Impact of Carbon Neutrality Goals on China’s Coal Industry: Mechanisms and Evidence

Abstract

China’s coal industry is reckoned as one of the topmost contributors to global carbon emissions, and as a result, poses severe challenges both to human health and climate change mitigation efforts. Achieving carbon neutrality requires thorough analyses of mechanisms driving the coal sector’s transition. This study employs a structural model to investigate the transmission pathways through/by which the “dual carbon” goals influence the coal industry, using a policy text quantification approach to assess specific carbon reduction measures. Findings reveal that the impact of the “dual carbon” target on the coal industry operates through multiple pathways. Carbon reduction policies significantly enhance technical advancements, social and economic factors, energy-saving measures, and alternative energy development, all of which indirectly affect coal supply. Notably, the pathway from coal demand to coal supply shows a high path coefficient of 1.121, far surpassing the path coefficient from factor input to coal supply, measured at 0.169. This highlights coal demand as the pivotal intermediary variable in determining the “dual carbon” target’s impact on the coal industry. While current technologies and alternative energy sources have limited immediate effects on coal supply, they hold significant potential as transformative factors in the future.

1. Introduction

Climate change is a global issue linked to extreme weather events and significant impacts on the human environment [1]. Governments worldwide have placed a premium on carbon reduction and neutrality, to tackle the escalating threats of climate change [2,3]. Carbon peaking and carbon neutrality (hereinafter referred to as the “dual carbon” target) has become a global consensus, with over 90% of countries worldwide actively researching and formulating a series of carbon reduction policies for specific actions [4], such as rigorous development of renewable energy solutions, to promote substitution of fossil fuels [4,5,6], promoting advanced energy-saving and carbon-reducing technologies [7,8], improving energy utilization efficiency, by demonstrating and applying carbon capture and storage (CCS) technologies [9,10].

The energy sector is a key focus area to achieving the set “dual-carbon” goal with every conscious policy that may be conceived [11,12]. By 2023, global consumption of fossil fuels such as coal, oil, and natural gas is projected to generate 36.8 billion tons of carbon dioxide, marking a 1.1% increase, if compared to 2022. This accounts for approximately 90% of global carbon emissions, with coal consumption alone representing over 40% [13]. In recent years, the trend in global coal consumption has continually grown. In 2023, global coal consumption hit a record high, exceeding 8.5 billion tons, by an unprecedented 1.4% increase from 2022. China, India, and Southeast Asia accounted for 75% of global coal consumption, while China remarkably contributed 56% to global coal consumption [14]. Even as an accelerated low-carbon energy transition is expected to either cause stability or a slight decrease in global coal consumption, the status of coal in the global energy structure is not expected to fundamentally change in the short term, especially in industrialized countries on the Asian continent, particularly China and India, where coal consumption is projected to increasingly grow [15,16].

As the world’s largest producer and consumer of coal, coal phaseout is critical for China to achieve carbon neutrality and contribute to global decarbonization [17,18]. China’s dominance in global coal consumption, its status as the largest carbon emitter, and its ambitious carbon neutrality goals make it a crucial case study for understanding the impact of the ‘dual carbon’ target on the coal industry. The country’s policy decisions and transition strategies not only shape its own energy future but also influence global climate governance and serve as a model for other coal-dependent economies. Coal is the main energy source for China; thus, it has been dominant in the energy consumption structure for a long period. In 2023, China’s total energy consumption is 5.72 billion tCO₂e, an increase of 5.7% year-on-year, with coal accounting for 55.3% of the energy consumption structure. China’s coal development and utilization account for ~60–70% of total carbon emissions, with over 90% of carbon emissions related to coal utilization and less than 10% related to coal development [19,20]. Although carbon emissions during coal development process are relatively lower, emissions of methane released during coal mine development have attracted widespread attention due to their significant warming potential [21]. Wang et al. opined that effective control of methane emissions from coal mines can be used to achieve triple benefits in safety, energy, and environment, which can help achieve carbon neutrality goals [22]. Liu et al. [23] explored strategies for enhancing methane emission reductions during the gradual phase-out of coal which emphasize the importance of prioritizing the closure of small-scale coal mines characterized by a high gas content that is relative to high operational costs and increased health burdens.

The achievement of the “dual carbon” target will bring uncertainty to China’s future energy security [24], and phasing out coal completely might have serious economic consequences [25]. Balancing the reduction in coal consumption to promote environmental sustainability and achieve better economic growth [17]. Limited qualitative studies exist on policy transmission mechanisms in other industries [7,26,27]. Hence, there is an urgent need to study the transmission mechanism of the impact of the “dual carbon” target on the coal industry and provide a reference basis for relevant policy formulation. As the world’s second largest economy and the largest carbon emitter, China’s emission reduction actions play a demonstrative and leading role in global climate governance. Hence, this study analyzes the impact of the “dual carbon” target on the Chinese coal industry, constructs a structural equation model of the impact of carbon reduction policies on the coal industry, identifies the main pathways through which the “dual carbon” target influence the coal industry, and reveals the transmission mechanism of the impact of carbon reduction policies on the coal industry, by providing research references for other coal-dependent countries or regions worldwide. Previous studies have primarily focused on achieving carbon neutrality across various stages of coal development and utilization, as well as proposing implementation pathways. However, there is a lack of qualitative research examining policy transmission mechanisms in other industries [7,27,28,29]. This study comprehensively analyzes the main pathways in which the “dual carbon” target affects the coal industry, constructs a transmission mechanism model of the impact of the “dual carbon” target on the Chinese coal industry from a policy transmission perspective, and verifies the reliability of the model. The impetus of this lies in the quantitative identification of the main pathways through which the “dual carbon” target affects the coal industry, filling the research gap on the basic logic and impact of the transmission mechanism of the “dual carbon” target on the coal industry. This study also provides support for policy measures and recommendations for the coal industry in other countries and regions.

