Analysis of the Impact of Industrial Structure Upgrading and Energy Structure Optimization on Carbon Emission Reduction

Abstract

In China, there has been a significant increase in carbon emissions in the new era. Therefore, evaluating the influence of industrial structure upgrades and energy structure optimization on reducing carbon emissions is the objective of this research. Based on the provincial panel data of 30 provinces and cities across China from 1997 to 2019, this paper builds up a fixed-effect panel quantile STIRPAT model to investigate the differences in the impact of industrial structure on carbon emission intensity at different quantile levels from the provincial perspective, and as a way of causality test, the mediation effect model is adopted to empirically test the transmission path of “industrial structure upgrading—energy structure optimization—carbon emission reduction”. The research results show that: (1) Both industrial structure upgrades and energy structure optimization have significant inhibitory effects on carbon emissions, and there are regional heterogeneities. (2) The upgrading of industrial structure has a significant positive effect on optimizing energy structure. (3) The upgrading of industrial structure can not only directly restrain carbon emissions but also indirectly have a significant inhibitory effect on carbon emissions by promoting the optimization of energy structure. Based on the above conclusions, corresponding policy recommendations are proposed to provide suggestions for China to achieve the goal of carbon neutrality.

1. Introduction

Issues such as global warming caused by carbon emissions have gradually attracted widespread attention from the international community in the new era. Although global carbon emissions have experienced a short-term decline due to the influence of the COVID-19 pandemic [1], the urgency and importance of addressing the challenge of climate change have not changed. According to “WMO Provisional State of the Global Climate 2022”, the latest report issued by The World Meteorological Organization (WMO), the global mean temperature in 2022 is currently estimated to be 1.15 ± 0.13 °C above the 1850–1900 average. The eight years 2015 to 2022 are likely to be the eight warmest years on record, with 2022 most likely to be 5th or 6th warmest. To control rising temperature and carbon dioxide concentration, greenhouse gas emissions must be reduced immediately, rapidly, and on a large scale. At the same time, the current global climate governance model has shown a state of governance failure, disorder, and ineffectiveness under the “anti-globalization” trend. The original climate cooperation process and the international climate order have been disrupted, so a transformation is urgently needed.

As the world’s largest developing country, China is in a critical period of modernization and faces more severe challenges in coordinating economic growth and reducing carbon emissions. With financial and social development, increasing urbanization rate, continuous population growth, and rapid growth of total energy consumption, environmental disasters, resource shortages, climate change, and other issues have become increasingly prominent, and China’s carbon emissions have risen rapidly in the past two decades. In 1997, China’s total carbon emissions were 3.308 billion tons. In 2007, China’s total carbon dioxide emissions surpassed that of the United States and ranked first in the world. Countries whose share of carbon emissions is still growing among the major countries in the world. However, China has always actively assumed the international responsibility of reducing carbon emissions, pursued a favorable global governance policy against the background of anti-globalization, closely combined the transformation of domestic development mode with global climate governance, and actively participated in and made significant contributions to the change of global climate governance, a key hub and a leading role in the future, and continuously improve the degree of nationally determined contributions. The 18th National Congress of the Communist Party of China proposed to build a new system of green and low-carbon modern industrial development in China. In 2020, China announced the “30·60” dual-carbon goals and successively released vital areas and industries’ carbon peak implementation plans and a series of supporting and safeguard measures to build a carbon peak, carbon neutral “1 + N” policy system. However, achieving the goal of “emission peaking and carbon neutrality” is a profound economic and social systemic change that is currently constrained by technology, complexity, and frontiers. In addition, China has a vast territory, and each province has a resource endowment and production factor endowment. There are regional heterogeneities in industrial structures and low-carbon economic development. Therefore, exploring the impact and path mechanism of industrial structure upgrades on the carbon emission intensity of various provinces and cities in China is of great significance for scientifically formulating carbon emission reduction policies.

In the early stage of economic development, traditional industrial structures led to the rapid growth of carbon emissions, and this extensive growth model no longer matches the goal of low-carbon development at this stage [2]. With the continuous development of research relating to industrial structures and carbon emissions in the 21st century, scholars believe that upgrading industrial structures provides scientific guidance for industrial upgrade and transformation, which is an effective way to reduce carbon emissions [3]. Adjustment and change are of great significance for formulating carbon emission reduction policies. The relationship between industrial structures and carbon emission has been regarded as an important research subject by international organizations and academic circles.

Compared with the existing research, the possible marginal contributions of this paper are as follows: (1) Use the panel quantile regression model instead of the mean regression to make up for the lack of accuracy of the mean regression estimation and the fact that it can only examine the influence of covariates near the mean on the dependent variable and to investigate the heterogeneity of the impact of the industrial structure of each province on carbon emissions under different carbon emission levels. (2) Use the mediation effect model to empirically analyze the mediation transmission path of the impact of industrial structure upgrading on carbon emissions through energy structure optimization. (3) According to the research conclusions, propose a practical path to upgrade the industrial structure to promote the optimization of the energy structure and reduce carbon emissions and provide a new perspective and valuable reference for the scientific formulation of carbon emission reduction policies.

The rest of this paper is organized as follows: Section 2 is a literature review and research hypothesis statement and theoretically analyzes the mechanism of an industrial structure affecting carbon emissions through the energy structure. Model construction, variable selection, and data are discussed in Section 3. Section 4 lists a panel-based analysis of the practical results of quantile regression and the mediation effect model. Section 5 summarizes the paper, provides conclusions, proposes theoretical implications and relevant policy suggestions, and makes a recommendation for future research in view of limitations.

2. Literature Review and Research Hypotheses

2.1. Literature Review

Scholars have studied carbon emissions from various perspectives, especially industrial structures, which are considered to be one of the main factors affecting carbon emissions and have received extensive attention from the academic community [4,5,6]. The most frequent ways are modeling carbon emissions and conducting quantitative studies on its influencing factors in previous literature, which includes estimating the Environmental Kuznets Curve (EKC) [7,8], using the STIRPAT model, studying the Kaya equation and so on. Dong et al. [9] verified the Environmental Kuznets Curve hypothesis. They believed that the upgrade of industrial structures not only directly affects CO₂ emissions but also indirectly affects global CO₂ emissions by promoting technological innovation. Liu et al., Zhao et al., and Gu et al. used the STIRPAT model to verify the significant effect of industrial structure transformation and upgrades on reducing carbon emissions [10,11,12]. Developing the Tapio decoupling model, Kaya’s equation and LMDI decomposition method, Han et al. [13] found that the industrial structure and energy structure effect play negative roles in driving carbon emissions. The degree of coupling coordination among industrial structure, carbon emissions, and the regional innovation of Shandong Province were calculated, and their spatial characteristics and the aggregation effects of both coupled and coordinated development were explored by Wang et al. [14]. However, in the existing literature studies focusing on the influencing factors of carbon emissions, most of them have proposed policy suggestions based on the average impact of each element on carbon emissions during the sample period, which does not effectively reflect the differences in the degree of influence with different quantile points of carbon emission levels.

Furthermore, current research mostly takes the impact of industrial structures on energy structures as an implicit assumption, only defining industrial structures as a mediation variable. Gao et al. [15] verified that the upgrade of industrial structures plays a vital role in green technology innovation to diminish carbon emissions and realized that accelerating both green technology innovation and industrial structure upgrades to fulfill the synergy of both factors is crucial. Zhang et al. [16] regarded industrial upgrades as the basis for examining if there was a competitive or cooperative relationship between the carbon emission effects generated and promoting market integration and found that a negative impact of market integration on carbon emissions could be diluted by industrial upgrades. Research using the “industrial structure—energy structure—carbon emissions” analysis framework is relatively scarce.

Among them, upgrading industrial structures, which is the main driving force of economic development, is gradually being integrated with two other trends of high-quality economic development: the optimization of energy consumption structures and green and low-carbon development. Grasping new growth points in the process of optimizing the industrial structure and adjusting the energy structure is of great significance for the realization of a green, low-carbon, sustainable, and high-quality economic and social development stage. Research by Brannlund and Persson [17] demonstrated that optimizing industrial structures inhibits carbon emissions by affecting changes in energy consumption structures and efficiency. Currently, scholars’ research on the relationship between industrial structures, energy structures, and carbon emissions mainly focuses on the following three aspects. The first is the impact of industrial structures on carbon emissions; as part of the core content of supply-side structural reform, the industrial structure is intrinsically associated with carbon emissions [18,19]. Until 2020, the shares of the output value of the three industries in China were 7.7%, 37.8%, and 54.5%, with industrial structures advancing toward an advanced level and realizing the development mode of mainly secondary and tertiary industries. The upgrade of industrial structures contributes to directing economic development and low-carbon governance, a significant channel influencing carbon emissions [20]. Moreover, Zheng et al. [21] found that variations in local development patterns have contributed to industrial structure upgrades at the regional level, inhibiting carbon emissions in most areas. Zhang et al. [22] developed a comprehensive framework to assess the impact of industrial structure and technological progress on carbon intensity, suggesting that industrial structure upgrades indirectly increase carbon intensity by promoting technological change. Zeng et al. [23] use epsilon-based measure (EBM) and data envelope analysis (DEA) models to evaluate carbon emission efficiency and its differences among 30 Chinese provinces, indicating that industrial structures, technological innovation, and carbon emission efficiency are significantly and positively correlated. The second is the impact of industrial structures on energy structures. Kambara [24] believed that the structural efficiency caused by the adjustment of industrial structures had an effect of more than 50% in improving energy use efficiency in China from 1980 to 1990. The adjustment of industrial structures during the transformation of intensive and labor-intensive industries has promoted improving energy use efficiency. Xue et al. [25] used the super-efficiency slack measure (SBM) model and the coupling coordination model to measure the coupling and coordination degree of energy efficiency and industrial development in key industries. The results show that the impact of industrial structures on energy efficiency is periodic, and the coupling imbalance between energy efficiency and industrial development is the most apparent feature of critical industries. The third is the impact of energy structure on carbon emissions. According to statistics, 95% of the increase in greenhouse gases comes from energy consumption. Wen et al. [26] used the LMDI method to analyze the relevant data from 30 provinces in China from 2000 to 2014. They concluded that the contribution of energy structures to carbon emissions changed from negative to positive and generally increased. Soares and Tolmasquim [27] and Siitonen et al. [28] showed that both energy efficiency and energy mix improvements are effective in reducing CO₂ emissions.

