EN中文

Journal of Resources and Ecology >

2023 , Vol. 14 >Issue 3: 445 - 453

DOI: https://doi.org/10.5814/j.issn.1674-764x.2023.03.002

Carbon Emissions

Industrial Upgrading, Total Factor Energy Efficiency and Regional Carbon Emission Reduction in China

  • ZHU Meifeng , * ,
  • HAN Zeyu
Expand
  • School of Economics and Management, North University of China, Taiyuan 030051, China
*ZHU Meifeng, E-mail:

Received date: 2022-02-15

  Accepted date: 2022-08-19

  Online published: 2023-04-21

Supported by

The Philosophy and Social Science Project of Shanxi Province(2021YY040)

The Philosophy and Social Sciences Key Research Base of Higher Education Institutions of Shanxi(2022J022)

The Humanities and Social Science Research Project of the Ministry of Education of China(21YJA790062)

Fold

Abstract

Based on the panel data of 30 provinces in China from 2000 to 2018, the mutual relationships and mechanisms of influence between industrial upgrading, total factor energy efficiency and regional carbon emission were investigated. The results show that the sophistication of industrial structure has a significant inhibitory effect on carbon emissions in all regions. The intensity of inhibition in different regions shows a sequence of “western > central > eastern”. The inhibitory effect of the rationalization of industrial structure on carbon emissions varies greatly among the different regions, with a significant restraining influence in the central and western regions, but much less influence in the eastern region. The inhibition of carbon emissions through the improvement of total factor energy efficiency is significant in all regions, and the inhibition intensity shows the sequence of “western > eastern > central”. Furthermore, the mediating effect test shows that the total factor energy efficiency in different regions has either a partial or complete mediating effect on the influence of industrial upgrading on carbon emission, so it can promote and strengthen the inhibitory effect of industrial upgrading on carbon emissions. Therefore, upgrading the industrial structure and improving the total factor energy efficiency are effective means to promote carbon emission reduction. Reducing carbon emissions by relying solely on industrial upgrading is not ideal, and it needs to be combined with improvements in the total factor energy efficiency to effectively promote carbon emission reduction.

Key words： carbon emissions; sophistication of industrial structure; rationalization of industrial structure; total factor energy efficiency

Cite this article

ZHU Meifeng , HAN Zeyu . Industrial Upgrading, Total Factor Energy Efficiency and Regional Carbon Emission Reduction in China[J]. Journal of Resources and Ecology, 2023 , 14(3) : 445 -453 . DOI: 10.5814/j.issn.1674-764x.2023.03.002

