Synergistic Impacts of Clean Energy Demonstration Policy on Air Pollution and Carbon Reduction

Abstract

The development of clean energy is of great significance for achieving air pollution and carbon reduction. China has launched clean energy demonstration province (CEDP) construction as a pilot program to accelerate cleaner alternatives and promote synergies between air pollution and carbon reduction. Existing studies have focused on the carbon abatement effects of this clean energy demonstration policy but are inconclusive as to whether the policy also reduces air pollution. In this paper, we empirically assess the synergistic impact of the CEDP policy on air pollution and carbon reduction using the difference-in-differences method by treating the CEDP pilot as a quasi-natural experiment using provincial-level panel data from 2005 to 2020 in China. We find that the PM2.5 and carbon intensity in the eastern and central regions show a tendency to move towards a lower gradient compared to the western region, and the CEDP pilot has a synergistic effect on air pollution and carbon reduction, as the implementation of the policy significantly increases synergy between air pollution and carbon reduction and reduces the intensity of PM2.5 and carbon emissions in the pilot provinces, which remains a robust finding after multiple tests. In terms of regional differences, the policy’s effects in the central and western regions are more significant than those in the eastern regions, which suggests that the central and western regions have made more progress in environmental improvement after the implementation of the relevant policies, and this finding provides new ideas and possibilities for reducing regional pollution. Further mechanism tests find that industrial upgrading and energy efficiency improvement are important channels for the CEDP policy to achieve the synergistic effects of air pollution and carbon reduction. Accordingly, we put forward policy recommendations to expand the coverage of CEDP pilots in an orderly manner, strengthen the transmission role of industrial upgrading and energy efficiency improvement, and improve the evaluation system of CEDP construction and operation.

1. Introduction

The large-scale urbanization and industrialization process, especially in developing countries, has led to new records in energy consumption demand annually [1]. Fossil fuels accounted for 82% of global energy consumption in 2022, and the use of this energy directly generates a variety of air pollutants [2] that not only endanger human health [3] but also exacerbate the situation of climate change [4]. As one of the world’s largest energy consumers [5], China’s energy structure has long relied on fossil fuels such as coal, which has contributed to the country’s severe air pollution problem and caused mainland China to surpass the United States as the world’s largest carbon emitter in 2005 [6]. According to data from China’s Ecological and Environmental Status Bulletin (https://www.mee.gov.cn/hjzl/sthjzk/zghjzkgb/202406/P020240604551536165161.pdf, accessed on 27 October 2024), around 40.1% of China’s cities had exceeded air ambient quality standards in 2023, with PM2.5 being one of the main exceedances.

A healthy ecological environment is an important foundation for building a beautiful China. In July 2024, the Chinese government issued a document entitled ‘Opinions of the Central Committee of the Communist Party of China and the State Council on Accelerating the Promotion of Comprehensive Green Transformation of Economic and Social Development’, which explicitly emphasized that, by 2035, it will have made significant progress in pollutant emission reduction, carbon emission reduction, and synergistic efficiency enhancement, and will have basically realized the goal of a beautiful China. Therefore, the synergistic promotion of air pollution and carbon reduction is the new target and mission of China’s ecological civilization construction and green development at the present stage. In fact, both greenhouse gases and air pollutants originate heavily from the combustion of fossil fuels [7], implying that energy transitions centered on the development of cleaner energy have a synergistic effect in controlling carbon emissions and air pollutants [8,9].

Clean energy sources, including solar, wind, hydro, and geothermal energy, are characterized by their harmlessness or minimal harm to the environment and their wide distribution of resources [10]. China has incorporated the development of clean energy to achieve synergistic control of air pollutants and greenhouse gases into its national planning and has successively issued pilot programs in key areas and industries, as well as a series of supportive and safeguard measures, to accelerate clean substitution and improve environmental quality. Among the many initiatives, the creation of the clean energy demonstration province (CEDP) project has received much attention. To date, China has initiated the implementation of the CEDP policy in six provinces. Specifically, Ningxia was set up as China’s first new comprehensive energy demonstration area in 2012, taking clean energy development as a major thrust to reorient the province’s energy structure and promote energetic transformation. In 2014, Zhejiang became a pilot province, with Sichuan (2016), Tibet (2016), Gansu (2016), and Qinghai (2018) as successive CEDP pilots (

Table A1. Policy sources for building clean energy demonstration provinces in each pilot province.

Pilot Provinces	Policy Year	Policy Sources
Ningxia	2012	https://www.ccchina.org.cn/Detail.aspx?newsId=16993&TId=57, accessed on 27 October 2024.
Zhejiang	2014	https://www.zj.gov.cn/art/2016/1/12/art_1229697772_2443817.html, accessed on 27 October 2024.
Sichuan	2016	https://www.sc.gov.cn/10462/10464/10465/10574/2016/9/13/10395702.shtml, accessed on 27 October 2024.
Tibet	2016	https://www.ndrc.gov.cn/xxgk/zcfb/ghwb/201701/W020190905497899281430.pdf, accessed on 27 October 2024.
Gansu	2016	https://fzgg.gansu.gov.cn/fzgg/c106094/201306/c4cd78576cb64a4ba7294853827a2106.shtml, accessed on 27 October 2024.
Qinghai	2018	http://www.qinghai.gov.cn/xxgk/xxgk/qhzb/qhzb2018/201903/P020190304589259565147.pdf, accessed on 27 October 2024.

). Pilot areas combined with planning concepts have been developed and introduced specific work programs, for example, in Zhejiang Province, through the vigorous development of renewable energy such as photovoltaic energy, optimizing the layout of energy development, and promoting clean and low-carbon energy production and living. Qinghai Province, focusing on the large-scale development of new energy sources, has made it clear that by 2020, the proportion of clean energy production will reach 51 percent, and the proportion of consumption will reach 41 percent, initially becoming an important new energy industry base in the country. Gansu Province, in order to carry out new energy comprehensive demonstration work as an opportunity to significantly increase the proportion of new energy in energy consumption and gradually establish a new energy-led, diversified, low-carbon, safe energy supply system, by 2020 initially built an energy system with a high proportion of renewable energy.

The CEDP pilot is an innovative environmental policy to promote clean energy development, which is of great significance to improve energy structure, protect the ecological environment, and achieve sustainable economic and social development. A large number of studies have used theoretical methods such as fixed-effects models, difference in differences, and synthetic control to confirm that environmental policy tools such as environmental protection law [11], green credit [12], the carbon emissions trading pilot project [13], and low-carbon cities [14] have a significant role in reducing air pollution and lowering carbon emissions, but almost all of them analyzed the effect of pollution reduction and the effect of carbon reduction separately. In terms of influencing factors, Cheng et al. [15], Wang et al. [16], and Chen et al. [17] found that the level of economic development, government intervention, degree of openness to the outside world, and population density are important influencing factors affecting the effect of environmental governance policies on pollution reduction and carbon reduction. In addition, some scholars measured the level of pollution and carbon reduction synergy through the coupled synergy degree, the indicator system method, and the synergistic emission reduction cross elasticity method [18,19,20]. These studies provide a useful frame of reference for this paper, but there is some room for expansion. Firstly, in terms of research objects, most of the existing literature uses air pollutant and carbon dioxide emissions to discuss the pollution reduction or carbon reduction effects of environmental policies from a single perspective, ignoring the isotropic nature of pollutant and greenhouse gas emission reduction, and lacking an answer to the degree of synergy of emission reduction between the two. Secondly, although the pollution emission index can simply and intuitively reflect the level of regional pollution, it does not incorporate the impact of economic activities, which severs the correlation between the environment and the economy and cannot effectively reflect the connotation of sustainable development. Thirdly, in terms of research content, some studies have shown that reducing coal consumption and developing cleaner energy sources can significantly reduce air pollution and carbon emissions [21,22,23,24]. However, in the current discussion on the environmental benefits of the CEDP policy, most of the research topics focus on carbon emission reduction, ignoring the characteristics of air pollutants and carbon dioxide that have the same roots and origins, so there is no conclusive conclusion that this energy–environment policy promotes synergies with pollution and carbon reduction.

