The impact of eco-environmental regulation on green energy efficiency in China

Abstract

To figure out at what extent of China's investment in energy technology in recent years has helped economic growth and improved the environment. From the basic theory of environmental regulation’s influence on energy consumption efficiency, this study uses the super-efficiency slack base model to measure the green energy efficiency of 30 provinces in China from 2010 to 2019, and uses the kernel density function to analyze evolution trajectory. Then, the spatial Dubin model is used to test the influence of some important economic variables. The results of value show that although China's green energy efficiency improved significantly from 2010 to 2019, the overall performance was lacking. However, the overall efficiency value was significantly improved during the study period. Moreover, the efficiency value was closely related to the performance of economic development. From the dynamic trajectory perspective, efficiency values of all years show the type of "two peaks".Altough the width between the two peaks did not change noticeably, but the efficiency still had a significantly improved. In addition, according to the results of spatial economic regression analysis, eco-environmental constraints, foreign direct investment, and industrial agglomeration were not conducive to the efficiency value, while industrial structure and manufacturing enterprise scale had positive effects. In addition, the influence effects had obvious spatial phenomena. Therefore, for the Chinese government, enterprises and society, the next step is not only to vigorously exploit energy-saving and emission reduction technologies to improve energy consumption efficiency but also to establish reasonable environmental regulation policies to push enterprises to produce in a cleaner way. Additionally, a reasonable regional development plan should be established to promote the rational distribution and flow of the energy industry across different places. This paper provides a novel perspective for the study of energy input and usage.

Keywords

Green energy efficiency environmental regulation slacks-bases measure model spatial economic high quality development

Introduction

Energy Is an important factor in economic and social development. The history of human economic and social development is also a history of energy development.¹ Since the first industrial revolution, the discovery and use of carbon-based energy (coal, oil and gas) has greatly increased productivity. However, the serious crisis of the economy and environment caused by energy consumption has gradually become a popular research topic. Studies have confirmed that energy consumption causes serious environmental pollution.^2–5 Since the 1970s, environmental protection and regulation have been widely considered by countries around the world.^6,7 As a fast-growing economy, China's extensive economic growth model has wreaked havoc on environmental quality. For example, in 2016, China's CO₂ emissions from fossil energy accounted for 1/3 of world emissions, far exceeding that of other countries.⁸ The conflict between resource and environmental constraints and accelerated industrialization is becoming fierce. Currently, the carrying capacity of resources and the environment in most areas of China has reached or is approaching its upper limit.^9,10 The development pattern of the past no longer supports China's rapid economic growth and sustainable development. Environmental protection has become an important part of people's livelihood.

Recently, the Chinese government has been deeply aware of this problem. On September 22, 2020, Chinese President Xi Jinping, had proposed at the general conference of the 75th United Nations that China would adopt more powerful policies and measures to strive for the peak of carbon dioxide emissions by 2030 and achieve carbon neutrality by 2060.¹¹ However, this is not an easy task. This will be a huge transformation of economic structure, social cognition, energy transformation and technological innovation, and even a change of the humankind, because energy is not only the driving force of economic and social development, but also the main source of carbon emissions.^1,12 Now China's energy structure is relying too much on coal. From the perspective of China's energy consumption structure in 2020, coal consumption accounts for 56.8%, oil for 18.9%, and clean energy such as gas and non-fossil accounts for 24.3%.¹ Under the premise of continuous industrialization, China's demand for energy will continue increased in the future. Therefore, the core of how to ensure China's energy security should be to vigorously reduce the proportion of coal and other fossil energy, and develop non-fossil energy dominated by renewable category.¹³ Therefore, it is necessary to commit to improving energy efficiency and pushing the transformation of China's economy to a green economy with low energy consumption, low emissions, and high output.¹⁴ Energy utilization is the key factor to economic development and environmental governance promotion. Therefore, under the condition of energy shortages and ecological destruction, environmental regulation and energy efficiency are both important objectives of China's economic development and transformation. Whether the two types of policy, energy and environment, can achieve a better combination is the key to the realization of China's future green development transformation.

This paper establishes a theoretical framework for the relationship between environmental regulation and energy efficiency, bringing the possible environmental damage caused by energy consumption into the calculation of energy efficiency, and constructs the concept of “green energy efficiency”. This study also seeks to further enrich the literature of energy and environment economics and to clarify the relationship between environmental regulation and energy efficiency. Additionally, the results can aid in assessing the applicability of the “Porter Hypothesis” under the background of the Chinese economy entering a high-quality period while also helping to clarify the magnitude and direction of environmental regulation of energy efficiency.