1 Introduction

In the Paris Climate Agreement, China pledged that its carbon emissions will peak by around 2030 and set a target of reducing carbon dioxide emissions per unit of GDP by 60%-65% relative to the 2005 levels. With the continuous growth of China’s economy, the pressure of carbon emission reduction has been gradually increasing. In 2019, China’s carbon emissions account for 28% of the global total, and the per capita emissions are 46% higher than the world average. Carbon emission reduction is important for mitigating risks that cannot be ignored for the survival of enterprises and sustainable economic development in the future. Therefore, the issue of carbon emission reduction has long been of particular concern to academia and the real economy. Energy related industries in the secondary industry have been the main source of carbon emissions, especially the thermal power industry. For example, 42% of China’s carbon emissions in 2018 came from the thermal power industry. So making adjustments to the economic structure and the energy structure are the most direct ways to achieve carbon emission reduction. Therefore, determining the extent to which the change in industrial structure will have an impact on carbon emission intensity, through what means, and whether there is heterogeneity among different provinces are significant for guiding China’s carbon emission reduction.
The upgrading of industrial structure is the fundamental driving force for the low-carbon development of the industrial system (Zhang et al., 2016a). In recent years, the research on industrial structure and carbon emission reduction at home and abroad includes three main aspects. 1) The relationship between industrial structure upgrading and carbon emissions has been analyzed from the perspectives of capital investment (Wang and Wang, 2019), energy consumption structure (Andersson and Karpestam, 2013; Margarita and Victor, 2013; Cao and Karplus, 2014), and financial asset allocation (Zhou and Ji, 2019). Most of the literature on this topic argues that there is a significant relationship between industrial structure and carbon emissions. The change in carbon emission intensity has been affected by the energy carbon emission density, energy intensity and industrial structure of the three industries. The low-carbon development of the industrial system has been driven by the energy structure. Secondary industry has a great impact on energy use efficiency, so improving energy efficiency will help to reduce carbon emissions (Qiu, 2016; Zhang et al., 2016a; Zhou and Ji, 2019). There are significant differences in the relationships between industrial structure and carbon emission in different regions. 2) In terms of the research on the coupling and coordinated development of industrial structure and carbon emissions, this topic has been analyzed from the perspectives of multiple coupling and space, and the coupling development degrees of the research areas have been quantitatively analyzed. The research conclusions are presented by a dynamic spatial econometric model, and the path for improving the coupling degree was designed (Tian et al., 2020; Zhou et al., 2020). 3) In terms of carbon emission impact mechanisms, such mechanisms have been analyzed from the perspectives of urbanization (Bi, 2015; Zhang et al., 2016b; Fan and Zhou, 2019; Wang and Cheng, 2020), environmental regulation (Meng and Han, 2017), government intervention (Qiu and Yuan, 2019), and economic agglomeration (Ren et al., 2020). This research from different angles has reached the same conclusions: urbanization, environmental regulation, government intervention and economic agglomeration all have different effects on promoting carbon emission reduction, but there are significant differences in different regions. In addition, some scholars have studied the mechanism affecting carbon emissions from the perspectives of population structure (Tian et al., 2015) and trade opening (Zhan, 2017).
Previous studies have shown that the larger the scale of economic development, the greater the carbon emissions. However, the carbon emissions of provinces with similar economic scales often differ to some extent, which is related to the differences in their degrees of industrial structure optimization, that is, the adjustment of industrial structure can have an impact on carbon emissions. Through the optimization of industrial structure, resource allocation promotes the flow of employment and the optimization of production capacity. Therefore, the production efficiency of all industries will be improved. From the perspective of industrial efficiency, the research to date has not been able to highlight the specific mechanisms of the impact of industrial structure on carbon emissions. Furthermore, the impacts of industrial structure optimization on carbon emissions in different regions are not consistent. Considering that energy consumption is the direct cause of carbon emissions, this study considers the mechanism by which industrial structure optimization influences carbon emissions from the perspective of energy consumption efficiency. This study supplements and improves on those cited above from the following two aspects. Firstly, among the existing literature on the relationship between industrial structure and carbon emissions, few studies have examined carbon emissions from the perspective of industrial structure upgrading and subdivision. Industrial structure upgrading is a multi-dimensional process, and the impacts on carbon emissions of industrial structure upgrading in different dimensions are inconsistent. Therefore, this study subdivides the upgrading of industrial structure into the upgrading of industrial structure and the rationalization of industrial structure, and determines their impacts on carbon emissions in the eastern, central and western regions of China separately. Secondly, in terms of the impact mechanism, few previous studies have considered the mechanism of the impact of industrial structure on carbon emissions from the perspective of total factor energy efficiency, even though the upgrading of industrial structure has a direct impact on energy efficiency, and energy efficiency is closely related to low-carbon development. Therefore, this study calculates the total factor energy efficiency and determines its role in the whole impact mechanism. Through the mediating effect model, this study reveals the interactive mechanism among industrial structure upgrading, energy efficiency and carbon emissions.

2 Model setting and variable selection

2.1 Model setting

In order to study the relationship between industrial structure upgrading and carbon emission intensity, the following econometric model was established:
$\begin{align} & CAR{{I}_{it}}={{\alpha }_{0}}+{{\alpha }_{1}}IS{{R}_{it}}+{{\alpha }_{2}}IS{{S}_{it}}+\sum{{{\beta }_{i}}}Control{{s}_{it}}+ \\ & \begin{matrix} \begin{matrix} {} & {} \\ \end{matrix} & {} \\ \end{matrix}\begin{matrix} {} \end{matrix}{{\omega }_{i}}+{{\varphi }_{i}}+{{\varepsilon }_{it}} \\ \end{align}$
Where $CAR{{I}_{it}}$ is the carbon emission intensity, $IS{{R}_{it}}$ is the rationalization level of industrial structure, $IS{{S}_{it}}$ is the advanced level of industrial structure, and $Control{{s}_{it}}$ is control variable. Subscripts $i$ and $t$ indicate provinces and years, respectively. ${{\omega }_{i}}$ is the individual effect, ${{\varphi }_{i}}$ is the time effect. ${{\varepsilon }_{it}}$ is the random interference term, and ${{\varepsilon }_{it}}\tilde{\ }N(0,{{\delta }^{2}})$. ${{\alpha }_{0}},\ {{\alpha }_{1}},\ {{\alpha }_{2}}$ and ${{\beta }_{i}}$ are parameters of the function.