Looking at the existing literature, domestic and foreign literature that independently study industrial structures, energy structures, and carbon emissions or the relationship between them is abundant. However, few works of literature include the three in the same causal chain and build a research framework based on this for analysis. There is still room for further research relating to the relationship between them. To summarize, this paper constructs a panel quantile and mediation effect model, empirically analyzes the impact of industrial structure upgrades on carbon emissions from the perspective of provincial differences, and further tests the mediation effect of energy structure optimization.

2.2. Research Hypothesis

2.2.1. Impact Mechanism of Industrial Structure on Carbon Emissions

During China’s current industrial development, a situation of unreasonable industrial layout and high investment and pollution from secondary industries has gradually emerged. The differences in the level of industrial structures in different regions have led to different carbon emission efficiencies [29]. The impact path of industrial structures on carbon emissions can be summarized in the following two aspects: first, there are apparent differences in the energy demand of various industries. The carbon emission coefficient of each energy source is different, and the carbon emission generated in the production process is also different, especially in secondary industries, which include most industries with high energy consumption [30]. The energy consumption per unit output value of various industries is different, and the development ratio of various industries directly affects the total demand for energy, which in turn has an indirect impact on carbon emissions. Upgrading industrial structures will increase the proportion of the service industry and gradually gather in low-energy sectors, which will gradually “decouple” economic development from carbon emissions. Second, upgrading industrial structures has realized the free flow of production factors among industries, which is conducive to the effective allocation of resources, making capital or labor factors flow to enterprises faster and promoting the green transformation of enterprises.

Based on the above analysis, Hypothesis 1 is proposed as follows:

Hypothesis 1.

The upgrading of industrial structure has a significant negative inhibitory effect on carbon emissions.

2.2.2. Impact Mechanism of Industrial Structure on Energy Structure

The differences in the resource allocation and utilization efficiency of various industries lead to changes in industrial structures that can affect the level of carbon emissions by affecting the structure of energy consumption. The impact of industrial structures on energy structures mainly includes the following two aspects: first, upgrading the industrial structure can optimize the energy consumption structure [31]. From a macro point of view, upgrading the industrial structure will promote the use of clean energy and even help stimulate the development and utilization of clean energy; from a micro point of view, natural gas and solar energy will be used to replace traditional high-carbon emission energy such as gas and coal used by residents to reduce carbon emissions. Second, upgrading the industrial structure is conducive to technological progress [32]. On the one hand, technological progress can lead to the development of new low-carbon technologies that reduce carbon emissions in the production process; on the other hand, low-carbon energy can be researched and improved to make it a better substitute.

Based on the above analysis, Hypothesis 2 is proposed as follows:

Hypothesis 2.

The upgrading of industrial structure has a significant negative inhibitory effect on the proportion of coal energy consumption. That is, the upgrading of industrial structure has a positive impact on the optimization of energy structure.

2.2.3. Impact Mechanism of Energy Structure on Carbon Emissions

Since the distribution of resources in Chinese territory has the characteristics of “rich coal, little oil, and poor gas”, the resource endowment conditions under the attribute of natural resources have a specific impact on the distribution characteristics of greenhouse gas emissions in various regions. There are three main pathways for the energy structure to promote the process of “carbon neutrality”. First, with the implementation of China’s energy-saving and emission-reducing measures, especially the continuous increase in control in the industrial sector, the growth rate of energy consumption has slowed down to lower than the growth rate of the GDP. China’s total energy consumption has been controlled, resulting in the decreased volume and intensity of carbon emissions. Economic development is gradually decoupling from high energy consumption and high carbon emissions. It has entered a stage of high-quality and stable economic development supported by relatively low energy consumption growth. Second, the energy structure has changed from coal-based to diversified, the structure of consumption products has improved, and the proportion of coal has continued to decrease. The CO₂ emissions of coal are 1.6 times that of natural gas and 1.2 times that of petroleum. At the same time, the consumption of energy from nuclear power, wind power, hydropower, solar energy, and other renewable energy sources does not emit CO₂. The proportion of carbon emissions from coal and petroleum has decreased, the proportion of carbon emissions from natural gas has increased, and the utilization efficiency of clean energy has improved. Together, the effects of slowing down the increase in carbon emissions and reducing carbon emissions are apparent. Third, carbon emissions from the power sector account for 85% of China’s carbon emissions. The continuous optimization of the energy structure promotes a decline in the proportion of coal-fired power. Changes in energy use methods expand the application of renewable energy in the power industry and build a clean-energy-based structure. New power systems will play a vital role in the deepening carbon reductions in terms of energy.

According to the previous analysis, optimizing and upgrading industrial structures have a positive impact on reducing carbon emission intensity. Based on the above analysis, Hypothesis 3 and Hypothesis 4 are proposed as follows:

Hypothesis 3.

The decline in the proportion of coal energy consumption has a significant positive effect on carbon emission reduction; that is, the optimization of energy structure has a significant positive impact on carbon emission reduction.

Hypothesis 4.

Energy structure has a significant mediating effect between industrial structure and carbon emissions.

3. Model and Data

3.1. Model

3.1.1. Quantile Regression Model

In most previous regression models, the focus has been on the influence of explanatory variables on the conditional expectations of the explained variables, which is mean regression. However, sometimes, for the conditional distribution, the mean does not reflect the whole picture of the distribution well. In 1978, Roger Koenker and Gilbert Bassett [33] proposed a quantile regression method that could solve this problem very well. The basic idea is to minimize the difference between the explanatory variable and the fitted value. Quantile regression can observe the tail of the dependent variable and more accurately reflect the influence of the independent variable on the shape of the conditional distribution of the dependent variable. Additionally, it does not make any assumptions about the distribution of random error items, the results are not easily affected by extreme values, the regression is more robust, and it can reflect the data more comprehensively. In the past, when scholars analyzed the influencing factors of carbon emissions, most of them divided China into regions according to different geographical locations. Considering the actual situation in China, this analysis method ignored the apparent differences in the economic structures and development levels in various provinces in China. For regions with different development models and development processes, the same factors have different influences on carbon emissions. Therefore, this paper adopts the method of quantile regression and takes carbon emission intensity as the dependent variable, sets five quantile points (0.10, 0.25, 0.50, 0.75, and 0.90) to divide China’s 30 provinces, municipalities, and autonomous regions (except the Tibet Autonomous Region, Hong Kong, Macao, and Taiwan) into six identifiable carbon emission areas. The impact of industrial structures on carbon emission intensity is explored by region. According to the results of the analysis, carbon emission reduction policies suitable for the area are proposed for areas with different carbon emission levels.

Ehrlich et al. [34] first proposed the IPAT model to assess environmental pressure, and the model has been widely used by researchers to study the influencing factors of environmental pollution. The expression of the IPAT model is as follows:

(1)$I=P\times A\times T$

where $I$ represents the degree of environmental pollution, $P$ represents population size, $A$ stands for the degree of the economic prosperity of a country, and $T$ denotes the level of technological development.

However, many scholars have pointed out that there are some defects in the traditional IPAT model. Firstly, Formula (1) is a mathematical identity that does not allow statistical parameter estimation and hypothesis testing [35]. Secondly, the model assumes that the elastic coefficient of $P,A$, and $T$ is 1, but it is not easy to satisfy in the actual situation [36]. Based on the IPAT model, Dietz et al. [18] proposed the STIRPAT model; its specific form is as follows:

(2)$I=a{P}^b{A}^c{T}^{d}e$

where $I,P,A$, and $T$ are defined in equation (1), $a$ indicates the intercept item and $b,c$, and $d$ are the influence coefficients of $P,A$, and $T$, respectively, the subscript $t$ means different years, and $e_t$ depicts the random error item.

On this basis, York et al. [37] proposed that taking the logarithm of the STIRPAT model can eliminate heteroscedasticity to a certain extent, and the formula after taking the logarithm is more convenient for subsequent parameter estimation and hypothesis testing, as follows:

(3)$\ln I=\ln a+b\ln P+c\ln A+d\ln T+\ln e$

where $I,P,A,T,a,b,c,d$, and $e$ are defined in equation (2).

For a given $\tau (0<\tau <1),$ denoted by $Q_{\tau }\left(y|x\right)$, the τth conditional quantile function of $y$ is given as $x.$ The quantile regression model assumes that the conditional quantile function $Q_{\tau }\left(y_i|x_i\right)$ is expressed as follows:

(4)$Q_{\tau }\left(y_i|x_i\right)={\beta }_0\left(\tau \right)+{\beta }_1\left(\tau \right)x_{i1}+\cdots +{\beta }_p\left(\tau \right)x_{ip}, i=1,\dots ,n.$

where $y_i$ is a response variable, $x_i={\left(x_{i1},\cdots ,x_{ip}\right)}^T$ is a $p$-dimensional random vector, and ${\beta }_0\left(\tau \right), \dots ,{\beta }_p\left(\tau \right)$ are unknown regression parameters.

