Effect of Scale and Structure Changes of China’s High-Carbon Industries on Regional Carbon Emissions

Abstract

China’s carbon emissions have a stable industrial concentration. In recent years, the carbon emissions of the six major high-carbon industries have accounted for approximately 80% of the national total and are thus priority areas for emission reduction. With the promotion of energy-saving and emission-reduction policies, the structure and scale of high-carbon industries in various regions have undergone changes, but their carbon reduction effects show significant regional differences. Based on China’s provincial panel data from 2006 to 2020, this study discusses the structural characteristics of high-carbon industries with their proportion of energy-based industries and measures their scale characteristics with their output values. On this basis, a fixed-effects model is used to analyze the single and synergistic effects of the scale and structure of high-carbon industries on carbon emissions in each province. The results indicate that changes in the scale and structure of high-carbon industries significantly affect carbon emissions but show regional differences in both the single and synergistic effects. When considering these synergistic effects, the single effect of high-carbon industries on carbon emissions will be weakened. In regions with large-scale high-carbon industries, the increase in the proportion of energy-based industries significantly increases carbon emissions, but this effect gradually weakens as the overall scale expands. In areas with small-scale high-carbon industries, the increase in the proportion of energy-based industries has a relatively small effect on carbon emission growth that gradually increases with the overall scale. In addition, the implementation of the carbon emission trading policy has a significant moderating effect on the carbon emissions of high-carbon industries and strongly promotes its reduction.

1. Introduction

Since its reform and opening up, China has made remarkable achievements in economic growth and social development. The economy increased from RMB 0.37 trillion in 1978 to RMB 121.02 trillion in 2022, and the quality of people’s lives has rapidly improved. However, this economic growth is accompanied by a rapid increase in carbon emissions. Energy conservation and emission reduction have become important work of the government, aiming to achieve the goal of peaking carbon dioxide emissions through implementing the “differentiation” plan, which will deepen the low-carbon clean transformation in energy, industry, construction, transportation, and other areas, strictly control the consumption of fossil energy, especially coal, vigorously develop non-fossil energy, thereby reducing the carbon emissions per unit GDP of the “differentiated” industries by 18%. Fossil energy accounts for a large proportion of China’s energy structure, and therefore, maintaining economic growth produces a large amount of carbon dioxide [1]. China thus faces huge pressure to reduce emissions.

A high-carbon industry is characterized by high energy consumption and carbon emissions, which are the key areas of carbon emission reduction. However, academic circles have no unified classification of high-carbon industries. Liu (2019) combined the statistical bulletin of national economic and social development listed in energy-intensive industries, non-metallic mineral products, chemical raw materials and products manufacturing, non-ferrous metal smelting and rolling processing, ferrous metal smelting and rolling processing, petroleum processing coking and nuclear fuel processing, power heat production, and supply industries as high-carbon industries [2]. In the calculation of 2006–2015 accumulated carbon emissions of manufacturing industries, Wang (2020) found that the carbon emissions of manufacturing industries such as ferrous metal processing, non-metal mineral products, petroleum smelting, chemical raw materials, and non-ferrous metal smelting processing in various provinces were in the top eight, so they were defined as high-carbon manufacturing industries [3]. With a comprehensive review of the high-carbon industry classification, this study adopts the China industry classification. Using the Intergovernmental Panel on Climate Change (IPCC) carbon emission coefficient method, the energy consumption and carbon emissions of the industrial segment are calculated, and six high-carbon industries are identified: coal mining and washing; oil, coal, and other fuel processing; chemical raw materials and products manufacturing; non-metallic mineral production; ferrous metal smelting and rolling processing; and electricity and heat production and supply. As high-polluting entities, high-carbon industries must face their own structural changes and development. On the one hand, high-carbon industries have inherent obstacles to low-carbon transformation, which have an innate dependence on energy consumption. On the other hand, these industries are of great significance to economic development, which supports the operation of the national economy.

At present, many studies examine the factors influencing carbon emissions, including industrial structure, economic growth, and technological progress [4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]. The focus is on the correlation between high-carbon industry development and carbon emissions, as well as the influencing factors and path design of their low-carbon transformation. Cui (2022) believes that the international perspective of senior executives can help promote innovation in high-carbon industries [19]. Zhao (2018) examined the effect of environmental regulation on high-carbon industries and revealed that the production and supply of electricity and heat have exceeded the optimal environmental regulation intensity [20]. Hou and Fang (2012) stated that, from a long-term perspective, strengthening environmental regulation promotes technological and management innovation in high-emission industries and thus reduces costs and carbon emissions [21]. Guo and Sun (2016) believed that the proportion of non-fossil energy used is an important factor in promoting energy conservation and emission reduction in high-emission industries and realizing low-carbon transformation [22]. Zhang et al. (2016) believed that improving the level and reducing the cost of low-carbon technology while improving laws and regulations can promote emission reduction in high-carbon industries [23].

In summary, many scholars have examined carbon emissions from the perspective of the overall high-carbon industry structure, measured with the industry output value ratio, but neglect the internal structure differences, structure, and scale. Therefore, this study divides high-carbon industries into two categories, namely, energy and non-energy. High-carbon energy industries include coal mining and washing and oil, coal, and other fuel processing industries. High-carbon non-energy industries include chemical raw materials and products manufacturing; non-metallic mineral production; ferrous metal smelting and rolling processing; and power and heat production and supply industries. The ratio of output value of high-carbon energy and non-energy industries is used to represent the structure of high carbon industries. Based on the data of 30 provinces, municipalities, and autonomous regions in China from 2006 to 2020, this study empirically analyzed the effect of changes in the scale and structure of the high-carbon industry by region to determine any heterogeneity. The purpose is to deepen the understanding of the relationship between the development of the high-carbon industry and carbon emissions and thus provide a scientific basis for the formulation of relevant policies and measures.

