Measurement and Driving Factors of Carbon Emissions from Coal Consumption in China Based on the Kaya-LMDI Model

Abstract

As the top emitter of carbon dioxide worldwide, China faces a considerable challenge in reducing carbon emissions to combat global warming. Carbon emissions from coal consumption is the primary source of carbon dioxide emissions in China. The decomposition of the driving factors and the quantification of regions and industries needs further research. Thus, this paper decomposed five driving factors affecting carbon emissions from coal consumption in China, namely, carbon emission intensity, energy structure, energy intensity, economic output, and population scale, by constructing a Kaya-Logarithmic Mean Divisia Index (Kaya-LMDI) decomposition model with data on coal consumption in China from 1997 to 2019. It was revealed that the economic output and energy intensity effects are major drivers and inhibitors of carbon emissions from coal consumption in China, respectively. The contribution and impact of these driving factors on carbon emissions from coal consumption were analyzed for different regions and industrial sectors. The results showed that carbon emissions from coal consumption increased by 3211.92 million tons from 1997 to 2019. From a regional perspective, Hebei Province has the most significant impact on carbon emissions from coal consumption due to the effect of economic output. Additionally, the industrial sector had the most pronounced influence on carbon emissions from coal consumption due to the economic output effect. Finally, a series of measures to reduce carbon emissions including controlling the total coal consumption, improving the utilization rate of clean energy, and optimizing the energy structure is proposed based on China’s actual development.

1. Introduction

Global warming is among the most serious threats confronting human society in the 21st century. High levels of carbon dioxide have been retained in our atmosphere by human activities such as the burning of fossil fuels including coal and oil, industrial processes, and changes in land use since the industrial revolution. These human activities have caused the sea-level rise and global warming, putting human survival and global economic growth at risk. According to the latest figures published by the International Energy Agency (IEA), the global carbon dioxide emissions from fossil energy consumption reached 32.318 billion tons in 2020, 50% higher than that in 1990. In 2019, global carbon emissions reached their highest value to date at 34.356 billion tons (from the IEA), three times as much as in 1965. In 2005, China contributed 21% of the world’s total carbon emissions, taking over from the United States as the world’s largest emitter of carbon dioxide. In 2020, China was responsible for 30.7% of the world’s carbon emissions, while its energy-related carbon dioxide emissions amounted to 9.899 billion tons [1]. With an energy endowment rich in coal and scarce in oil and gas and a long-standing coal-based energy consumption structure, carbon emissions from coal consumption have become the main contributor to China’s carbon dioxide emissions (Figure 1). To curb the fast growth of carbon emissions and effectively control the overall carbon dioxide emissions, China has proposed a commitment to a substantial carbon emission reduction target by 2030. Carbon dioxide emissions per unit of gross domestic product (GDP) will be cut by 65% or more when compared to 2005. At the same time, the overall control of coal consumption and the coal transformation and expansion will be the foothold and primary task of China’s energy revolution strategy. Therefore, it is necessary to evaluate the carbon emissions generated by China’s coal consumption, quantify the driving factors, and analyze the changing characteristics and laws governing coal consumption carbon emissions. It has important theoretical and practical implications to formulate more scientific and reasonable emission reduction policies and alleviate China’s increasingly severe carbon emissions trend.

In the research areas of energy consumption and its environmental effects, decomposition analysis methods are often utilized to quantify the influence of various factors on carbon dioxide emissions [2]. Structural decomposition analysis (SDA) and index decomposition analysis (IDA) are two major decomposition analysis methods. Leontief initially used the SDA method to study the input–output structure and analyzed the mechanism of factors affecting the final demand side [3]. The IDA method is mainly adopted to analyze energy systems, uses the aggregate data of energy sources of various departments for analysis, and is widely applied in energy environmental research [4]. Here, the study of China’s carbon emissions from coal consumption and its driving factors was mainly based on the coal consumption and carbon dioxide emissions data at the provincial level and various industrial sectors, so the IDA method is more appropriate. The Laspeyres and Divisia methods are commonly used in the IDA method. The Laspeyres method has residual items that cannot be eliminated during factor decomposition, which influences the decomposition results. Wang et al. [5], Zhang et al. [6], and Wang et al. [7] proposed the Logarithmic Mean Divisia Index (LMDI) decomposition method on the basis of the Divisia method, which is widely used to construct energy-related and carbon-emission-driven decomposition models as no unexplained residuals occur after decomposing the objects [8].

Many academics have studied carbon dioxide emissions and their influencing factors on energy consumption from the national, regional, and industrial sector levels based on the LMDI model. Xu et al. [9], Ma et al. [2], Yang et al. [10], and Wu et al. [11] all employed the LMDI index decomposition analysis method to decompose the factors affecting China’s carbon emissions from energy consumption into the energy structure, energy intensity, industrial structure, economic output, and population-scale effect. They found that the economic output effect was the dominant driving factor of carbon emissions. Similar decomposition methods have also been used in other countries. For example, Cansino et al. [12] decomposed the driving factors of carbon dioxide emissions in Spain between 1995 and 2009 through the IDA method and quantitatively analyzed their contribution to carbon emissions. Furthermore, Chen et al. [13] utilized the LMDI model to address the factors affecting carbon emissions from fossil energy in 30 Organization for Economic Co-operation and Development (OECD) countries from 2001 to 2015. It turned out that the GDP per capita and energy intensity were the principal causes of the rise and fall in carbon dioxide emissions, respectively. Additionally, Wójtowicz et al. [14] researched the implications of public spending on carbon emissions in Poland from a fiscal policy perspective. Zhao et al. [15] also researched carbon emissions from energy consumption in 50 economies from the perspective of debt and revealed that GDP was the largest contributor to carbon emissions. In parallel, studies conducted by Zhang et al. [16], Engo [17], and Yasmeen et al. [18] in different countries have confirmed that economic activities are the predominant drivers of carbon dioxide emissions.