Literature review

Energy is an important guarantee for social and economic development. From the static perspective of traditional economic theory, environmental regulations will force enterprises to increase investment in pollution treatment, resulting in increased production costs, reduced business benefits and energy efficiency. However, Porter¹⁵ and Porter & Vander Linde¹⁶ proposed the Porter hypothesis, which is different from traditional theory and holds that environmental regulation promotes enterprise innovation, optimizes the energy structure, and urges technological progress, thereby improving energy efficiency. Academic studies have confirmed that technological progress is the key to improving energy efficiency.^17–20 Some representative studies include, Cagno et al.²¹ who studied Italian foundry enterprises and found that enterprises with better innovation activity had higher energy efficiency, were more willing to adopt existing energy saving technologies, and had fewer cognitive barriers to energy efficiency improvement, Otsuka et al.²² who used data from Japan and found that the productivity improved by technological progress promoted energy efficiency, Ramirez-portilla et al.²³ who also confirmed the positive correlation between enterprise innovation and energy efficiency, Li & Du²⁴ who confirmed that the deep integration of the internet and manufacturing technology is conducive to promoting the development of the manufacturing industry towards intelligent energy saving, which is conducive to improving energy efficiency and Chen & Lin²⁵ who find that infrastructure development can reduce regional differences in green total factor energy efficiency in the manufacturing industry and that the impact is heterogeneous across transport and power infrastructure. Over time, the marginal effect of infrastructure diminished.

However, for environmental regulation, there are two opposing views. Some scholars have calculated that environmental regulation has an inhibitory effect on energy efficiency.^19,26 However, additional studies have confirmed that environmental regulation has a positive effect on energy efficiency.^27,28 For example, as early as 1977, the National Research Council and EPA conducted a study on the relevant data of the 1960s and 1970s in the United States and found that environmental regulation can lead to an increase in energy efficiency. The economic benefit is much greater than the cost of the environmental regulations.

Researchers contend that the effect of environmental regulation on energy efficiency depends on the combined effects of both positive and negative aspects²⁹). The strengthening of environmental regulations increases the production and operation costs of enterprises, which forces enterprises to spend part of their funds on pollution prevention and control or pay pollutant discharge taxes and fees, thereby increasing production costs.³⁰ In addition, with the increase of the intensity of environmental regulation, enterprise pollution treatment cost increases gradually, which, to an extent, is due to productive investment, resulting in a decrease in the level of enterprise production. If the energy input levels remain the same, the output is reduced to a certain extent, reducing the overall energy efficiency.^31,32 Conversely, forced by the government's environmental regulations, some enterprises must improve and upgrade their production methods to maintain market advantages, which may promote the overall production performance of enterprises and energy efficiency.^33,34 Additionally, new enterprises are usually required to install more advanced energy and pollution control equipment than old enterprises, and existing enterprises and new enterprises are required to set different pollution prevention standards, thereby establishing a barrier for new enterprises to develop. This type of regulation is known as “grandfather rules”.³⁵ Therefore, under strict environmental regulations, the overall enterprise has higher energy efficiency and pollution control ability.^36,37

In addition to environmental regulations, energy efficiency is also affected by other factors. For example, economic agglomeration is the primary manifestation of the spatial distribution of economic activities and elements, and the spatial distribution of economic activities will undoubtedly have a significant impact on energy demand and pollution (carbon) emissions.³⁸ It also has a significant impact on the productivity of output and additional factors.³⁹ Foreign direct investment has an obvious impact on energy consumption in both the long and short term.⁴⁰ Industrial scale change, structural adjustment and industrial upgrading all have great significance for improving energy efficiency, realizing energy conservation and emission reduction, and regional economic development.^41,42

In conclusion, there are certain differences in the verdicts about the impact of environmental regulations on energy efficiency. Therefore, further verification of the relationship between the two has certain significance for solving the conflict between the environment and economy in China. Considering this issue, this study uses the super-efficiency SBM model to measure total factor energy efficiency, uses the spatial econometrics method to incorporate it into the analysis framework, and further analyses the relationship between the two. The results of this study provide some reference for the formulation of relevant environmental regulation policies and further supplement the literature in this field.