2. Theoretical Mechanism and Research Hypothesis

Under the constraints of low-carbon economic development, the influence of the high-carbon industry on carbon emission was mainly taken through three paths: structural, resource utilization, and policy effects, as shown in Figure 1.

2.1. Structural Effects

The structural effect on carbon emissions is mainly discussed from two aspects, namely, industrial and energy.

The theory of industrial structure upgrading refers to the theoretical model and strategic guiding principle of transforming traditional low-value-added industries into high-value and innovative ones with economic development. The effect of industrial structure upgrading on carbon emissions is examined through two aspects: one is to reduce the proportion of energy-type high-carbon industries, which have high energy consumption and large carbon emissions. Thus, reducing and adjusting its proportion can reduce carbon emissions. Second, this study promotes technological innovation and energy efficiency. By increasing investment in research and development (R&D), promoting scientific and technological innovation, and mastering core technologies, high-carbon industries can better adapt to market changes and improve their competitiveness. The improvement of production technology, such as the intelligent development of coal mines, can reduce the carbon emission intensity (carbon dioxide emissions per unit of GDP) of high-carbon industries and realize its separation from economic growth.

From the perspective of the industrial structure effect, in the early stage of the scale expansion of high-carbon industries, the high-carbon energy industry accounted for a high proportion, with great energy investment. The promotion of low-carbon transformation limits the carbon emissions of enterprises, leading to differences in production cost input across high-carbon industries, the adjustment range of industrial structure by capital accumulation, and production factors that gradually flow to high-productivity sectors and drive structural upgrading. Thus, the carbon emission levels change. With the development of urban modernization, many provinces are constantly reducing the proportion of energy-type and high-carbon industries. The intelligent development of coal mines has also considerably promoted the transformation and upgrading of the upstream and downstream coal industries. Seven key coal enterprises, including the China National Coal Group, have built an intelligent production capacity of 1.393 billion tons per year, accounting for 74.7% of its total. Intelligent upgrading improves labor efficiency and accelerates structural transformation to a certain extent. The proposal of the carbon emission reduction policies causes great pressure on the emission reduction of high-carbon industries at the initial stage, but industries with high technology and low-energy intensity may show a rapid development trend, stimulating the transformation of the industrial structure toward energy conservation. China’s high-carbon industrial structure has shown a downward trend, from 0.28 in 2006 to 0.23 in 2020.

Green development theory is a paradigm that emphasizes environmental protection and sustainability in economic growth. The theory aims to provide a sustainable and environmentally friendly path for development, balancing economic, environmental, and social aspects. This theory encourages high-carbon industries to achieve low-carbon and environmental protection development through technological innovation and transformation. High-carbon industries need to introduce and promote low-carbon and clean energy technologies and circular economy thinking to reduce carbon emissions, such as by using renewable energy sources.

In terms of the energy structure effect, different industries have varying energy consumption demands and structures, and thus the carbon emission intensity also varies. In recent years, with the use of clean energy, such as wind and solar, the energy structure of high-carbon industries has been optimized, and the carbon emission intensity has been reduced [24]. According to the National Bureau of Statistics, from 2006 to 2020, the carbon emission intensities decreased from 7.15 to 5.57 tons/ RMB 10,000 for high-carbon energy industries and decreased from 5.27 tons/RMB 10,000 to 2.70 tons/RMB 10,000 for high-carbon non-energy industries. China has vigorously promoted the transformation of green and low-carbon energy, achieved breakthroughs in the development of renewable energy, and entered a new stage of large-scale and high-quality development. In 2022, the total installed capacity of renewable energy power generation historically exceeded that of coal power. Wudongde Hydropower Station, located at the junction of Yunnan and Sichuan provinces, generates an average annual power generation of 38.91 billion KWH, saving 12.2 million tons of standard coal and reducing 30.5 million tons of carbon dioxide emissions. Now, all the units of Wudongde Hydropower Station have been produced locally. From Gezhouba, Three Gorges, Wudongde, Xiluodu, Xiangjiaba to Baihetan Hydropower Station, China has built the world’s largest clean energy corridor along the Yangtze River, realizing its transformation from “following” to “leading” hydropower. The Taratan Ecological Photovoltaic Park, located in Qinghai province, is the largest centralized photovoltaic power station group in the world. Renewable energy development, of course, still faces challenges. The energy system reform has achieved certain results despite the mechanism in “pipes”, but the reform process remains slow. Several areas, constrained by geographical location and investment, face more difficulties in carrying out clean energy construction. Moreover, promoting the whole society’s enthusiasm for consuming renewable energy is a necessity [25]. In the future, China’s clean energy can continue to develop and gradually complete the deep transformation of the energy system.

2.2. Resource Utilization Effect

The theory of circular economy aims to realize sustainable development by maximizing the service life of resources, reducing the generation of waste, and promoting the efficient utilization of resources. This theory emphasizes reducing resource consumption through recycling and reuse, which can realize the internal closed loop of resources and the supply chain. The effect is a maximization of the use of limited resources to achieve the expected goal through reasonable planning and effective management.