2. Methodology

2.1. Model Architecture

There are many influencing factors affecting coal supply and demand, and many studies chose one or more influencing factors to conduct quantitative research based on different research purposes [30,31]. However, variables selected by different influencing factors are characterized by serious collinearity and complex causality. Conventional statistical methods such as multiple regression are limited to the numerical analysis of the relationship between each factor and coal’s supply and demand [32]. It is impossible to analyze the interaction between cofactors and the process and path of transmission to coal demand and supply. The “dual carbon” target will eventually be translated into a series of carbon reduction policies, collectively known as “carbon reduction” factors. On the basis of systematic analysis of the impact mechanism of carbon reduction policy transmission on coal supply and demand, it was revealed that “dual carbon” target mainly impacts the coal industry as follows: (1) by increasing investment in research and development, setting macroeconomic goals for social and economic development, energy conservation and carbon reduction goals and other policies to guide the capital input, manpower and other factors to achieve the adjustment of coal supply; (2) through policy regulation and control, to guide the iteration of technological progress, industrial structure adjustment, energy consumption patterns and structural changes, to achieve total coal demand and structural optimization, and ultimately affect coal supply [33]. Carbon reduction policy refers to a series of policies related to carbon reduction issued for the development of the coal industry on technical factors, social and economic factors, energy conservation and emission reduction, alternative energy, factor input and coal demand through the regulation of policy intensity to achieve the regulation of coal supply and demand [33].

In this paper, six factors, including “carbon reduction” policy factor, socio-economic factor, energy-saving “carbon reduction” factor, alternative energy factor, coal demand factor and factor input factor, are selected as latent variables to build a structural equation theoretical model of the impact of “dual carbon” target on the coal industry from the perspective of policy transmission. To verify the transmission path and main factors of the impact of carbon reduction policy on coal supply and demand. The logical relationship of different potential variables of the structural equation and theoretical model of the transmission mechanism of the coal industry by “carbon reduction” policy is shown in Figure 1.

As shown in Figure 1, “carbon reduction” policy refers to a series of policies related to carbon emission reduction, such as technical factors, social and economic factors, energy conservation and emission reduction, alternative energy, factor input and coal demand, which are introduced for the development of the coal industry [33]. The regulation of coal supply can be realized through regulation of policy intensity.

In summary, the structural equation model established in this paper selects eight second-order latent variables such as carbon reduction policy, technical factor, social and economic factor, energy-saving and emission reduction factor, factor input, coal demand and coal supply, and three first-order latent variables such as oil supply, natural gas supply, and alternative energy to study the relationships between carbon reduction policy and coal supply and different latent variables. Among them, carbon reduction policies are exogenous latent variables, while technical factors, social economy, energy saving and carbon reduction, alternative energy, oil supply, natural gas supply, factor input, coal demand, and coal supply are endogenous latent variables.

2.2. Model Indicator Selection and Parameter Values

The number of policy items was selected as the observed variables of the carbon reduction policy factors. Coal consumption per kilowatt-hour, coal consumption per ton of steel, coal consumption per ton of cement and energy consumption per ton of ethylene are selected as the observed variables of the technological progress level of electric power, steel, building materials and chemical industry. The above indicators can reflect the coal consumption demand of the industry and have an important impact on the coal demand. Gross domestic product (GDP) and the proportion of urban population were selected as the observed variables of socio-economic factors. Carbon emission intensity of coal development, energy intensity, carbon emission intensity per unit GDP and SO₂ emission were selected as the observed variables of the energy conservation and emission reduction. Oil, natural gas and renewable energy supplies were selected as observed variables of the alternative energy factors. For oil supply, domestic oil supply and oil import volume were selected as observed variables, and for natural gas supply, domestic gas supply and gas import volume were selected as observed variables. The amount of coal resources, the number of employees in coal mining and selection industry and the comprehensive energy consumption of raw coal were selected as the observed variables of input factors. Electricity, steel, building materials and chemical consumptions from various industries were selected as the coal demand factors’ observed variables. Coal production and import volume were selected as the coal supply factors observed variables. The corresponding observed variables were selected for different latent variables, as shown Figure 2 and in

Table 1. Definition of coal supply and demand factors and measurement indicators.