The form of the STIRPAT model is flexible and allows for the proper decomposition of each factor. Therefore, based on this model, this paper conducts a proper decomposition of each influencing factor. We choose industrial structure (IS) and energy structure (ES) (seen in Section 2.1), technological innovation (TE), population (PL), level of opening-up (OP), urbanization rate (UR), and carbon emission trading policy (CM) (seen in Section 3.2.4) as independent variables, with reference to previous studies. Additionally, the two-way fixed effect model for regression analysis is used according to the following tests shown in Section 4. As a result, the benchmark model of this paper is obtained by combining the STIRPAT model and the quantile regression model, as follows:

(5)$$\begin{gathered} \ln C{G}_{it}={\mu }_0+{\mu }_1\ln I{S}_{it}+{\mu }_2\ln E{S}_{it}+{\mu }_3\ln O{P}_{it}+{\mu }_4\ln T{E}_{it}+{\mu }_5\ln P{L}_{it}+ \\ {\mu }_6\ln U{R}_{it}+{\mu }_7\ln C{M}_{it}+u_i+v_t+{\epsilon }_{it} \end{gathered}$$

where ${\mu }_0,{\mu }_1,\cdots ,{\mu }_7$ are unknown regression parameters, $i$ and $t$ represent the region and time, respectively, $u_i$ means the province fixed effects, $v_t$ is the time fixed effect, and ${\epsilon }_{it}$ is random error whose $\tau$th quantile conditional on the covariates equals 0.

3.1.2. Mediation Effect Model

For the transmission path between the industrial structure, energy structure, and carbon emission intensity, this paper mainly selects the stepwise regression method to verify the relationship between them.

First, according to theoretical analysis, it is known that there may be a negative correlation between the industrial structure and carbon emission intensity. This paper constructs a double fixed effect model for empirical testing. The specific model for testing the impact of industrial structure upgrades on carbon emission intensity is shown in the following:

(6)$\ln C{G}_{it}={\alpha }_0+{\alpha }_1\ln I{S}_{it}+{\alpha }_2{X}_{it}+u_i+v_t+{\epsilon }_{it}$

In Formula (6), $C{G}_{it}$ is the carbon emission intensity at the provincial level, $I{S}_{it}$ is the industrial structure upgrade index at the provincial level, and $X_{it}$ is the control variable at the provincial level, including technological innovation (TE), population (PL), level of opening-up (OP), urbanization rate (UR), and carbon trading market pilot policy (CM). In addition, $i$ and $t$ indicate the province and the year, respectively, ${\alpha }_0$ means the coefficients of the constant term, ${\alpha }_1$ and ${\alpha }_2$ depict the regression coefficients of the industrial structure and control variables, $u_i$ means the province fixed effects, $v_t$ is the time fixed effect, and ${\epsilon }_{it}$ is a random disturbance term.

Secondly, to test the mediating effect of the energy structure, this paper draws on the research of Baron and Kenny [38]. The specific model is shown in Formulas (6), (7), and (8) as follows:

(7)$\ln E{S}_{it}={\beta }_0+{\beta }_1\ln I{S}_{it}+{\beta }_2{X}_{it}+u_i+v_t+{\epsilon }_{it}$

(8)$\ln C{G}_{it}={\gamma }_0+{\gamma }_1\ln I{S}_{it}+{\gamma }_2\ln E{S}_{it}+{\gamma }_3{X}_{it}+u_i+v_t+{\epsilon }_{it}$

where $I{S}_{it}$ is the mediation variable of the energy structure, $E{S}_{it}$ signifies the energy structure, ${\beta }_0$ and ${\gamma }_0$ signify the coefficients of the constant term, ${\beta }_1$ and ${\beta }_2$ depict the regression coefficients of the industrial structure and the control variables, ${\gamma }_1$ and ${\gamma }_2$ express the regression coefficients of the industrial structure and the control variables, and other parameters consistent with Formula (6). The test principle of the mediation effect is as follows: the first step is to test whether the negative effect of the industrial structure on carbon emission intensity exists; that is, a test to determine whether the value in Formula (6) ${\alpha }_1$ is significantly negative. In the second step, after the verification of Formula (7), we tested whether an advanced industrial structure has an inhibitory effect on the proportion of coal energy consumption of the mediator variable; that is, to check whether the value in Formula (7) ${\beta }_1$ is significantly positive. In the third step, after Formula (7) is verified, the carbon emission intensity and the mediation variable are included in the same model to test whether the mediation effect of the mediation variable exists. If ${\gamma }_2$ and ${\gamma }_1$ are significant, it indicates that the mediation variable is partial mediation; if ${\gamma }_2$ is significant but ${\gamma }_1$ is not significant it suggests that the mediating variable is fully mediating. Moreover, if ${\gamma }_1{\gamma }_2$ is insignificant, it indicates that the mediating variable does not have a mediating effect.

3.2. Variable Selection and Measurement

3.2.1. Interpreted Variables

Carbon emission intensity refers to the amount of carbon dioxide emissions per unit of gross national product [39], and its calculation formula is as follows:

Carbon emission intensity = carbon emissions/gross national product.

This indicator is mainly used to measure the relationship between a country’s economy and carbon emissions. If a country’s economic growth is accompanied by a decline in carbon dioxide emissions per unit of gross national product, it means that the country has realized a low-carbon development model. Carbon intensity is a relative indicator, that is, the amount by which carbon emissions will increase or decrease relative to economic growth. The pure total carbon emissions can visually display the historical emissions of various provinces and cities in China, but it does not take economic development into account and is not enough to accurately describe the critical issues involved in the process of carbon emissions. Therefore, carbon emission intensity, which considers economic output factors, is chosen as the carbon emission assessment index in this paper.

3.2.2. Core Explanatory Variables

Upgrading industrial structures essentially refers to the process or trend of the transformation of industrial structures from a low-level to a high-level form. Its ultimate direction is the high-tech and high intensification of industries. Industrial structure upgrades are usually divided into two aspects: industrial structure upgrades and rationalization, and upgrading is the fundamental goal of industrial rationalization. Therefore, this paper chooses the industrial structure upgrading index (IS) as the index of industrial structure upgrading. Refer to the method of Chunhui Gan [40,41,42,43] to calculate the advanced industrial structure index. Its calculation formula is as follows:

Index of industrial structure upgrading = added value of the tertiary industry/added value of the secondary industry.

The larger the value of $IS$, the more advanced the industrial structure [44], and vice versa.

3.2.3. Mediator Variable

In this paper, energy consumption structure (ES) is used to measure the energy structure; it is a control variable in the panel quantile model and a mediator variable in the mediation effect model. Referring to the research on the change law of China’s energy consumption structure, this paper defines the energy consumption structure as the proportion of coal consumption in the national economy to the total energy consumption within a certain period [45,46]. Its calculation formula is as follows:

Energy consumption structure = coal energy consumption/total energy consumption.

A smaller ES value means that the energy consumption structure is continuously optimized, the proportion of clean energy consumption is continually increasing, and the energy structure is developing in a clean, efficient, and low-carbon direction. Chinese energy consumption is dominated by coal. Studying the energy consumption structure will help to grasp the sources of energy consumption, provide a scientific basis for the rational allocation and utilization of energy, and lay the foundation for balancing energy supply and demand.

3.2.4. Control Variables

Referring to the practical condition of China and previous studies, the following variables are selected as control variables: level of opening-up (OP) [47], technological innovation (TE) [48], population (PL) [49], urbanization rate (UR) [50], and carbon trading market pilot policy (CM) [51].

To eliminate the influence of heteroscedasticity that may exist in the variables, this paper takes the logarithm of all variables.

3.3. Data Sources and Statistical Characteristics

From 1997 to 2019, 30 provinces, autonomous regions, and municipalities directly under the Central Government (excluding Tibet, Hong Kong, Macao, and Taiwan) were used as the research sample. The data relating to the gross national product, population, urbanization rate, the added value of the secondary industry, the added value of the tertiary industry, and the total import and export data were taken from the “China Statistical Yearbook” and the Reiss database, and the number of patent authorizations was taken from the “China Science and Technology Statistics Yearbook. Coal consumption and total energy consumption were taken from the China Energy Statistical Yearbook. Carbon emissions were taken from the “CEADs—China Carbon Accounting Database” [52,53,54]. Some of the missing data were filled through linear interpolation. Descriptions of the related variables and their descriptive statistics are shown in Table 1 as follows:

4. Empirical Test

4.1. Regional Division

4.1.1. Current Status of Carbon Emissions in China

Figure 1 represents the total carbon emissions from 1997 to 2019 in China. From a national perspective, with the rapid development of China’s economy and urban construction, China’s carbon emissions have generally shown an increasing trend. In 1997, China’s total carbon emissions were 3.308 billion tons, which increased to 12.290 billion tons in 2019, a four-fold increase in 23 years. In the past, the development of China’s economy relied on a large amount of fossil energy consumption, and an economic growth mode showed extensive characteristics. Social and economic development and modernization were in a period of acceleration. From 1997 to 2011, Chinese carbon emissions rapidly increased. With the advancement of production technology, limited resources, and the pressure of environmental degradation, although the overall national carbon emissions have shown a general upward trend after 2011, the growth rate has slowed down significantly. Especially in 2015, there was a slight decline, and the country’s total carbon emissions have been significantly controlled.