High-carbon industries have the problems of excessive resource consumption and low utilization rate. Developing a circular economy is the most effective method to improve the resource output rate and realize the cycle of “resource–product–waste–resource”. China actively promotes the development of a circular economy. China passed the Circular Economy Promotion Law in 2008, and the National Development and Reform Commission and other departments jointly issued the Circular Development Leading Action in 2017. This act pointed out that the output rate of China’s main resources has steadily improved, the development of a circular economy shows a positive trend, and the recycling rate of major wastes can reach 54.6% in 2020. The resource utilization effect on carbon emissions is mainly achieved through the following two aspects. The first is the improvement of the resource utilization rate, including resource recycling and the reduction of carbon emissions in production links. Technological progress can continuously broaden the depth and breadth of comprehensive resource utilization. Resource recycling refers to turning waste into treasure and harm into profit. Reducing carbon emissions in production mainly refers to those from physical and chemical reactions. Specific measures include optimizing the structure of raw materials needed for production, improving waste recycling, and strengthening the remanufacturing of industrial products. Second, we optimize the distribution of production factors. Under the constraints of low-carbon economic development and to avoid the negative effects brought by the increase of environmental costs, enterprises escape from high-polluting industries, reduce the scale of high-polluting areas, release capital, form new capital rent-seeking, enter industries with relatively low environmental costs, and expand the scale of industries with low-carbon emission intensity. Meanwhile, carbon emission reduction-related policies limit the loan amount of highly polluting enterprises, increase their financing costs, and reduce the scale of new investment.

2.3. Policy Effect

The Coase theorem holds that after clear property rights, spontaneous market behavior can achieve the optimal resource allocation and thus realize the Pareto optimal of the whole society. Meanwhile, the Porter hypothesis states that proper environmental regulation stimulates technological innovation. Carbon emission trading policy is an environmental regulation with clear property rights, which includes dual attributes of the Coase theorem and the Porter hypothesis. The government provides carbon emission quotas to enterprises that take their carbon emission rights as a tradable commodity. Companies with actual carbon emissions lower than the quota can be sold in the market to obtain income, while those with actual carbon emissions higher than the quota need to buy extra to avoid punishment. Thus, the production or the environmental cost increases. Its essence is that the government sets a mandatory total carbon emission target and uses the market mechanism to reasonably allocate carbon quotas.

High-carbon industries comprise the demand side of carbon emission rights and are highly affected by policies. In general, the policy effect promotes green technology innovation and energy structure optimization and affects carbon emissions through cost optimization incentives, profit compensation drive, and risk fluctuation avoidance.

With the cost optimization incentive, high-carbon enterprises need to buy carbon emissions, increase production costs, reverse transmission of green and low-carbon technology innovation, upgrade production technologies, and improve energy efficiency. Thus, enterprises can remove part, or all of the hedge policy of negative externality, namely, the purchase of carbon emissions environmental cost can be optimized through technological innovation. Second, enterprises are driven by revenue compensation. The emission reduction effect of green and low-carbon technologies enables enterprises to have excess carbon quota, which can be directly sold in the market. Income compensation then drives enthusiasm for such technologies and carbon emission reduction. Third, as a risk fluctuation avoidance, carbon emission right has commodity attributes and follows the law of value; its price is unstable and affected by supply and demand. This factor also aggravates the instability of production costs, and to avoid risk fluctuations, companies focus more on technological innovation and improving energy efficiency. At present, the academic community has discussed the emission reduction and macroeconomic effects of the carbon emission trading policy, which is generally believed to reduce carbon emissions. Li et al. (2021) believed that the carbon trading policy changes the energy structure of enterprises by reducing the use of primary energy, thus improving efficiency [26]. Du et al. (2021) found that the carbon emission trading policy plays a significant role in promoting green innovation in pilot areas [27]. Zhang et al. (2019) found that carbon emissions significantly decreased after the launch of its trading policy [28].

Based on the theoretical analysis and research results, the following hypothesis is proposed:

H1: 

The scale expansion of high-carbon industries and the increase of the proportion of high-energy consuming and high-carbon industries will lead to an increase in carbon emissions. This effect weakens if the scale expansion is accompanied by industrial structure optimization, clean energy construction, waste recycling rate improvement, and policy boost effect.

H2: 

The pilot of carbon emissions trading is conducive to the development of the high-carbon industry and reduces carbon emissions.

3. Study Design

3.1. Model Construction

Based on the stochastic impacts by regression on population, affluence, and technology (STIRPAT) model, this study discusses the differential effect of the scale and structural change of the high-carbon industry on regional carbon emissions. The model is expanded as follows:

$$InC{E}_{it}={\alpha }_0+{\alpha }_{1}In{S}_{it}+{\alpha }_{2}In{V}_{it}+{\alpha }_3\left(In{S}_{it}\times In{V}_{it}\right)+{\alpha }_{4}Contro{l}_{it}+{\epsilon }_{it}$$

(1)

In Equation (1), CE is the carbon emission, $i$ is the region, $t$ is the time, ${\alpha }_0$ is the intercept term, $S$ is the high-carbon industry structure, $V$ is the high-carbon industry scale, $Control$ is the control variable, ${\alpha }_1$~${\alpha }_4$ is the regression coefficient of each variable, and ${\epsilon }_{it}$ is the random error term.