2.2 Variable selection

2.2.1 Core explanatory variables

(1) Rationalization of industrial structure
The rationalization of industrial structure can measure the upgrading of industrial structure from the perspective of industrial integration, so it can reflect the effect of the optimal allocation and utilization of resources. The indicators for measuring the rationalization of industrial structure mainly include technological progress rate, deviation index, Theil index, and others. Because the Theil index considers the importance of various industries and can reflect the coupling between output structure and employment structure, the Theil index was selected to measure the rationalization of industrial structure in this study.
(2) Sophistication of industrial structure
The sophistication of industrial structure measures the degree of transformation of the industrial structure from labor-intensive to knowledge- and technology-intensive industrial structure. The proportions of the three industries are gradually adjusted. The process of upgrading is represented by a shift in the industrial structure focus from primary industry to secondary industry and tertiary industry. Generally, the higher the proportion of tertiary industry, the higher the degree of industrial structure upgrading and the higher the degree of industrial structure sophistication.
(3) Total factor energy efficiency
There is a certain gap in energy efficiency between China and developed countries, and China’s energy consumption intensity is 2-3 times that of developed countries. The large scale of energy consumption and low utilization efficiency have a direct and significant pressure on China’s carbon emissions. Generally, total factor energy efficiency can be measured comprehensively by using the data envelopment model with a constant input-oriented return to scale on the premise of the constant input of other factors. Here, total factor energy efficiency is used to measure the energy efficiency.

2.2.2 Control variables

In order to consider the influences of other factors on the model estimation, four types of control variables were added to the model: Economic development, Urbanization, Investment environment, and Energy consumption intensity.
The economic development level represents the development scale, which is closely related to carbon emissions (Zhang et al., 2020). The level of urbanization has a significant impact on regional carbon emissions, so the ratio of urban population to rural population was selected as one of the control variables. The foreign direct investment was selected to measure the regional investment environment. The total energy consumption per unit GDP was chosen to measure the degree of control of energy consumption. The definitions of these variables are shown in Table 1.
Table 1 Definitions of main variables
Variable symbol Variable name Variable calculation process or formula
CARI Carbon emission intensity Carbon emissions / GDP
ISR Rationalization of industrial structure Theil index
ISS Sophistication of industrial structure Output value of tertiary industry / output value of secondary industry
TFEE Total factor energy efficiency Data envelopment analysis calculation
ED Economic development level Per capita GDP (10000 yuan)
UR Urbanization rate Urban population / rural population
FDI Investment environment Foreign direct investment (USD 100 million)
ENI Energy consumption intensity Total energy consumption / GDP

3 Data sources and calculations

The panel data of 30 provinces in China (excluding Hong Kong, Macao, Taiwan and Tibet due to the lack of data) from 2000 to 2018 was selected as the research object. The data were obtained from the local statistical yearbooks.

3.1 Rationalization and Sophistication of industrial structure

The rationalization of industrial structure can be represented by the Theil index, and its calculation is shown in formula (2).
$TL=\sum\limits_{i=1}^{n}{\left[ \left( \frac{{{Y}_{i}}}{Y} \right)\ln \left( {\frac{{{Y}_{i}}}{{{L}_{i}}}}/{\frac{Y}{L}}\; \right) \right]}$
where Y and L represent the total output value and labor force of each industry, respectively. ${{Y}_{i}}$and ${{L}_{i}}$ represent the output value and labor force of every industry, respectively, where i is from 1 to n, and n is the number of industries. The Theil index values of 30 provinces from 2000 to 2018 were calculated and used to replace their industrial structure rationalization index values. From equation (2), when the economy is in equilibrium, TL is zero, that is, the more reasonable the industrial structure and the closer the TL value to zero. On the contrary, the more TL deviates from zero, the worse the rationality of the industrial structure.
The ratio of the output value of the tertiary industry to the output value of the secondary industry was selected to measure the sophistication of the industrial structure (Gan et al., 2011).

3.2 Total factor energy efficiency

The calculation of total factor energy efficiency (TFEE) involves a variety of indicators such as the labor, energy and capital in the production process. Stochastic frontier function analysis (SFC) and data envelopment analysis (DEA) are often used to calculate TFEE. However, SFC cannot deal with the problem of multiple input-output data, so this study used DEA to measure TFEE.
The DEA creates a mathematically linear programming model when the output or input index of the decision-making unit (DMU) remains unchanged, in order to obtain the relative effective value. If it is not on the production frontier, the relative effectiveness can be evaluated by observing the deviation between the DMU and the frontier. If the DMU is on the leading edge, the efficiency value is 1, otherwise, the efficiency value is between 0 and 1. This study adopted the CCR model with a constant return to scale, and the selection of indicators mainly included the input and output.