The IPCC report pointed out that the combustion of fossil energy is the main cause of CO₂ emissions in the form of greenhouse gases and that most non-fossil energy is clean energy. China has a relatively prominent feature in the energy utilization structure; that is, the energy consumption structure in China is dominated by the utilization of raw coal. Figure 2 reports the total carbon emission intensity in each province from 1997 to 2019, from which it can be seen that the carbon emission intensity of Shanxi Province was the highest in the country. The main reason is that the development of Shanxi Province mainly relies on coal-based energy-intensive industries, which generate excessive CO₂ emissions. Ningxia Province ranks second, followed by Inner Mongolia, Henan Province, and Guangdong Province. On the whole, Chinese carbon dioxide emissions are low in the west and southwest but high in the east and north, which is consistent with the economic development level and resource endowment of each province.

4.1.2. Regional Grouping

Considering the actual situation, China has a vast territory, and there are significant differences in factors such as economic development levels, resource endowments, and energy consumption structures between the provinces [55]. If regional differences are not taken into account and all development models are identical, it will be difficult to find an optimal emission reduction path to achieve energy saving and emission reduction goals. In this paper, according to the order of the annual average carbon emission intensity of 30 provinces from small to large from 1997 to 2019, five representative quantile points (0.10, 0.25, 0.50, 0.75, and 0.90) were investigated, and the 30 provinces were divided into six groups (see Table 2). It can be seen from Figure 2 that there are obvious differences in the carbon emission intensity of different provinces, which shows that the panel quantile model used in this paper is reasonable and applicable.

4.2. Quantile Regression Analysis

4.2.1. Panel Unit Root Test and Cointegration Test

To avoid spurious regression and ensure the validity of the regression estimation results, it is necessary to perform a unit root test on the panel data. Due to the low effectiveness of one test method and to ensure the validity of the regression estimation results, this paper adopts four unit root test methods: LLC test, IPS test, ADF test, and PP test. A variable need to pass these four tests at the same time to be considered stationary. The unit root test results obtained by Stata show that all variables are non-stationary at zero order. Table 3 shows that after the variables are differentiated, the first-order difference sequence rejects the null hypothesis at the 1% significance level, and the difference sequence is stable.

Given the first-order stationarity of the variables, a cointegration test can be carried out to judge whether there is a long-term equilibrium relationship between the variables. In this paper, the Pedroni test, Kao test, and Westerlund test are used to test the panel cointegration of the other variables, except for the dummy variables. As shown in Table 4, all cointegration test results reject the null hypothesis at the 1% significance level; therefore, it can be concluded that there is a long-term cointegration relationship between the variables, and the possibility of pseudo-regression problems in the model can be excluded.

4.2.2. Model Form

To choose a specific form of the model, make the regression estimate more accurate. First, this paper uses the F-test in the fixed effects model to judge whether to use the fixed effects model or the mixed effects model. According to the regression analysis results, the p-value of the F test is 0.0000, which strongly rejects the null hypothesis, and it is believed that the fixed effect model is better than the mixed effect model. Secondly, the LM test is used to judge whether to adopt the mixed effect model or the random effect model. According to the analysis results, the p-value of the LM test is 0.0000, which strongly rejects the original hypothesis of “there is no individual random effect”, and it is believed that the random effect model is better than the mixed effect model. Finally, this paper uses the Hausman test to judge whether the model is a fixed effect or a random effect. According to the test results, the p-value is 0.0000, strongly rejecting the original hypothesis of “random effect is the most efficient”, and it is believed that a fixed effects model should be used. In this section, the above tests are not listed in detail due to limited length.

4.2.3. Analysis of Regression Results

Table 5 and Figure 3 describe the quantile regression results. It can be seen from Table 5 and Figure 3a that for the core explanatory variable industrial structure, the sign of the 10% quantile point regression coefficient is positive. The rapid development of the carbon emission industry has increased the carbon emission intensity overall. At the 25%, 50%, 75%, and 90% quantile points, the industrial structure and carbon emission intensity have a negative relationship. The direction of the two changes is opposite, indicating that as the regional economy further enters a new and high-quality development stage, higher requirements were proposed for the optimization of industrial structures, the proportion of high-energy-consuming industries has gradually decreased, and the industrial focus shifted to the sustainable development of the tertiary industry [56]. The regression coefficient generally shows a “U-shaped” trend. After taking the absolute value of the coefficient, the regression coefficient decreases first, then increases, and finally declines with the change in quantile points: from 5% to 20% quantile points, the impact of industrial structures on carbon emissions. The positive promotion effect gradually weakens from the 20% to the 40% quantile point, the negative inhibitory effect of the industrial structure on carbon emissions gradually increases, and the coefficient shows a “sag” at the 40% quantile point in the point in Figure 3a; from the 40% to the 95% quantile, the negative inhibitory effect of industrial structure on carbon emissions gradually weakens. In the early stage, the economic structure changed from agriculture to energy-intensive heavy industry, which increased pollution emissions. Then, the industry shifted to low-pollution services and knowledge-intensive industries, the emission level per unit of output decreased, and the environmental quality gradually improved.

Table 5 and Figure 3b above show that for the energy structure, the signs of the quantile regression coefficients are all positive, and the coefficients are relatively large, indicating that the proportion of coal energy consumption and carbon emission intensity is significantly in the same direction, and the direction of the change of the two is the same. That is, when the proportion of coal consumption is high, the carbon emission intensity is high. Moreover, as the ratio of clean energy continues to increase, the carbon emission intensity will show a downward trend. Additionally, from the coefficient point of view, from 1.2656 at the 10% quantile level to 0.8192 at the 90% quantile level, the impact of the energy structure on carbon emission intensity gradually weakens. Due to the underdeveloped industries in low-carbon emission areas, the consumption of fossil energy is relatively small, and the changes in energy structure caused by technological progress have little impact on low-carbon emission areas. The energy consumption structure has been slightly optimized, and the emission reduction effect is remarkable. The pillar industries in Chinese high-carbon emission areas are industrial-based, and technological innovation can effectively reduce the proportion of coal energy consumption. Additionally, a large amount of energy consumption makes the unit carbon emission intensity lower than that of low-carbon emission areas. The potential and space for energy structure optimization are gradually decreasing, and the emission reduction effect is not as significant as that of low-carbon emission areas. The impact on carbon emissions begins to show a gradual downward trend.

It can be seen intuitively from Figure 3c that for the total amount of imports and exports, the signs of the quantile regression coefficients are all positive, indicating that the increase in imports and exports during the sample period has a significant positive impact on the increased carbon emission intensity. The development of export trade can promote economic growth, and the expansion and development of economic scale will introduce very high requirements for the resource elements that need to be invested in when developing the economy. It will require more energy consumption, such as from coal, and carbon emissions will also increase accordingly. As the quantile level increases, the regression coefficient fluctuation decreases. The continuous expansion of export trade in low-quantile provinces will significantly increase carbon emissions, so export trade has a significant positive impact on carbon emission intensity through the scale effect. However, provinces with high quantile levels have relatively strong technological advantages and more technical talents. Relying on technological advantages to expand the economic scale, the resulting increase in carbon emission intensity is relatively tiny.

As shown in Figure 3d, for the technology level, the signs of the quantile regression coefficients are all negative. That is, technology level and carbon emission intensity have an inverse relationship, and the direction of the change of the two is opposite. Technological progress, especially low-carbon technological progress, can make energy development and utilization more advanced and reasonable, thereby significantly improving energy utilization efficiency, reducing high-carbon energy consumption, and ultimately reducing carbon emission intensity. After taking the absolute value of the regression coefficient, it can be seen that the impact of the technology level on carbon emissions gradually weakens. The reason is that in the early stage of scientific and technological progress, the increase in the number of patent authorizations has prompted enterprises to improve processes, transform equipment, and increase production efficiency. Scientific and technological innovation provides technical support for reducing carbon emissions and has a significant effect. However, with the gradual increase in the number of patent authorizations, the resistance to scientific and technological innovation is increasing. We are faced with problems such as insufficient basic research, insufficient transformation of achievements, and a lack of motivation for transformation [57]. At the same time, the marginal utility of innovation is growing weaker and weaker. Innovation is substantial, but the quality has gradually declined. The decline in the quality of innovation leads to wasted capital investment, slowly leading to a decrease in production efficiency, and the carbon emission reduction effect of technological progress is thus weakened.

Judging from Figure 3e, for the population size, the signs of the quantile regression coefficients are all positive; that is, the population size and carbon emissions change in the same direction. Directly speaking, residents are the main body of production, consumption, and energy use. With improved living standards, the per capita energy consumption continues to increase. The increasing pursuit of improved material living standards by residents leads to an increase in the living energy consumption of residents, affecting energy consumption and causing a rapid increase in carbon dioxide emissions. The regression results show that the influence coefficient of population size on carbon emission intensity fluctuates and decreases, indicating that the impact of population size on carbon emission is gradually weakening. In the initial stage of rising carbon emission intensity, the expansion of population size drives increased consumer demand, and the rapid development of secondary and tertiary industries places tremendous pressure on carbon emission reduction. When the carbon emission intensity reaches a certain quantile level, the lifestyle of residents will transform, gradually forming a green and low-carbon consumption pattern and lifestyle, and the impact of the population on carbon emissions will be reduced.