To explore the effect of policy on carbon emissions, we introduce the virtual variable of carbon emission trading pilot policy to explore whether its implementation produces a significant emission reduction effect. The experimental group includes six provinces, namely Tianjin, Shanghai, Guangdong, Hubei, Beijing, and Chongqing. The remaining 24 provinces serve as the control group. In this study, the implementation year of the carbon trading pilot policy in 2013 is taken as the policy impact time, and the following dual difference model is constructed:

$$\begin{array}{cc}\hfill InCE& ={\alpha }_0+{\alpha }_1\left(trea{t}_{it}\times tim{e}_{it}\right)+{\alpha }_{2}In{S}_{it}+{\alpha }_{3}In{V}_{it}\hfill \\ & +{\alpha }_4\left(In{S}_{it}\times In{V}_{it}\right)+{\alpha }_{5}Contro{l}_{it}+u_{it}+v_{it}+{\epsilon }_{it} \hfill \end{array}$$

(2)

where $treat$ is the virtual variable that groups provinces and is equal to 1 for the experimental group and 0 for the control group; $time$ is the time dummy variable of policy impact, equal to 1 after 2013 and 0 before 2013; $u_i$ is region fixed effect; $v_t$ is time fixed effect; and the other variables are the same as in Equation (1).

3.2. Variable Selection and Data Description

3.2.1. Interpreted Variables

Carbon emissions (CEs) refers to China’s total carbon emissions. The consumption of primary energy and electricity provided by the Statistical Yearbook is estimated according to the method provided by the IPCC. The specific formula is as follows:

$$CE={\displaystyle \sum }_i^n\left(E_i\times E{F}_i\right)$$

(3)

where $E_i$ refers to the physical quantity of energy consumption; $E{F}_i$ represents the $i$ kind of energy carbon emission factor; and $i$ represents eight types of fossil energy (coal, coke, crude oil, gasoline, kerosene, diesel, oil, fuel, natural gas) and electricity, including the electricity emission factor from the National Development and Reform Commission 2019 China regional power grid baseline emission factor. Figure 2 shows the trend of the proportion of CEs of high-carbon industries to the total CEs. Although the high-carbon industries show constant changes in CEs from 2006 to 2020, their CEs account for approximately 80% of the national CEs. The data are shown in Figure 2.

3.2.2. Core Explanatory Variables

The core explanatory variables in this study are as follows:

(1) High-carbon industrial structure (S), expressed by the ratio of output value of high-carbon energy and non-energy industries.

(2) Scale of high-carbon industry (V), expressed by the total output value of the high-carbon industry. Figure 3 shows the scale and composition of high-carbon industries. Among them, the scale of high-carbon non-energy industries grew rapidly, and their proportion to the total scale increased from 77.96% in 2006 to 80.93% in 2020.

(3) Policy virtual variables (treat × time): China has launched the carbon emission rights trading markets in six provinces, including Beijing, Shanghai, and Tianjin. This pilot carbon trading policy aims to control CEs by using the market mechanism.

3.2.3. Control Variables

The control variables in this study are as follows:

(1) Urbanization (Urban) refers to the ratio of the urban population to the permanent resident population at the end of the year. Energy consumption and CEs concentrate in and are affected by cities. In the early stage of urbanization, economic development is the top priority; therefore, industries—even high-pollution ones—are vigorously developed. At the same time, large numbers of people migrate to the cities, putting great pressure on the local environment. With the improvement of urbanization, on the one hand, people need to use considerable energy (such as electricity and coal) to meet the needs of work and life, which increases energy consumption [29]. On the other hand, the government and people pay more attention to environmental protection issues, thus significantly improving energy efficiency and pollution treatment.

(2) Energy intensity (Ei) is the energy consumption per unit of GDP or the proportion of total energy consumption to regional GDP. Energy intensity reflects the level of technology. Technological progress can encourage enterprises to use more green and ecological production methods, optimize production, reduce pollutant emissions [30], and improve energy efficiency, thus reducing CEs.

(3) Trade openness (Open) is expressed by the ratio of total imports and exports to regional GDP. More economically developed provinces tend to introduce advanced equipment and technology to reduce CEs. By contrast, economically backward provinces tend to introduce relatively backward equipment and technology and even receive transfers of capital-intensive enterprises to the region from developed countries, producing a “pollution shelter effect” and increasing CEs.

(4) Environmental regulation (Er) is expressed by the proportion of the completed investment in industrial pollution control in the industrial added value. Environmental regulation has both “competitive” and “crowding-out” effects on CEs [31]. On the one hand, environmental regulation increases the living cost of heavily polluting industries, prompting the elimination of more polluting enterprises, thus reducing CEs. On the other hand, environmental regulation enables enterprises to invest more resources to reduce pollution and squeeze out green innovation investment, which is not conducive to reducing CEs.

(5) Foreign direct investment (Fdi) is the proportion compared with GDP. Foreign direct investment has a dual effect on CEs. On the one hand, foreign direct investment can bring a large amount of capital and advanced production technology, improve the level of industrial technology and management level, and thus reduce CEs. On the other hand, as China is the largest developing country, the relatively low labor force and loose environmental control policies introduce high energy consumption and high pollution industries, which is not conducive to the reduction of CEs.