Calculation and dynamic analysis of green energy efficiency

Super-efficiency SBM model

Data envelopment analysis (DEA), a nonparametric estimation method created by the famous operational research scientists A. Charnes and E. Rhodes, is frequently used in the interdisciplinary areas of operations research, management science and econometrics. Its principle is to construct a relatively effective efficiency frontier by mathematical programming and then judge the relative effectiveness of each production decision unit (DMU) according to its deviation from the frontier, which is a decision-making method to evaluate efficiency. The greatest advantage of this method is that it does not need to consider the functional relationship between input and output indices or estimate parameters and weight assumptions, thereby avoiding the influence of subjective judgement.⁴³ The slack-based measure (SBM) model is a non-radial DEA model based on slack variables that was proposed by Kaoru Tone in 2002.⁴⁴ This model directly adds slack variables into the objective function, and the result is no longer just the maximization of the efficiency ratio but the actual profit. The super-efficiency data envelopment analysis (SE-DEA) was modified on the traditional DEA model. In practical applications, the classical DEA model may produce multiple DMUs at the same production frontier (that is, the efficiency value is 1), which creates difficulties for further comparative analysis and sequencing. The SE-DEA model was improved based on the BCC model and proposed by P. Andersen and N. C. Petersen in 1993. It can make the efficiency value of the effective DMU greater than 1 so that all decision units can be further evaluated and compared.

Suppose there are n decision units, and each decision unit consists of three parts: input m, desired output $r_{1}$ , and undesirable output $r_{2}$ . We can use the vector to express it: $x \in R^{m}$ ， $y^{d} \in R^{r_{1}}$ ， $y^{u} \in R^{r_{2}}$ ; $X, Y^{d}$ and $Y^{u}$ are vectors, which can be described as $X = [X_{1} \dots X_{n}] \in R^{m \times n}$ ， $Y^{d} = [y_{1}^{d} \dots y_{n}^{d}] \in R^{r_{1} \times n}$ and $Y^{u} = [y_{1}^{u} \dots y_{n}^{u}] \in R^{r_{2} \times n}$ . The super-efficiency SBM model constructed in this study is as follows⁴⁰:

m i n = \frac{\frac{1}{m} \times \sum_{i = 1}^{m} (\frac{\bar{x}}{x_{i k}})}{\frac{1}{(r_{1} + r_{2})} \times (\frac{\sum_{s = 1}^{r_{1}} \bar{y^{d}}}{y_{s k}^{d}} + \frac{\sum_{q = 1}^{r_{2}} \bar{y^{u}}}{y_{q k}^{u}})}

\begin{aligned} s . t . \bar{x} \geq \sum_{j = 1, \neq k}^{n} x_{i j} λ_{j} i = 1, \dots, m \\ \bar{y^{d}} \leq \sum_{j = 1, \neq k}^{n} y_{s j}^{d} λ_{j} s = 1, \dots, r_{1} \\ \bar{y^{d}} \geq \sum_{j = 1, \neq k}^{n} y_{q j}^{u} λ_{j} q = 1, \dots, r_{2} \\ λ_{j} \geq 0 j = 1, \dots, n; j \neq 0 \\ \bar{x} \geq x_{k} k = 1, \dots, m \\ \bar{y^{d}} \leq y_{k}^{d} q = 1, \dots, r_{1} \\ \bar{y^{u}} \geq y_{k}^{u} u = 1, \dots, r_{2} \end{aligned}

Variable definition and data sorting

Energy consumption is a complex production system with many input and output variables, so efficiency evaluation should be considered comprehensively. Referring to previous academic studies on energy efficiency and considering important conditions such as data consistency, the variable indices selected in this paper are all large-calibre statistical variables related to energy consumption, as shown in Table 1.

Table 1.

The inputn-output system of green energy efficiency.

Indicator type	Indicator name	Indicator description and unit
Input indicators	Energy input	Total energy consumption（10,000 tons standard coal）
	Labour input	Employment numbers at the end of the year（Man-year）
	Capital stock	Fixed investments（10,000 yuan）
	Technology input	Technology market turnover（10,000 yuan）
Output indicators	Expect output	GDP per capita（10,000 yuan）
Output indicators	Unexpected output	The comprehensive value of CO₂ emission calculated by entropy method

This study considers the research years of 2010∼2019, and the data primarily originates from the China Statistical Yearbook, the China Energy Statistical Yearbook and the China Environmental Statistical Yearbook from 2011 to 2020. Cooper et al. concluded in 2001 that the number of DEA decision units in the algorithm should meet the condition of $n \geq m a x {m \times s, 3 (m + s)}$ to ensure the accuracy of the results, where n represents the number of decision units (i.e. the number of DMUs) and m and s represent the number of input and output variables, respectively. This research meets this demand.