According to Figure 3f, for the level of urbanization, the signs of the quantile regression coefficients are all positive; that is, the change direction of the urbanization level and carbon emission intensity is the same. With the acceleration of China’s industrialization processes, the level of urbanization has also continued to increase. The advancement of urbanization not only requires the consumption of a large amount of reinforced concrete but also has a particular impact on people’s living habits, resulting in a substantial increase in energy consumption and the generation of more CO₂. However, the regression coefficients of different quantile points showed a trend of first rising briefly and then fluctuating down, indicating that the urbanization level of different quantiles of carbon emission intensity has various impacts on carbon emissions. When the carbon emission intensity is low, there is ample space for urbanization development, there are too many industrial production activities, the distribution of residential and living service resources is unbalanced, public transportation cannot cover primary commuting, residents turn to private transport, and real estate is over-developed. With high-carbon energy and infrastructure investment, it is more challenging to clear carbon emissions. As the proportion of the urban population increases, the marginal cost of reducing emissions in cities increases, and the impact coefficient on carbon emissions gradually decreases.

As seen in Table 5 and Figure 3g, for the carbon trading pilot policy, a dummy variable, the regression coefficients of both quantiles are insignificant, and the implementation of the carbon trading pilot policy lacks explanatory power in terms of carbon emissions. The reason is that provinces and municipalities with high and low carbon emission intensity have not incorporated carbon trading pilot policies into their carbon emission reduction frameworks. For the provinces in the middle of the quantile, the impact of the policy on carbon emission intensity is not stable, and the policy reduces the carbon emission intensity of the 25%, 50%, and 75% carbon emission areas by 1.92%, 3.58%, and 9.23%, respectively. During the sample period, eight provinces and cities, including Beijing, Tianjin, Shanghai, Chongqing, Hubei, Guangdong, Shenzhen, and Fujian, successively launched carbon emission trading pilot work, explored the use of market mechanisms to control greenhouse gas emissions and promoted China’s realization of green and low-carbon development strategies. However, judging from the impact coefficients of each quantile point, the carbon trading pilot policy has little impact on carbon emissions. During the inspection period, China’s carbon emissions trading market is still in the initial stage of construction, the types of transactions and trading mechanisms are far from being able to meet international standards, regulations and market legislation lag behind, and the accuracy of carbon emission data statistics, verification, and quota allocation plans are relatively large. There is room for improvement, and the carbon trading mechanism will still face challenges in the future.

4.2.4. Robustness Test

First, replace the explained variable. In the baseline regression analysis, carbon emission intensity is used as the explained variable. Here, the logarithm of per capita carbon emissions (lnPC) is used as the explained variable for robustness testing. The regression results are shown in column (1) in Table 6. Second, replace the explanatory variables. Referring to the research of Liang et al. [58] and Gu et al. [12], the logarithm of the industrial structure rationalization index (lnIR) is used to measure the industrial structure. The regression results are shown in column (2). It can be found that the estimated value of the coefficient of industrial structure is still significantly negative, and it has a significant negative inhibitory effect on both carbon emission intensity and per capita carbon emission, which is consistent with the baseline regression results. Therefore, the conclusion of this paper is robust.

4.3. Mediating Effect Analysis

4.3.1. Stepwise Regression Test

After controlling the relevant variables, the benchmark regression results of the three models are obtained by using the stepwise regression method, as shown in Table 7. Among them, Model 1 to Model 3 are the direct effect model of the industrial structure on carbon emission intensity, the direct effect model of the industrial structure on energy structure, and the indirect effect model of the energy structure on carbon emission.

According to the empirical results, it can be seen that the core explanatory variable industrial structure and the mediation variable energy structure in models 1 to 3 have passed the significance test at the 1% level, and the goodness of fit of the models is good. From the regression results of Model 1 and Model 3, the regression coefficients of the industrial structure are all negative; that is, optimizing and upgrading industrial structures have significantly suppressed carbon emissions and have had a significant negative impact on it, verifying Hypothesis 1. In the context of a low-carbon economy, promoting the optimization and upgrade of industrial structures can not only help optimize the economic structure and transform growth drivers but also help reduce carbon dioxide emissions and the impact of economic development on the ecological environment. The industrial structure has gradually become a “connector” and “speaker” between economic growth and carbon emission reduction. On the one hand, advanced industrial structures encourage new economic growth momentum; on the other hand, fewer high-energy-consuming industries improve the service industry ratio, creating a more reasonable industrial layout for the development of a low-carbon economy. From the regression results of Model 2, the regression coefficient of the energy structure is negative, indicating that the optimization and upgrade of the industrial structure will reduce the proportion of coal consumption in total energy consumption, thus verifying Hypothesis 2. The rapid development of high-energy-consuming industries will inevitably lead to strong demand for energy supply, and the expansion of the proportion of the tertiary industry in the national economy will bring about a reduction in energy consumption, thereby slowing down the growth of carbon emissions, which is crucial to the realization of “double carbon”. Expected goals have a positive, stimulating effect. From the regression results of Model 3, the regression coefficient of the energy structure is positive, indicating that the optimization of the energy structure has a significant positive impact on reducing carbon emission intensity, thus verifying Hypothesis 3.

Based on the above results, further explore the mediating effect of the energy structure in the process of industrial structure’s impact on carbon emission intensity. The regression results of Model 1 show that the total effect coefficient of the industrial structure on carbon emission intensity is −0.3670; the regression results of Model 2 show that upgrading industrial structures plays a significant role in promoting the optimization of energy structures; the regression results of Model 3 show the upgrading industrial structures. Both energy structure optimization and energy structure optimization have a negative inhibitory effect on carbon emission intensity. At this time, the impact coefficient of industrial structures on carbon emission is −0.2943. Compared with the coefficient in Model 1, the absolute value is slightly lower, indicating the energy structure index. After adding the regression equation, the inhibitory effect of the industrial structure on carbon emission intensity is reduced. From the mediation effect β1γ2/α1, it can be seen that the mediation effect of industrial structure upgrades is 19.83%; that is, in the process of industrial structure restraining carbon emission intensity, industrial structure upgrades exert a 19.83% mediation effect. Based on the above analysis, Hypothesis 4 can be verified.

4.3.2. Bootstrap Mediation Effect Test

In order to further verify the mediating effect of the industrial structure, this paper uses the basic Bootstrap re-sampling technique to test the significance of the mediating effect of energy structure upgrades. The test results are shown in Table 8. The total effect, direct effect, and indirect effect have all passed the significance test at the 1% level, and the sign and value of the effect coefficient are almost the same as the results in the above stepwise regression test method. Based on the above analysis, Hypothesis 4 is still confirmed.

5. Discussion

This paper takes the panel data of 30 provinces, autonomous regions, and municipalities directly under the central government in China from 1997 to 2019 as the research sample. It uses the panel quantile regression model to study the regional differences in the impact of industrial structure on carbon emissions in the 30 provinces and analyze the dynamic change process under different quantiles. Furthermore, the mediation effect model was used to investigate the relationships and transmission mechanisms between upgrading industrial structures, energy structure optimization, and carbon emission reduction. The main conclusions are as follows: (1) Overall, both regional industrial structure upgrades and energy structure optimization have significant inhibitory effects on carbon emissions. (2) As the quantile increases, the regression coefficient of the impact of industrial structures on carbon emissions generally presents a “U-shaped” trend. After taking the absolute value of the coefficient, the regression coefficients first take on a downward and then upward and finally downward trend with the changing quantile. (3) The elastic coefficients of the seven influencing factors, including industrial structure, were significantly different in different quantiles. That is, the impact on the carbon emission intensity of different provinces was significantly different in magnitude and direction. The main factors that inhibit carbon emissions are industrial structure and technological progress. On the contrary, the factors that promote the growth of carbon emission intensity are the energy structure, import and export trade, population size, and urbanization level. (4) Upgrading industrial structures plays a significant role in promoting the optimization of energy structures. (5) Upgrading industrial structures can not only directly restrain carbon emissions but also indirectly have a significant inhibitory effect on carbon emissions by promoting the optimization of energy structures.

According to the findings above, the theoretical implications can be given as follows: (1) When comparing our results to those of older studies, it must be pointed out that we focused on the differences in the effects of factors, such as the industrial structure, between provinces and different carbon emission intensities on carbon emission reduction, which may enrich research results regarding the application of the quantile model in the field of carbon emission reduction. (2) Based on the mediation effect, previous studies believed that industrial structures would have an impact on carbon emissions, but there has been little empirical research on the impact path of “Industrial Structure Upgrading—Energy Structure Optimization—Carbon Emission Reduction”. This study uses the mediation effect model to verify this and supplements the research literature on the mechanisms of industrial structures in relation to carbon emissions.