Considering the period of 2006–2020, this empirical study included panel data from 30 provincial administrative regions (Shandong, Shanxi, Liaoning, Inner Mongolia, Shaanxi, Guangdong, Hebei, Henan, Jiangsu, Zhejiang, Hubei, Sichuan, Fujian, Hunan, Anhui, Tianjin, Shanghai, Jiangxi, Guangxi, Guizhou, Yunnan, Xinjiang, Hainan, Qinghai, Ningxia, Chongqing, Gansu, Heilongjiang, Jilin, Beijing). The required data were derived from China Statistical Yearbook, China Industrial Statistical Yearbook, China Energy Statistical Yearbook, and Provincial Statistical Yearbook. In several provinces, the output value data of high-carbon industries in 2018–2020 were missing and were completed by using the linear interpolation method. The descriptive statistics of the variables are shown in

Table 1. Descriptive statistics of relevant variables from 2006 to 2020.

Variable	Observations	Mean	Std. Dev.	Min	Max
CE	450	386.96	282.62	19.79	1541.12
S	450	0.3829	0.3676	0.0305	1.7522
V	450	8837.39	8310.66	239.58	44,905.82
Urban	450	0.5577	0.1373	0.2745	0.8958
Ei	450	0.997	0.655	0.170	4.099
Open	450	0.2959	0.3552	0.0076	1.7991
Er	450	0.0041	0.0037	0.0001	0.0310
Fdi	450	0.0215	0.0198	0.0001	0.1210

.

4. Empirical Analysis

4.1. Unit Root Inspection

Non-stationarity of the sequence can affect the magnitude of the model fit, and if the unit root exists in the sequence, then a series of problems, such as pseudo-regression, may occur. Therefore, to verify the validity of the model, we carried out a stationarity test by using the LLC method. From the results in

Table 2. Results of the unit root test.

Variable	LLC Test	First-Order Differential LLC Test
lnCE	0.0003 ***	0.0000 ***
lnS	0.0000 ***	0.0000 ***
InV	0.0000 ***	0.0000 ***
lnUrban	0.0000 ***	0.0000 ***
lnEi	0.0026 ***	0.0000 ***
lnOpen	0.0000 ***	0.0000 ***
lnEr	0.0377 **	0.0000 ***
lnFdi	0.0000 ***	0.0000 ***

Note: *, **, and *** indicate significance at the significance level of 10%, 5%, and 1%, respectively.

, Er was significant at the 5% level, and the remaining variables were significant at the 1% level. The first-order differential LLC tests were all significant at the 1% level, rejecting the null hypothesis with no unit root present, indicating that the panel data to be regressed passed the stationarity test.

4.2. Multiple Collinearity Test

To test for collinearity among model variables, we carried out multicollinearity tests, and the results are shown in

Table 3. Results of the multicollinearity test.

Variable	VIF	1/VIF
lnS	1.25	0.798574
lnV	1.31	0.765428
lnUrban	1.86	0.537268
lnEi	2.31	0.433342
lnOpen	1.83	0.547277
lnEr	1.72	0.582415
lnFdi	1.16	0.862921
Mean VIF	1.63

. The variance extension lead (VIF) of each variable was much less than 10, indicating that there was no multicollinearity among the selected variables.

4.3. Hausman-Test

F, LM, and Hausman tests were carried out to determine the fixed, random, or mixed effects models. The test results are shown in

Table 4. F, LM, and Hausman test results.

Test	p	The Results Indicate That
F-test	0.0000	The fixed effect was better than the mixed effects.
LM-test	0.0000	Random effects outperformed mixed effects.
Hausman-test	0.0000	The fixed effect was better than the random effects.

.

From this result, a fixed-effects model was used.

4.4. Empirical Results Analysis Based on a Fixed-Effects Model

Fixed-effects regression was carried out based on the model Formula (1), and the results are shown in

Table 5. Results of the fixed-effects model regression.

Explanatory Variable	Explained Variable lnCEs
	(1)	(2)	(3)
lnS	0.0775 *** (4.46)	0.0888 *** (5.70)	0.166 *** (7.86)
lnV	0.395 *** (29.84)	0.407 *** (18.98)	0.424 *** (20.13)
lnS × lnV			−0.0361 *** (−5.23)
lnUrban		0.833 *** (9.27)	0.773 *** (8.79)
lnEi		0.298 *** (9.86)	0.295 *** (10.09)
lnOpen		−0.0458 * (−2.39)	−0.0393 * (−2.12)
lnEr		0.00575 (0.64)	0.00303 (0.35)
lnFdi		−0.0200 (−1.87)	−0.0257 * (−2.46)
_cons	2.364 *** (20.72)	2.712 *** (11.88)	2.725 *** (12.31)

Note: *, **, and *** indicate significance at the significance level of 10%, 5%, and 1%, respectively.

.