In view of this, this paper explains the policy effects of the CEDP pilot comprehensively from the perspective of synergistic reduction in air pollution and carbon dioxide and tries to reveal the mechanism of the pilot policy. This paper has three valuable contributions at the level of the existing literature. First, this study is one of the first articles to validate the synergistic promotion of air pollution and carbon reduction by CEDP policies through data, which makes up for the lack of the existing literature in this regard. In the previous studies on CEDP policies, most of the literature focuses on the strategic background of CEDP construction [25], the implementation path [26,27], the initial effect [28], and the pilot experience [29,30], etc. However, there are only a few policy assessments based on empirical analyses, and they only focus on a single carbon emission reduction effect [31,32,33], lacking discussion of the degree of emission reduction synergy. We assessed the environmental benefits of the CEDP pilot by providing empirical data, confirming that the policy improved the synergies between air pollution and carbon emissions and helped to reduce the intensity of air pollution and carbon emissions in the pilot provinces.

Second, this study uses the pollution intensity indicator instead of the pollution emission indicator to reflect the regional pollution level while taking into account the impact of economic activities, revealing the positive impact of the clean energy transition programs represented by the CEDP pilot on environmental protection and economic growth, which further enriches the theoretical study of air pollution and carbon emission reduction.

Third, this study further clarifies the channels through which the CEDP policy works to achieve synergy of air pollution and carbon reduction. Previous studies have shown that environmental regulatory policies can contribute to green economic growth in a variety of ways [34,35,36,37,38]. Based on the existing literature, we discuss the positive effects of industrial upgrading effects and energy use efficiency in the process of CEDP policies to improve environmental quality. In addition, we find that the positive environmental effects of CEDP policies are more pronounced in the central and western parts of the country. This study provides rich references and guidance for policymakers to further improve the CEDP policy and promote other environmental protection policies.

Figure 1 shows the research framework of this article. The subsequent sections will follow the structure of a literature review, theoretical analysis and research hypotheses, research design, empirical results, and policy recommendations.

2. Literature Review

As energy and environmental issues have received widespread attention, creating a sustainable economic and social future through the development of clean energy is an important issue. This paper provides a rough overview of the domestic and international literature, and the research views are broadly divided into two categories; one is that the development of clean energy can help to solve the energy and environmental problems, and the other emphasizes the damage to ecological and environmental quality caused by the development of clean energy. Under the first viewpoint, scholars basically make it clear in their studies that fossil energy sources have environmental pollution problems [39,40,41], while the development of cleaner energy systems has a positive impact on air pollution reduction [42,43].

In terms of research strategies, domestic scholars mainly explore the emission reduction effects of renewable energy development based on energy structure optimization, energy conservation, and emission reduction policies, especially in improving air quality and reducing carbon emissions. For example, Chi et al. [44] empirically demonstrated that a greener energy structure leads to lower carbon emissions, highlighting the beneficial effects of energy transformation on climate change. Tang et al. [45] identified energy structure optimization as a key factor that influences both air pollution and carbon abatement, particularly in China. Zhang et al. [46] conducted empirical research showing that “coal to gas” energy transition initiatives from China significantly reduced the air quality index by 20.4%, PM2.5 concentration by 18.59%, and sulfur dioxide concentrations by 68.4% during winter heating. Furthermore, there is compelling evidence of the emission reduction benefits from clean energy development at various levels, including those of countries [47], provinces [48], cities [49], industries [50], and sectors [51]. The answers of foreign scholars to the question of whether clean energy development helps to reduce environmental pollution are mainly discussed through the development of renewable energy in specific countries. For example, Said [52] studied sub-Saharan African countries and found that the use of renewable energy helped to reduce CO₂ emissions. Bekhet and Othman [53] emphasized that renewable energy is one of the important factors in improving the quality of the environment in Malaysia. Mehrdad and Azin [54] advocated the wide application of wind energy in urban area planning for ecological sustainability in the form of a case study. Esdras and Wang [55] highlighted geothermal energy as one of the promising renewable energy sources.

However, some works in the literature have opposing views; for example, Yang et al. [56] systematically discussed the connotations, driving mechanisms, and multidimensional effects of an energy transition, emphasizing that the massive exploitation and application of wind and solar energy and key minerals can create new environmental hazards and have extensive ecological impacts. Based on data from Africa, Sai et al. [57] found that expanding investment in clean energy negatively affects carbon productivity, but this effect decreases as economic development advances. Li et al. [58] empirically showed that efficiency improvements in the renewable energy production sector can have both economic expansion and energy rebound effects that increase carbon emissions. Nonetheless, considering the multiple correlates of lower pollution and carbon emissions, increasing the share of clean energy is the most effective and least costly management measure to address the pressure of long-term growth in energy demand and to create a sustainable future [13]. Therefore, this paper deserves to be discussed in depth on the issue of synergies between clean energy development policies for air pollution and carbon reduction, which will help to deepen our understanding and reflection on the existing literature.

When the focus of our literature search fell on the CEDP policy, we found that some scholars have made slight progress in evaluating the effectiveness of CEDP construction. Zhou et al. [31] discussed the effect of CEDPs upon CO₂ emission and economic growth in pilot areas, based on Chinese provincial data. Li et al. [32] evaluated the effects of the formulation of a CEDP project in Zhejiang and found period heterogeneity in the carbon abatement performance over pilot policies. Chen et al. [33] suggested that in a more micro-observational dimension, the pilot policy improved the carbon emissions and carbon intensity of CEDP county units. These varying conclusions highlight the need for an unbiased assessment of the impact of CEDP policies as a practical tool. Furthermore, insights from research on other energy policies such as new energy demonstration cities [59], low-carbon pilot cities [60], and clean heating programs [61] offer valuable references for this study. Although there is no energy policy similar to the CEDP policy in foreign countries, projects in some countries targeting the promotion of renewable energy can help us investigate the emission reduction effect of the CEDP policy. For example, the United Kingdom built a zero-carbon community in 2002 to promote a green and low-carbon lifestyle in the community and implement energy-saving and emission reduction policies, which effectively reduced the cost of living in a low-carbon and green environment. The city of Freiburg in Germany advocates zero-carbon regional energy and vigorously promotes the development of renewable energy sources such as whole-roof photovoltaic panels, energy storage, and solar thermal energy, which lead to a year-on-year reduction in per capita carbon dioxide emissions.

Therefore, based on a comprehensive review of the existing literature, we have to emphasize that this study is one of the first articles to validate the synergistic promotion of air pollution and carbon reduction by the CEDP policy through data, which helps policymakers identify and quantify the actual effects of the CEDP policy in achieving air pollution and carbon reduction, extends theoretical studies on the effectiveness of environmental regulatory policies, and provides strong evidence from a policy perspective to prove that clean energy development has environmental benefits.

