonstrating substantial robustness in the baseline regression outcomes.

3.

Placebo Test

We constructed placebo policy dummy variables and re-estimated the model after advancing the assumed policy implementation year to 2011. The estimated coefficient for the placebo policy variable DID showed a direct impact of −0.00376 and an accompanying indirect influence of 0.0126. Statistical tests revealed that neither coefficient deviated significantly from zero. Advancing the policy implementation year to 2011 effectively nullified any causal relationship between the placebo intervention and the outcome variable. This null result strongly supports the validity of our main regression findings and confirms the original model’s conclusions. These conclusions are resilient to false-positive biases from temporal misalignment.

4.3.4. Mechanism Test

The preceding analysis has confirmed that the NEDC policy exerts a positive impact on local IGG while generating positive spillover effects on neighboring cities. This part will continue to center on the NEDC policy as the key explanatory variable, with GTI set as the dependent variable, to empirically explore the specific mechanisms at play.

According to the regression outcomes presented in Column 3 of Table 4, the NEDC policy significantly promotes urban IGG through GTI as a key mechanism. Specifically, the direct effect coefficient of the NEDC is 0.252 (p < 0.01), signifying that the policy has effectively spurred enterprises to intensify investments in green technology innovation and adoption, enhancing urban GTI capabilities. The indirect effect coefficient stands at 1.612 (p < 0.01), indicating that the pilot cities’ policy has substantially impacted green technology innovation in surrounding cities, thereby driving IGG. These findings support Hypothesis 2.

5. Further Analysis

5.1. Characteristics of Spatial Spillover Effects

Prior analysis has established the NEDC policy’s substantial spatial spillover effects. To delve deeper into the policy’s specific regional boundaries regarding spatial spillover, we will define the threshold geographical distance spatial weight matrix. In this matrix, spatial units i and j are deemed to interact if their distance is below the threshold, with no interaction otherwise. The initial threshold is set at 50 km, increased stepwise by 50 km, to test policy spillover effects within 250 km of pilot cities. The details are as follows:

$$W_d|d=d_{min},d_{min}+r, d_{min}+2r,\dots ,d_{max}$$

(3)

$$W_{ij}=\left\{\begin{array}{c} 1, 0<d_{ij}\le d\\ 0,{ d}_{ij}>d\end{array}\right.$$

(4)

Here, the threshold spatial weight matrix is denoted by $W_d$. $d_{min}$ defines the shortest possible interval of the geographic centroid distance between any two cities under consideration. This value sets the lower boundary for measuring spatial proximity. On the contrary, $d_{max}$ represents the upper limit, capturing the greatest extent of the centroid-to-centroid distance among city pairs. $r$ stands for the increment or step size used to discretize the distance range between $d_{min}$ and $d_{max}$. $d$ acts as a critical cut-off, functioning as the geographic threshold against which actual distances are compared. $d_{ij}$ precisely measures the straight-line geographic distance between the centroids of city i and city j, serving as the foundational metric for evaluating spatial relationships.

Table 5. Spatial spillover effect characteristics results.

	(1)	(2)	(3)	(4)	(5)
	50 km	100 km	150 km	200 km	250 km
Direct effect	0.0350 ***	0.0261 ***	0.0164 **	0.00506	0.00468
	(0.00761)	(0.00746)	(0.00733)	(0.00730)	(0.00725)
Indirect effect	0.00425	0.0122	0.0533 ***	0.0367 *	0.0442
	(0.00309)	(0.00790)	(0.0147)	(0.0199)	(0.0270)
Control variables	Yes	Yes	Yes	Yes	Yes
time	Yes	Yes	Yes	Yes	Yes
individual	Yes	Yes	Yes	Yes	Yes
Observations	4828	4828	4828	4828	4828

Robust t-statistics in parentheses: *** p < 0.01, ** p < 0.05, * p < 0.1.

depicts the results of model regression using different threshold geographical distance spatial weight matrices. The neighboring effect assessment reveals the spatial spillover characteristics of NEDC policies at different geographical distances, providing an important perspective for understanding the diffusion mechanism of policies. The analysis shows that the NEDC’s spatial spillover effect exhibits a pattern of first rising and then declining in an inverted “U” shape.