As regional economies grow, the LMDI model can also be used to decompose the factors affecting regional carbon emissions. Jung et al. [19] probed the characteristics of energy-related carbon emissions in eco-industrial parks in South Korea. Chen et al. [20] adopted a multi-regional LMDI model to simulate the driving forces of changes in carbon emissions in manufacturing in China’s Pearl River Delta region. Zou [21] analyzed the differences in the drivers of carbon emissions in the Beijing–Tianjin–Hebei region and uncovered that the per capita GDP and population positively affected the carbon emissions. Moreover, Jiang et al. [22], Tan et al. [23], Jiang et al. [24], Wang et al. [8], Gu et al. [25], Yang et al. [26], and Xin et al. [27] explored the drivers of carbon dioxide emissions in different regions with LMDI. The findings illustrated that economic expansion was the primary driver of carbon emissions from a regional perspective.

In addition, at the industrial sector level, Yang et al. [28] carried out LMDI and scenario analysis to assess the potential impact of emission reductions in China’s power industry on carbon dioxide emissions. Furthermore, Wu et al. [29] analyzed the decoupling relationship between economic output and carbon emissions within China’s construction industry. Du et al. [30] determined the carbon emission factors of China’s six energy-intensive industries with the LMDI decomposition method. Raza et al. [31] investigated the carbon dioxide emissions from Pakistan’s transportation sector with the aid of the decoupling method of LMDI and Tapio. Dolge et al. [32] adopted the LMDI decomposition method to probe the impact of varying factors on the carbon emissions of industrial manufacturing in the EU and revealed that improving industrial energy efficiency could effectively reduce carbon emissions. Meanwhile, Zhang et al. [33] analyzed and quantified the association between industrialization, energy system, and carbon emissions. The above literature shows that although carbon emissions from energy consumption are increasingly being studied, there is little research on carbon emissions from a single energy, particularly in coal consumption. The drivers of carbon emissions from China’s coal consumption have been studied [34]. However, the contribution of each driving factor to carbon dioxide emissions has not been quantified, and the driving factors have not been analyzed from the regional and industrial sector levels. As China’s leading source of carbon emissions, coal is a critical material basis for socio-economic development. Therefore, it becomes essential to study the driving factors of coal consumption and analyze its contribution and impact on carbon emissions from regional and industrial perspectives.

Taking China as an example, this paper decomposed the factors influencing carbon emissions in 30 provincial-level administrative units and seven consumption departments with the LMDI method, analyzed their trends and intensity, and put forward relevant policy suggestions. The second part of this article introduces the research methods and data sources including the accounting process of coal carbon emissions, the construction of the decomposition model based on the Kaya-LMDI index, and the data sources of each index. The third part of this paper discusses the contribution and impact of driving factors on carbon emissions from 30 provinces and seven industries. Finally, the fourth part of the article presents the study findings and policy recommendations.

2. Methods and Data

2.1. Calculation Method of Carbon Emissions

The Intergovernmental Panel on Climate Change (IPCC) is widely adopted by international agencies to estimate carbon dioxide emissions from fossil fuels [35]. Therefore, this paper refers to the IPCC method to estimate the carbon emissions of coal consumption, which is calculated by the following formula:

$$C={{\displaystyle \sum }}_{i=1}^{30}{{\displaystyle \sum }}_{j=1}^7{{\displaystyle \sum }}_{r=1}^{6}E{C}_{ijr}·{\rho }_r$$

(1)

The subscripts i, j, and r represent provinces, industrial sectors, and coal types, respectively. C denotes the total carbon emissions from coal consumption; $E{C}_{ijr}$ indicates the consumption of the type of coal, $r$, in the $j$ industry sector, in $i$ province. ${\rho }_r$ represents the carbon emission coefficient of the $r$ coal types. The carbon emission coefficient is related to the quality, attributes, and combustion efficiency of fuel, as shown in Equation (2):

$${\rho }_r=LC{V}_r·C{C}_r·O_r·\frac{44}{12}$$

(2)

$\mathrm{where} LC{V}_r$ refers to the mean low calorific value of the $r$ coal types; $C{C}_r$ represents the carbon content per unit calorific value of the $r$ coal types; and $O_r$ stands for the carbon oxidation rate of the $r$ coal types.

2.2. Model Construction

Japanese scholar Kaya [36] first proposed the Kaya equation at the IPCC Conference in 1989 and analyzed the influence of energy, economy, population, and other factors on greenhouse gas emissions at the national level, calculated as follows:

$$GHG=\frac{GHG}{TOE}\times \frac{TOE}{GDP}\times \frac{GDP}{POP}\times POP$$

(3)

$\mathrm{where} GHG$ represents the greenhouse gas emissions; $TOE$ stands for the total energy consumption; $GDP$ refers to the gross domestic product; and $POP$ indicates the population size. The ratio on the right-hand side of the Equation $\frac{GHG}{TOE}$ represents the greenhouse gas emission intensity, reflecting the greenhouse gas emissions per unit of energy consumption. $\frac{TOE}{GDP}$ represents the energy intensity per unit of $GDP$, reflecting the energy consumption per unit of GDP. $\frac{GDP}{POP}$ represents the per capita GDP and is indicative of the extent of economic growth. These three factors and the total population make up the four influencing factors of greenhouse gas emissions.