3. Mechanism Analysis and Research Hypotheses

This section picks up on the theme from the previous chapter and continues to elaborate on the reasons why the formulation of the CEDP policy can achieve positive impacts on air pollution and carbon abatement. Carbon dioxide and conventional pollutants have the same origins, which means that it is practicable to address synergistic management of air pollution and carbon abatement by reducing fossil energy use. First, establishing a CEDP pilot will help the region build a clean and diversified energy system, cut down fossil fuels use, and increase the proportion of clean energy from the energy supply side, thereby reducing the production of pollutants and greenhouse gases. Second, CEDP formulation will encourage the pilot areas to innovate methods of energy use, accelerate the transformation of cleaner production in key industries, widen cleaner energy consumption, and improve the efficiency of resource and energy use, thus synergistically promoting a reduction in pollution from the energy consumption side. Therefore, the CEDP pilot program is expected to reduce air pollution and carbon emission intensities.

Hypothesis 1. 

CEDP formulation facilitates synergies for air pollution and carbon abatement in pilot provinces.

3.1. Industrial Upgrading Effect

The priority target of the CEDP program is to reduce coal consumption as well as boosting non-fossil energy utilization, which directly forces the pilot province to curb unreasonable energy consumption, promote the upgrading of heavy industrial enterprises, eliminate backward production capacity, and promote the exploitation and industrial application of clean energy and new energy storage technologies, so as to ultimately translate into passive upgrading of the conventional industry. The clean energy industry is considered a strategic emerging sector [62], and energy transformation is expected to shift the energy industry supply chain from fossil-fuel-centered to clean-energy-centered industrial clusters. This shift is expected to drive the development of new energy vehicles, photovoltaic (PV) power generation, low-energy buildings, and other green and low-carbon industries. Industrial upgrading raises entry barriers for enterprises, guiding existing industrial entities to adapt and survive in the long term. This process will also encourage the adoption of clean technology, upgrades to production processes, and a shift towards high-end environmental protection and green practices to ultimately reduce pollutant emissions.

3.2. Energy Efficiency Improvement Effect

The development and utilization of clean energy are primarily based on distributed power generation methods [63]. The comprehensive implementation of the CEDP program will facilitate the transition to cleaner energy sources in the pilot provinces on both the supply and consumption sides, leading to a significant increase in the ratio of electric energy in the end-use energy consumption structure. The data indicate that the efficiency of electric energy utilization at the terminal is 3.2 times that of oil and 17.3 times that of coal. Moreover, for every 1.0 standard coal increase in electricity substitution, the energy consumption required for a GDP of CNY 10,000 is reduced by 0.76 standard coal [64]. To comply with traditional energy consumption reduction requirements, pilot provinces should enhance energy efficiency management in key industries such as steel and mining through technological upgrades and process optimization. They will also raise energy efficiency and environmental protection standards for energy-using equipment, expedite clean projects such as industrial energy conservation and green transportation, and enhance the overall energy efficiency of the industrial supply chain in these provinces. Improving energy efficiency can reduce energy consumption and the waste of raw materials, lower energy costs for enterprises, stimulate energy-saving potential, and reduce pollutant emissions.

Hypothesis 2. 

Industrial upgrading and energy efficiency improvement are the mechanism channels through which CEDPs can exert synergistic impacts on air pollution and carbon abatement.

4. Research Design and Data Sources

4.1. DID Model

Classical DID designs typically involve two cohorts and periods. Given that the CEDPs are established gradually, we constructed a multiperiod difference-in-differences model using Bacon’s [65] research. Previous studies using the DID model have often only accounted for time differences that do not vary with individuals and individual differences that do not vary with time in the sample, ignoring individual differences over time. Therefore, we referred to Bai [66] to incorporate time-varying individual differences by introducing interaction fixed effects. This addition helped to improve the control of variable factors at the individual level. The base model was designed as follows:

$$Y_{i,j,t}={\alpha }_j+\beta {CEDP}_{i,t}+\lambda X_{i,t}+{\mu }_i+{\pi }_t+{\varphi }_{it}+{\epsilon }_{i,j,t},$$

(1)

where Y_i,j,t represents the jth variable in tth year of the ith province, and j = 1, 2, 3 where j = 1 indicates the PM2.5 intensity, j = 2 is the carbon emission intensity, and j = 3 is the synergy between the two. CEDP_i,t is a policy dummy variable with CEDP_i,t = 1 if a province is identified as a CEDP pilot in a given year and 0 otherwise. X_i,t is the set of control variables, with μ_i denoting the fixed province, π_t denoting the fixed time, and ϕ_i,t denoting the fixed interaction variables.

4.2. Variable Selection

4.2.1. Explained Variables

The explained variables in this paper are PM2.5 emission intensity, CO₂ emission intensity, and the synergy between them. Although some studies have used SO₂ and CO₂ as explained variables [19], PM2.5 is a more complex source than a single pollutant and is one of the main influences on ambient air quality exceedances in China, making it more valid to use both PM2.5 and carbon dioxide data to measure emissions reductions. As China’s significant inter-provincial economic disparities highlight the importance of promoting ecological improvement and economic growth to ensure sustainable development, the use of a pollution intensity indicator (pollution emissions per unit of GDP) is more consistent with China’s long-term goal of achieving both quantitative growth and qualitative improvement in economic development.

In addition, the synergy of pollution and carbon reduction requires that pollution reduction be accompanied by carbon reduction and that carbon reduction be accompanied by pollution reduction, which is an interrelated and mutually reinforcing process between the two systems of pollution and carbon reduction [67]. Some scholars have used the coupled coordination degree approach to measure the level of synergy between carbon and pollution reduction; this is because the coupling degree is an indicator that expresses the strength of the interaction between subsystems [68], and the coordination degree is the degree of synchronized change between subsystems [69]. Referring to Yang’s approach [18,70], we used the coupled coordination degree model to construct the pollutant’s synergistic degree indicator, which is constructed as follows:

$$C={\displaystyle \frac{2\times \sqrt{Z_1\times Z_2}}{{(Z}_1+Z_2)}}$$

(2)

$$T=a\times Z_1+b\times Z_2$$

(3)

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

(4)

where Z₁ and Z₂ are the normalized values of the PM2.5 emission intensity and the CO₂ emission intensity, and C denotes the degree of coupling between the two. T is the composite coordination index, where pollution reduction and carbon reduction play equal roles, and a,b are treated as 0.5. Synergy is the degree of synergism between pollution reduction and carbon reduction indicators, which reflects both the coupling relationship and the level of synergism.

Some studies on the effects of environmental policies based on the coupled coordination degree approach and double difference models provide literature support for the writing of this paper. For example, Luo and Lei [71] constructed a coupled coordination degree model with CO₂ and SO₂ and used the fixed-effects double-difference method to analyze the synergistic effect of pollution and carbon reduction of China’s carbon market policies. Therefore, PM2.5 emission intensity (PM2.5 EI), carbon emission intensity (Carbon EI), and the synergism (Synergy) are used as explained variables below, respectively.

4.2.2. Core Explanatory Variables

The dichotomous dummy variable CEDP was the core explanatory variable used for this study. The variable CEDP = 1 in a given year if a province is identified as a pilot province in that year and thereafter, and 0 in all other cases. In this paper, these pilot provinces are in the following time windows with CEDP = 1 (Ningxia in 2012 and later, Zhejiang in 2014 and later, Sichuan, Tibet, Gansu in 2016 and later, Qinghai in 2018 and later).