3.2.1 Input indicators

(1) Energy input. The energy input was replaced by the total energy consumption, in units of tons of standard coal.
(2) Labor input. The number of employees in China from 2000 to 2018 was selected as the labor input index. It was obtained by totaling the number of employees in the tertiary industry in each province for each year.
(3) Capital stock. The “perpetual inventory method” was used to estimate the capital stock, and was calculated as:
$\begin{align} & {{K}_{i,t}}={{I}_{i,t}}+(1-{{\delta }_{i,t}}){{K}_{i,t-1}},\ i=1,2,\cdots,30;\ t=2000, \\ & \text{ }2001,\cdots,2018 \end{align}$
Where ${{K}_{i,t}}$ is capital stock, i represents the province, t represents the year,${{I}_{i,t}}$ is fixed asset investment, and ${{\delta }_{i,t}}$ is the depreciation rate of fixed assets. The year 2000 was taken as the base period for the calculation, the fixed asset investment price index was adjusted by the base period, and the fixed asset depreciation rate was uniformly taken as 9.6% (Zhang et al., 2004).

3.2.2 Output indicators

In this study, the regional GDP of 30 provinces in China from 2000 to 2018 was selected as the output index of energy efficiency. In order to eliminate the impact of price, the regional GDP was reduced at the constant price in 2000.

3.2.3 Calculation model

According to the purpose of decision-making, this study used four decision-making indicators. Of the four, the real GDP is the only output indicator, and the other three are input indicators: the total energy consumption, the capital stock calculated by the “sustainable inventory method” and the number of employees. In this study, the input factor is mainly considered for the total factor energy efficiency of various provinces in China. Therefore, the CCR model of minimizing energy input was adopted, that is, under the condition of constant return to scale (CRS), the input-oriented method was adopted for calculation. The CCR model was calculated as follows:
DMUj(j=1,2,...,n) is the decision-making unit. The objective function and constraints were calculated as:
$\min [\theta -\varepsilon (S_{1}^{-}+S_{2}^{-}+\cdots +S_{p}^{-})-\varepsilon (S_{1}^{+}+S_{2}^{+}+\cdots +S_{p}^{+})]$
$\text{s}\text{.t}\text{.}\left\{ \begin{align} & \sum\limits_{j=1}^{m}{{{x}_{j}}{{\lambda }_{j}}+{{S}^{-}}}=\theta {{x}_{0}} \\ & \sum\limits_{j=1}^{n}{{{y}_{j}}{{\lambda }_{j}}-{{S}^{+}}}={{y}_{0}} \\ & {{\lambda }_{j}}\ge 0,\begin{matrix} {} \\ \end{matrix}{{S}^{+}}\ge 0,\begin{matrix} {} \\ \end{matrix}{{S}^{-}}\ge 0 \\ \end{align} \right.$
where θ is a scalar. The calculated θ is the efficiency value of the DMU, which is generally θ≤1. If θ=1, it means that the DMU is technically efficient and at the forefront. S+ and S- are slack variables of the input and output variables. ε is the Archimedes infinitesimal. xj and yj represent the input and output variables of the j-th DMU to the element, respectively. λj is an n×1 constant vector, which represents the corresponding weight.
The optimal target value can be obtained according to the CCR model. The target value of energy consumption can be calculated according to formula (5):
${{E}_{\text{adjustment}}}={{E}_{\text{actual}}}-{{E}_{\text{target}}}$
where Eadjustment is the energy adjustment, Etarget is the target value of energy consumption, and Eactual is the actual value of energy consumption. So, the energy consumption target values of the 30 provinces in China from 2000 to 2018 can be calculated, and then the total factor energy efficiency was calculated according to formula (6):
$TFEE=\frac{{{E}_{\text{target}}}}{{{E}_{actual}}}$

3.3 Carbon emission intensity

The carbon emission intensity (CARI) was calculated according to IPCC coefficient method:
$C(t)=\sum{E_{j}^{{}}}(t)\times {{\phi }_{j}}$
where C(t) is the carbon dioxide emission, Ej(t) is the j-th energy consumption, ϕj is the carbon emission coefficient of the j-th energy, j is the energy type, and t is the measurement period. See Table 2 for the carbon emission coefficients of various fossil energy and oil refining types.
Table 2 Carbon emission coefficients of major energy types
Energy type Coefficient Energy type Coefficient Energy type Coefficient
Raw coal and washed coal 0.7561 Other coal washing 0.8472 Coke 0.8552
Crude oil 0.5859 Gasoline 0.5539 Kerosene 0.5923
Diesel oil 0.5923 Refinery Gas 0.4602 Natural gas 0.4484
Liquified natural gas 0.5131 Coking link 0.1527 Refining link 0.0987

Note: These data were calculated according to IPCC national greenhouse gas guidelines and general rules for the calculation of comprehensive energy consumption (GB-T2489).