Based on Kaya’s theory and the findings of various researchers on the factors influencing carbon emissions from energy consumption (Ouyang et al. [37], Yu et al. [38], and Zhang et al. [39]), this paper summarizes the factors affecting carbon emissions from coal consumption as the carbon emission intensity of coal consumption, energy structure, energy intensity, economic output, and population size. In addition, the Kaya equation is only applicable to the analysis at the national and regional levels. To comprehensively analyze the contribution of driving factors to carbon emissions from coal consumption at the Industrial sector level, the Kaya Equation (3) is extended to where:

$$C={{\displaystyle \sum }}_{i=1}^{30}{{\displaystyle \sum }}_{j=1}^7\frac{C_{ij}}{E{C}_{ij}}\times \frac{E{C}_{ij}}{E_{ij}}\times \frac{E_{ij}}{GD{P}_i}\times \frac{GD{P}_i}{P_i}\times P_i$$

(4)

$\mathrm{where} C$ indicates the total carbon emission of coal consumption; $C_{ij}$ represents the carbon emissions of coal consumption of the $j$ industrial sector in $i$ province; $E{C}_{ij}$ refers to the coal consumption of the $j$ industrial sector in $i$ province; $E_{ij}$ stands for the energy consumption of the $j$ industrial sector in $i$ province; $GD{P}_i$ denotes the gross regional product of $i$ province; and $P_i$ signifies the total population of $i$ province ($i$ = 1, 2, 3… 30 represents province, $j$ = 1, 2, 3 ... 7 represents the industry sector). $\frac{C_{ij}}{E{C}_{ij}}$ refers to the carbon emission intensity of coal consumption in the $j$ industrial sector of $i$ province, and $\frac{E{C}_{ij}}{E_{ij}}$ indicates the energy consumption structure of the $j$ industrial sector of $i$ province. $\frac{E_{ij}}{GD{P}_i}$ reflects the energy consumption intensity of the $j$ industrial sector of $i$ province, and $\frac{GD{P}_i}{P_i}$ signifies the per capita GDP of $i$ province.

The results were consistent, since the LMDI decomposition analysis has two decomposition forms: addition and multiplication (Ang, B. et al. [40]). This paper employed the additive decomposition method and further decomposed the change in China’s total carbon emissions from coal consumption into:

$$\Delta C=C^T-C^0={{\displaystyle \sum }}_{i=1}^{30}{{\displaystyle \sum }}_{j=1}^7\left(\Delta C_{\frac{C_{ij}}{E{C}_{ij}}}+\Delta C_{\frac{E{C}_{ij}}{E_{ij}}}+\Delta C_{\frac{E_{ij}}{GD{P}_i}}+\Delta C_{\frac{GD{P}_i}{P_i}}+\Delta C_{P_i}\right)$$

(5)

Among them, $C^0, C^T$ represents the base period and T period; $\Delta C$ signifies the total change in carbon emissions from coal consumption from the base period to cycle T. $\Delta C_{\frac{C_{ij}}{E{C}_{ij}}}$ stands for the carbon emission intensity effect of coal consumption, and $\Delta C_{\frac{E{C}_{ij}}{E_{ij}}}$ refers to the energy structure effect. $\Delta C_{\frac{E_{ij}}{GD{P}_i}}$ represents the energy intensity effect, $\Delta C_{\frac{GD{P}_i}{P_i}}$ represents the economic output effect, and $\Delta C_{P_i}$ represents the population size effect.

Therefore, the influence of each driving effect on the carbon emissions of coal consumption in different provinces and industrial sectors can be expressed as follows:

Carbon emission intensity effect of coal consumption:

$$\Delta C_{\frac{C_{ij}}{E{C}_{ij}}}=w\left(C_{ij}^T,C_{ij}^0\right)·\left[ln{\left(\frac{C_{ij}}{E{C}_{ij}}\right)}^T-ln{\left(\frac{C_{ij}}{E{C}_{ij}}\right)}^0\right]$$

(6)

Energy structure effect:

$$\Delta C_{\frac{E{C}_{ij}}{E_{ij}}}=w\left(C_{ij}^T,C_{ij}^0\right)·\left[ln{\left(\frac{E{C}_{ij}}{E_{ij}}\right)}^T-ln{\left(\frac{E{C}_{ij}}{E_{ij}}\right)}^0\right]$$

(7)

Energy intensity effect:

$$\Delta C_{\frac{E_{ij}}{GD{P}_i}}=w\left(C_{ij}^T,C_{ij}^0\right)·\left[ln{\left(\frac{E_{ij}}{GD{P}_i}\right)}^T-ln{\left(\frac{E_{ij}}{GD{P}_i}\right)}^0\right]$$

(8)

Economic output effect:

$$\Delta C_{\frac{GD{P}_i}{P_i}}=w\left(C_{ij}^T,C_{ij}^0\right)·\left[ln{\left(\frac{GD{P}_i}{P_i}\right)}^T-ln{\left(\frac{GD{P}_i}{P_i}\right)}^0\right]$$

(9)

Population size effect:

$$\Delta C_{P_i}=w\left(C_{ij}^T,C_{ij}^0\right)·\left[ln{P}_i{}^T-ln{P}_i{}^0\right]$$

(10)