The indirect effect is not significant within the distance range (100 km). This can be attributed to two factors: (1) insufficient spatial interaction units in the weight matrix, and (2) the potential aggregation shadow. The central city has a strong attraction to resource elements, resulting in an economic shadow area with a lower density of economic activity, a net outflow of resource elements, and less developed industries in the surrounding area [49]. Within the 150 km range, the indirect effect strengthens significantly to 0.0533 (p < 0.01). This outcome demonstrates that the NEDC policy has produced significant spatial spillover effects within this distance range, significantly promoting inclusive green growth in non-pilot cities. For the 200 km distance class, the indirect effect is estimated at 0.0367 (p < 0.10), showing that the policy’s spatial spillover impact remains within this scope, yet its significance is less pronounced than at the 150 km scale. This may be because the intensity of the spillover effect gradually diminishes as the distance increases, but it still has some impact on non-pilot cities. This finding is in line with the first law of geography, which states that the intensity of spatial interactions decreases with distance but remains strong within a certain range [50]. Finally, at a distance of 250 km, the indirect effect amounts to 0.0442 but lacks statistical significance, suggesting the policy’s spatial spillover effect gradually diminishes within this range. In the process of spillover effects caused by urban imitation and technology diffusion, there is a loss of transmission due to the distortion and redundancy of information caused by the increase in distance, and the time obtained will also be delayed and delayed due to the increase in distance [51]. Other factors, such as regional economic development tiers and policy execution rigor, moderate the impact of pilot cities’ policies toward neighboring cities.

5.2. Heterogeneity Analysis

Policy effects can vary by region based on different resource endowments, geographical locations, innovation environments, and population sizes [52,53]. The influence of NEDC policies on urban IGG can also differ across cities, contingent on specific attributes like developmental stage, population distribution patterns, degree of regional integration, and natural geographic positioning [54]. Parameter estimates based on overall samples do not reveal the regional characteristics of NEDC policies on IGG. Based on the regional classification standards of China’s National Bureau of Statistics, cities are categorized into four regions: eastern, central, western, and northeastern. This classification system by the National Bureau of Statistics is mainly based on the significant differences among the regions in terms of economic development stage, industrial structure characteristics, resource endowment conditions, etc. Through this scientific division of regions, researchers can systematically analyze the differences in development gradients among different regions and monitor the process of coordinated regional development. This classification reflects significant regional development imbalances in China as a super-large economy. Its core research value helps governments identify regional development gaps, optimize resource allocation, and achieve nationwide balanced and coordinated development.

Notable regional disparities in the impact of the NEDC policy on urban IGG are evident in

Table 6. Results of heterogeneity analysis.

	(1)	(2)	(3)	(4)
	Eastern	Central	Western	Northeastern
Direct effect	0.0111	0.0106	0.0559 ***	0.0329
	(0.0151)	(0.00964)	(0.0142)	(0.0274)
Indirect effect	0.00847	0.0592 ***	0.0834 ***	0.275 ***
	(0.0317)	(0.0204)	(0.0272)	(0.0820)
Control variables	Yes	Yes	Yes	Yes
Time	Yes	Yes	Yes	Yes
Individual	Yes	Yes	Yes	Yes
Observations	1462	1360	1428	578

Robust t-statistics in parentheses: *** p < 0.01.

, with western China being the most prominent. These disparities manifest in both the policy’s direct and indirect impacts. Western China has shown significant positive effects in both direct and indirect aspects of policy. This suggests that although western China has a weak development foundation, it has a “latecomer advantage” when driven by policy. Despite the lower levels of IGG in the western region, policies have effectively acted as a “catalyst for growth” by compensating for market failures and guiding the reallocation of resources. As shown in Column 1, the NEDC policy’s effect in eastern China is insignificant. The marginal utility of the policy is diminishing, primarily due to high marketization and advanced green technology levels. This finding is in line with the “policy effect convergence hypothesis” [55].