Calculation of super-efficiency of energy consumption

MaxDEA Pro 6.6 was used to realize the operation of the model, and the energy efficiency of 30 provinces and autonomous regions in China from 2010 to 2019 was calculated. The results are shown in Table 2 and Figure 1. The drawn of Figure 1 is based on the average value.

Figure 1.

Spatial differentiation of regional green energy efficiency in China.

Table 2.

The assessing results of green energy efficiency.

GEE	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019	Average	Ranking
Beijing	1.178	1.008	0.644	1.007	1.051	1.141	1.109	1.245	1.245	2.183	1.181	2
Tianjin	0.46	1.08	1.04	1.045	1.03	1.015	1.033	1.048	1.036	0.436	0.922	5
Hebei	0.07	0.082	0.061	0.095	0.078	0.08	0.082	0.087	0.077	0.047	0.076	27
Shanxi	0.109	0.14	0.111	0.126	0.113	0.115	0.12	0.125	0.124	0.086	0.117	19
Inner Mongolia	0.084	0.111	0.111	0.141	0.137	0.154	0.121	0.156	0.118	0.087	0.122	18
Liaoning	0.167	0.259	0.223	0.263	0.241	0.26	0.291	0.203	0.254	0.182	0.234	8
Jilin	0.175	0.227	0.218	0.268	0.252	0.261	0.268	0.296	0.238	0.139	0.233	9
Heilongjiang	0.124	0.157	0.146	0.163	0.161	0.156	0.154	0.191	0.153	0.096	0.15	12
Shanghai	0.333	0.407	1.012	0.349	0.384	0.395	0.403	1.009	1.111	1.118	0.652	6
Jiangsu	3.479	3.805	4.205	3.707	4.259	3.28	4.587	1.559	4.852	7.632	4.137	1
Zhejiang	1.388	1.027	1.022	1.328	1.042	1.067	1.078	1.05	1.084	1.034	1.112	3
Anhui	0.063	0.08	0.074	0.088	0.087	0.088	0.093	0.109	0.093	0.082	0.086	25
Fujian	0.127	0.166	0.147	0.201	0.201	0.213	0.222	0.264	0.226	0.208	0.198	11
Jiangxi	0.088	0.108	0.1	0.122	0.113	0.118	0.119	0.157	0.122	0.091	0.114	20
Shandong	0.27	0.314	0.196	0.377	0.2	0.259	0.256	0.179	1.092	1.017	0.416	7
Henan	0.052	0.06	0.049	0.071	0.062	0.069	0.072	0.074	0.072	0.066	0.065	30
Hubei	0.073	0.094	0.094	0.112	0.114	0.12	0.121	0.158	0.123	0.104	0.111	21
Hunan	0.074	0.095	0.086	0.102	0.102	0.11	0.109	0.168	0.113	0.086	0.105	22
Guangdong	1.043	1.116	1.114	1.093	1.094	1.069	1.053	1.196	1.034	0.45	1.026	4
Guangxi	0.153	0.169	0.12	0.172	0.115	0.152	0.128	0.156	0.117	0.089	0.137	16
Hainan	0.105	0.137	0.122	0.162	0.153	0.158	0.15	0.176	0.152	0.129	0.145	15
Chongqing	0.124	0.182	0.186	0.204	0.209	0.235	0.217	0.331	0.23	0.237	0.216	10
Sichuan	0.048	0.065	0.062	0.069	0.067	0.074	0.07	0.105	0.07	0.059	0.069	29
Guizhou	0.122	0.123	0.11	0.126	0.11	0.13	0.143	0.195	0.127	0.093	0.128	17
Yunnan	0.094	0.11	0.087	0.094	0.088	0.099	0.096	0.147	0.096	0.089	0.1	23
Shaanxi	0.103	0.139	0.142	0.16	0.158	0.154	0.156	0.187	0.156	0.114	0.146	14
Gansu	0.126	0.155	0.149	0.156	0.153	0.15	0.152	0.18	0.147	0.104	0.147	13
Qinghai	0.059	0.079	0.079	0.099	0.102	0.109	0.105	0.134	0.112	0.09	0.097	24
Ningxia	0.064	0.079	0.076	0.08	0.083	0.082	0.076	0.097	0.079	0.07	0.079	26
Xinjiang	0.049	0.064	0.059	0.082	0.079	0.078	0.077	0.082	0.077	0.055	0.07	28
Average	0.347	0.388	0.395	0.402	0.401	0.38	0.422	0.369	0.484	0.542	0.413	-