The carbon emission intensity is the carbon emission per unit of GDP, and calculated by formula (8):
CARI=C(t)/GDP

4 Empirical results and analysis

4.1 Industrial structure upgrading and carbon emissions

4.1.1 Benchmark regression analysis

Based on the panel data of the 30 provinces in China from 2000 to 2018, this study analyzed the relationship between China’s industrial structure upgrading and carbon emissions. Firstly, the Hauseman test was carried out, and the P value was zero, so the fixed effect panel regression model was selected. In order to test the dynamic impact of industrial structure on carbon emissions, model Ⅰ and Ⅱ were set up. The dependent variable and control variables of the two models are consistent, but the core explanatory variables are different. In modelⅠ, the core explanatory variables are ISR and ISS. In model Ⅱ, the core explanatory variables are L.ISR and L.ISS, which are the lag items of ISR and ISS, respectively. The regression results are shown in Table 3. Model Ⅰ gives the regression result of the explanatory variable in the current period, and model Ⅱ gives the regression result of the explanatory variable lagging by one period. The empirical analysis results show that models Ⅰ and Ⅱ both have good fitting, the models are significant, and the regression results are approximately consistent. In other words, in the long run, the impact of ISR on carbon emissions is not significant, but the ISS and all factor energy efficiency can significantly inhibit carbon emissions. The elasticity coefficients of the ISS and TFEE on carbon emissions are -0.244 and -1.279, respectively. The inhibitory effects of ISS and TFEE on carbon emission are obvious, and as the ISS and TFEE increase by 1 percentage point, the carbon emission level decreases by 0.244 and 1.279 percentage points, respectively.
Table 3 Regression results
Variable ln CARI ln CARI
Model Ⅰ Model Ⅱ
ln ISR -0.130
(-0.92)
ln L.ISR ‒0.064
(0.43)
ln ISS ‒0.244***
(‒6.16)
ln L.ISS ‒0.237***
(‒5.77)
ln TFEE ‒1.279***
(‒10.32)		 ‒1.235***
(‒9.97)
ln ED 0.009***
(2.98)		 0.008***
(2.73)
ln UR 0.008***
(6.54)		 0.008***
(6.18)
ln FDI 0.108***
(7.91)		 0.099***
(7.42)
ln ENI ‒0.238***
(‒9.61)		 ‒0.224***
(‒9.03)
Constant 8.261***
(34.04)		 8.302***
(34.14)
Observations 540 510
R2 0.759 0.747
F-Statistic 226.23 199.29

Note: ***, **, * respectively indicate significance at 1%, 5% and 10% significance levels, and the T statistics are in brackets.

4.1.2 Heterogeneity analysis

In order to test the differences in the impacts of different regional industrial structures on carbon emission, the 30 provinces were divided into the eastern region (Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong and Hainan), the central region (Shanxi, Inner Mongolia, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei, Hunan and Guangxi) and the western region (Sichuan, Chongqing, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang). From the perspective of industrial structure upgrading, the mean of ln ISR values in the eastern, central and western regions are 0.128, 0.287 and 0.409, respectively; while the mean of ln ISS values are 1.156, 0.818 and 0.882, respectively. That is, the ISR and ISS in the eastern region are the highest, the ISR in the western region is the lowest, and the ISS in the central region is the lowest. From the perspective of total factor energy efficiency, the eastern region has the highest TFEE, followed by the central region and then the western region.
Table 4 Descriptive statistics of the main variables
Variable Mean Max Min Standard
deviation
Whole country (N=540) East (N=198) Central (N=180) West (N=162)
ln CARI 8.455 8.587 8.691 8.031 10.061 5.374 0.853
ln ISR 0.265 0.128 0.287 0.409 3.417 0.017 0.212
ln ISS 0.961 1.156 0.818 0.882 4.237 0.494 0.487
ln TFEE 0.764 0.859 0.769 0.643 1.000 0.351 0.148
ln ED 3.442 5.075 2.778 2.183 41.320 0.282 3.968
ln UR 49.135 58.915 46.291 40.343 89.600 13.885 15.145
ln FDI 12.018 13.286 12.130 10.344 15.090 7.310 1.806
ln ENI 1.285 0.889 1.262 1.795 5.229 0.255 0.814
The results of the subregional regression analysis are shown in Table 5. On the whole, the inhibitory effect of industrial structure upgrading on carbon emissions is the most significant in the central and western regions, but relatively weak in the eastern region. This difference is probably due to the high degree of industrial structure optimization in the eastern region, which makes it difficult to further curb carbon emissions through additional optimization. The optimization of industrial structure in the central and western regions is relatively low, so there still is room to promote the improvement of industrial efficiency and effectively curb the scale of carbon emissions. ISS has a significant inhibitory effect on carbon emissions in different regions, but the inhibition intensity in different regions shows a pattern of “West > Central > East”, which is just the opposite of the regional pattern of the ISS. These results show that the weaker the ISS, the stronger the inhibitory effect. From the empirical results, if the ISS in the eastern, central and western regions increase by 1 unit, the carbon emission intensities decrease by 0.224, 0.233 and 0.472 units, respectively. The inhibitory effect of ISR on carbon emissions varies greatly in the different regions. There is a significant inhibitory effect in the central and western regions, but it is not obvious in the eastern region. Therefore, improving ISR in the central and western regions is an effective way to curb carbon emissions, but this path has failed in the eastern region. In the different regions, the improvements of TFEE have shown significant inhibitory effects on carbon emissions, and the inhibition intensity shows a pattern of “West > East > Central”. Therefore, improving regional energy efficiency nationwide is an effective means to curb carbon emissions.
Table 5 Empirical results of the heterogeneity analysis
Variable National ln CARI Eastern region ln CARI Central region ln CARI Western region ln CARI
Model I Model III Model IV Model V
ln ISR -0.130
(-0.92)		 0.033
(-1.14)		 0.497**
(2.08)		 0.369*
(1.92)
ln ISS -0.244***
(-6.16)		 -0.224***
(-4.48)		 -0.233***
(-2.98)		 -0.472***
(-3.90)
ln TFEE -1.279***
(-10.32)		 -1.131***
(-5.76)		 -1.090***
(-5.51)		 -1.811***
(-5.78)
ln ED 0.009***
(2.98)		 0.0004
(0.16)		 0.006
(0.94)		 -0.004
(-0.27)
ln UR 0.008***
(6.54)		 0.005***
(4.21)		 -0.005**
(-2.20)		 0.028***
(4.05)
ln FDI 0.108***
(7.91)		 0.110***
(4.82)		 0.208***
(7.70)		 0.039*
(1.72)
ln ENI -0.238***
(-9.61)		 -0.243***
(-3.07)		 -0.139***
(-3.71)		 -0.202***
(-4.32)
Constant 8.261***
(34.04)		 8.731***
(21.00)		 7.438***
(16.73)		 8.277***
(15.31)
Observations 540 198 180 162
R2 0.759 0.825 0.850 0.806
F-Statistic 226.23 120.82 132.11 86.75