Second Order Latent Variable	Short for Second-Order Latent Variable	First Order Latent Variable	Short for First-Order Latent Variable	Observed Variables	Units	Observation Index
Carbon reduction policy	CRP	/		Technology	/	A1
				Economy	/	A2
				Energy conservation	/	A3
				Population	/	A4
				Alternative energy	/	A5
Technical factor	TF	/		Coal consumption for power supply	gce/kW·h	B1
				Coal consumption of tons of steel	gce/ton	B2
				Coal consumption of tons of cement	tCO2e/ton	B3
				Coal consumption of ethylene per product	tCO2e/ton	B4
Social economy	SCE	/		GDP	Trillion dollars	C1
				Proportion of urban population	/	C2
Energy saving and carbon reduction	ESCR	/		Carbon intensity of coal development	kgCO2/ton	D1
				Energy intensity	ton/104 dollars	D2
				Carbon emission intensity per unit of GDP	tCO2/104 dollars	D3
				SO2 emission	104 ton	D4
Alternative energy	ALE	Oil supply	ALE_OS	Domestic oil supply	108 ton	F1
				Oil import volume	108 ton	F2
		Natural gas supply	ALE_GS	Domestic gas supply	108 m3	F3
				Natural gas imports	108 m3	F4
		/		Renewable energy supply	108 tCO2e	F5
Factor input	FI	/		Coal resources	108 ton	G1
				Number of employees	104 people	G2
				Energy consumption of raw coal production	kgce/ton	G3
Coal demand	CD	/		Coal consumption in power industry	108 ton	H1
				Coal consumption in the steel industry	108 ton	H2
				Coal consumption in building materials industry	108 ton	H3
				Coal consumption in Chemical industry	108 ton	H4
Coal Supply	CS	/		Coal production	108 ton	I1
				Coal import	108 ton	I2

.

2.3. Model Data Processing Methods

This paper uses policy text analysis to quantify the intensity of coal-related carbon reduction policies. The methods of policy text analysis can be divided into two categories: (1) qualitative analysis of policy text, and (2) put forward policy quantification standards and quantify policies. Policy intensity is an indicator to describe the effectiveness of a policy, which is usually determined by the subject and type of policy [34,35]. As a scientific method to quantify the effect and intensity of policy, policy text analysis has been widely used in different industries [36,37,38]. Below are the steps followed to model data for policy intensity.

(1)

The number of articles on regulations promulgated by different institutions was obtained by searching keywords of different dimensions in Peking University Talisman database [36].

(2)

According to the level and type of effectiveness of the policy and regulation promulgation agency, the intensity of each policy can be assigned 1 to 5 points. The larger the score, the higher the effectiveness, as shown in

Table 2. Grading table for different policies and regulations.

Category	Merit Points	Scoring Standard	Issuing Authority
Policy intensity rating	5	Administrative regulations	The State Council
	4	Departmental regulations	Regulations and normative documents of ministries and departments
	3	Inner-party regulations	Departments and institutions of the Communist Party of China Central Committee
	2	Group regulation	Other institutions
	1	Industry regulation

.

(3)

After assigning points to various policies and regulations, Equation (1) is used to calculate the policy intensity (PI) of the policies and regulations implemented each year, and the policy intensity (API) value of the policies and regulations in previous years is summed up.

$$AP{I}_t={{\displaystyle \sum }}_{i=1}^{N}P{I}_{ti}={{\displaystyle \sum }}_{i=1}^N{{\displaystyle \sum }}_{j=1}^{M}NP{R}_{tij}\times {\delta }_{ij}$$

(1)

where $AP{I}_t$ represents the annual policy intensity score of the policies and regulations promulgated and implemented in year $t$; $P{I}_{ti}$ represents the policy intensity of policy category $i$ in year $t$; $NP{R}_{tij}$ represents the number of policies issued and implemented by Category $j$ institutions in category $i$ in year $t$; ${\delta }_{ij}$ represents the assignment of the $i$ policy category to be promulgated and implemented by the $j$ institutions as shown in Table 1; $N$ represents the number of policy categories; $M$ represents the number of institutions that issue and implement different policies; $t$ indicates the year.

(4)

Standardization policy intensity (SPI) refers to the proportion of the different types of $PI$ in the total policy intensity value. According to Equation (2), the SPI in different categories is calculated, respectively, according to the $\mathit{API}$.

$$SP{I}_{ti}=P{I}_{ti}/AP{I}_t$$

(2)

$SP{I}_{ti}$ represents the SPI corresponding to policy category $i$ in year $t$.

The policy text method is used to determine the PI of technical factors, economic factors, population factors, energy-saving and carbon reduction factors and alternative energy factors. Technical factors, social and economic factors, alternative energy factors and factor input, coal supply and other corresponding observation indicators are mainly sorted through the National Energy Statistical Yearbook, the National Statistical Yearbook, the Annual Report of Coal Industry Development, the National Mineral Resources Report and other statistical documents. Observation indicators such as carbon emission intensity of coal development, energy intensity, carbon emission intensity of coal per unit GDP, comprehensive energy consumption of raw coal production is mainly derived through further calculations. The consumption data of power, iron and steel, building materials and chemicals used by the coal industry are obtained from China Coal Market website [10].

Time series data and information of relevant observation indicators were also collected from 2005 to 2023. However, there was a large gap between dimensions of different observation variables, such as items, percentages, billion yuan, billion tons, million people, etc. At the same time, the data of observation variables showed similar trends in time change, and there were multiple nonlinear relationships that made the data impossible to directly analyze. Therefore, it became imperative to carry out dimensionless processing of observation index data before simulation of structural equation model [39,40]. The data standardization processing can be directly standardized for the forward indicators, and it is necessary to make the backward indicators forward via reciprocity before standardizing the processing [41]. The commonly used data dimensionless methods include the vector gauge method, range transformation method, interval number transformation method, and logarithmic transformation method [42].