Without the effect of control variables on CEs, Column (1) shows that the expansion of the scale of high-carbon industries and the increase of the proportion of energy-type and high-carbon industries significantly increase CEs. In Column (2), considering the control variables, the regression coefficient of S is 0.0888, which passes the significance test of 1%, indicating that the increase in the proportion of high-carbon energy industries can lead to an increase in CEs. High-carbon energy industries have greater demand for energy and high-carbon emission intensity. When their proportion increases, CEs also increase. The regression coefficient of V is 0.407, which passes the 1% significance test, indicating the scale expansion and significantly increasing CEs. As the scale increases, energy consumption and CEs increase. Column (3) introduces the cross of S and V; considering the scale and structure of the synergistic effect, the expansion of scale and proportion increase of the high-carbon energy industry still significantly increase CEs. However, the lnS × lnV regression coefficient of −0.0361 is significant at a 1% level, indicating that with the scale expansion, the CEs of high-carbon industries weakens. Therefore, Hypothesis 1 is verified. The reason is that, on the one hand, in terms of the internal structure of high-carbon industries, the scale of non-energy ones is much larger while its carbon emission intensity is smaller than those of energy sectors. In 2020, the proportion of high-carbon non-energy industries increased to 80.93% and inhibited the increase of CEs. On the other hand, the optimization of industrial structure, clean energy construction, and improvement of waste recycling and policy boost reduce the effect of high-carbon industries on CEs. In recent years, China has issued a series of carbon emission reduction strategies, mainly targeting high-energy consumption and high emissions, including industrial repellent of inefficient capacity, speeding up the start of the national carbon market construction, improving the energy consumption double control index management and energy conservation, and promoting the high-carbon industry management. All of these cause great significance to carbon emission reduction [32]. These measures are similar to other countries. For example, Switzerland’s high-carbon industries include energy production, oil refining and chemicals, transportation, and construction, which contribute more to Switzerland’s carbon emissions. Switzerland’s high-carbon industry is developing in a similar direction to China, focusing on reducing its dependence on fossil fuels and promoting transformation in areas such as renewable energy, energy efficiency, sustainable transportation, and construction. These initiatives will help reduce carbon emissions and push Switzerland towards a more sustainable and low-carbon future.

With increasing global attention on climate change, China will strengthen regulation and control of carbon emissions in the future. The Chinese government promises to achieve carbon dioxide emissions by 2030 and carbon neutrality by 2060. This could lead to high-carbon industries gradually adopting more environmentally friendly technologies and processes in the future to reduce carbon emissions. At the same time, the lower cost and increased adoption of renewable energy could prompt high-carbon industries to switch to cleaner energy sources, thus reducing carbon emissions. In addition, consumers and investors’ focus on environmental sustainability could also spur more actions to cut emissions. It is expected that in the future, the synergistic effect of the scale and structure of the high-carbon industry will be reduced; that is, with the expansion of the scale, the increasing effect of the high-carbon industries on carbon emissions will be weakened.

From the perspective of control variables, the regression coefficient of the urbanization level is 0.773 and passes the significance test of 1%, indicating a significant increase in CEs. Given that the urban population needs to use much energy for daily work and living, the CEs increase. Energy intensity significantly increases CEs but also represents the technological level; that is, technological progress can reduce energy consumption per RMB 10,000 of GDP. Specifically, for each unit increase in technology level, CEs decrease by 0.295 units. With China entering the post-industrial era, the energy intensity has been declining in recent years, which is conducive to the suppression of CEs. Trade openness slightly suppresses CEs, possibly because of the introduction of advanced equipment and technology. The regression coefficient of environmental regulations is positive, indicating that in the face of strict environmental regulations, enterprises need to invest more funds and effort in pollution control and emission reduction, thereby reducing the improvement speed of production technology innovation [33]. The “crowding-out effect (Environmental regulation increases the cost of enterprises, which may make inefficient and highly polluting enterprises gradually phased out)” is greater than the “competitive effect (Environmental regulation enables enterprises to speed up technology research and development, improve energy efficiency, and reduce environmental costs)”, which also verifies the “green paradox (Under the policy measures to address climate change and environmental problems, the environmental effects are opposite to those expected)” hypothesis. The regression coefficient of foreign direct investment is negative. Foreign direct investment brings a large amount of capital and advanced production technology, which improves the industrial technology and management levels, thereby reducing CEs.

4.5. Heterogeneity Analysis

China has vast land and abundant resources, with different conditions across regions and resource endowment, and in the scale and structure of high-carbon industries in different provinces and cities. According to the scale and structure of high-carbon industries, this study divides 30 provinces into rich high-carbon energy industries, high-carbon non-energy industries, and small high-carbon industries. Based on the regional results in

Table 6. Sub-region results.

Division Basis	Region
Rich in high-carbon energy industries	Shandong, Shanxi, Liaoning, Inner Mongolia, Shaanxi, Guangdong, Hebei, Henan
Rich in high-carbon non-energy industries	Jiangsu, Zhejiang, Hubei, Sichuan, Fujian, Hunan, Anhui, Tianjin, Shanghai, Jiangxi, Guangxi, Guizhou, Yunnan, Xinjiang
Sparse in high-carbon industries	Hainan, Qinghai, Ningxia, Chongqing, Gansu, Heilongjiang, Jilin, Beijing

, the regression results of the empirical test based on the model Equation (1) are shown in

Table 7. Sub-regional fixed-effects model regression results.

Explanatory Variable	Explained Variable lnCEs
	(4)	(5)	(6)
lnS	0.439 *** (8.74)	0.300 *** (8.28)	0.0703 * (2.47)
lnV	0.444 *** (15.08)	0.372 *** (13.18)	0.487 *** (13.87)
lnS × lnV	−0.0764 *** (−8.24)	−0.154 *** (−9.86)	0.0017 (0.18)
lnUrban	0.778 *** (5.55)	0.206 (1.66)	1.556 *** (10.43)
lnEi	0.216 *** (3.42)	0.0909 * (2.50)	0.483 *** (8.78)
lnOpen	−0.0477 (−1.55)	−0.0577 (−1.93)	0.0170 (0.69)
lnEr	−0.00053 (−0.04)	0.00862 (0.64)	−0.00452 (−0.40)
lnFdi	−0.0743 *** (−4.83)	0.0285 (1.46)	0.0368 ** (2.67)
_cons	3.205 *** (10.70)	3.459 *** (10.44)	2.314 *** (6.62)

Note: *, **, and *** indicate significance at the significance level of 10%, 5%, and 1%, respectively.