In Equations (6)–(10), $w\left(C_{ij}^T,C_{ij}^0\right)$ is the weight coefficient of each variable, and its expression is:

$$w\left(C_{ij}^T,C_{ij}^0\right)=\{\begin{array}{c}\left(C_{ij}^T-C_{ij}^0\right)/\left(ln{C}_{ij}^T-ln{C}_{ij}^0\right),C_{ij}^T\ne C_{ij}^0\\ C_{ij }^T, C_{ij}^T=C_{ij}^0\\ 0,C_{ij}^T=C_{ij}^0=0\end{array}$$

(11)

Since the carbon emission coefficient of coal consumption is fixed and $\Delta C_{\frac{C_{ij}}{E{C}_{ij}}}$ is always 0 in the decomposition process, Equation (5) can be rewritten as:

$$\Delta C=C^T-C^0={{\displaystyle \sum }}_{i=1}^{30}{{\displaystyle \sum }}_{j=1}^7\left(\Delta C_{\frac{E{C}_{ij}}{E_{ij}}}+\Delta C_{\frac{E_{ij}}{GD{P}_i}}+\Delta C_{\frac{GD{P}_i}{P_i}}+\Delta C_{P_i}\right)$$

(12)

Therefore, Equations (7)–(12) can be used to measure and analyze the contributions and influence of different drivers on carbon emissions from coal consumption.

This paper focused on carbon emissions from coal in 30 provincial administrative units and seven industries sectors. Based on the Kaya-LMDI decomposition model, the contribution and influence of driving factors on carbon emissions from coal consumption from 30 provincial administrative regions and seven industrial levels were discussed. The analysis framework is shown in Figure 2.

2.3. Data Sources

Based on data availability, this paper selected 30 provincial-level administrative regions (as energy data for Hong Kong, Macau, Taiwan, and the Tibet Autonomous Region are missing, they will not be discussed in this paper) and seven consumption sectors in China from 1997 to 2019 as research objects. The 30 provincial-level administrative regions were grouped into six regions in the manner of the “China Statistical Yearbook”: North China (Beijing, Tianjin, Hebei, Shanxi, Inner Mongolia), Northeast China (Liaoning, Jilin, Heilongjiang), East China (Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong), South-Central China (Henan, Hubei, Hunan, Guangdong, Guangxi, Hainan), Southwest China (Chongqing, Sichuan, Guizhou, Yunnan), and Northwest China (Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang). Again, referring to the “China Energy Statistical Yearbook,” the seven consumption sectors were categorized into agriculture, forestry, animal husbandry, fishery, industry, construction, transportation, storage and postal services, wholesale and retail, accommodation and catering, residential life, and others. Concerning types of energy, there were five main categories of conventional energy currently included in the statistical accounts: coal, oil, gas, electricity, and other fuels. The integrity of the data at the provincial level for electricity consumption, nuclear energy, and renewable energy does not reference carbon dioxide directly. Therefore, the type of energy that this paper studied was limited to coal and oil. Coal includes washed and other coal, coke, coke oven gas, and other gas. In parallel, oil refers to crude oil, gasoline, kerosene, diesel oil, fuel oil, gas, liquefied petroleum gas, dry refinery gas, and natural gas. Data were mostly sourced from the “China Energy Statistical Yearbook” (1998–2020) [41], and the carbon emission coefficients of various energy sources are exhibited in

Table 1. Carbon emission coefficients of various energy sources.

Types of Energy	Average Low Calorific Value(kJ/kg, kJ/m3)	Standard Coal Coefficient (kgce/kg, kgce/m3)	Carbon Content per Unit Calorific Value (tC/TJ)	Carbon Oxidation Rate	Carbon Emission Coefficient (kgCO2/kg, kgCO2/m3)
Raw coal	20,908	0.7143	26.37	0.94	1.9003
Washed coal	26,344	0.9000	25.41	0.98	2.4054
Other coal washing	8363	0.2857	25.41	0.96	0.748
Coke	28,435	0.9714	29.5	0.93	2.8604
Coke oven gas	17,981	0.6143	13.58	0.99	0.8864
Other gas	5227	0.1786	12.2	0.99	0.2315
Crude oil	41,816	1.4286	20.1	0.98	3.0202
Gasoline	43,070	1.4714	18.9	0.98	2.9251
Kerosene	43,070	1.4714	19.5	0.98	3.0179
Diesel	42,652	1.4571	20.2	0.98	3.0959
Fuel oil	41,816	1.4286	21.1	0.98	3.1705
Liquefied petroleum gas	50,179	1.7143	17.2	0.98	3.1013
Refinery dry gas	45,998	1.5714	18.2	0.98	3.0082
Natural gas	38,931	1.33	15.3	0.99	2.1622

Note: (1) The average low calorific value was from the “China Energy Statistical Yearbook 2020” (National Bureau of Statistics, 2020). (2) The standard coal coefficient was converted according to the 20 °C card, and the low calorific value of 1 kg of standard coal (1 kGCE) equals 29,271.2 kJ. (3) The carbon content and carbon oxidation rate per unit calorific value were derived from the “Provincial Greenhouse Gas Inventory Guidelines” (National Center for Climate Change Strategy and International Cooperation, 2020) [43].

. Other indicators such as population and GDP were obtained from the “China Statistical Yearbook” (1998–2020) [42]. To remove the impact of price changes, GDP was calculated at constant prices in 1978.

3. Results and Discussion

3.1. Overall Analysis

This paper investigated the implications of energy structure, energy intensity, economic output, and population on China’s carbon emissions from coal consumption with the help of the Kaya-LMDI decomposition model. The outcomes are presented in

Table A1. Decomposition results of the coal consumption carbon emission drivers in China, 1997–2019 (unit: million tons).