(1)

Vector gauge method

The vector norm method changes the index value between (0,1). Because of its simple calculation method and convenient operation, it is the most used method after testing. The transformation equation for the benefit and cost data is as follows:

$$x_{ij}^{\prime }={\displaystyle \frac{x_{ij}}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{x_{ij}}^2}}}, x_{ij}^{\prime }={\displaystyle \frac{\frac{1}{x_{ij}}}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(\frac{1}{x_{ij}}\right)}^2}}}$$

(3)

For convenience, the normalized data $x_{ij}^{\prime }$ is still referred to as $x_{ij}$.

(2)

Logarithmic conversion method

$$x_{ij}^{\prime }=\log \left(x_{ij}\right) or \ln \left(x_{ij}\right)$$

(4)

For convenience, the logged data, $x_{ij}^{\prime }$ is still referred to as $x_{ij}$.

(3)

Range transformation method

The best index value is changed to 1, the worst index value is changed to 0, and the transformation formula for the benefit and cost data is as follows:

$$x_{ij}^{\prime }={\displaystyle \frac{x_{ij}-\underset{i}{min}{x}_{ij}}{\underset{i}{max}{x}_{ij}-\underset{i}{min}{x}_{ij}}}, x_{ij}^{\prime }={\displaystyle \frac{\underset{i}{max}{x}_{ij}-x_{ij}}{\underset{i}{max}{x}_{ij}-\underset{i}{min}{x}_{ij}}}$$

(5)

The closer the value after range transformation is to 1, the better the index value is. The closer it is to 0, the worse the index value.

(4)

Interval number transformation method

Let the original data be $x_{ij}=\left[{\underset{\_}{x}}_{ij}, {\stackrel{\_}{x}}_{ij}\right]$, and the transformed data be $x_{ij}^{\prime }=\left[{\underset{\_}{x}}_{ij}^{\prime }, {\overline{x}}_{ij}^{\prime }\right]$. The methods for standardizing interval index values can be divided into the following two categories:

When $j\in$ is benefit type,

$${\underset{\_}{x}}_{ij}^{\prime }={\displaystyle \frac{{\underset{\_}{x}}_{ij}}{\sqrt{{{\displaystyle \sum }}_{i=1}^n({\stackrel{\_}{x}}_{ij}{)}^2}}}, {\overline{x}}_{ij}^{\prime }={\displaystyle \frac{{\stackrel{\_}{x}}_{ij}}{\sqrt{{{\displaystyle \sum }}_{i=1}^n({\underset{\_}{x}}_{ij}{)}^2}}}$$

(6)

When $j\in$ is cost type,

$${\underset{\_}{x}}_{ij}^{\prime }={\displaystyle \frac{1}{{\stackrel{\_}{x}}_{ij}\times \sqrt{{{\displaystyle \sum }}_{i=1}^n(\frac{1}{{\underset{\_}{x}}_{ij}}{)}^2}}}, {\overline{x}}_{ij}^{\prime }={\displaystyle \frac{1}{{\underset{\_}{x}}_{ij}\times \sqrt{{{\displaystyle \sum }}_{i=1}^n(\frac{1}{{\stackrel{\_}{x}}_{ij}}{)}^2}}}$$

(7)

The magnitude of degree of the three standardization methods is compared and analyzed. The vector normalization method has a great advantage among the three methods. In addition, compared with the range transformation method, the vector normalization method does not change the data distribution.

By comparing different methods, this paper finally determined the basic steps of data preprocessing as follows: (1) all data were processed according to Equation (4) and logarithmic processing was carried out to eliminate the heteroscedasticity of time series data and improve the stability of time series data; (2) data after logarithmic processing were taken and forward and reverse processing were carried out according to Equation (5) to achieve the purpose of dimensionless processing. The data used by the model in this paper is the data file after completing the above two steps of preprocessing based on the original data file.

2.4. Model Tool Selection

Structural equation modeling tools based on covariance matrix, such as AMOS and Mplus, are known for shortcomings such as (1) non-positive definite matrix with coefficient greater than 1 or 0 cannot be identified by the model; (2) the model cannot operate normally due to the large or small sample size of observation indicators; (3) the observation index variables are seriously collinearity, resulting in unacceptable model results and serious deviation in parameter estimation; (4) serious non-normal distribution of measurement data leads to problems in parameter estimation; (5) the function of processing the latent variable fraction of the formative measurement model is insufficient. SmartPLS (http://www.smartpls.com, 10 February 2022) has strong compatibility and can solve the above problems, especially for the measurement model formed by non-normal distribution data and small sample data, and the processing of a single measurement indicator has a unique advantage. Through comprehensive comparison, this paper uses SmartPls software for numerical analysis and calculation.

2.5. Methodological Limitations

While the method provides a comprehensive analysis of how the “dual carbon” target influences China’s coal industry, several limitations should be acknowledged:

  i.

Omitted Variables: The model does not account for all potential factors that may influence coal demand and supply, such as political stability, global energy market fluctuations, and behavioral responses to policy changes. These omitted variables could introduce bias into the findings. Additional variables such as geopolitical risks, global energy price trends, and behavioral responses to policy changes in future studies to provide a more nuanced understanding of the transition dynamics.

 ii.

Data Constraints: The study relies on available statistical data, which may not fully capture informal coal production, regional variations, or emerging trends in energy consumption. Alternative data sources like satellite imagery and remote sensing to estimate informal coal production and regional energy consumption patterns can be leveraged to identify localized trend and challenges.

iii.