6. Discussion, Conclusions, and Policy Implications

6.1. Discussion

Firstly, this paper finds that the NEDC policy significantly boosts pilot cities’ IGG, aligning with Wang and Yi (2021) [6]. Economically, it uses tax incentives and subsidies to foster green industrial clusters, creating a self-reinforcing growth cycle. Environmentally, it promotes renewable energy, improving air quality, and forming sustainability feedback loops. Socially, it lowers clean energy access costs, narrowing gaps and ensuring inclusive benefits. The paper confirms spatial spillover effects on surrounding cities’ IGG, showing an inverted “U” trend. Within 100 km, the spillover is insignificant, linked to the “agglomeration shadow” effect and insufficient spatial interaction—core cities attract resources strongly, creating low-activity shadow areas, while limited interaction units in the weight matrix hinder factor flow measurement [49]. At 100–150 km, the agglomeration shadow weakens, and spillovers rise: surrounding cities absorb factors via infrastructure networks, and proximity strengthens policy learning, pushing spillovers to a peak as nearby cities imitate pilot tools. Beyond 150 km, spillovers decay with distance, reflecting “distance decay” and combined spatial factors [50,56]. These factors—including distorted policy information, delayed technology diffusion, and sparse infrastructure—jointly worsen the decay.

Secondly, the policy’s spatial spillover shows striking regional heterogeneity, resonating with debates on geographic context shaping outcomes [57]. Western China sees the strongest impact, with significant positive direct and indirect effects, while eastern China shows no statistical significance. This divergence stems from resource endowments, development stages, and institutions. Western China’s abundant renewable energy creates “resource-policy synergy,” lowering implementation costs—wind power subsidies achieve higher utilization due to reduced adoption barriers. Its “late-comer advantage”—weaker economic foundations but less high-carbon path dependency—lets policies act as “growth catalysts,” aligning with theories of leapfrogging pollution stages [58,59]. Western governments’ stronger strategic industry authority amplifies spillovers, as seen in Gansu: photovoltaic investments clustered enterprises, boosting employment and reducing pollution. In contrast, eastern China’s mature markets and advanced green tech leave little room for improvement, reflecting the “policy effect convergence hypothesis” [42].

Finally, this paper identifies green technology innovation as the policy-driven mechanism linking NEDCs to IGG, consistent with Wang and Yi (2021) [6]. It clarifies the pathway for inclusive green growth, supporting policy practice. NEDC benefits and funding motivate governments and enterprises to train talent, increase R&D investment, and advance energy and emission technologies, lowering specific energy consumption. Green technology—a system aiding conservation, pollution control, efficiency, and recycling—drives pollution reduction, carbon cuts, social inclusion, and sustainable development [34,35].

While this study yielded certain insights, several limitations remain that warrant further refinement in subsequent research. First, constrained by data accessibility, the study covers a period from 2006 to 2022, which may not fully reflect the latest policy effects after the implementation of the “dual carbon” goals. Subsequent research could expand the observation timeframe to 2025 and beyond, aiming to capture the policy’s long-term dynamic impacts. Secondly, the traditional method of dividing the four major regions may mask the heterogeneity among cities within the regions. It is recommended that subsequent studies combine urban agglomeration planning or economic belt division standards and introduce high-resolution remote sensing data to more accurately identify spatial differences. Third, the current study fails to fully reveal the dynamic evolution of policy effects. A dynamic spatial difference-in-differences model can be constructed to examine the differentiated effects over different periods, such as 1–3 years and 3–5 years of policy implementation. Fourth, the lack of analysis of micro-subjects such as enterprises limits the depth of mechanism research. Subsequently, the transmission mechanism can be verified from dimensions such as green patents and R&D investment by matching the industrial enterprise database. Finally, a further limitation of the research lies in its omission of synergies between new energy policy frameworks and complementary environmental policy initiatives. It is recommended to use the policy combination analysis method to quantify its interaction with policies such as the carbon market and environmental inspections. By constructing a multi-scale and multi-dimensional analytical framework, it will help to formulate more precise and effective regional green development policies.