Year	The Total Effect	Energy Structure Effect	Energy Intensity Effect	Economic Output Effect	Population Size Effect
1997–1998	28.41	22.57	283.80	254.64	23.32
1998–1999	76.39	26.47	301.45	229.37	22.16
1999–2000	63.40	66.64	267.46	237.79	32.91
2000–2001	86.19	3.09	186.76	253.14	16.71
2001–2002	181.22	5.49	129.56	297.10	19.17
2002–2003	424.41	77.58	18.02	342.48	22.36
2003–2004	399.05	18.65	27.14	419.08	25.76
2004–2005	929.44	42.06	351.27	543.10	7.00
2005–2006	447.85	1.23	198.77	611.33	34.05
2006–2007	343.76	12.76	381.03	702.69	34.86
2007–2008	493.56	78.58	251.85	622.78	44.06
2008–2009	399.51	26.65	333.22	661.46	44.62
2009–2010	380.32	10.92	448.09	771.18	68.15
2010–2011	591.03	9.81	265.77	828.91	37.70
2011–2012	251.17	34.76	581.90	755.05	43.26
2012–2013	645.27	64.70	1302.72	680.05	42.10
2013–2014	116.25	11.73	466.66	554.05	40.60
2014–2015	188.51	49.48	1560.41	1378.35	43.02
2015–2016	318.85	16.97	93.59	439.01	43.54
2016–2017	281.23	58.81	706.94	445.18	39.35
2017–2018	340.53	75.98	715.54	416.86	34.13
2018–2019	110.73	27.81	285.05	395.23	28.36
1997–2019	3211.92	214.82	8267.30	10,960.84	733.20

. The overall impact of economic output and population scale on carbon emissions from coal consumption was positive from 1997 to 2019, resulting in an uplift in carbon dioxide emissions from 10,968.4 million tons to 733.2 million tons. However, the overall effect of the energy structure and energy intensity on carbon emissions from coal consumption was negative, reducing carbon dioxide emissions by 214.82 million tons and 8267.3 million tons in the same period, respectively. The cumulative effect of the drivers in Figure 3 revealed that economic output was the primary driving factor of carbon emissions from coal consumption from 1997 to 2019, contributing as much as 341.25%, followed by the population scale, which had a contribution of 22.83%. In contrast, energy intensity was the primary restraining factor on carbon emissions from coal consumption, contributing to −257.39%. The energy structure, on the other hand, provided a relatively slight disincentive, with a contribution of -6.9%.

These findings highlight that reducing energy intensity and boosting energy efficiency were effective in cutting carbon emissions from coal consumption. The progress of the energy structure diminished the carbon emissions of coal consumption to a certain extent. However, as natural gas and alternative energy sources are still not effective substitutes for coal, China’s energy structure has not been improved, with the contribution from reducing carbon emissions from coal consumption being negligible.

A horizontal comparison with other countries shows that the key drivers affecting the increase or reduction in carbon emissions in different countries are basically the same. For example, Koilakou [44] and Trotta (2019) [45] both hold that the economic output effects and energy intensity effects are the dominant driving factors for the increase and decrease in carbon emissions in the United States and Germany, respectively. Yasmeen (2020) [18], Winyuchakrit (2016) [46], and other studies on developing countries have also come to this conclusion. These studies are similar and consistent with the findings of the study on China in this paper.

As the driving factors differed significantly in different periods, the following specific analysis was conducted to study the dynamic changes of the driving factors. As the principal driver of carbon emissions from coal consumption, economic output directly impacts carbon emissions from coal consumption. Figure 4 shows that from 1997 to 2015, the economic output effect fluctuated, but the overall carbon emission from coal consumption was elevated. China’s economy reflected high energy consumption and extensive development during this period. Large amounts of energy resources were invested to pursue rapid economic growth, steadily increasing the carbon emissions from coal consumption yearly. China’s economic growth rate decelerated to 6.9% in 2015, the slowest rate since 1990 [37], which led to a dramatic decline in economic output and a drop in carbon emissions from coal consumption.

From 2016 to 2019, as China’s economy entered a new phase of high-quality development from high-speed development, the economic output effect gradually rose to a gentle level. The pace of growth in carbon emissions from coal consumption also slowed down under strict state control. As depicted in Figure 4, China was in a phase of rapid industrialization and urbanization during the years 1997 to 2019. Population growth would increase the demand for energy through various means including production and living, which led to greater energy consumption and positively promoted carbon emissions from coal consumption [47].

The energy intensity effect is related to the total energy consumption and GDP, and it is not difficult to see from Figure 4 that the energy intensity effect fluctuated greatly.

From 1997 to 2005, as China’s economy grew rapidly, the dampening effect of energy intensity on carbon emissions gradually decreased. Furthermore, the suppressive effect on carbon emissions was converted into a driving effect between 2004 and 2005. In 2005, the state launched a new policy system for energy saving and consumption control, strictly limiting the mindless expansion of industries involving high energy consumption, high emissions, and overcapacity. As a result, the overall energy intensity effect trended downwards, and energy efficiency continued to improve. As China’s economy enters a new normal, its growth mode will bid farewell to the era of relying on a massive input of energy and resources and shift to the pursuit of economic quality and efficiency. The energy intensity effect will then show a decreasing trend year by year.