At the national level, the average number for green energy efficiency in China from 2010 to 2019 was 0.413, which means that the overall performance was poor. However, the average efficiency value increased from 0.347 in 2010 to 0.542 in 2019. Although there were many fluctuations during that period, the overall efficiency value was significantly improved. The results were generally in accordance with China's development situation. In recent years, the Chinese government has accelerated the implementation of energy conservation and emission reduction policies and achieved remarkable results. During this period, the country effectively implemented energy conservation and emission reduction by strengthening constraints, promoting transformation, controlling increments, optimizing the stock system, and promoting innovation,⁴⁵ thereby promoting green energy efficiency.

According to the average efficiency values of different provinces, Jiangsu (4.137), Beijing (1.181), Zhejiang (1.112) and Guangdong (1.026) were the four provinces with the best performance (their efficiency values are all greater than 1). Tianjin (0.922) was the only province that had an efficiency value between 0.9 and 1. No province had an efficiency value between 0.7 and 0.8, and only Shanghai’s efficiency value (0.652) was between 0.6 and 0.7. The above province number was six, which means that only one-fifth of the 30 provinces were above 0.6, and the other four-fifths all had poor performance. It is particularly worth noting that the performance of Shanghai, China's most developed city, has been disappointing.

The worst-performing provinces (all with efficiency values less than 0.1) were Qinghai (0.097), Anhui (0.086), Ningxia (0.079), Hebei (0.076), Xinjiang (0.07), Sichuan (0.069) and Henan (0.065). These are all important manufacturing provinces in China. The poor efficiency performance of these manufacturing provinces means that China's industrial development was based on energy and the environment. From the perspective of space (which can be seen from Figure 1), most provinces with better efficiency values were located in developed regions in the southeast coast, while most provinces with poor performance were located in less developed regions in the inland area, which shows that the performance of regional green energy efficiency was significantly correlated with the level of regional economic development.

Dynamic evolution analysis of green energy efficiency

In this study, the dynamic evolution and trend of China's industrial green energy efficiency is portrayed through the nonparametric estimation method -- kernel density analysis. The basic principle is to suppose the probability density estimation formula f(x) of the variable X at point x is as follows:

f_{n}^{Λ} (x) = \frac{1}{n h^{p}} \sum_{i = 1}^{n} K [\frac{x - X_{i}}{h}]

In the formula,

k (·)

is the kernel function, h is the bandwidth, and n is the number of samples. The kernel function and bandwidth should be selected in the specific estimation. The key to the fitting result is the setting of the optimal bandwidth, and the basic idea is to minimize the mean square error. There are many forms of kernel functions. This study chooses the most common function—the Epanechnikov kernel function. In addition, the four years 2010, 2013, 2016 and 2019 were chosen as the investigation profiles. The distribution of the kernel density function in each year is shown in Figure 2.

Figure 2.

Kernel density diagram of industrial green energy efficiency.

From Figure 2, the dynamic evolution of China's industrial green energy efficiency distribution presents two features. First, the distribution peak of China's industrial green energy efficiency did not change much from 2010 to 2019, indicating that the improvement of green energy efficiency in most Chinese provinces was limited during this period. Second, the value of industrial green energy efficiency in the four years all belonged to the “twin peaks” type, and the width between the two peaks did not change noticeably. The values of the two equilibrium points were near 0.1 and 1, which means that China's industrial green energy efficiency presented an obvious polarization phenomenon, and the gap between provinces has not narrowed. In addition, from the figure above, the wave peak near 0.1 was significantly higher than that near 1, indicating that there were significantly more provinces with low efficiency than with high efficiency. In addition, it is worth mentioning that the kurtosis of the two peaks of 2019 was significantly lower than the crest of the other three years and appears to have a small crest near 2.1, which indicates that in 2019, the low values of the data series were relatively few, the polarization phenomenon was relatively slow, and there were more efficient values. This result indicates that the energy efficiency of 2019 has significantly improved compared with previous years.