Based on the above conclusions, this paper proposes the following suggestions: (1) Adhere to the principle of the scientific layout of industrial structures and implement differentiated industrial development strategies. In provinces where the industrial structure has a strong inhibitory effect on carbon emissions, it is necessary to continue to deepen industrial reform, take advantage of knowledge-intensive and technological innovation, vigorously promote the development of high-tech industries, and promote the further convergence of carbon emissions. In provinces where the industrial structure has a relatively weak inhibitory effect on carbon emissions, the adjustment of the industrial structure and the transformation of the regional economy will require a long period of transition and a change in driving force. During this period, it is necessary to seize development opportunities and change the weak state. Actively encourage the development of strategic emerging industries and service industries and continuously reduce the proportion of high-energy-consuming industries in the regional economy. When necessary, a plan to eliminate backward production capacity can be implemented, a mechanism for exiting backward production capacity can be established and improved, and the internal structure of the primary, secondary, and tertiary industries can be continuously optimized in order to effectively and accurately empower carbon emission reductions. (2) Reasonably divide regions, identify key provinces, and adopt targeted energy policies. For provinces whose energy structure has a strong inhibitory effect on carbon emissions, the energy policy should continue to adjust the energy structure as the primary task: promote the development of energy diversification, reduce the proportion of fossil energy in the energy structure, and accelerate the development and utilization of new energy. Develop renewable energy sources, such as wind energy, solar energy, biomass energy, and geothermal energy, according to the local conditions, increase the replacement ratio of renewable energy and new energy, and build a safer, more efficient, cleaner, low-carbon, and green energy supply and consumption system. For provinces whose energy structure is relatively weak in inhibiting carbon emissions, policy directions should be adjusted to improve energy efficiency, as follows: increase the introduction, digestion and absorption, and innovation of high-tech and advanced applicable technologies, graft and transform traditional industries, and accelerate the technological upgrade of traditional industries. Promote reforming the system and mechanisms, strengthen the management of the energy industry and energy strategies, plans, policies, measures, supervision, and services. (3) Adhere to the principle of combining industrial structure adjustment and energy structure optimization. While actively adjusting the industrial structure, vigorously promote improving the internal energy efficiency of enterprises, develop and promote the utilization of renewable clean energy, and optimize the energy structure; promote the transformation and upgrade of traditional industries and encourage existing high energy-consuming enterprises to accelerate the transformation to low-consumption. Improve energy development strategies, planning, and industrial policies and develop low-carbon transportation and green buildings, etc.; build a public research and development platform for industrial technology and energy technology and promote collaborative innovation. (4) The adjustment of industrial structure has its own evolution law, and there are also interrelationships between industries. Therefore, while reducing carbon emissions, it is necessary to maintain stable and rapid economic growth, ensure the level of urbanization and other strategic goals, and coordinate planning from a global perspective. R&D expenditure and personnel investment in emission reduction technology should be further expanded to encourage technological progress; incentive and restrictive policies should be formulated to slow down population growth and housing demand and avoid irreversible environmental damage in some provinces; scientifically plan urban construction and promote the construction of new infrastructure; promote import and export trade to green trade; coastal provinces should give full play to their port advantages, establish a trade system compatible with the low-carbon economy, and improve the corresponding import and export trade mechanisms; inland provinces should adhere to technological innovation. On the basis of innovation and management innovation, actively play the positive effect of import and export trade and promote the expansion of the national carbon emission trading market, accelerate the inclusion of major energy-intensive industries, balance the relationship between carbon emission reduction and economic growth, and realize the goal of “double carbon” with the smallest investment.

Limitations and future research directions include the following: (1) In the quantile regression analysis, it was found that there were insignificant points in the carbon trading mechanism. One possible reason is that the subject and duration of policy implementation are insufficient, and it is difficult to find regularity. At present, the geographical and industry scope of carbon market transactions is being promoted, and it is believed that more sufficient data will be included in future research. (2) The impact of the industrial structure on carbon emissions may have other impact mechanisms. In the future, research will continue to focus on other impact paths to build a more efficient carbon reduction mechanism by optimizing the industrial structure. (3) This paper uses the industrial structure as one of the key variables, but our data mainly divide the industries of each province into primary, secondary, and tertiary industries and does not conduct a more detailed study on the carbon emission impact mechanism under each industry. The following research can shift the research perspective from industry to industry by conducting research on the factors affecting carbon emissions in key industries with relatively unreasonable energy consumption structures and heavy carbon usage and comparing various industries to test this research again.

Author Contributions

Conceptualization, A.Z. and G.F.; methodology, A.Z. and G.F.; software, A.Z.; validation, A.Z.; formal analysis, A.Z.; visualization, A.Z.; resources, A.Z.; data curation, A.Z.; writing—original draft preparation, A.Z.; writing—review and editing, G.F.; supervision, H.X.; project administration, H.X.; funding acquisition, H.X. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1. National Total Carbon Emissions from 1997 to 2019.

Figure 2. Distribution map of total carbon emission intensity in each province from 1997 to 2019.

Figure 3. Quantile regression result graph.

Table 1

Variable description and descriptive statistics of sample data.

Variable Name	Economic Meaning	Metrics	AVG	Std.	Min	Max
CG	Carbon Intensity: the amount of carbon dioxide emitted per unit of gross national product	Carbon emissions/Gross national product (million tons/billion yuan)	342.9253	270.0545	15.7056	1564.8340
IS	Industrial Structure	The added value of the tertiary industry/the added value of the second industry	1.1374	0.5521	0.5271	5.2340
ES	Energy Consumption Structure	Regional coal consumption/regional total energy consumption (%)	0.6853	0.2701	0.0177	1.7578
OP	Level of Opening-up	Total import and export volume of domestic destinations and sources of goods (100 million yuan)	7920.0000	17,500.0000	15.2940	128,000.0000
TE	Technological Innovation	Number of patent authorizations (pieces)	24,627.4300	53,679.7700	56.0000	527,390.0000
PL	Population	Total population (10,000 people)	4387.5790	2683.5330	495.6000	12,489.0000
UR	Urbanization level	Urban population/total population (%)	0.2707	0.1798	0.0197	0.8028
CM	Carbon Trading Pilot Policy	0~1 dummy variable			0	1

Table 2

The regional grouping of sample areas based on carbon emission intensity.

Group	Area
≤0.10 quantile group	Jiangxi, Hunan, Guangxi
0.10~0.25 quantile group	Beijing, Fujian, Shanghai, Hainan
0.25~0.50 quantile group	Zhejiang, Jiangsu, Sichuan, Chongqing, Hubei, Yunnan, Tianjin, Anhui
0.50~0.75 quantile group	Shaanxi, Hebei, Qinghai, Jilin, Liaoning, Heilongjiang, Shandong
0.75~0.90 quantile group	Gansu, Xinjiang, Guizhou, Guangdong, Henan
≥0.90 quantile group	Inner Mongolia, Ningxia, Shanxi

Table 3

Panel unit root test.

Variable	LLC	IPS	ADF-FisherChi-Square	PP-FisherChi-Square	Conclusion
D_lnCG	−18.4329 ***	−16.4297 ***	213.9499 ***	538.7272 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnIS	−13.4503 ***	−11.4533 ***	198.5181 ***	246.8631 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnES	−19.0036 ***	−16.7741 ***	250.4381 ***	529.6426 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnOP	−16.0971 ***	−14.7765 ***	263.9364 ***	452.4454 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnTE	−16.8806 ***	−14.6179 ***	161.3194 ***	410.4722 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnPL	−14.8879 ***	−12.1741 ***	223.7696 ***	327.3083 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnUR	−17.8013 ***	−18.5908 ***	320.9167 ***	631.6294 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
D_lnCM	−19.0570 ***	−15.3160 ***	147.0966 ***	407.6537 ***	smooth
	(0.0000)	(0.0000)	(0.0000)	(0.0000)

Note: *** represents a significance level of 1%.

Table 4

Cointegration test.

Testing Method	INSPECTION FORM	Statistics	p-Value
Kao	Modified Dickey–Fuller t	−5.4478	0.0000
	Dickey–Fuller t	−4.7316	0.0000
	Augmented Dickey–Fuller t	−5.2597	0.0000
Pedroni	Modified Phillips–Perron t	8.1915	0.0000
	Phillips–Perron t	−2.9919	0.0014
	Augmented Dickey–Fuller t	−4.4612	0.0000
Westerlund		−2.3487	0.0094

Table 5

Quantile regression results.

Variable	lnIS	lnES	lnOP	lnTE	lnPL	lnUR	lnCM	_cons
10%	0.2950 ***	1.2656 ***	0.2134 ***	−0.5537 ***	0.4081 ***	0.8036 ***	0.0228	−3.528 ***
	(0.0077)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.9010)	(0.0040)
25%	−0.0114 ***	1.0875 ***	0.126 ***	−0.5365 ***	0.3393 ***	0.7888 ***	0.0192 *	−1.3342 *
	(0.0091)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0850)	(0.0530)
50%	−0.2105 ***	1.0122 ***	0.136 ***	−0.5087 ***	0.1996 ***	0.4597 ***	0.0358 *	1.3491 **
	(0.0040)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0656)	(0.0110)
75%	−0.1938 ***	0.8755 ***	0.0961 ***	−0.4518 ***	0.1566 ***	0.3627 ***	0.0923 **	2.6937 ***
	(0.0090)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0258)	(0.0000)
90%	−0.1442 ***	0.8192 ***	0.0671 ***	−0.4442 ***	0.1749 ***	0.4226 ***	0.0434	2.8686 ***
	(0.0015)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.5040)	(0.0000)

Note: ***, **, and * represent significance levels of 1%, 5%, and 10%, respectively.

Table 6

Robustness check.

	Replace the Explained Variable	Replace Explanatory Variables
	(1)	(2)
	lnPC	lnCG
lnIS	−0.2668 ***
lnIR		−0.0261 ***
lnES	0.7018 ***	0.6332 ***
lnOP	0.2329 ***	0.0858 *
lnTE	−0.1707 ***	−0.2168 ***
lnPL	0.5678 ***	0.8548 ***
lnUR	0.4206 ***	0.2920 ***
lnCM	−0.0603	−0.2925 ***
_cons	3.0882 ***	−2.0662
adjust R2	0.8984	0.7372

Note: *** and * represent significance levels of 1% and 10% respectively.