Model Assumptions: The structural equation model assumes linear relationships between variables, which may not always reflect the complex, nonlinear nature of energy transitions. Scenario analysis could also be used to explore how different policy pathways and technological developments interact in nonlinear ways and nonlinear modeling techniques like machine learning algorithms can be employed.

iv.

Policy Implementation Variability: The effectiveness of carbon reduction policies can vary based on enforcement, regional adaptations, and technological feasibility; factors not explicitly modeled in this study. Agent-based policy simulation models could be used to explore how enforcement mechanisms and regional adaptations affect policy outcomes.

3. Research Hypothesis and Model Test

3.1. Research Hypothesis

There are many factors affecting coal supply. This paper studies the transmission path and main factors of carbon reduction policy’s impact on coal supply and demand from the perspective of policy transmission, that is, how the carbon reduction policy formulated and implemented by the government affects factors such as technology, social economy, energy conservation and emission reduction, alternative energy, and so on. Influences on factor input and coal demand of the coal industry were also evaluated, to determine their effects on coal supply. Based on the theoretical analysis of the influencing factors of coal supply and demand in China, the following hypotheses are proposed for the influencing relationships between the above variables, as shown in

Table 3. Model assumptions regarding the influence of relationships between different variables.

Number	Assumed Content	Number	Assumed Content
H1	Carbon reduction policy has a significant positive impact on technology.	H11	Alternative energy has a significant positive impact on coal demand.
H2	Carbon reduction policy has a significant positive impact on social economy.	H12	Factor input has a significant positive effect on coal supply.
H3	Carbon reduction policy has a significant positive impact on energy conservation and carbon reduction.	H13	Coal demand has a significant positive impact on coal supply.
H4	Carbon reduction policy has a significant positive impact on alternative energy.	H14	Carbon reduction policy has a significant positive effect on factor input.
H5	Technology has a significant positive effect on factor input.	H15	Carbon reduction policy has a significant positive impact on coal demand.
H6	Technology has a significant negative impact on coal demand.	H16	Carbon reduction policy has a significant positive impact on coal supply.
H7	Social economy has significant positive influence on factor input.	H17	Technology has a significant negative impact on coal supply.
H8	Social economy has a significant positive impact on coal demand.	H18	Social economy has significant positive influence on coal supply.
H9	Energy saving and carbon reduction have significant negative effects on factor input.	H19	Energy saving and carbon reduction have a significant negative impact on coal supply.
H10	Energy conservation and carbon reduction have a significant negative impact on coal demand.	H20	Alternative energy has a significant positive impact on coal supply.

.

3.2. Reliability Test

Reliability test refers to the consistency, stability, and reliability of the measurement results obtained when the same method is used to measure similar object repeatedly [42]. Before analyzing the sample data of the model, it is necessary to test for the reliability of the obtained data. Only when the reliability is accepted are the data analysis results of the model credible [41]. To avoid the problem of reliability underestimation due to errors, this study measures the reliability of model samples based on the combined reliability (CR) index which belongs to the same model and has more accurate test results. When CR > 0.7, it means that the combined reliability of the model sample data is good [43]. The software SmartPls was used to calculate and check the reliability and aggregate validity of the model data, and the calculation results are shown in

Table 4. Latent variable combination reliability value with reflective measurement index.

Latent Variable Name	Reliability Test	Validity Test
	Combined Reliability (CR)	Average Variance Extracted (AVE)
CRP	0.923	0.717
SCE	0.999	0.997
ESCR	0.989	0.958
ALE	1.000	1.000
ALE_OS	0.945	0.851
ALE_GS	0.976	0.910
FI	0.989	0.979
CS	0.923	0.717

.

It can be seen from Table 4 that all latent variables with reflective measurement indicators in the model meet the requirement of combination reliability (CR) greater than 0.7. That is, the model has preliminary reliability and can be used for further analysis [44].

3.3. Validity Test

Validity test refers to the degree of consistency between the relationship between variables reflected in the actual measurement data obtained by measurement tools or means and the relationship between variables assumed by oneself. In the validity analysis, the structure validity of the evaluation is usually reflected by the aggregate validity index and the discriminative validity index [45].

3.3.1. Aggregation Validity

Aggregate validity refers to the degree of agreement between different measures of the same latent variable [46]. The aggregate validity of the model can be evaluated by averaging the percentage of variance extracted for each factor (the average extracted variance, AVE). The larger the value of AVE, the stronger the coupling between the measurement index of the latent variable and its corresponding latent variable. When AVE > 0.5 is satisfied, the model is said to have aggregation validity [47]. The AVE of latent variables with reflective measurement indicators can be calculated using the model to test for aggregation validity, as shown in Table 4.

3.3.2. Discriminative Validity

Discriminative validity refers to the degree of difference between different measurement indicators of different latent variables. The discriminative validity of the model can be evaluated by Fornell matrix [48]. Fornell matrix is a matrix composed of the square root of AVE and the coefficient correlation between the latent variables. In the Fornell matrix, if the square root of AVE is greater than all other data values in the column, it indicates that the latent variable is different from other latent variables, that is, the latent variable has good discriminative validity. Otherwise, it indicates that there is no obvious difference between the latent variable and other latent variables, that is, the latent variable does not have discriminative validity [49]. The software SmartPls was used to calculate the Fornell matrix of the test model, and the calculation results are shown in

Table 5. Fornell matrix with reflective measures.