4.2.3. Control Variables

This study introduced several indicators as control variables that can affect the evaluation of policy effects in CEDPs. These indicators are as follows:

• Level of economic development (lnDIPC), measured as the logarithmic form of per-capita disposable income. Per capita disposable income and per capita GDP are both important indicators of the level of economic development, but they reflect different emphases. Although GDP per capita can reflect the economic strength of a region, it is not directly equivalent to the income level of residents. The effectiveness of clean energy policy implementation depends largely on the acceptance and consumption behavior of residents. For example, when promoting clean energy, groups in high-income areas may be more likely to bear the initial installation and usage costs, while groups in low-income areas may require more financial subsidies and support. By controlling for per capita disposable income, the study can better identify and address these potential inequalities, and thus more effectively assess the implementation effects and social impacts of clean energy policies.
• Population density (lnPD), measured in the logarithmic form of the population per square kilometer. In vast geographical areas, the total population may be high, but the distribution can be very scattered. Population density directly reflects the degree of population concentration within a specific area. In densely populated regions, energy demands, air pollution, and carbon emissions may be more concentrated, leading to a greater willingness to promote clean energy.
• Degree of openness to the outside world (lnFIA), measured as the logarithmic form of the number of foreign-registered enterprises. The number of enterprises registered by foreign investors can reflect the degree of foreign capital participation and the openness of the market. An increase in the number of foreign-registered enterprises in a region typically signifies more foreign capital entering the area. This not only brings in funds but may also introduce technology and management experience, which can help promote the application and development of clean energy technology.
• Energy consumption demand (lnREC), measured as the logarithmic form of the total social electricity consumption in a region. Total societal electricity consumption covers the power consumption of all industries, including industrial, commercial, and residential sectors, thereby providing a more comprehensive reflection of a region’s energy demands.
• Government intervention (PFE), measured as the proportion of fiscal expenditure to regional GDP. The government’s fiscal policies and financial support are key factors in promoting the development of clean energy. By providing fiscal subsidies, tax incentives, and government procurement, they encourage businesses and the public to adopt clean energy. Controlling the extent of fiscal intervention helps to accurately identify the effectiveness of clean energy policies.
• Green lifestyle tendency (lnEPA), measured as the logarithmic form of the number of urban parks. The increase in the number of parks is usually associated with the degree of emphasis on a clean environment by the government and the public. More parks imply greater ecological investment in cleanliness and greening. Compared to some environmental protection tendency indicators that may be difficult to quantify, the number of parks is a concrete and easily countable metric, which can help our readers to understand and analyze.
• The level of industrialization (lnIDL), expressed in the logarithmic form of the per capita industrial output value. Regions with higher levels of industrialization often have more heavy industry and manufacturing, which are typically major emitters of pollutants and carbon emissions. The population base varies greatly in different regions. If only the total industrial output value is used as the control variable, it is difficult to control the impact of population size on industrial development. It is more reasonable to use per capita industrial output value to compare the industrialization level of different regions.

4.3. Data Sources

Considering data availability, the data used in this study were collected from 30 provinces in China from 2005 to 2020, excluding Hong Kong, Macao, Taiwan, and Tibet. Economic and social indicators for the provinces were primarily sourced from the China Statistical Yearbook and provincial statistical bureaus, whereas data on pollutant emissions such as PM2.5 and CO₂ were acquired from the Multi-resolution Emission Inventory for China model.

Table 1. Descriptive statistics of the main variables used in the study.

Variable	Observations	Mean	Standard Deviation	Minimum Value	Maximum Value
PM2.5 EI	480	32.0403	33.2016	0.8527	246.0764
Carbon EI	480	2.5317	2.2012	0.1830	13.5984
Synergy	480	0.8115	0.2323	0	0.9995
CEDP	480	0.0604	0.2385	0	1
lnDIPC	480	9.8668	0.4284	8.9861	11.1877
lnPD	480	7.8052	0.5190	5.2444	8.7496
lnFIA	480	8.7670	1.3522	4.8203	12.0966
lnREC	480	7.1422	0.7652	4.4188	8.8452
PFE	480	0.2284	0.0983	0.0870	0.6430
lnEPA	480	5.5161	0.9739	2.8519	8.3736
lnIDL	480	9.4796	0.6693	7.5427	10.8105

presents the descriptive statistics for the main variables.

5. Results

5.1. Provincial Differences in Pollution and Carbon Emission Intensity

In this paper, a quantile classification is used to describe the pollution intensity and carbon dioxide intensity of Chinese provinces in 2005, 2010, 2015, and 2020 in a hierarchical manner, with a total of five categories, where the color of the markers changing from green to blue implies that a province is at a higher grassroots level of air pollution or carbon dioxide intensity in the country as a whole (Figure 2). From the perspective of China’s four major economic regions, the eastern and central regions show a tendency to move towards the lower end of the scale in terms of both air pollution and carbon intensity, while the western and northeastern regions show the opposite.

This phenomenon reflects the essential differences between regions in terms of economic development, energy consumption, and policy implementation. Due to earlier industrialization and urbanization, strong economic fundamentals, industrial upgrading, and technological innovation in the eastern and central regions, coupled with policy impetus and increased public awareness of environmental protection, these regions have made significant progress in air pollution control and carbon emission reduction, with the air pollution and carbon dioxide intensity moving towards a lower gradient. Conversely, the western and northeastern regions are facing more environmental pressures in the course of their development due to their relatively slow economic development, and with the implementation of the Western Development Strategy and Northeastern Revitalization Strategy in recent years, industrialization and urbanization have been pushed forward in a big way, which has led to a gradual increase in the intensity of pollution and carbon emissions. This highlights the urgency and necessity of formulating a clean energy demonstration policy represented by the clean energy demonstration provincial pilot under the strategic requirements of China’s ecological civilization and the goal of achieving peak carbon and carbon neutrality. From the perspective of policy formulation, clean energy demonstration policies have multiple benefits, such as reducing fossil energy consumption, promoting clean energy technological innovation, and enhancing public awareness of environmental protection.

5.2. Benchmark Regression

This study quantitatively assessed CEDP formulation effects using a DID model. Before the empirical analyses, we tested the data for smoothness according to the LLC, IPS, Fisher-ADF, and Hadri-LM methods, and all the tests supported the assumption condition that the variables were smooth (

Table A2. Results of variable smoothness test.

	LLC Test	IPS Test	Fisher-ADF Test	Hadri Test
Variable	t Value	w-t-Bar Value	chi-Aquared Value	z Value
PM2.5 EI	−30.8033 ***	−13.5085 ***	261.0510 ***	9.1429 ***
Carbon EI	−14.1033 ***	−5.7767 ***	305.7824 ***	6.6782 ***
Synergy	−10.4971 ***	−1.3258 *	96.6962 ***	4.2700 ***
lnDIPC	−4.3430 ***	−2.1741 **	166.0780 ***	4.3655 ***
lnPD	−11.3186 ***	−9.2946 ***	112.3761 ***	8.3065 ***
lnFIA	−5.8155 ***	−13.4757 ***	218.3640 ***	7.2267 ***
lnREC	−6.1137 ***	−3.4011 ***	165.9829 ***	6.1011 ***
PFE	−1.9241 **	−2.0539 ***	218.4706 ***	7.9200 ***
lnEPA	−20.4257 ***	−4.6930 ***	186.2338 ***	8.2746 ***
lnIDL	−5.5160 ***	−7.7631 ***	132.4283 ***	8.1844 ***

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, and * p < 0.1.