Global spatial autocorrelation analysis of green energy efficiency

Global spatial autocorrelation analysis primarily studies the overall spatial distribution of the object through its location and performance in the whole space. Scholars often use global Moran's I to express this, and the value range is [-1, 1]. Generally, when the significance level is certain, if Moran's I value is greater than 0, it indicates that units with similar green innovation efficiency show agglomeration, and the larger the value is, the higher the agglomeration degree is. If Moran's I value is less than 0, it indicates that the units with similar green innovation efficiency show dispersion, and the smaller the value is, the higher the dispersion degree is. If Moran's I value is equal to 0, it indicates that units with similar green innovation efficiency present a completely uniform distribution, and there is no spatial correlation at all, which only exists in theoretical assumptions. In practice, the closer the Moran's I value is to 0, the more uniform the distribution and the weaker the spatial correlation. The test results of the global spatial autocorrelation analysis of the green energy efficiency of provinces and cities in China are shown in Table 3.

Table 3.

Spatial autocorrelation test of the dynamic panel model.

Year	Moran’s value	Standard error	Year	Moran’s value	Standard error
2010	0.052*	0.083	2015	0.082**	0.085
2011	0.077***	0.076	2016	0.064***	0.067
2012	0.049*	0.072	2017	0.082**	0.116
2013	0.077**	0.081	2018	0.083**	0.074
2014	0.065**	0.07	2019	0.045*	0.059

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

From Table 3, the green energy efficiency of all provinces and cities in China has an obvious positive spatial agglomeration phenomenon, which is significant at the 10% level. This means that the spatial distribution of the energy efficiency of provinces and cities in China does not show a completely random state but a spatial agglomeration between similar values, that is, regions with higher efficiency are adjacent to each other, and regions with lower efficiency are the same. Therefore, China's energy efficiency research cannot ignore the factor of the objective existence of “economy - geographical” spatial distribution. The spatial econometric analysis method for China's energy efficiency research can consider the spatial effect, namely, spatial dependence and heterogeneity. Therefore, using the spatial econometric model to research China's energy efficiency was inevitable.

Spatial economic analysis of the impact of eco-environmental regulation on energy efficiency in China

Research method selection

The theory of spatial economics is based on the first law of geography, that is, a certain economic geographical phenomenon or an attribute value in a spatial unit of a region is related to the neighbouring region.⁷ The spatial lag model (SLM), spatial error model (SEM) and spatial Durbin model (SDM) can be used to show the spatial effects of spatial correlation.⁴⁶ Among these models, the spatial Durbin economic model considers the spatial correlation of not only dependent variables but also independent variables. The dependent variables are affected not only by the independent variables of the local region but also by lagged independent and dependent variables of other regions. Therefore, to test the impact of eco-environmental regulation on energy efficiency and the spatial spillover effect in various regions of China more scientifically, this study uses the spatial Durbin measurement model. The formula is as follows:

Y = ρ W Y + X β + W X θ + ε

In the formula, Y is the value of green energy efficiency, W is the spatial weight matrix, X is all the influencing factors,

W X

is the lag terms, and

ε

is the random disturbance term independent of region and time. With the introduction of the spatial weight matrix, the spatial econometric model has a nonlinear structure, so the regression coefficient no longer reflects only the influence of independent variables on dependent variables. Lesage & Pace⁴⁷ gave the definition of the parameters of the spatial Durbin econometric model in the form of a partial derivative matrix and proposed the concepts of total effect, direct effect, and indirect effect. The total effect represents the overall impact of independent variables on dependent variables in all regions, the direct effect represents the impact of local regions, and the indirect effect represents the impact of other regions, which can also be called the spillover effect.

Next, we introduce the establishment of the spatial weight matrix. Research on the spatial dependence or autocorrelation of green energy efficiency among provinces and cities in China is based on the geographical distance and adjacency relationship. The adjacency relationship is generally expressed by 0 and 1. If the number is 0, locations are not adjacent, and 1 indicates that they are adjacent. This assumption ignores the effect of actual geographical distance between regions. Therefore, scientists have also developed a spatial adjacency relationship based on the distance function. Under this definition, the spatial weight matrix is defined such that the influence of the two provinces is greater when the spatial distance between them is closer, while the influence is smaller when the distance is greater. The elements of the spatial weight matrix based on the distance function are defined as follows:

W_{i j}^{*} = {\begin{matrix} 0 i = j \\ \frac{1}{d_{i j}} i \neq j \end{matrix}

In the above formula,

d_{i j}

is the distance between the barycentre points of provinces i and j.