As exhibited in Figure 4, the range of change in the energy structure effect was small, with an overall negative dampening effect on carbon emissions. However, the amount contributed was far less than the energy intensity effect. In recent years, China has vigorously developed renewable energy sources such as solar, hydro, and wind energy. However, due to the technology and cost of renewable energy, it has not been easy to substitute them for coal, the primary energy source. Therefore, renewable energy will still make up a lower proportion of the total energy in the future for a long time.

3.2. Analysis of Industrial Sectors

For a long time, coal, oil, natural gas, and other fossil energies have made up 85% of China’s primary energy consumption, and heavy and chemical industries characterized the industrial structure. The carbon emissions from industrial energy consumption account for 65% of the national energy consumption carbon emissions (Guo shiyi et al. [48]). Figure 5 depicts the influence of seven industrial sectors on the carbon emissions from coal consumption in different periods. The carbon emissions and coal consumption trends for the industrial sector remained virtually unchanged from 1997 to 2019, indicating that industrial sectors are the key consumers of coal and the primary source of carbon emissions. In contrast, other industries had less impact on the carbon emissions from coal consumption.

To further calculate the impacts of different drivers on carbon emissions from coal consumption in seven industrial sectors, this paper conducted a decomposition analysis of the driving factors of coal consumption in these seven sectors, and the results are shown in

Table A2. Decomposition results of the coal consumption carbon emission drivers by industry sector in China, 1997–2019 (unit: million tons).

Industrial Sector	The Total Effect	Energy Structure Effect	Energy Intensity Effect	Economic Output Effect	Population Size Effect
Farming, forestry, animal husbandry, and fishery	1.62	15.01	188.72	194.42	10.93
Industrial	3328.82	267.69	7160.78	9567.60	654.31
The construction industry	3.91	26.10	33.10	59.38	3.73
Transportation, warehousing and postal services	13.97	83.03	45.35	109.83	4.57
Wholesale, retail and accommodation, catering	67.83	5.19	142.36	205.83	9.55
Other	34.03	67.38	116.81	204.15	14.06
Residents’ lives	210.31	285.81	580.18	619.63	36.05
Total industry	3211.92	214.82	8267.30	10,960.84	733.20

. In addition, Table A2 was transformed into Figure 5 to more intuitively exhibit the impact of various drivers on carbon emissions from coal consumption in diversified industrial sectors.

Figure 6a shows that the energy structure effect positively affected the carbon emissions from coal consumption in the industrial sector from 1997 to 2019. During that period, carbon dioxide emissions increased by 267.69 million tons, or 8.33% of total emissions, indicating an irrational energy structure in the industrial sector. In contrast, the contribution to the energy structure effect of carbon emissions from coal consumption in agriculture, forestry, animal husbandry, fishery, construction, transportation, storage, postal, wholesale trade, retail trade, accommodation, catering, and other industries and residents’ lives was negative. Among them, the energy structure of the residents’ lives had an apparent dampening effect on carbon emissions, reducing carbon dioxide emissions by 285.81 million tons and a contribution rate of 8.9% of total emissions. The energy structure effects of the other five industrial sectors had a restraining impact on the carbon emissions from coal consumption. This restraining effect shows that in addition to industrial sectors, the other five industrial sectors, mainly residential living sectors, have realized the reduction in carbon emissions from coal consumption through the improvement in energy structure. Overall, the energy structure effect negatively contributes to carbon emissions from coal consumption.

As presented in Figure 6b, the energy intensity effect adversely contributed to the carbon emissions from coal consumption in seven industrial sectors. First, the energy intensity effect of the industrial sector had the clearest dampening effect on the carbon emissions of coal consumption, reducing the carbon dioxide emissions by 7160.78 million tons, and contributing as much as −222.94%. This inhibitory effect fully demonstrates that improving the energy efficiency in the industrial sector can effectively reduce carbon dioxide emissions. The effect of the residents’ lives followed closely behind, reducing the carbon dioxide emissions by 580.18 million tons with a contribution rate of 18.06%. The energy intensity effect of the other five industries also exerted a specific inhibitory effect on carbon emissions from coal consumption, reducing the carbon dioxide emissions by 526.33 million tons.

Figure 6c shows that the economic output effect positively promotes carbon emissions from coal consumption in seven industrial sectors. First of all, the economic output effect of the industrial sector contributed the most to carbon emissions from coal consumption, increasing the carbon dioxide emissions by 9657.6 million tons and contributing to a rate of as much as 297.88%. Compared with the industrial sector, the economic output effect of the other six industrial sectors displayed a limited influence on the carbon emissions from coal consumption, with a total increase of 1393.24 million tons of carbon dioxide emissions. These outcomes demonstrate that the economic output of the industrial sector was a decisive factor in the rise in carbon emissions from coal consumption.

As can be seen in Figure 6d, the population size effect contributed positively to the carbon emissions from coal consumption in all seven industrial sectors. First, the industrial sector increased the carbon emissions from coal consumption by 654.31 million tons due to population size. The most significant contribution to carbon emissions was 20.37%. In their contribution to carbon emissions from coal consumption, the remaining six industrial sectors were residents’ lives, other industries, agriculture, forestry, animal husbandry and fisheries, wholesale and retail, accommodation and catering, transportation, storage, postal services, and construction. Overall, the population-scale effect of the seven industrial sectors was much weaker than the economic output effect on carbon emissions from coal consumption.

3.3. Regional Analysis

To further understand the spatial distribution of carbon emissions from coal consumption in China, this paper studied each driving factor’s influence on carbon emissions from coal consumption at the regional level. The results are shown in Appendix A 

Table A3. Decomposition of the coal consumption carbon emission drivers by province in China, 1997–2019 (unit: million tons).