).

In the actual regression equation (

Table 2. Benchmark regression results of the CEDP pilot with PM2.5 emission intensity, carbon intensity, and the synergy between the two.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	PM2.5 EI	Carbon EI	Synergy	PM2.5 EI	Carbon EI	Synergy
CEDP	−1.9531 **	−0.2465 **	0.0274 ***	−2.0051 **	−0.2937 ***	0.0278 ***
	(0.83)	(0.21)	(0.01)	(0.80)	(0.11)	(0.01)
lnDIPC				3.6827	0.2781	0.0254
				(3.32)	(0.37)	(0.03)
lnPD				−1.7863 ***	0.1354 **	−0.0023
				(0.53)	(0.07)	(0.00)
lnFIA				5.5912 ***	−0.1322	−0.0161 **
				(1.01)	(0.11)	(0.00)
lnREC				0.9437	0.7848 ***	−0.0250
				(1.79)	(0.28)	(0.02)
PFE				−9.1476	2.6097 ***	−0.5080 ***
				(7.42)	(0.81)	(0.06)
lnEPA				−0.5475	−0.2569 **	0.0194 **
				(1.08)	(0.12)	(0.01)
lnIDL				−3.8230 ***	−0.6690 ***	0.0418 ***
				(1.26)	(0.15)	(0.01)
_cons	−32.1586 ***	2.5466 ***	0.8098 ***	−4.5688	1.4614	0.5097 *
	(0.12)	(0.02)	(0.00)	(37.80)	(3.98)	(0.34)
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Pro-Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	480	480	480	480	480	480

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, and * p < 0.1.

), models (1), (2), and (3) are the regression results without adding control variables and models (4), (5), and (6) are the regression results with the addition of control variables. We found that regression coefficients for the key variables are significant regardless of whether control variables are included or not. Specifically, columns 4 and 5 in Table 2 highlight that the CEDP policy effectively reduced PM2.5 and carbon emissions in the pilot provinces, with statistical significance at the 5% level. The regression results in column 6 show that CEDP policy implementation increased the synergy between air pollution and carbon emission reductions in the pilot provinces; these findings confirmed Hypothesis 1, which stated that the CEDP formulation exerted synergistic pollution-controlling effects in the pilot regions.

5.3. Regional Heterogeneity

This study examined the varied environmental governance impacts of CEDP policies in the eastern and central–western regions using a group regression analysis (

Table 3. Regression results for subregional groupings.

Variable	The Eastern Area			The Midwest Area
	(1)	(2)	(3)	(4)	(5)	(6)
	PM2.5 EI	Carbon EI	Synergy	PM2.5 EI	Carbon EI	Synergy
CEDP	−1.0131	−0.0571	0.0030	−2.7652 **	−0.3115 **	0.0186 **
	(1.32)	(0.09)	(0.01)	(1.14)	(0.14)	(0.01)
_cons	−35.1900	−2.3932	1.4442 ***	−46.9687	50.9235 ***	1.1144 **
	(56.84)	(3.78)	(0.27)	(130.31)	(14.31)	(0.47)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Pro-Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	208	208	208	272	272	272

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05.

). Columns (1), (2), and (3) provide the results of the test for the eastern region, where the regression coefficients of the CEDP were not statistically significant, suggesting that the CEDP pilot program did not notably influence air pollution and carbon reduction in that region. Columns (4), (5), and (6) show the findings for the central and western regions, where the regression coefficients of the CEDP were significant, which indicated that CEDP implementation in these regions effectively promote synergy in pollution reduction and reduced the intensity of PM2.5 and carbon emissions in the pilot provinces.

Regional disparities were attributed to three main factors. The first is economic level and energy dependence. On the one hand, the eastern region is the most socio-economically developed region in China, bringing together major urban agglomerations such as Beijing–Tianjin–Hebei, the Yangtze River Delta, and the Pearl River Delta, and accounting for about 55% of the country’s gross regional product. On the other hand, China’s energy resources are unevenly distributed. In terms of coal resources and consumption, more than 86% of the country’s coal resources are concentrated in the northwestern regions such as Shanxi, Inner Mongolia, Xinjiang, and Shaanxi, while consumption is mainly concentrated in the eastern coastal regions, which has led to the energy flow patterns of ‘northern coal transported south’ and ‘western coal transported east’. In addition, most of the electricity production comes from coal power generation, and the eastern region is a major electricity load region, with an annual share of electricity consumption close to 50% of the national consumption and a perennially high demand for electricity and coal. As a result, the economic development of the eastern region is highly dependent on energy, and although the CEDP pilot can promote the development of clean energy, especially clean electricity, it is still unable to meet the needs of economic development, which causes the CEDP policy to have a non-significant positive impact on air pollution and carbon emissions in the eastern region. While the central and eastern regions have a lower level of economic development and rich natural resources, the development of clean energy is a greater alternative to traditional energy sources, which causes the policy implementation in the central and western regions to show more obvious synergistic effects in pollution reduction and carbon emission reduction.

The second factor is resource abundance and CEDP policy implementation. The abundance of water, wind, and solar energy resources in the central and western regions, as well as lower land use and development costs, means that the central and western regions are more conducive to clean energy infrastructure development and large-scale exploitation and utilization than the eastern regions, and the implementation of CEDP policies is more vigorous. For example, Qinghai Province is located in the core area of the Earth’s Third Pole, the birthplace of the Yangtze, Yellow, and Lancang rivers, and has the second-highest total annual solar radiation in China, with the lowest cost of PV power generation in the country [72]. Relying on the advantages of clean energy resources, Qinghai Province has integrated the construction of a clean energy demonstration province into the planning of economic and social development, strengthened the construction of energy infrastructure, vigorously implemented key energy projects, built the country’s first 100-megawatt photovoltaic power generation base, and a number of key projects have been selected for the national demonstration project, which has become an important new energy industry base in the country [73]. Sichuan Province is vigorously building a clean energy demonstration province, introducing relevant policies to support the development of renewable hydropower industry, and the installed capacity of hydropower is the first in the country [74]. Gansu Province has actively cultivated clean energy industry clusters, built the country’s first 10 million kilowatt wind power base, and orderly pushed forward the construction of multi-million-kilowatt wind power bases in Tongwei, Minqin, Huanxian, Baiyin, etc., as well as basically forming a complete industry chain of wind power, photovoltaic equipment manufacturing, and power generation and transmission. The eastern region CEDP pilot is dominated by Zhejiang Province, in contrast, where due to land scarcity and high population density, part of the construction of clean energy projects is subject to land acquisition and compensation, environmental impact, public sensitivity, and other factors, and in addition to solar power generation, most of the cost of production of renewable energy is higher than conventional energy products [75]. In addition, the implementation of CEDP policies in Zhejiang Province has been limited, mainly through the promotion of clean coal power conversion, equal or reduced energy use replacement in projects, green buildings and green transport, distributed photovoltaic power generation, and offshore wind power, etc., and these measures have had a limited impact on the realization of cleaner alternatives and the establishment of a diversified energy use system.