Table 7

Benchmark regression results.

	Model 1	Model 2	Model 3
	lnCG	lnES	lnCG
IS	−0.3670 ***	−0.1219 ***	−0.2943 ***
	(0.0082)	(0.0000)	(0.0055)
ES			0.5971 ***
			(0.0040)
OP	−0.0975	−0.0096	−0.1032
	(0.2920)	(0.9060)	(0.1630)
TE	−0.2808 ***	−0.0677 **	−0.2404 ***
	(0.0020)	(0.0450)	(0.0070)
PL	0.6057	−0.6529	0.9955 *
	(0.3220)	(0.1910)	(0.0700)
UR	0.3087 *	0.0783	0.2620
	(0.0970)	(0.4020)	(0.1470)
CM	−0.4407 ***	−0.2915 **	−0.2666 ***
	(0.0010)	(0.0190)	(0.0090)
_cons	2.5707	4.7036 ***	−3.1802
	(0.5960)	(0.0100)	(0.5130)
R2	0.8858	0.8286	0.9002

Note: The parentheses in the table are p-values; R² is the goodness of fit of the model; ***, **, and * indicate that the coefficients have passed the significance test at the 1%, 5%, and 10% levels, respectively.

Table 8

Bootstrap mediation effect test.

Mediator Variable	Path	Effect	Effect Coefficient	p-Value
energy structure	lnIS-lnES-lnCG	indirect	−0.0728	0.0000
	lnIS-lnES-lnCG	direct	−0.2943	0.0000
	lnIS-lnCG	total effect	−0.3670

References

1. Yang, C.; Hao, Y.; Irfan, M. Energy consumption structural adjustment and carbon neutrality in the post-COVID-19 era. Struct. Chang. Econ. Dyn.; 2021; 59, pp. 442-453. [DOI: https://dx.doi.org/10.1016/j.strueco.2021.06.017]

2. Zhang, S.; Li, H.; Zhang, Q.; Tian, X.; Shi, F. Uncovering the impacts of industrial transformation on low-carbon development in the Yangtze River Delta. Resour. Conserv. Recycl.; 2019; 150, 104442. [DOI: https://dx.doi.org/10.1016/j.resconrec.2019.104442]

3. Zhou, X.; Zhang, J.; Li, J. Industrial structural transformation and carbon dioxide emissions in China. Energy Policy; 2013; 57, pp. 43-51. [DOI: https://dx.doi.org/10.1016/j.enpol.2012.07.017]

4. Wang, J.; Song, X.; Chen, K. Which Influencing Factors Cause CO₂ Emissions Differences in China’s Provincial Construction Industry: Empirical Analysis from a Quantile Regression Model. Pol. J. Environ. Stud.; 2020; 29, pp. 331-347. [DOI: https://dx.doi.org/10.15244/pjoes/105239] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36262317]

5. Wang, K.M. The relationship between carbon dioxide emissions and economic growth: Quantile panel-type analysis. Qual. Quant.; 2013; 47, pp. 1337-1366. [DOI: https://dx.doi.org/10.1007/s11135-011-9594-y]

6. Sharif, A.; Raza, S.A.; Ozturk, I.; Afshan, S. The dynamic relationship of renewable and nonrenewable energy consumption with carbon emission: A global study with the application of heterogeneous panel estimations. Renew. Energy; 2019; 133, pp. 685-691. [DOI: https://dx.doi.org/10.1016/j.renene.2018.10.052]

7. Grossman, G.M.; Krueger, A.B. Environmental Impacts of a North American Free Trade Agreement; National Bureau of Economic Research: Cambridge, MA, USA, 1991; [DOI: https://dx.doi.org/10.3386/w3914]

8. Panayotou, T. Empirical Tests and Policy Analysis of Environmental Degradation at Different Stages of Economic Development; International Labour Office: Geneva, Switzerland, 1993.

9. Dong, J.; Dou, Y.; Jiang, Q.; Zhao, J. How Does Industrial Structure Upgrading Affect the Global Greenhouse Effect? Evidence From RCEP and Non-RCEP Countries. Front. Energy Res.; 2021; 9, 683166. [DOI: https://dx.doi.org/10.3389/fenrg.2021.683166]

10. Liu, Y.; Chang, X.; Huang, C. Research and Analysis on the Influencing Factors of China’s Carbon Emissions Based on a Panel Quantile Model. Sustainability; 2022; 14, 7791. [DOI: https://dx.doi.org/10.3390/su14137791]

11. Zhao, X.; Xi, Y. Threshold Effects of Urban Population Size and Industrial Structure on CO₂ Emissions in China. Front. Environ. Sci.; 2022; 10, 492. [DOI: https://dx.doi.org/10.3389/fenvs.2022.894442]

12. Gu, R.; Li, C.; Li, D.; Yang, Y.; Gu, S. The impact of rationalization and upgrading of industrial structure on carbon emissions in the Beijing-Tianjin-Hebei urban agglomeration. Int. J. Environ. Res. Public Health; 2022; 19, 7997. [DOI: https://dx.doi.org/10.3390/ijerph19137997]

13. Han, Y.; Zhang, J.; Yuan, M. Carbon emissions and economic growth in the Yellow River Basin: Decoupling and driving factors. Front. Environ. Sci.; 2022; 10, 2466. [DOI: https://dx.doi.org/10.3389/fenvs.2022.1089517]

14. Wang, J.; Wang, C.; Yu, S.; Li, M.; Cheng, Y. Coupling coordination and spatiotemporal evolution between carbon emissions, industrial structure, and regional innovation of counties in Shandong province. Sustainability; 2022; 14, 7484. [DOI: https://dx.doi.org/10.3390/su14127484]

15. Gao, P.; Wang, Y.; Zou, Y.; Su, X.; Che, X.; Yang, X. Green technology innovation and carbon emissions nexus in China: Does industrial structure upgrading matter?. Front. Psychol.; 2022; 13, 951172. [DOI: https://dx.doi.org/10.3389/fpsyg.2022.951172]

16. Zheng, K.; Deng, H.; Lyu, K.; Yang, S.; Cao, Y. Market Integration, Industrial Structure, and Carbon Emissions: Evidence from China. Energies; 2022; 15, 9371. [DOI: https://dx.doi.org/10.3390/en15249371]

17. Brannlund, R.; Persson, L. To tax, or not to tax: Preferences for climate policy attributes. Clim. Policy; 2012; 12, pp. 704-721. [DOI: https://dx.doi.org/10.1080/14693062.2012.675732]

18. Torvanger, A. Manufacturing sector carbon dioxide emissions in nine OECD countries, 1973–1987: A Divisia index decomposition to changes in fuel mix, emission coefficients, industry structure, energy intensities and international structure. Energy Econ.; 1991; 13, pp. 168-186. [DOI: https://dx.doi.org/10.1016/0140-9883(91)90018-U]

19. Guo, H.; Xie, S.; Pan, C. The impact of planting industry structural changes on carbon emissions in the three northeast provinces of China. Int. J. Environ. Res. Public Health; 2021; 18, 705. [DOI: https://dx.doi.org/10.3390/ijerph18020705]

20. Dong, B.; Xu, Y.; Fan, X. How to achieve a win-win situation between economic growth and carbon emission reduction: Empirical evidence from the perspective of industrial structure upgrading. Environ. Sci. Pollut. Res.; 2020; 27, pp. 43829-43844. [DOI: https://dx.doi.org/10.1007/s11356-020-09883-x]

21. Zheng, Y.; Peng, J.; Xiao, J.; Su, P.; Li, S. Industrial structure transformation and provincial heterogeneity characteristics evolution of air pollution: Evidence of a threshold effect from China. Atmos. Pollut. Res.; 2020; 11, pp. 598-609. [DOI: https://dx.doi.org/10.1016/j.apr.2019.12.011]

22. Zhang, F.; Deng, X.; Phillips, F.; Fang, C.; Wang, C. Impacts of industrial structure and technical progress on carbon emission intensity: Evidence from 281 cities in China. Technol. Forecast. Soc. Chang.; 2020; 154, 119949. [DOI: https://dx.doi.org/10.1016/j.techfore.2020.119949]

23. Zeng, L.; Lu, H.; Liu, Y.; Zhou, Y.; Hu, H. Analysis of regional differences and influencing factors on China’s carbon emission efficiency in 2005–2015. Energies; 2019; 12, 3081. [DOI: https://dx.doi.org/10.3390/en12163081]

24. Kambara, T. The energy situation in China. China Q.; 1992; 131, pp. 608-636. [DOI: https://dx.doi.org/10.1017/S0305741000046312]

25. Xue, L.; Li, H.; Xu, C.; Zhao, X.; Zheng, Z.; Li, Y.; Liu, W. Impacts of industrial structure adjustment, upgrade and coordination on energy efficiency: Empirical research based on the extended STIRPAT model. Energy Strategy Rev.; 2022; 43, 100911. [DOI: https://dx.doi.org/10.1016/j.esr.2022.100911]

26. Wen, L.; Huang, L. Analysis of influencing factors of Chinese provincial carbon emissions based on projection pursuit model and Markov transfer matrix. Int. J. Clim. Chang. Strateg. Manag.; 2018; 11, pp. 406-423. [DOI: https://dx.doi.org/10.1108/IJCCSM-05-2017-0116]

27. Borghetti Soares, J.; Tiomno Tolmasquim, M. Energy efficiency and reduction of CO₂ emissions through 2015: The Brazilian cement industry. Mitig. Adapt. Strateg. Glob. Chang.; 2000; 5, pp. 297-318. [DOI: https://dx.doi.org/10.1023/A:1009621514625]