Area	Provinces and Cities	The Total Effect	Energy Structure Effect	Energy Intensity Effect	Economic Output Effect	Population Size Effect
North China	Beijing	64.21	43.52	137.62	81.78	35.15
	Tianjin	26.45	11.91	171.90	163.63	46.63
	Hebei	609.02	4.76	751.61	1248.32	107.55
	Shanxi	174.82	15.53	439.79	571.46	58.67
	Inner Mongolia	379.71	2.62	147.19	503.42	20.87
Northeast	Liaoning	162.28	9.12	308.21	468.03	11.58
	Ji Lin	8.88	8.30	246.99	244.45	1.95
	Heilongjiang	74.72	50.69	155.70	182.04	2.31
East China	Shanghai	29.55	25.69	182.73	131.91	46.96
	Jiangsu	257.33	21.95	484.78	685.46	34.69
	Zhejiang	8.22	23.63	243.74	226.89	32.26
	Anhui	73.34	23.53	342.29	428.89	10.27
	Fujian	89.92	17.16	153.82	240.18	20.72
	Jiangxi	99.87	5.73	146.17	238.12	13.65
	Shandong	269.47	15.57	691.66	880.80	64.75
Central-South	Henna	58.10	49.69	601.73	701.53	8.00
	Hubei	66.93	13.17	519.48	591.84	7.73
	human	113.64	2.84	392.61	490.57	18.52
	Guangdong	88.12	43.82	309.42	283.78	69.94
	Guangxi	83.58	8.02	157.93	239.68	9.85
	Hainan	7.14	1.24	5.02	9.69	1.23
Southwest	Chongqing	15.34	23.64	247.37	279.05	7.29
	Sichuan	49.47	66.03	398.69	510.51	3.69
	Guizhou	56.35	36.58	311.34	407.30	3.03
	Yunnan	158.55	10.80	193.21	336.44	26.11
Northwest	Shaanxi	110.88	2.88	199.23	295.72	11.50
	Gansu	29.82	10.02	145.33	179.65	5.52
	Qinghai	25.67	3.75	27.48	51.58	5.32
	Mingxia	127.34	44.68	32.04	100.13	14.56
	Xinjiang	114.92	5.60	122.24	187.99	43.56
National		3211.92	214.82	8267.30	10,960.84	733.20

.

As shown in Figure 7a, from 1997 to 2019, the energy structure effect restrained the carbon emissions from coal consumption in Southwest, North, East, and South-Central China. First of all, Sichuan (−2.06%), Henan (−1.55%), and Beijing (−1.35%) were more prominent in improving the energy structure, reducing carbon emissions by 66.03 million tons, 49.69 million tons, and 43.52 million tons, respectively. In addition, the energy structure effect of Guizhou (−1.14%) and Chongqing (−0.74%) in Southwest China, and Shanghai (−0.8%), Zhejiang (−0.74%), Anhui (−0.73%), and Fujian (−0.53%) in East China had a significant dampening effect on the carbon emissions from coal consumption. Conversely, the energy structure effect of Heilongjiang (1.58%), Ningxia (1.39%), and Guangdong (1.36%) had a positive impact on the carbon emissions from coal consumption, which increased by 50.69 million tons, 44.68 million tons, and 43.82 million tons, respectively. The energy structure effect acted as a negative disincentive in carbon emissions from coal consumption.

The energy intensity effect (Figure 7b) exerted a dampening effect on the carbon emissions from coal consumption in all provinces in China. Among them, Hebei (−23.40%), Shandong (−21.53%), and Henan (−18.73%) made the largest contributions to the reduction in carbon emissions, reducing carbon emissions by 751.61 million tons, 691.66 million tons, and 601.73 million tons, respectively. In addition, the carbon reduction effects of Hubei (−16.17%) in South-Central China and Jiangsu (−15.09%) in East China were also more apparent, while Ningxia (−1.00%) and Qinghai (−0.86%) in Northwest China showed less inhibition of carbon emissions.

The economic output effect (Figure 7c) contributed positively to boosting the carbon emissions from coal consumption in all provinces in China. It generally showed a decreasing trend from north to south and east to west. In particular, east China, north China, and south-central China contributed significantly to carbon emissions from coal consumption owing to the economic output effect. In terms of provinces, Hebei (38.87%), Shandong (27.42%), Henan (21.84%), and Jiangsu (21.34%) saw an increase in carbon emissions by 1248.32 million tons, 880.80 million tons, 701.53 million tons, and 685.46 million tons, respectively, under the effect of economic output. In addition, Hubei (18.43%), Shanxi (17.79%), Sichuan (15.89%), Inner Mongolia (15.67%), and Hunan (15.27%) also displayed a marked positive impact on carbon emissions. In contrast, the economic output effect in the northeast and northwest regions had a relatively small contribution to coal consumption and carbon emissions.

The positive role of the population-scale effect (Figure 7d) on carbon emissions from coal consumption in North, East, and South-Central China was prominent. In particular, the first-tier developed cities were represented by Beijing (1.09%), Shanghai (1.46%), and Guangdong (2.18%), and the eastern coastal cities were represented by Hebei (3.35%), Shandong (2.02%), Tianjin (1.45%), Jiangsu (1.08%), and Zhejiang (1.00%), and inland cities in Central and Western China were represented by Shanxi (1.83%) and Xinjiang (1.36%). In contrast, the population-scale effect in Heilongjiang (−0.07%) and Guizhou (−0.09%) negatively affected the carbon emissions from coal consumption, reducing carbon emissions by 2.31 million tons to 3.03 million tons, respectively.