6.2. Conclusions

Using an SDID model and panel data from 284 Chinese cities (2006–2022), this study empirically analyzes how China’s NEDC policy impacts IGG and its mechanisms. The main conclusions drawn from the empirical analyses are as follows: (1) The policy is demonstrated to deliver a substantial enhancement in IGG for pilot cities, alongside generating positive spatial spillover impacts on neighboring urban areas. (2) The magnitude of the spatial spillover effect exhibits an inverted U-curve, characterized by initial growth followed by decline as distance increases, reaching a peak within 150 km and then gradually decaying, verifying the “distance decaying” pattern. (3) Notable regional heterogeneity is observed in the policy’s effect. In western China, the policy demonstrates peak efficacy, contrasting sharply with its more circumscribed influence in the east. (4) This differential outcome hinges on green technology innovation, the linchpin mechanism through which NEDC policies propel urban IGG by invigorating the dynamism of green innovation initiatives.

By examining the temporal dynamics, regional differences, and action paths of the policy, the findings emphasize the importance of balancing inclusiveness and sustainability in new energy policy design. It provides a Chinese experience for global policy evaluation on urban energy transition and green development, and offers practical references for balancing economic growth, environmental protection, and social inclusion.

Key scientific contributions of this study include: (1) Revealing the inverted U-shaped spatial spillover pattern: Unlike previous studies assuming linear or uniform spatial effects of energy policies, this study identifies an inverted U-shaped spillover effect of the NEDC policy on IGG in surrounding cities. It deepens the understanding of how geographical distance and regional contexts affect policy transmission, responding to the suggestion that spatial spillover effects should be interpreted with spatial economic logic. (2) Clarifying the mediating role of green technology innovation: This study verifies that green technology innovation is a key transmission path through which the NEDC policy impacts IGG, illustrating that the policy promotes the synergistic improvement of economic, environmental, and social benefits by incentivizing corporate R&D investment and green technology application, supplementing empirical evidence for policy mechanism research. (3) Integrating multi-dimensional evaluation methods: It evaluates IGG using the SBM model, identifies policy effects by combining the Spatial Durbin Model with traditional difference-in-differences (DID), and analyzes the mechanism through mediating effect analysis, forming a complete research framework and enhancing the accuracy of policy evaluation.

6.3. Policy Implications

To maximize the NEDC policy’s role in advancing IGG, strategies should integrate spatial collaboration within optimal radii, regionally differentiated measures, and synergistic innovation mechanisms.

Given the peak spillover effect of the policy within a 150 km radius, inter-municipal collaboration mechanisms should be established. For instance, a collaboration zone with a 150 km radius could be demarcated to promote the joint construction of shared renewable energy grids between pilot cities and surrounding areas. Local governments would take the lead in formulating electricity mutual-aid agreements, while public utility sectors would be responsible for infrastructure connection, so as to maximize the radiating benefits of the policy.

In view of the prominent policy efficacy in western regions, the central government could increase subsidies for green technology R&D in western pilot cities and build new energy industrial clusters based on their resource endowments. For eastern regions, there is a need to focus on industrial structure transformation; local governments should join hands with enterprises to promote green technological transformation in traditional manufacturing industries and reduce transformation costs through tax incentives.

Governments should set up special funds to support the transformation of green patents, enterprises should take the lead in the commercial application of technologies, and research institutes should provide technical support. Particularly in western pilot cities, a “patent sharing pool” model could be piloted, where public utility sectors coordinate the diffusion of technical resources to small and medium-sized enterprises, thereby strengthening the driving role of innovation in IGG.

These findings hold valuable implications for other countries. For instance, in large developing economies with fragmented urban systems like India, demarcating 150 km collaboration zones around renewable energy hubs could replicate spatial spillover benefits, with state governments able to negotiate cross-border power-sharing agreements. Resource-rich nations like South Africa, endowed with abundant solar and wind resources, might adopt the R&D subsidy and industrial clustering models used in western China to leverage their natural advantages. Meanwhile, Southeast Asian countries undergoing energy transition, such as Vietnam, could refer to the “patent sharing pool” mechanism to accelerate the diffusion of green technologies among small and medium-sized enterprises, overcoming barriers to innovation adoption in emerging markets.