28. Siitonen, S.; Tuomaala, M.; Ahtila, P. Variables affecting energy efficiency and CO₂ emissions in the steel industry. Energy Policy; 2010; 38, pp. 2477-2485. [DOI: https://dx.doi.org/10.1016/j.enpol.2009.12.042]

29. Zhang, Y.; Yu, Z.; Zhang, J. Analysis of carbon emission performance and regional differences in China’s eight economic regions: Based on the super-efficiency SBM model and the Theil index. PLoS ONE; 2021; 16, e0250994. [DOI: https://dx.doi.org/10.1371/journal.pone.0250994] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33951072]

30. Xu, S.C.; He, Z.X.; Long, R.Y.; Chen, H. Factors that influence carbon emissions due to energy consumption based on different stages and sectors in China. J. Clean. Prod.; 2016; 115, pp. 139-148. [DOI: https://dx.doi.org/10.1016/j.jclepro.2015.11.050]

31. Hong, Y.; Can, P.; Xiaona, Y.; Ruixue, L. Does change of industrial structure affect energy consumption structure: A study based on the perspective of energy grade calculation. Energy Explor. Exploit.; 2019; 37, pp. 579-592. [DOI: https://dx.doi.org/10.1177/0144598718784032]

32. Guo, Q.; Ding, C.; Wu, Z.; Guo, B.; Xue, Y.; Li, D. The impact of digital economy and industrial structure distortion on Xinjiang’s energy intensity under the goal of “double carbon”. Front. Environ. Sci.; 2022; 10, 1036740. [DOI: https://dx.doi.org/10.3389/fenvs.2022.1036740]

33. Koenker, R.; Bassett, G., Jr. Regression quantiles. Econom. J. Econom. Soc.; 1978; 46, pp. 33-50. [DOI: https://dx.doi.org/10.2307/1913643]

34. Ehrlich, P.R.; Ehrlich, A.H. Population, Resources, Environment: Issues in Human Ecology; Freeman: San Francisco, CA, USA, 1970; Volume 15, pp. 7-41. [DOI: https://dx.doi.org/10.1007/BF02207996]

35. Tursun, H.; Li, Z.; Liu, R.; Li, Y.; Wang, X. Contribution weight of engineering technology on pollutant emission reduction based on IPAT and LMDI methods. Clean Technol. Environ. Policy; 2015; 17, pp. 225-235. [DOI: https://dx.doi.org/10.1007/s10098-014-0780-1]

36. Huo, J.; Yang, D.; Zhang, W.; Wang, F.; Wang, G.; Fu, Q. Analysis of influencing factors of CO₂ emissions in Xinjiang under the context of different policies. Environ. Sci. Policy; 2015; 45, pp. 20-29. [DOI: https://dx.doi.org/10.1016/j.envsci.2014.09.006]

37. York, R.; Rosa, E.A.; Dietz, T. STIRPAT, IPAT and ImPACT: Analytic tools for unpacking the driving forces of environmental impacts. Ecol. Econ.; 2003; 46, pp. 351-365. [DOI: https://dx.doi.org/10.1016/S0921-8009(03)00188-5]

38. Baron, R.M.; Kenny, D.A. The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J. Personal. Soc. Psychol.; 1986; 51, 1173. [DOI: https://dx.doi.org/10.1037/0022-3514.51.6.1173]

39. Wang, Q.; Wang, S.; Jiang, X. Preventing a rebound in carbon intensity post-COVID-19–lessons learned from the change in carbon intensity before and after the 2008 financial crisis. Sustain. Prod. Consum.; 2021; 27, pp. 1841-1856. [DOI: https://dx.doi.org/10.1016/j.spc.2021.04.024]

40. Chen, X.; Chen, Y.E.; Chang, C.P. The effects of environmental regulation and industrial structure on carbon dioxide emission: A non-linear investigation. Environ. Sci. Pollut. Res.; 2019; 26, pp. 30252-30267. [DOI: https://dx.doi.org/10.1007/s11356-019-06150-6]

41. Shao, C.; Wei, J.; Liu, C. Empirical analysis of the influence of green credit on the industrial structure: A case study of China. Sustainability; 2021; 13, 5997. [DOI: https://dx.doi.org/10.3390/su13115997]

42. Zhao, X.; Shang, Y.; Song, M. Industrial structure distortion and urban ecological efficiency from the perspective of green entrepreneurial ecosystems. Socio-Econ. Plan. Sci.; 2020; 72, 100757. [DOI: https://dx.doi.org/10.1016/j.seps.2019.100757]

43. Gan, C.H.; Zheng, R.G.; Yu, D.F. An empirical study on the effects of industrial structure on economic growth and fluctuations in China. Econ. Res. J.; 2011; 5, pp. 4-16.

44. Hu, Y.; Jiang, H.; Zhong, Z. Impact of green credit on industrial structure in China: Theoretical mechanism and empirical analysis. Environ. Sci. Pollut. Res.; 2020; 27, pp. 10506-10519. [DOI: https://dx.doi.org/10.1007/s11356-020-07717-4] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31940144]

45. Ren, S.; Hao, Y.; Xu, L.; Wu, H.; Ba, N. Digitalization and energy: How does internet development affect China’s energy consumption?. Energy Econ.; 2021; 98, 105220. [DOI: https://dx.doi.org/10.1016/j.eneco.2021.105220]

46. Li, Y.; Yang, X.; Ran, Q.; Wu, H.; Irfan, M.; Ahmad, M. Energy structure, digital economy, and carbon emissions: Evidence from China. Environ. Sci. Pollut. Res.; 2021; 28, pp. 64606-64629. [DOI: https://dx.doi.org/10.1007/s11356-021-15304-4] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34318413]

47. Lu, L.; Liu, Z.; Mohsin, M.; Zhang, C. Renewable energy, industrial upgradation, and import-export quality: Green finance and CO₂ emission reduction nexus. Environ. Sci. Pollut. Res.; 2022; 13, pp. 13327-13341. [DOI: https://dx.doi.org/10.1007/s11356-022-22629-1]

48. Wang, X.; Zhang, T.; Nathwani, J.; Yang, F.; Shao, Q. Environmental regulation, technology innovation, and low carbon development: Revisiting the EKC Hypothesis, Porter Hypothesis, and Jevons’ Paradox in China’s iron & steel industry. Technol. Forecast. Soc. Chang.; 2022; 176, 121471. [DOI: https://dx.doi.org/10.1016/j.techfore.2022.121471]

49. Wu, L.; Jia, X.; Gao, L.; Zhou, Y. Effects of population flow on regional carbon emissions: Evidence from China. Environ. Sci. Pollut. Res.; 2021; 28, pp. 62628-62639. [DOI: https://dx.doi.org/10.1007/s11356-021-15131-7]

50. Duan, X.; Li, X.; Tan, W.; Xiao, R. Decoupling relationship analysis between urbanization and carbon emissions in 33 African countries. Heliyon; 2022; 8, e10423. [DOI: https://dx.doi.org/10.1016/j.heliyon.2022.e10423]

51. Zhang, W.; Li, J.; Li, G.; Guo, S. Emission reduction effect and carbon market efficiency of carbon emissions trading policy in China. Energy; 2020; 196, 117117. [DOI: https://dx.doi.org/10.1016/j.energy.2020.117117]

52. Shan, Y.; Huang, Q.; Guan, D.; Hubacek, K. China CO₂ emission accounts 2016–2017. Sci. Data; 2020; 7, 54. [DOI: https://dx.doi.org/10.1038/s41597-020-0393-y]

53. Shan, Y.; Liu, J.; Liu, Z.; Xu, X.; Shao, S.; Wang, P.; Guan, D. New provincial CO₂ emission inventories in China based on apparent energy consumption data and updated emission factors. Appl. Energy; 2016; 184, pp. 742-750. [DOI: https://dx.doi.org/10.1016/j.apenergy.2016.03.073]

54. Guan, Y.; Shan, Y.; Huang, Q.; Chen, H.; Wang, D.; Hubacek, K. Assessment to China’s recent emission pattern shifts. Earths Future; 2021; 9, e2021EF002241. [DOI: https://dx.doi.org/10.1029/2021EF002241]

55. Sun, P.; Liu, L.; Qayyum, M. Energy efficiency comparison amongst service industry in Chinese provinces from the perspective of heterogeneous resource endowment: Analysis using undesirable super efficiency SBM-ML model. J. Clean. Prod.; 2021; 328, 129535. [DOI: https://dx.doi.org/10.1016/j.jclepro.2021.129535]

56. Su, Y.; Liu, X.; Ji, J.; Ma, X. Role of economic structural change in the peaking of China’s CO₂ emissions: An input–output optimization model. Sci. Total Environ.; 2021; 761, 143306. [DOI: https://dx.doi.org/10.1016/j.scitotenv.2020.143306] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33190898]

57. Dai, Z.M.; Shen, X.; Guo, L. Technological innovation on economic growth from the perspective of investment-oriented environmental regulations: Considering the threshold effect of China human capital. Appl. Econ.; 2021; 53, pp. 4632-4645. [DOI: https://dx.doi.org/10.1080/00036846.2021.1904128]

58. Liang, G.; Yu, D.; Ke, L. An empirical study on dynamic evolution of industrial structure and green economic growth—Based on data from China’s underdeveloped areas. Sustainability; 2021; 13, 8154. [DOI: https://dx.doi.org/10.3390/su13158154]