Empirical analysis

Based on the understanding of green energy efficiency and referencing the previous literature, we believe that the main factors affecting regional green energy efficiency include industrial structure and concentration degree, foreign direct investment, and manufacturing scale, in addition to environmental regulation (see Table 4).

Table 4.

Variable description.

Type	Name	Code	Description
Explained variable	Green energy efficiency	GEE	Based on the above calculation results
Main explanatory variable	Environmental regulation	REGU	Proportion of investment in industrial pollution control to total manufacturing output value
Control variable	Industrial structure	STR	Proportion of GDP of secondary industry to whole
	Foreign direct investment	FDI	Proportion of total imports and exports to total output value of manufactured goods
	Degree of industrial agglomeration	AGGLO	Calculated by location entropy
	Scale of manufacturing enterprises	SCALE	The value of total manufacturing output divided the number of industrial enterprises

According to the above table, the spatial Durbin dynamic panel model of influencing factors of green energy efficiency is constructed in this paper as follows:

\begin{aligned} G E E_{i t} = β_{1} R E G U_{i t} + β_{2} S T R_{i t} + β_{3} F D I_{i t} \\ + β_{4} A G G L O_{i t} + β_{5} S C A L E_{i t} + ε_{i t} \end{aligned}

In addition,

β_{1} \sim β_{5}

are unestimated coefficients, and

ε_{i t}

is a random error term. If a common panel model is used for verification, the estimation process may have endogeneity issues. Therefore, to obtain more robust research results, this study adopts the spatial Durbin dynamic panel model for testing, which is very suitable for panel data with strong endogeneity, wide cross-sections and short time series.

The global Moran's I test showed that China's green energy efficiency had spatial correlation. Therefore, this paper adopts the method of spatial economics to test the relationship between the two. In addition, the following tests need to be done before determining the model. First, the autocorrelation results of the regression of the ordinary panel data model (OLS) showed that the p statistics of LM lag, robust LM lag, LM error and robust LM error all rejected the null hypothesis under the 5% assumption, which further supported the reasonable construction of the spatial econometric model. Second, through the LR likelihood ratio test, it was found that all variables passed the significance test under 5%, which means that the statistics of the spatial Durbin model (SDM) were more suitable than those of the spatial lag model (SLM) and spatial error model (SEM), and individual effects and time effects exist Finally, the Hausman test showed that the fixed-effect model was better than the random effect model. In summary, the spatial Durbin model with bidirectional fixed effects was selected to study the spatial effect. The results of the inspection process are shown in Table 5.

Table 5.

Spatial lag and error test of the dynamic panel model.

Test name	Statistic value	Low probability P value
LM-lag	5.5514	0
Robust LM-lag	4.8671	0
LM-error	0.03371	0.0218
Robust LM-error	0.037	0.0161
LR-SDM-SEM	9.51	0.016
LR-SDM-SAR	11.00	0.034
LR-both-time	153.66	0
LR-both-spatial	22.109	0.0414
Hausman	31.83	0

Then, based on the spatial Durbin model, this study uses the maximum likelihood method to estimate the impact of eco-environmental constraints on energy efficiency in China during 2010–2019. The estimated results are shown in Table 6.

Table 6.

SDM estimation results of spatial panel data.

Variable	Spatial fixed effects model		Time fixed effects model		Fixed effect of space and time model
Variable	Coefficient	T statistic	Coefficient	T statistic	Coefficient	T statistic
REGU	−6.96**	−0.87	−22.98**	−1.38	−43.46*	−1.93
STR	0.000000603***	1	0.00000426**	1.68	−0.00000166***	−0.09
FDI	−0.12	−3.46	0.20	7.71	0.29*	6.31
AGGLO	−1.65**	−0.62	−8.81**	−9.37	−7.65*	−3.5
SCALE	0.03**	1.53	0.056***	2.66	0.13*	4.29
_cons	0.36	3.02	0.091	0.54	−0.42	0.09
sigma_u	0.803		0.039		0.333
sigma_e	0.305		0.75		0.751
$ρ$	0.87		0.003		0.164
R²	0.7404		0.6281		0.6412

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

From Table 6, SDM coefficients ( $ρ$ ) were all positive and pass the significance level test of 5% in both time-fixed, spatial-fixed and spatial-time dual fixed models, indicating that green energy efficiency (GEE) presented a spatial agglomeration phenomenon. Among the three models, the spatial fixed model had the highest fitness. Therefore, we only analysed the spatial fixed model.