Thirdly, there is a difference between the industrial base and costs of technological upgrading of enterprises. Industrialization in the eastern region started earlier. The stock of large, rigid, traditional, high-energy-consuming enterprises has formed the production inertia of the existing facilities. The equipment used for cleaner production upgrade costs are higher, greatly limiting the rapid upgrading of the industrial structure and clean technology. By undertaking industrial transfer from the eastern region, the central and western regions have abundant renewable energy sources such as wind power, photovoltaic, photothermal, hydropower, etc., to achieve local consumption, which can transform the advantages of new energy into advantages of low-carbon development and promote long-term energy conservation and environmental protection of enterprises. These factors enable the central and western regions to benefit more directly and effectively from the implementation of CEDP policies.

5.4. Robustness Test

5.4.1. Parallel Trend Test

The theoretical prerequisite for the validity of the double-difference estimation was the fulfillment of the parallel trend assumption [76]. This assumption necessitates that there is no heterogeneous time trend in the key indicators of either the experimental or control groups before policy implementation. It is important to note that the parallel trend test requires both individual and joint significance. Individual significance requires that the difference between the treatment and control groups at a given point in time before the policy is implemented is not significant and that the test coefficient is not significantly different from zero. However, even if the differences at individual time points are not significant, the cumulative differences at multiple time points may be significant. Therefore, joint significance is a complement to individual significance tests to determine whether there are systematic differences between the treatment and control groups at all relevant time points before and after policy implementation. If the joint significance test fails, it means that the trends of the treatment and control groups had already begun to diverge before the implementation of the policy, which again violates the assumption of parallel trends. Only when both individual and joint significance tests are passed can we be more confident that the estimates from the DID model are reliable.

Figure 3 presents the outcomes of the parallel trend tests for PM2.5 intensity, carbon emission intensity, and synergy between the two, respectively (

Table A3. Regression results of the parallel trend test.

Test Variable	Relative to Policy Time
PM2.5 EI	pre8	pre7	pre6	pre5	pre4	pre3	pre2	current
	−8.9156	0.2927	1.5015	0.8728	−1.9874	−0.6819	2.4692	−2.9367 *
	(8.50)	(8.46)	(7.07)	(4.31)	(2.01)	(3.77)	(1.77)	(1.51)
	post1	post2	post3	post4	post5	post6	post7	post8
	−1.7358	−3.3012	−1.7706	−4.6286	−4.6992	−9.1714 *	−14.7977 ***	−18.3705 ***
	(2.10)	(2.19)	(3.23)	(3.97)	(3.66)	(4.60)	(5.15)	(5.11)
Carbon EI	pre8	pre7	pre6	pre5	pre4	pre3	pre2	current
	−0.0532	0.1854	0.4880	0.4693	0.1656	0.2163	0.0755	−0.1662
	(0.37)	(0.33)	(0.32)	(0.27)	(0.24)	(0.21)	(0.09)	(0.13)
	post1	post2	post3	post4	post5	post6	post7	post8
	−0.3911	−0.4655	−0.5266	−0.6036	−0.8299	−1.0056 **	−1.6879 ***	−1.8280 ***
	(0.29)	(0.36)	(0.49)	(0.58)	(0.54)	(0.47)	(0.27)	(0.29)
Synergy	pre8	pre7	pre6	pre5	pre4	pre3	pre2	current
	−0.0069	−0.0065	−0.0128	−0.0350	−0.0223	−0.0129	−0.0145	0.0015
	(0.04)	(0.05)	(0.05)	(0.05)	(0.05)	(0.05)	(0.05)	(0.05)
	post1	post2	post3	post4	post5	post6	post7	post8
	0.1538 ***	0.1616 ***	0.1929 ***	0.2043 ***	0.2862 ***	0.3035 ***	0.4907 ***	0.5160 ***
	(0.05)	(0.05)	(0.05)	(0.05)	(0.06)	(0.06)	(0.08)	(0.08)

Note: pre# indicates before the implementation of the policy, and post# indicates after the implementation of the policy. Standard errors in parentheses. *** p < 0.01, ** p < 0.05, and * p < 0.1.

provides detailed values for the regression coefficients). The x-axes in the three graphs represent the relative time of policy implementation in the CEDPs, and the y-axes represent the corresponding regression coefficients, and the red line indicates that the base period is picked to be 1 year before the policy. These coefficients reflect the variability in key indicators between CEDPs and non-CEDPs. Prior to policy implementation, the regression coefficients fluctuated around 0 and fell within the 95% confidence interval; the F-values of the pre-test of joint significance were 1.9323 (p-value was 0.1004), 0.8607 (p-value was 0.5482), and 0.1169 (p-value was 0.9972), respectively, and both were non-significant, suggesting that there were no significant differences between the two types of provinces and satisfying the parallel-trend assumption. After policy implementation, the negative regression coefficient gradually increased in absolute value; the F-values of the post-test of joint significance were 4.5027 (p-value was 0.0009), 8.3185 (p-value was 0.0000), and 10.0206 (p-value was 0.0000), respectively, and both were significant, indicating that the CEDP policy significantly contributed to the coordinated management of air pollution and carbon abatement in the pilot area.

5.4.2. Placebo Test

The baseline regression results indicated that the pilot provinces successfully reduced air pollution and carbon emissions. However, there may have been other unobservable factors influencing the robustness of the causal inference in the DID analysis that potentially led to a biased policy effect. To address this issue, a placebo test inspired by Ma et al. [77] was conducted using randomly generated experimental groups. Figure 4 shows kernel density distribution plots of the regression coefficients from 1000 random samples that illustrated that the estimated coefficients were predominantly centered around zero. The actual value from the benchmark regression was a rare occurrence in the placebo test, indicating that the DID policy effects observed in this study were not due to chance or external factors. This result confirmed the robustness of the benchmark regression results.

5.4.3. Instrumental Variable Regression

The benchmark regression added province-fixed effects, year-fixed effects, and province–year-interaction-fixed effects, which we used along with instrumental variables to address potential endogeneity. Some studies have used lagged terms or group means of endogenous variables as instrumental variables; however, these may not meet exogenous requirements [78,79]. Unlike traditional energy sources, renewable energy depends on climatic elements, such as rainfall, wind speed, insolation, temperature, and humidity; therefore, a single instrumental variable may not have sufficient explanatory power. Thus, using methods described by Li and Cheng [80] and Bahar [81], we introduced synthetic instrumental variables as instrumental variables for the CEDP pilot provinces.

Theoretically, factors such as the number of sunshine hours, terrain relief, and the air circulation coefficient can indicate the climatic conditions of a region. As we know, the choice of instrumental variables needs to satisfy both relevance and exogeneity. Firstly, in terms of relevance: (1) The pilot clean energy demonstration province is closely related to the use of renewable energy, especially solar energy. The number of sunshine hours directly affects the potential and efficiency of solar power generation; therefore, the correlation between the number of sunshine hours and the clean energy demonstration policy is high. (2) The air circulation coefficient affects the distribution and efficiency of wind energy resources, and wind energy is an important part of clean energy. Therefore, the correlation between air circulation coefficients and clean energy demonstration policies is also high. (3) Terrain relief affects the distribution of wind and solar energy; for example, wind power generation may be more effective in areas with greater terrain relief, and solar power generation can also obtain more hours of sunshine under certain terrain conditions; therefore, the correlation between terrain relief and clean energy demonstration policies is also high. Secondly, in terms of exogeneity: (1) The number of sunshine hours is mainly determined by geographic location and climatic conditions, which are usually independent of local economic activities and environmental pollution and thus can be considered exogenous. (2) Air circulation coefficients are affected by atmospheric circulation patterns, which are usually determined by the global climate system and are not easily affected by changes in policies in a single region and thus are also exogenous. (3) Topographic relief refers to natural geographic features that are not easily influenced by human activities and are exogenous. We introduced sunshine hours, terrain relief, and air circulation coefficient as potential instrumental variables, and the first- and second-order lag terms of CEDPs were applied to interact with the time dimension and were included in an ordinary least squares (OLS) regression to determine whether a province was a CEDP pilot province or not. The fitted values were then used as instrumental variables for CEDPs.