. Columns (4), (5), and (6) represent the regression results of the fixed-effects model of areas rich in high-carbon energy industries, rich in high-carbon non-energy industries, and sparse in high-carbon industries, respectively. The scale and structure of high-carbon industries in different regions have different single and synergistic effects on CE.

In areas rich in high-carbon energy and non-energy industries, the influence coefficients of the change of high-carbon industrial structure on CEs are 0.439 and 0.300, respectively. This result shows that in areas rich in high-carbon energy industries, the change of high-carbon industrial structure has a higher effect on CEs than that in high-carbon non-energy industries. The coefficients of lnS × lnV are all negative, indicating that the single effect of the high-carbon industry in causing CEs is weakened when considering the synergistic effect of scale and structure. China has been vigorously developing clean coal utilization technology. Through the intelligent transformation of mines, the level of equipment, the main transportation system, and the recycling efficiency of coal resources have improved [34]. The scale of coal mining and washing industries in many provinces has also decreased year by year. From 2006 to 2020, the scale of the coal mining and washing industry decreased from RMB 7.51 to 0.5 million in Zhejiang province, from RMB 1.234 billion to RMB 0.107 million in Hubei province, and decreased from RMB 2.449 billion to RMB 0.963 million in Tianjin. From the areas rich in high-carbon energy industries, most of them are the central and western regions, whose economic development is highly dependent on energy, with a low-energy utilization rate [35], relatively extensive urban land expansion, and the transfer of several high-carbon industries in the eastern region, and many high-carbon enterprises. Existing research results show that there is a close relationship between high-carbon industries and carbon emissions, which are the main source of carbon emissions in China. Measures such as implementing emission reduction policies and the promotion of clean energy and energy efficiency technologies can significantly reduce the carbon emissions of high-carbon industries. On this basis, this paper found that there are differences in carbon emission intensity among different industries of high-carbon industries, and high-carbon energy industries have high carbon emission intensity. In areas with abundant energy-based high-carbon industries, considering the synergistic effect of scale and structure, the single effect of weakened carbon emissions caused by high-carbon industries is greater than that in areas rich in non-energy high-carbon industries. This may be because the areas rich in non-energy and high-carbon industries are mostly economically developed areas, accompanied by higher levels of energy conservation and emission reduction technologies in the process of scale expansion.

In areas with small-scale high-carbon industries, the influence coefficient of S on CEs is 0.0703, indicating that the increase in the proportion of energy industries in this region has less effect on the growth of CEs. The coefficient of lnS × lnV is positive, indicating that this effect gradually increases as the overall size increases, but the increase is not significant. The reason may be that these regions have few high-carbon industries, most of the geographical locations and natural resources are not suitable for the development of high-carbon industries, and they do not pay attention to the R&D of low-carbon technologies and lack promotion of carbon emission reduction policies. The blind scale expansion strengthens the role of high-carbon industries in increasing CEs.

4.6. Empirical Results Analysis Based on Double Difference Model

Through the dual difference model (2), we verified the effectiveness of the carbon emission rights trading policy. The empirical results after controlling time and individual two-way fixed effects are shown in

Table 8. Results of the dual differential model regression.

Explanatory Variable	Explained Variable lnCEs
	(7)	(8)
treat × time	−0.245 *** (−3.10)	−0.079 * (−1.87)
lnS		0.151 ** (2.61)
lnV		0.426 *** (8.78)
lnS × lnV		−0.033 * (−1.89)
lnUrban		0.630 ** (2.17)
lnEi		0.272 ** (2.40)
lnOpen		−0.042 (−1.13)
lnEr		−0.0017 (−0.16)
lnFdi		−0.026 (−1.52)
_cons	5.563 *** (290.32)	2.547 *** (4.92)
Area fixation effect	Yes	Yes
Time fixed effect	Yes	Yes
R-sq	0.3736	0.7820

Note: *, **, and *** indicate significance at the significance level of 10%, 5%, and 1%, respectively.

. Column (7) shows the regression results without considering the control variables, and the regression coefficient of the policy dummy variable (treat × time) is significant and negative at the 1% level. In Column (8), the control variable is introduced, and the regression coefficient remains significant and negative. This result shows that the provinces implementing the carbon emission trading policy can effectively reduce CEs compared with the control group. Therefore, Hypothesis 2 is verified. The possible reason is that with the implementation of the carbon emission trading policy, enterprises in the pilot provinces need to buy shares of CEs and production costs increase. To maximize profits and reduce production costs, enterprises promote their technological innovation and use clean energy as an effect of the carbon emission rights trading policy [36]. At the same time, the carbon emission trading policy accelerates the green and low-carbon industry transformation to reduce the overall CEs of the pilot cities.

4.7. Robustness Test

To verify the reliability of the impact of the scale and structure of the high-carbon industry on CEs and avoid the impact of different indicators of the explanatory variables on the regression results, we replaced the explanatory variables with per capita carbon emission and carbon emission intensity and then reran the regression data under the whole sample. The results are shown in

Table 9. Results of the robustness test.