Note: ***, **, * indicate significance at 1%, 5% and 10% significance levels, respectively, and the T statistics are in brackets.

4.2 Mechanisms of total factor energy efficiency

The impact of industrial structure upgrading on carbon emissions is affected by energy efficiency. This study empirically analyzed the mediating effects of total factor energy efficiency in the eastern, central and western regions by using the mediating effect model.

4.2.1 Analysis of the mediation effect in which ISR is the core variable

All variables were centralized, and the model is shown in equations (9-11).
$CAR{{I}_{it}}=\alpha IS{{R}_{it}}+\sum{\beta }Control{{s}_{it}}+{{\varepsilon }_{it}}$
$TFE{{E}_{it}}=\gamma IS{{R}_{it}}+\sum{\delta }Control{{s}_{it}}+{{\varepsilon }_{it}}$
$CAR{{I}_{it}}=\mu IS{{R}_{it}}+\nu TFE{{E}_{it}}+\sum{\varpi }Control{{s}_{it}}+{{\varepsilon }_{it}}$
where $CAR{{I}_{it}}$ is the carbon emission intensity, $IS{{R}_{it}}$ is the rationalization level of the industrial structure, $TFE{{E}_{it}}$ is the total factor energy efficiency, and $Control{{s}_{it}}$ is the control variables. Subscripts i and t indicate provinces and years, respectively. ${{\varepsilon }_{it}}$ is a random interference term, and${{\varepsilon }_{it}}\tilde{\ }N(0,{{\delta }^{2}})$. $\alpha $ is the influence coefficient of ISR on CARI; $\beta $ is the influence coefficient vector of the control variables on CARI; $\gamma $ is the influence coefficient of ISR on TFEE; $\delta $ is the influence coefficient vector of the control variables on TFEE; $\mu $ is the influence coefficient of ISR on CARI considering the mediator; $\nu $ is the influence coefficient of TFEE on CARI; and $\varpi $ is the influence coefficient vector of the control variables on CARI considering the mediator.
Among the mediating effect test methods, the Bootstrap method has high statistical effectiveness and can replace the Sobel method for testing the product coefficient directly (Preacher and Hayes, 2008; Fang and Zhang, 2013; Wen and Ye, 2014). Here, this method was used to test the mediating effect (with the sampling times set as 500), and the discriminant condition is whether the 95% confidence interval of the indirect effect contains 0. The empirical analysis results are shown in Table 6.
Table 6 Empirical results of mediating effect for Model I
Region Effect Observed coefficient Bias Bootstrap standard error 95% Confidence interval
East Direct effect 0.580*
(0.06)		 -0.011 0.317 [0.005, 1.262]
Indirect effect 2.612***
(0.00)		 0.018 0.814 [1.121, 4.318]
Central Direct effect 0.203*
(0.08)		 -0.004 0.118 [-0.034, 0.447]
Indirect effect 1.752***
(0.00)		 0.003 0.343 [1.086, 2.478]
West Direct effect 0.169
(0.21)		 -0.004 0.136 [-0.073, 0.450]
Indirect effect -1.157***
(0.00)		 0.062 0.396 [-1.972, -0.508]

Note: ***, **, * indicate significance at 1%, 5% and 10% significance levels, respectively, and the P values are in parentheses. ISR is the core variable for the result.