To mitigate overfitting issues, we used instrumental variables–Lasso regression in our analysis. The results presented in columns (1), (2), and (3) of

Table 4. Results of robustness testing using instrumental and replacement variables.

Variable	Instrumental Variable Regression			Substituting Key Variables
	(1)	(2)	(3)	(4)	(5)	(6)
	PM2.5 EI	Carbon EI	Synergy	PM2.5 EI	Carbon EI	Synergy
CEDP	−6.4814 **	−0.9823 ***	0.0811 ***	−0.0660 **	−0.0971 ***	0.0190 **
	(3.07)	(0.17)	(0.02)	(0.03)	(0.02)	(0.01)
_cons	365.2312 ***	13.4551 ***	0.9451 ***	−4.9250 **	5.0532 ***	1.7947 ***
	(29.29)	(1.36)	(0.09)	(1.11)	(7.77)	(0.33)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Pro-Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	480	480	480	480	480	480

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05.

indicate that even after accounting for endogeneity concerns, the regression coefficients of the key explanatory variables were significantly negative, suggesting that CEDP implementation effectively reduced PM2.5 and carbon emission intensities in the pilot provinces, reinforcing the robustness of our research findings.

5.4.4. Substituting Key Variables

This study re-evaluated the aggregate indicators of PM2.5 and carbon dioxide emissions and presented the regression results in columns (3)–(5) of Table 4. We found that the regression coefficients of the core explanatory variables were still significantly negative after the intensity indicators of the explanatory variables were replaced with the total indicators, indicating that CEDP implementation reduced the total amount of PM2.5 and carbon emissions and also improved synergies between pollution and carbon abatement in the pilot provinces. Furthermore, these findings indicated that the pilot policy played a crucial role in continuously enhancing China’s ecological quality and promoting the development of an aesthetically pleasing environment.

5.4.5. Controlling the Expected Effects of Policy

Some provinces may have experienced a long process from the proposal, declaration, and political review to the final establishment of the CEDP pilot, and these regions carry out sufficient preparation and anticipation work for the implementation of the pilot policy. To avoid estimation bias due to policy expectations in the pilot provinces, referring to the research methodology of Song et al. [82], we add dummy terms for one year prior to the pilot policy and two years prior to the pilot policy to the base regression equation. The regression results are shown in

Table 5. Robustness test results controlling for expected effects.

Variable	(1)	(2)	(3)
	PM2.5 EI	Carbon EI	Synergy
CEDP	−2.6530 ***	−0.3244 ***	0.0180 **
	(0.95)	(0.11)	(0.01)
Expectancy effect (pre_-1)	−1.5420	−0.2041	0.0101
	(1.31)	(0.14)	(0.01)
Expectancy effect (pre_-2)	−1.0200	−0.1525	0.0013
	(1.28)	(0.13)	(0.01)
_cons	3.7271	13.2693 ***	0.6910 ***
	(38.51)	(3.72)	(0.22)
Controls	Yes	Yes	Yes
Province FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Pro-Year FE	Yes	Yes	Yes
N	480	480	480

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05.

below, where it can be found that the core explanatory variables remain significant, and the coefficients of the policy dummy terms for one year and two years prior to the pilot are not significant, indicating that there are no expected benefits from the pilot policy.

5.4.6. Controlling the Impact of Non-Randomness

In this paper, CEDP pilot selection may be closely related to geographic location, existing economic level, energy consumption demand, etc., and thus provinces that are more influenced by these factors are more inclined to implement CEDP pilots. To exclude the non-randomness of sample selection, we use the propensity score matching method to predict the probability of provinces to carry out CEDP pilot projects and choose the nearest neighbor matching method to match samples to obtain more representative samples. The relevant regression results are shown in columns (1)–(3) of

Table 6. Test results for controlling non-randomization pilots.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	PM2.5 EI	Carbon EI	Synergy	PM2.5 EI	Carbon EI	Synergy
CEDP	−3.2610 ***	−0.1890 **	0.0463 ***	−1.2324 **	−0.1824 **	0.0358 ***
	(1.12)	(0.08)	(0.01)	(0.57)	(0.09)	(0.01)
_cons	43.6034	−0.0109	−1.4996 ***	80.8406	5.1516	1.2897 ***
	(51.64)	(5.04)	(0.77)	(29.72)	(3.65)	(0.48)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Controls × Year	No	No	No	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Pro-Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	197	197	197	480	480	480

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05.

, and it can be found that the regression coefficients of CEDPs on pollutant intensity are significantly negative, and the regression coefficients of pollutant synergy are significantly positive, and the conclusions of this paper are further verified. In addition, drawing on the methodology of Wang and Ge [83] and Edmonds et al. [84], the estimation bias of non-randomized choices of pilot provinces is mitigated by adding an interaction term between the baseline factor and the time linear trend in the regression, and the related regression results are shown in columns (4)–(6) of Table 6. It can be found that after considering the non-randomness of the pilot policy selection, the CEDP can still realize the synergies of pollution and carbon reduction in the pilot provinces, and the results obtained from the baseline regression estimation are robust.

5.4.7. Re-Examination of Synthetic Control Methods (SCM)

The SCM introduced by Abadie [85] and widely used in policy evaluation involves leveraging data from a control group to create a counterfactual scenario for the experimental group in the absence of policy intervention. The treatment effect can be estimated by comparing the experimental group with the synthetic control group. Figure 5 presents the fitted graphs of PM2.5 intensities, carbon emission intensities, and synergy for the actual experimental provinces (i.e., CEDP pilot provinces) and synthetic control provinces, respectively. The vertical line in the figure denotes the current period when the pilot provinces initiated CEDP implementation, with the left and right sides representing the periods before and after implementation, respectively. The horizontal axis indicates the period during which the pilot provinces implemented the CEDPs. Before policy implementation, the clean energy demonstration and synthetic control provinces exhibited a close fit with similar trends. However, following policy implementation, the PM2.5 and carbon emission intensities of the real experimental provinces gradually decreased compared to those of the synthetic control provinces, and the synergy between the two increased significantly. This indicated that CEDP implementation led to a reduction in PM2.5 and carbon emission intensity in the pilot provinces when compared with the non-pilot provinces.

Table 7. Difference and significance of core indicators between experimental and control groups after policy intervention.