Category	ESCR	TF	ALE	CS	CD	SCE	FI	JTCRP
ESCR	0.979
TF	0.960	0.996
ALE	0.975	0.986	1.000
CS	0.769	0.908	0.862	0.989
CD	0.840	0.953	0.912	0.984	0.954
SCE	0.978	0.996	0.992	0.880	0.932	0.999
FI	−0.965	−0.936	−0.926	−0.740	−0.811	−0.944	0.923
JTCRP	0.936	0.976	0.976	0.903	0.947	0.976	−0.878	0.847

.

As described in Table 5, all latent variables with reflective measurement indicators have discriminative validity. In other words, the model passes the aggregate validity test and the differential validity test at the same time, and the model has preliminary validity.

3.4. Goodness of Fit Test

The value of the latent variable of structural equation model is determined by the weight in essence, and the fitting effect of the model can be evaluated by the same method as the regression model [50]. Therefore, the goodness of fit ($R^2$) index can be used to evaluate the fit degree of the model when evaluating the model. The value of $R^2$ is between 0 and 1. The closer the value is to 1, the better the model fitting effect is, or otherwise. Generally, $R^2$ > 0.7 is required, indicating a good degree of model fitting, which can be used for subsequent analysis. The $R^2$ of the test model was calculated based on SmartPls software, and the results are shown in

Table 6. Goodness of fit for endogenous latent variables.

Latent Variable Definition	Latent Variable Name	${\mathit{R}}^2$	$\mathbf{Adjusted} {\mathit{R}}^2$
Technical factor	TF	0.981	0.980
Social economy	SCE	0.951	0.947
Energy saving and carbon reduction	ESCR	0.928	0.923
Alternative energy	ALE	0.982	0.981
Factor input	FI	0.974	0.962
Coal supply	CS	0.978	0.975
Coal demand	CD	0.971	0.958

.

It can be seen from Table 6 that the $R^2$ of all endogenous latent variables is greater than 0.9, that is, the indicators selected by the model are reasonable and the model has a good fitting effect. In summary, through testing the reliability, validity, and goodness of fit of the model, the established model of the transmission mechanism of the impact of carbon reduction policies on the coal industry has reliability and validity, and the model has a good fitting effect, which is statistically acceptable and can be used for further studies.

4. Empirical Results and Analysis of the Model

Through path coefficient and load coefficient, the structural equation model can accurately reveal the structural relationship. The direct influence from cause variable (endogenous latent variable or exogenous latent variable) to result variable is usually called direct effect, which is measured by the path coefficient from cause variable to result variable [51]. The Bootstrap method was first proposed by Efron [52]. In essence, it is a method of repeated sampling; that is, the original sample is regarded as the sampling population, where many Bootstrap samples are extracted through repeated random sampling with replacement, and the required statistics are obtained through calculation [53].

The Bootstrap method was applied to add auxiliary variables (to solve the problem of incomplete analysis), and SmartPls software was applied to automatically calculate and give output of the estimated value of the path coefficient and intermediary effect among each latent variable, the corresponding standard error, $t$-value and $p$-value, to evaluate the significance of each path and intermediary effect. Given this, the significance of the relationship between each variable is determined ($p$ < 0.05, significant).

4.1. Path Coefficient Analysis

In this study, both direct and indirect effects were analyzed to understand the transmission mechanisms of the “dual carbon” target on China’s coal industry. Figure 3 presents the overall direct and indirect effects, showing the strength and significance of relationships, but it does not specify through which intermediary variables these effects occur. Figure 4 addresses this by breaking down the indirect effects into specific transmission pathways, clarifying whether carbon reduction policies influence coal demand and supply via socio-economic factors, energy conservation measures, or technological advancements.

The model operation results and reliability results, different path coefficients, and significance are shown in Figure 3.

According to Figure 3, the significance under different paths can be determined, and the following conclusions can be drawn:

➀

The path coefficients of carbon reduction policy factors on technical factors, social economic factors, energy-saving and emission reduction factors and alternative energy factors are 0.976, 0.976, 0.936, and 0.976, respectively, indicating that carbon reduction policy factors have significant positive effects on technical factors, social economic factors, energy-saving and carbon reduction factors and alternative energy factors, assuming H1 to H4 are all true. This aligns with previous research, such as [7], which highlighted the role of policy inducement in driving energy efficiency improvements in China. Similarly, Jia et al. [6] emphasized the importance of policy support in promoting renewable energy and reducing coal dependency. In addition, the effect of carbon reduction policy factors on technical factors, socio-economic factors, and alternative energy factors is more obvious than that of energy conservation and emission reduction factors. The findings of this study align with the study on biomass co-firing in Jiangsu, China as it is supported by appropriate carbon reduction policies, can enhance technical feasibility [54].

➁

The path coefficients of technical factors and energy-saving and carbon reduction factors on factor input are −3.206 and −2.030, respectively, which have significant negative effects. This suggests that technical factors have the greatest negative effects on factor input, if H5 and H9 are assumed to be valid. The path coefficient of social and economic factors on factor input is 4.235, which has a significant positive impact, if hypothesis H7 is also assumed to be valid. The influence of economic and social factors on factor input includes both the input of fixed assets and the number of employees. Therefore, compared with technical factors and energy-saving and carbon reduction factors, the path coefficient of social and economic factors on factor input is greater. This finding suggests that as technology advances, the need for traditional inputs (like labor and capital) in the coal industry decreases. This finding is consistent with Liu et al. [23], who argued that technological advancements in coal mining could reduce the need for labor-intensive practices, especially in small-scale mines.