The results in Table 6 show that the total factor energy efficiency in eastern and central China has a significant indirect effect on the relationship between ISR and carbon emissions, and there is a partial intermediary effect, which accounts for 81.83% and 89.62% of the total effect, respectively. This means that in eastern China, 81.83% of the inhibitory effect of ISR on carbon emissions comes from the improvement in the total factor energy efficiency. However, this proportion rises to 89.62% in the central region, indicating that TFEE plays a more significant role in promoting carbon emission reduction by ISR in the central region.
However, the direct effect of ISR on carbon emissions in western China is not significant, but the indirect effect is significant. In other words, TFEE is the only intermediary variable for strengthening the effect of ISR on carbon emissions.

4.2.2 Analysis of the mediation effect in which ISS is the core variable

The model is just the same equations (9-11). The empirical analysis results are shown in Table 7.
Table 7 Empirical results of mediating effect for Model II
Region Effect Observed coefficient Bias Bootstrap standard error 95% Confidence interval
East Direct effect -0.051**
(0.04)		 -0.002 0.026 [-0.110, -0.012]
Indirect effect -0.514***
(0.00)		 0.008 0.085 [-0.663, -0.330]
Central Direct effect 0.003
(0.96)		 -0.007 0.619 [-0.137, 0.109]
Indirect effect -0.447***
(0.00)		 -0.019 0.118 [-0.739, -0.267]
West Direct effect -0.019
(0.65)		 -0.006 0.043 [-0.139, 0.043]
Indirect effect 0.702***
(0.00)		 0.016 0.252 [0.234, 1.246]

Note: ***, **, * indicate significance at 1%, 5% and 10% significance levels, respectively, and the P values are in parentheses. ISR is the core variable for the result.

The data in Table 7 show that the indirect effects of TFEE on ISS and carbon emissions in eastern China are significant, there is a partial intermediary effect, and the proportion of the intermediary effect is 90.97%. That is, TFEE in this region greatly strengthens the inhibitory effect of ISS on carbon emissions. The direct effect of ISS in the central and western regions is not significant, but the indirect effect brought by total factor energy efficiency is significant, so the total effect of ISS on carbon emissions is significant. In the central and western regions, the effect of ISS on carbon emissions is reflected in the indirect effect of TFEE, that is, TFEE plays a full intermediary effect. Therefore, strengthening the tertiary industry while reducing the development of the secondary industry cannot directly bring about a significant impact on carbon emissions. Improving the TFEE is an effective way to increase the inhibitory effect of ISS on carbon emissions.

5 Conclusions and policy recommendations

Carbon emission constraints on China’s economic development have been becoming more and more obvious. According to the Paris Agreement, China’s carbon emissions will peak in 2030, with CO2 emissions dropping by 60%-65% compared with the 2005 level. Therefore, achieving the low carbon emissions will be the greatest challenge for China’s economy in the future. The inhibitory effect of industrial structure upgrading on carbon emissions is not obvious in many regions, so this study examined its internal mechanism of influence from the perspective of TFEE.
Through an empirical study of panel data for 30 provinces from 2000 to 2018, we identified four key aspects of this system. 1) In the long run, the ISR has no significant impact on carbon emissions. However, the ISS and TFEE can inhibit carbon emissions significantly. 2) The inhibition intensity of ISS on carbon emissions shows a pattern of “western > central > eastern”. 3) The effect of ISR on carbon emissions is significant in the central and western regions, but not in the eastern region. 4) TFEE can strengthen the inhibitory effect of industrial structure upgrading on carbon emissions, and its mediating effect is significant.
Based on these conclusions, two suggestions are put forward. 1) The degrees of ISR and ISS in the central and western regions are relatively weak, so enhancing ISS can effectively restrain carbon emissions. 2) In order to achieve low-carbon development，the governments should consider TFEE as an important influencing factor, and promote the coordinated development of industrial structure upgrading and TFEE as a development goal within the process of industrial structure upgrading.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Andersson F N G, Karpestam P. 2013. CO2 emissions and economic activity: Short- and long-run economic determinants of scale, energy intensity and carbon intensity. Energy Policy, 61: 1285-1294.