Post-Treatment Results	PM2.5 EI	Carbon EI	Synergy
	Estimation (p-Value)	Estimation (p-Value)	Estimation (p-Value)
Period 1	−0.039 * (0.079)	−0.065 * (0.089)	0.036 (0.156)
Period 2	−0.052 (0.199)	−0.149 ** (0.015)	0.548 *** (0.000)
Period 3	−0.075 * (0.094)	−0.168 *** (0.009)	0.554 * (0.073)

Note: p-values are in parentheses.*** p < 0.01, ** p < 0.05, and * p < 0.1.

presents a comparison between CEDPs and synthetic control provinces following policy intervention, along with the corresponding p-values. This analysis highlighted the impact of the CEDP formulation on the core observation indicators of the experimental and control groups. The results showed a notable increase in the difference between the experimental and control provinces over time, with statistical significance for most periods. In addition, this paper conducts a counterfactual test of the SCM approach using random generation of treatment groups and policy times (Figure A1), and the counterfactual test results show that the virtual CEDP policy does not have a pollution and carbon reduction effect after changing the sample. This implies that the empirical results of the original paper did not appear randomly and that the CEDP policy significantly contributed to the coordinated management of pilot provinces in air pollution and carbon reduction efforts.

5.5. Impact Mechanism Testing

It is theorized that the CEDP can have a synergistic effect on air pollution and carbon alleviation through industrial upgrading and energy efficiency improvement, which was verified in our stepwise regression analysis. Industrial upgrading (indUP) is measured by the ratio of the value added by the tertiary industry to the value added by the secondary industry. In other words, the larger the ratio, the more advanced the industrial structure. Energy efficiency upgrading (eneUP) is measured by billion yuan of electricity consumption, where the lower the electricity consumption intensity, the higher the level of energy efficiency.

The results presented in the first and last four columns of

Table 8. Test results of the mechanism for industrial upgrading and energy efficiency improvement.

Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	indUP	PM2.5 EI	Carbon EI	Synergy	eneUP	PM2.5 EI	Carbon EI	Synergy
CEDP	0.1343 ***	−2.1720 ***	−0.2427 **	0.0264 **	−0.0059 **	−1.6750 **	−0.1664 **	0.0259 **
	(0.05)	(0.84)	(0.11)	(0.01)	(0.00)	(0.80)	(0.09)	(0.01)
indUP		−4.9285 **	−0.5295 ***	0.0367 **
		(1.92)	(0.16)	(0.02)
eneUP						34.7155 **	14.9137 ***	0.2857 **
						(15.44)	(0.99)	(0.14)
_cons	7.1025 ***	−120.2463 *	8.2598	0.6491	−0.4885 ***	2.5458	4.0379	0.7501 *
	(1.95)	(62.23)	(6.96)	(0.53)	(0.08)	(39.88)	(3.24)	(0.41)
Province FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pro-Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	480	480	480	480	480	480	480	480

Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, and * p < 0.1.

outline the outcomes of the mechanism test of the impact of industrial upgrading and energy efficiency improvement. When industrial upgrading was used as the mediating variable, the positive and significant regression coefficient of CEDPs on industrial upgrading (column (1)) suggested that CEDP implementation significantly promoted industrial upgrading in the pilot provinces. The results in columns (2) and (3) demonstrate that industrial upgrading effectively reduced PM2.5 and carbon emission intensities, indicating that industrial upgrading is one of the influencing mechanisms of the CEDP policy, confirming Hypothesis 2. In the mechanism test, when energy efficiency improvement was used as the mediating variable, CEDP implementation caused a significant reduction in the intensity of electricity use in the pilot provinces, indicating a notable improvement in energy efficiency levels when compared with that of the non-CEDP provinces. The regression results presented in columns (6) and (7) further revealed that energy efficiency upgrading significantly decreased PM2.5 and carbon emission intensities in pilot provinces, reinforcing our hypothesis that CEDP implementation achieves pollution and carbon reduction by enhancing the energy efficiency level of the pilot provinces. The regression results in columns (4) and (8) emphasize that industrial upgrading and energy efficiency upgrading significantly increase the synergy between air pollution and carbon reduction. Overall, these outcomes suggested that industrial upgrading and energy efficiency improvement are crucial mechanism channels that promote the synergistic effects of air pollution and carbon reduction in the pilot provinces.

6. Conclusions and Recommendations

This study quantitatively assessed the synergistic effects of the CEDP project on air pollution and carbon abatement using provincial-level panel data from 2005 to 2020. A multi-temporal difference-in-differences model was used to reveal the emission reduction mechanism of the pilot policy from the perspectives of industrial upgrading and energy efficiency improvement. The main findings suggest that CEDP implementation promotes a synergistic reduction in air pollution and carbon emissions. Policy implementation significantly decreased PM2.5 and carbon emission intensities in the pilot provinces, with robustness tests supporting these results. Regional analysis indicated varying sensitivity to the CEDP policy across the eastern, central, and western regions, with the policy effectively reducing emissions in the central and western regions but showing less of an impact in the east. The mechanistic tests highlighted industrial upgrading and energy efficiency enhancement as crucial pathways for achieving the synergistic effects of air pollution and carbon abatement via CEDP implementation.

In view of the above findings, this study is putting forward a number of recommendations:

• First, there is a need to focus on top-level design and enhance the evaluation system for CEDP implementation. Energy transformation requires a comprehensive approach that integrates practicality, robust policy planning, and targeted management. This necessitates refining and enhancing CEDP formulation and operational evaluation standards to holistically assess the development levels, cost–benefit, and sustainability of the provinces in the clean energy sector. The evaluation outcomes should serve as a crucial foundation for adjusting policies, allocating capital, and determining project layouts.
• Second, it is recommended to consolidate implementation experience and gradually expand the implementation of pilot programs to more provinces. Our findings indicate that CEDP implementation positively affects air pollution and carbon reduction synergy. Therefore, encouraging additional provinces and regions to establish CEDP policies can facilitate nationwide adoption of clean energy technologies and prompt localities to devise specialized clean energy development models that are tailored to their unique resource endowments. In addition, from a cost–benefit perspective, the expansion of the policy also needs to consider the economic differences and industrial characteristics between regions. The cost of implementing CEDP policies in the central and western regions is relatively low because of the availability of abundant local natural resources and the high technological starting point for new projects. The eastern region, with its high intensity of economic activities, needs to overcome more technical and economic barriers, such as eliminating outdated production capacity and upgrading existing facilities, all of which require substantial investment. Therefore, the government also needs to guide localities to strengthen synergistic cooperation to ensure the effective supply and efficient use of clean energy on a wider scale, providing strong support for ecological and environmental governance at the national level.
• Third, the transmission mechanism should be enhanced to strengthen the connection between industrial upgrades and energy efficiency improvements. Our analysis demonstrates that the CEDP policy achieved synergistic emissions reduction by combining industrial upgrading and energy efficiency improvement. Therefore, pilot provinces should expedite industrial technological transformation and upgrading; promote the adoption of new energy-efficient equipment, processes, and technologies; facilitate the transition of high-energy-consuming industries towards lower energy consumption and higher-value-added products; and achieve overall structural optimization of the industrial chain along with energy efficiency enhancement. Additionally, it is crucial to establish a robust feedback and adjustment mechanism to regularly evaluate the actual impact of industrial upgrading and energy efficiency improvements and to dynamically adjust and enhance relevant policy tools to ensure the effectiveness and adaptability of the transmission mechanism.

This study quantitatively assessed the synergistic effects of China’s CEDP pilot policy on air pollution and carbon abatement, and its findings offer a practical approach for developing China’s ecological civilization in an environmentally friendly manner. Although this study can serve as a valuable reference for environmental governance in other global economies, the provincial indicators used for analysis are only from China between 2005 and 2020, suggesting the need for further research at a more granular level, such as city or county dimensions, as well as a comprehensive examination of the construction impact, implementation costs, promotion value, and policy transmission mechanism of the CEDP pilot.