➂

Technical factors and alternative energy factors have no significant impact on coal demand, when H6 and H11 are assumed not valid. Although the influence of technical factors and alternative energy factors on coal demand is not significant, they both have negative effects. The possible explanation is that the stronger the technical factors, the higher the production efficiency, and the corresponding coal demand is reduced. Therefore, the influence of technical factors on coal demand is negative, which infers that the influence is not significant because there are more factors affecting coal demand; hence, the negative effect of technical factors is relatively weak. The path coefficient of energy-saving and carbon reduction factors on coal demand is −1.739, and has a significant negative impact, assuming that H10 is valid. The better the effect of energy saving and carbon reduction, the lower the demand for coal under the same conditions. In recent years, China has adopted strict energy conservation and emission reduction policies to accelerate the reduction in energy intensity, so that energy-saving and carbon reduction factors have a significant negative effect on coal demand and production capacity, and further verify the significant effect of national policies. Social and economic factors have a significant positive impact on coal demand, and the path coefficient is 3.373, if hypothesis H8 is also assumed valid. This reflects China’s coal-dependent energy structure, where rapid economic growth drives coal consumption. China’s coal-based energy structure suggests that the rapid economic and social development cannot be separated from the support of coal. This is in line with previous studies which opined that economic growth in China is still heavily reliant on coal, despite efforts to transition to cleaner energy sources [55,56]. The simulation further verifies the significant positive effect of social and economic factors on coal demand.

➃

The path coefficient of factor input on coal supply is 0.169, which is a positive effect, but the effect is not significant, that is, the hypothesis H12 is not valid. The path coefficient of coal demand to coal supply is 1.121, and has a significant positive impact, when hypothesis H13 is assumed valid. China’s coal supply is mainly demand-driven, with rapid economic growth promoting the rise in demand, driving up coal prices, and stimulating coal production enterprises to increase supply. Coal supply in China is primarily driven by demand, which is consistent with previous studies which established that coal production in China is highly responsive to market demand and price fluctuations [54,57].

4.2. Total Indirect Effect Analysis

Effects that may influence the outcome variable by affecting one or more mediating variables are often referred to as indirect effects [9]. When there is only one intermediate variable, the indirect effect strength is the product of two path coefficients. When there are multiple mediating variables, the indirect effect strength is equal to the product of all direct effect path coefficients. The corresponding hypothesis was verified by total indirect effect analysis, and the results are shown in Figure 3.

➀

The total indirect effect of carbon reduction policy factors on factor input is −0.893, which has a significant negative effect. In other words, assuming that H14 is not valid, it indicates that policies can regulate coal supply through influencing factor input. The total indirect effects of policy factors on coal demand and coal supply are 0.939 and 0.901, respectively, both of which have significant positive effects, when H15 and H16 are both assumed to be valid. This finding buttresses the previous suggestion that carbon reduction policies could lead to a decline in coal production capacity by reducing the economic viability of coal mining [9,42].

➁

The total indirect effect of technical factors and alternative energy factors on coal supply is −0.898 and −0.478, respectively, both of which are negative, but the influence is not significant, that is, the hypothesis H17 and H20 are not valid. This finding aligns with He et al. [18], who noted that China’s economic growth continues to rely on coal, despite efforts to transition to cleaner energy sources. The stronger the technical factor or alternative energy factor, the weaker the coal demand, the weaker the coal supply demand. Therefore, the technical factor and alternative energy factor have a negative and insignificant impact on the coal supply. Another reason for the insignificant negative impact of alternative energy factors on coal supply is that, although new energy, represented by renewable energy, has been developing rapidly, its proportion in the energy consumption structure is still small [58]. Under the target of “dual carbon”, with the rapid development of new energy, the proportion of energy that can replace coal in the energy consumption structure will increase rapidly, and the impact on coal supply will become negative and significant [59,60].

➂

The total indirect effect of socio-economic factors on coal supply is 4.496, which has a significant positive impact, that is, hypothesis H18 is valid. The total indirect effect of energy-saving and carbon reduction factors on coal supply is −2.292, which has a significant negative impact, that is, the hypothesis H19 is valid. This value indicates that these measures effectively reduce coal demand and, consequently, coal supply. This is consistent with Wang et al. [22], who found that energy efficiency improvements and stricter emission standards could significantly reduce coal consumption in China. Social and economic factors have a positive and significant impact on coal supply by regulating capital, manpower, and other factors. The energy-saving and carbon reduction factors mainly affect the coal supply by regulating the energy consumption requirements of coal production technology. The stronger the energy-saving and carbon reduction factors are, the more restricted the use of high-energy coal production technology and thus have an impact on coal supply [60,61,62].

4.3. Specific Indirect Effect Analysis

Through specific indirect effect analysis, the internal mechanism of each path is further analyzed and studied, to clarify the effect of the influencing factors of the multivariable path. The results are shown in Figure 4.