4. Conclusions and Suggestions

4.1. Conclusions

In this paper, the driving factors influencing the carbon emissions of China’s coal consumption from 1997 to 2019 were analyzed with the Kaya-LMDI index decomposition model. The driving factors affecting the alterations in carbon emissions were probed from the perspectives of countries, regions, and industrial sectors, and the following main conclusions were drawn:

(1) As per the decomposition and analysis of factors at the national level, the economic output effect and the population-scale effect positively drove the carbon emissions from coal consumption. In contrast, the energy structure and energy intensity negatively affected the carbon emissions from coal consumption. The economic output and energy intensity effects were the principal driving factors and inhibitors of coal consumption and carbon emissions, respectively. Overall, the negative drivers were not enough to offset the positive drivers of carbon emissions from coal consumption.

(2) Based on the decomposition and analysis of factors at the industrial sector level, the energy intensity effect of the industrial sector was the leading force in restraining carbon emissions from coal consumption. Improving the energy efficiency of the industrial sector can effectively reduce carbon emissions from coal consumption. However, it is far less effective than the influence of the economic output effect of the industrial sector on the carbon emissions of coal consumption. Therefore, the industrial sector remains a major contributor to the rise in carbon emissions from coal consumption. In addition, the energy structure effect on residents’ lives can offset the carbon emissions from the energy structure effect of the industrial sector. This offset is a result of the national policy directing the promotion of energy saving and emissions reduction. The amount of coal used in the residents’ daily lives was dramatically reduced. Additionally, there has also been an increase in the consumption of clean energy sources such as natural gas and electricity, with residents using cleaner, more efficient, and greener energy sources.

(3) The decomposition and analysis of factors at the regional level illustrate that emissions reduction measures vary from region to region because of the differences in economic development, energy structure, energy intensity, population size, etc.

4.2. Suggestions

The results disclose that East China, as the most economically developed area in China, contributes the most to the carbon emissions of coal consumption due to the economic output effects. At the same time, due to the early implementation of energy transition policies, the energy intensity effect on coal consumption carbon emissions is also the most prominent in East China.

Second, the energy structure adjustment in the Southwest has achieved concrete results, which has had a relatively large effect in curbing the carbon emissions from coal consumption.

However, there remains much room for improvement in energy restructuring in the Northeast and Northwest. Finally, the demographic advantage in North and East China has great implications for carbon emissions from coal consumption.

Given the findings of this study, we propose the following recommendations to enable China to further achieve its carbon reduction targets in the future.

(1) The energy structure effect should not be ignored, although it only has a slight dampening effect on coal carbon emissions. Local governments should give due attention to the dominant role of energy management and speed up the reform of the energy structure adjustment of the industrial sector without impeding economic development. Given the national situation of coal as a major energy source, we should pay close attention to the clean and efficient use of coal, increase the consuming capacity of renewable energy sources, and facilitate an optimal mix of coal and other renewable energy. The total amount of coal consumption should be reasonably controlled through technological progress. The substitution process of coal should be gradually completed to ensure the safety and reliability of renewable energy in achieving recyclable economic development. In addition, there should be the active promotion of industrial clusters in the northeast region, the extension of energy advantages in the industrial chain, and the simultaneous transfer of excess capacity to expand foreign trade in energy commodities, especially electricity. Full use should be made in the northwest region of its unique advantage in resources such as land, light, and wind power and by actively developing advantageous industries such as solar, thermal, optoelectronics, and wind power while promoting the transformation of resource advantages into industrial and economic benefits.

(2) As a major factor in restraining carbon emissions from coal consumption, the energy intensity effect is crucial to cutting carbon emissions. Over the years, the state has introduced a number of policies to conserve energy and cut emissions, which have allowed for energy utilization to be distinctly improved. Nevertheless, there is still a high possibility of a future energy intensity drop in the industrial sector with high energy consumption and heavy pollution.

Through technological renovation, the upgrade of production processes, product structure optimization, and other measures such as fully tapping the energy-saving potential in the production process, energy utilization efficiency can be improved and energy intensity reduced.

For example, regions such as Hebei, Shandong, and Henan should fully reconcile the relationship between energy utilization and economic growth and maximize the emissions reduction effect of the energy intensity effect. The northwest regions represented by Ningxia and Qinghai should further optimize and improve the energy utilization rate to reduce the energy intensity.

(3) The economic output effect is the dominant driver affecting the rise in carbon emissions from coal consumption in China. As a developing country striving to ensure sound and rapid growth of the national economy, China should accelerate economic growth and transformation while establishing a green, low-carbon, and circular economic system. Regions such as Hebei in North China and Shandong and Jiangsu in East China, who produce significant amounts of carbon emissions from coal consumption due to economic output effects, should pay more attention to social quality of economic growth while paying attention to economic development. It is necessary to make full use of the region’s energy resource advantages while facilitating the transformation of the mode of economic growth and vigorously developing a low-carbon economy featuring low energy consumption, minimal pollution, and reduced emissions.

(4) The population-scale effect contributes less to the carbon emissions from coal consumption. However, as a country with a large population, a small increase in the total population will entail more energy demand, which leads to a rise in carbon emissions. Therefore, Hebei in North China and Shandong in East China should reasonably control the size of their populations. Furthermore, local governments should strengthen the publicity and education on ecological civilization, support the citizens’ awareness of energy conservation and environmental protection, and guide the residents’ lifestyle and consumption patterns to transition to sustainable and low-carbon models. These actions will engender a positive trend for the entire society in health and the common good.