[2]
Bi X H. 2015. The impact mechanism of urbanization on carbon emissions. Shanghai Economic Review, 27(10): 97-106. (in Chinese)

[3]
Cao J, Karplus V J. 2014. Firm-level determinants of energy and carbon intensity in China. Energy Policy, 75: 167-178.

[4]
Fan J S, Zhou L. 2019. The mechanism and effect of urbanization and real estate investment on carbon emissions in China. Scientia Geographica Sinica, 39(4): 644-653. (in Chinese)

[5]
Fang J, Zhang M Q. 2013. Assessing point and interval estimation for the mediating effect: Distribution of the product, nonparametric bootstrap and Markov Chain Monte Carlo methods. Acta Psychologica Sinica, 44(10): 1408-1420. (in Chinese)

[6]
Gan C H, Zheng R G, Yu D F. 2011. An empirical study on the effects of industrial structure on economic growth and fluctuations in China. Economic Research Journal, 46(5): 4-16, 31. (in Chinese)

[7]
Margarita R A, Victor M. 2013. Decomposition analysis and Innovative Accounting Approach for energy-related CO2 (carbon dioxide) emissions intensity over 1996-2009 in Portugal. Energy, 57: 775-787.

[8]
Meng F S, Han B. 2017. Research on impact of government environmental regulation on enterprises’ low carbon technology innovation behavior. Forecasting, 36(1): 74-80. (in Chinese)

[9]
Preacher K J, Hayes A F. 2008. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3): 879-891.

[10]
Qiu L X, Yuan S. 2019. Government intervention, time-space effects and carbon emission reduction of typical cities. Soft Science, 33(5): 123-128. (in Chinese)

[11]
Qiu Z Z. 2016. The countermeasures of speeding up the industrial transformation and upgrading in China under the vision of low-carbon economy. Economic Review, (5): 57-60. (in Chinese)

[12]
Ren X S, Liu Y J, Zhao G H. 2020. The impact and transmission mechanism of economic agglomeration on carbon intensity. China Population, Resources and Environment, 30(4): 95-106. (in Chinese)

[13]
Tian C S, Hao Y, Li W J, et al. 2015. Population age structure effects on carbon emission in China. Resources Science, 37(12): 2309-2318. (in Chinese)

[14]
Tian Z, Jing X D, Xiao Q W. 2020. Coupling coordination degree and spatio-temporal evolution analysis of carbon emissions-industrial structure-regional innovation in the Yangtze River Economic Belt. East China Economic Management, 34(2): 10-17. (in Chinese)

[15]
Wang X J, Cheng Y. 2020. Research on the influencing mechanism of urbanization on carbon emission efficiency—Based on an empirical study of 118 countries. World Regional Studies, 29(3): 503-511. (in Chinese)

[16]
Wang Z, Wang L H. 2019. A study on the relationship among R & D input, upgrading industrial structure and carbon emission. Journal of Industrial Technological Economics, 38(5): 62-70. (in Chinese)

[17]
Wen Z L, Ye B J. 2014. Analyses of mediating effects: The development of methods and models. Advances in Psychological Science, 22(5): 731-745. (in Chinese)

[18]
Zhan H. 2017. Threshold effects of trade openness on China’s carbon emissions. World Economy Study, (2): 38- 49, 135-136. (in Chinese)

[19]
Zhang J, Wu G Y, Zhang J P. 2004. The estimation of China’s provincial capital stock: 1952-2000. Economic Research Journal, (10): 35-44. (in Chinese)

[20]
Zhang T F, Yang J, Sheng P F. 2016a. The impacts and channels of urbanization on carbon dioxide emissions in China. China Population, Resources and Environment, 26(2): 47-57. (in Chinese)

[21]
Zhang W, Zhu Q G, Gao H. 2016b. Upgrading of industrial structure, optimizing of energy structure, and low carbon development of industrial system. Economic Research Journal, 51(12): 62-75. (in Chinese)

[22]
Zhang Y, Zhang S Q, Zeng L J. 2020. The decoupling relationship between carbon emission, coal industry and economic development. Statistics and Decision, 36(3): 109-113. (in Chinese)

[23]
Zhou D, Zhang X R, Wang X Q. 2020. Research on coupling degree and coupling path between China’s carbon emission efficiency and industrial structure upgrading. Environmental Science and Pollution Research, 27: 25149-25162.

[24]
Zhou Y J, Ji P. 2019. Research on the impact of industrial upgrade and financial resource allocation efficiency on carbon emissions—Based on provincial spatial panel data analysis. East China Economic Management, 33(12): 59-68. (in Chinese)

Options
Outlines

/

〈 〉