➀

Carbon reduction policy factors have a significant negative impact on factor input through technical factors and energy-saving and emission reduction factors, respectively, with corresponding specific indirect effects of −3.127 and −1.901. They further maintain a negative impact on coal supply, with corresponding specific indirect effects of −0.529 and −0.321, but the impact is not significant. Carbon reduction policy factors have a significant positive impact on factor input through social and economic factors, and the corresponding specific indirect effect is 4.134. This is consistent with previous findings that argued that energy efficiency measures and renewable energy development could reduce coal demand in China [16,63]. Furthermore, the impact on coal supply also maintains a positive effect, and the corresponding specific indirect effect is 0.699, but the impact effect is not significant. Carbon reduction policy factors are transmitted to factor input through technical factors, social and economic factors, energy-saving and carbon reduction factors, and the reason why the impact on coal supply is not significant may be caused by the superposition of positive and negative effects of various factors [63].

➁

Carbon reduction policy factors have a significant negative impact on coal demand through energy conservation and carbon reduction factors, and the corresponding specific indirect effect is −1.628. Meanwhile, it also has a significant negative impact on coal supply, and the corresponding specific indirect effect is −1.825. Carbon reduction policy factors have a significant positive impact on coal demand through social and economic factors, and the corresponding specific indirect effect is 3.293. At the same time, it also has a significant positive impact on coal supply, and the corresponding specific indirect effect is 3.691. This suggests that, while alternative energy sources are growing, their impact on coal supply is still limited. This finding aligns with previous research that opined that the share of renewable energy in China’s energy mix is still relatively small, and its impact on coal consumption is not yet fully realized [58,64].

➂

The specific indirect effect of carbon reduction policy factors on coal demand through technical factors and alternative energy factors is −0.310 and −0.416, respectively, and the impact effect is not significant. Further, the specific indirect effect on coal supply is −0.347 and −0.467, respectively, and the impact effect is also not significant, but they all maintain a negative impact. The study’s findings on the strong influence of carbon reduction policies on technical and socio-economic factors are consistent with a previous study that highlighted the role of policy in driving technological innovation and economic restructuring in China’s energy sector [4].

➃

Socio-economic factors and energy-saving and carbon reduction factors further have significant positive and negative impacts on coal supply through coal demand, corresponding to specific indirect effects of 3.780 and −1.949, respectively. The specific indirect effects of technical factors on coal supply through coal demand and factor input are −0.356 and −0.542, respectively, which are not significant. The specific indirect effect of alternative energy factors on coal supply through coal demand is −0.478, and the effect is not significant. The specific indirect effects of social and economic factors and energy-saving and carbon reduction factors on coal supply through factor input are 0.716 and −0.343, respectively, and the effects are not significant. This result of this study supports the proposition that coal demand is the primary driver of coal supply and that coal production in China is highly responsive to market demand, particularly in the power and industrial sectors [30]. From the supply capacity of domestic oil and natural gas resources and the acquisition capacity of the international market, as well as the development technology and potential analysis of other energy sources, the substitution of fossil energy for coal is limited, and the substitution of new energy for coal is gradually increasing under the “dual carbon” target [65,66].

5. Conclusions

China’s coal industry is undergoing a significant transformation under the influence of the “dual carbon” target, with multiple pathways shaping its future. Among these, coal demand emerges as the most critical transmission mechanism. The path coefficient from coal demand to coal supply (1.121) is significantly higher than that of factor input (0.169), underscoring the significant role of coal demand in determining how carbon reduction policies affect the industry. While energy conservation measures aim to curb demand, socio-economic factors continue to sustain coal consumption, creating a complex and sometimes contradictory dynamic.

Despite the potential of technological advancements and alternative energy sources, their current impact remains limited due to slow adoption rates, high transition costs, and the small share of renewables in China’s energy mix. However, with strategic policy incentives, increased investment in research and development, and improved infrastructure for alternative energy integration, these factors could become key drivers of transformation in the future.

Beyond China, these findings offer valuable insights for other coal-dependent nations working toward carbon neutrality. The transition away from coal is not a straightforward process; it requires a balanced approach that considers energy security, economic stability, and environmental sustainability. Policymakers must recognize that coal demand is the primary lever in this shift. Progress will depend on implementing policies that not only restrict coal consumption but also make clean energy adoption economically viable. Strengthening carbon reduction strategies, fostering technological innovation, expanding access to renewable energy, and supporting workforce transitions in coal-dependent regions will be essential steps in achieving long-term sustainability.

The analysis of direct, indirect, and specific indirect effects reveals the complex mechanisms through which carbon reduction policies influence China’s coal industry. Direct effects show strong positive impacts on technical advancements, socio-economic factors, and alternative energy. Indirect effects highlight coal demand as pivotal, with socio-economic factors sustaining consumption and energy-saving measures curbing it. Specific indirect effects clarify that economic growth drives coal supply, while energy conservation reduces it. These findings align with previous research, emphasizing the challenges of transitioning to a low-carbon economy. Policymakers must balance economic growth, energy security, and environmental sustainability to achieve carbon neutrality.

The path to carbon neutrality presents both challenges and opportunities. While coal will remain a key part of China’s energy landscape in the near term, the long-term shift toward a low-carbon economy is inevitable. By focusing on targeted policy measures and accelerating the development of cleaner energy solutions, China—and other nations—can make meaningful progress toward global decarbonization goals while maintaining economic resilience and energy security.

The findings of this study offer important direction for other coal-dependent countries working towards carbon neutrality, emphasizing the need for a balanced approach that considers energy security, economic stability, and environmental sustainability.