Explanatory Variable	Explained Variable lnPCEs		Explained Variable lnCI
	Coef.	t	Coef.	t
lnS	0.191 ***	8.51	0.277 ***	10.24
lnV	0.368 ***	16.45	0.042 *	1.55
lnS × lnV	−0.0354 ***	−4.82	−0.0365 ***	−4.12
lnUrban	0.865 ***	9.25	−0.599 ***	−5.31
lnEi	0.313 ***	10.05	0.646 ***	17.19
lnOpen	0.0066	0.33	0.0780 **	3.27
lnEr	0.0061	0.66	0.0166	1.49
lnFdi	−0.0252 *	−2.27	0.0165	1.23
_cons	−4.787 ***	−20.33	−3.607 ***	−12.68

Note: *, **, and *** indicate significance at the significance level of 10%, 5%, and 1%, respectively.

. The significance and positive and negative directions of the core explanatory variables are basically unchanged, and only the coefficients slightly vary, indicating that the main regression results are robust.

5. Study Conclusions and Recommendations

In this study, regression analysis on the panel data of 30 provinces in China from 2006 to 2020 was conducted using a fixed-effects model of high-carbon industry scale and structure of provincial CEs and single influence. The dual difference model was adopted to build a natural experiment and measure the effect of the CE trading pilot policy on CEs. The empirical results show that the changes in the scale and structure of high-carbon industries significantly affect CEs, but regional differences still occur in the single and synergistic effects of scale and structure. When considering the synergies of size and structure, the single role of high-carbon industries in leading to CEs is weakened. In areas with large high-carbon industries, the increase in the proportion of energy industries significantly increases CEs. However, this effect gradually weakens as the overall size expands. In areas with small-scale high-carbon industries, the increased proportion of energy industries has less effect on the growth of CEs but gradually increases as the overall size increases. However, it was also considered that some economically developed provinces attract a large amount of investment with their own advantages, so that the technology industry gathers in this region, expands the scale of technology-intensive industries, and then transfers the high emission and high pollution industries to the adjacent backward areas, and increases the carbon emissions of the adjacent areas. In addition, the implementation of the carbon trading policy exerts a significant regulatory effect on the role of CEs in high-carbon industries and effectively promotes carbon emission reduction.

On this basis, we propose the following policy suggestions:

First, we should continue to promote a low-carbon energy structure and increase the proportion of clean energy. In addition, this study suggests reducing the level of coal consumption, accelerating the development of alternative energy, selecting appropriate renewable energy for technology R&D, and combining the advantages of regional natural resources, from wind, water, solar, ecological, and other clean energy. We should improve the capacity to replace non-fossil energy sources securely and reliably and create a diversified clean energy supply system with wind, light, water, raw, nuclear, and hydrogen. At the same time, technologies can be developed for the clean utilization of coal and other fossil fuels to improve energy efficiency. We can promote the flexible upgrading of coal power generation, accelerate the development of pumped storage, peak–peaking gas and electricity, and new energy storage, and strengthen the upgrading of distribution networks to support access to a high proportion of new energy sources [37]. Enterprises can be encouraged to use energy-saving technologies, strengthen resource conservation, and pay attention to resource recycling. The government must supervise the adjustment of the energy structure of relevant enterprises, reduce the cost of clean energy through financial subsidies and other means, and encourage relevant enterprises in high-carbon industries to adopt clean and efficient energy.

Second, regional cooperation must be strengthened and coordinated development must be promoted among high-carbon industries. Pilot cities can use their advantages in green innovation technology and governance experience to drive the joint upgrading of related industries and share management experience with adjacent regions. Regions with large-scale high-carbon industries can learn from the development experience of pilot cities, improve their industrial technology innovation [38], optimize the structure of high-carbon industries, and move forward from traditional to high-end industries. At the same time, a regional carbon emission reduction cooperation mechanism can be established to achieve effective interregional docking in energy policies such as new energy technology R&D and energy conservation subsidies and promote the sharing of resources and low-carbon technologies among provinces. Regions rich in energy-based and high-carbon industries can strengthen their awareness of environmental protection, improve their capacity for independent innovation, and strengthen cooperation with advanced regions in low-carbon technologies [39]. When including high-carbon industries, neighboring regions must consider their differences in natural resources and promote rational resource allocation.

Third, the layout of high-carbon industries requires a rational plan and formulation of relevant policies according to local conditions. The government can implement a series of measures such as high-carbon industry planning and policies, rationally plan the layout of high-carbon industries, and accelerate the progress of emission reduction work. First, policies can be introduced to accelerate the R&D of green and low-carbon technologies and provide these with financial support by establishing carbon funds and supporting investment institutions so as to diversify individual risks and increase their enthusiasm for R&D. A green technology innovation system can be promoted, with enterprises as the main body, research institutes as the basis, and the government as the support [40], to accelerate the transformation and upgrading of traditional high-carbon industries. Second, we can introduce more scientific and reasonable policies for elimination, incentives, and compensation. Enterprises with high emissions and backward production capacity can be resolutely eliminated, encouraging them to actively transform through appropriate compensation. Meanwhile, enterprises that adopt low-carbon technologies and low emissions can be rewarded in terms of capital and technology when necessary. Third, to strengthen law enforcement and market supervision, we can link effective policy implementation with the performance assessment of government officials and end asylum behavior.