The estimated results of the Durbin spatial fixed effect in Table 6 show that the coefficient of eco-environmental regulation (REGU) was negative and passes the significance level test of 5%, indicating that eco-environmental constraints were not conducive to the improvement of regional green energy efficiency, which further means that the compliance cost effect of China's current environmental regulations was greater than the innovation compensation effect of industrial energy enterprises. Therefore, government environmental regulations reduced the productivity and profitability of enterprises. From the perspective of total factor production, enterprise energy efficiency declined. Therefore, improving energy efficiency and reducing environmental pollution are dilemmas for China's future development.

The coefficients of industrial structure (STR) and manufacturing enterprise scale (SCALE) were both positive, indicating that they have positive effects on green energy efficiency. The reason may be that with the increase of the industry proportion and enterprise scale, some important elements like capital, labour, energy and others can be better coordinated. In particular, the potential of machinery and equipment has been further developed, which has led to increasing returns to scale and energy intensification, and energy efficiency has been improved.

The coefficients of foreign direct investment (FDI) and industrial spatial agglomeration (AGGLO) were negative, indicating that they had a negative effect on green energy efficiency. The former proved the existence of the “pollution haven” phenomenon in China, in which economically less developed areas often choose loose environmental regulation, creating a comparative advantage in energy trade products and promoting high pollution industries to use relatively low environmental control technology for production to save costs, which further drives high pollution industry to accumulate in the less developed areas. This trend shows that vicious competition among local governments in China for liquid resources comes at the expense of energy efficiency and environmental deregulation, which is alarming. The latter indicates that the agglomeration of industrial enterprises does not help energy conservation and emission reduction. It can be hypothesized that faced with drastic competition, local governments in China attract enterprises to form neck-deep industrial agglomerations by administrative measures to pursue “policy rent”. In addition, clustered enterprises tend to pursue economic benefits rather than environmental benefits, which leads to more repetitive construction and low-level capacity expansion and results in increased energy consumption and a decline in efficiency. In addition, we believe that China's low-level industrial agglomeration does not produce a good technology spillover effect, leading to agglomeration that cannot be good for energy efficiency.

In the nonspatial econometric model, the regression coefficient of the explanatory variable indicates its influence on the explained variable. However, in the spatial econometric model, the regression coefficient of the explanatory variable includes not only direct influence but also indirect influence (namely, the spillover effect). The impacts were decomposed based on the SDM partial differential method (Table 7). The total effects can be divided into two parts. The first is the direct effect, namely, the local effect, which represents the impacts of explanatory variables in the region on its own green energy efficiency. The second is the indirect effect, also named the spillover effect, which represents the influence of explanatory variables in this region on green energy efficiency in adjacent regions.

Table 7.

Direct and indirect (spillover) effects of SDM results.

Effect	GEE	Coef.	Std. Err.	Z	P > \|z\|	[95% Conf. Interval]
LR_Direct	REGU	6.472	11.4139	−0.57	0.571	−28.8427	15.8988
	STR	0	0	0.19	0.85	0	0
	FDI	−0.1193	0.0711	−1.68	0.093	−0.2587	0.02
	AGGLO	−5.7435	5.2945	−1.08	0.278	−16.1206	4.6336
	SCALE	0.0584	0.0185	3.17	0.002	0.0223	0.0946
LR_Indirect	REGU	−9.4451	24.6444	0.38	0.702	−38.857	57.7473
	STR	0	0	0.6	0.548	0	0
	FDI	−0.0196	0.1375	−0.14	0.887	−0.2892	0.25
	AGGLO	14.9218	7.5825	1.97	0.049	0.0603	29.7832
	SCALE	−0.1321	0.0306	−4.32	0	−0.192	−0.0722
LR_Total	REGU	−2.9731	25.2458	0.12	0.906	−46.5077	52.454
	STR	0	0	0.87	0.384	0	0
	FDI	−0.1389	0.1503	−0.92	0.355	−0.4334	0.1556
	AGGLO	9.1783	6.2437	1.47	0.142	−3.0592	21.4157
	SCALE	−0.0737	0.0306	−2.41	0.016	−0.1337	−0.0136

The following observations can be made according to the decomposition results in Table 7. ① The impact benefit of environmental regulation (REGU) on local areas was 6.472, the spillover effect on adjacent areas was − 9.4451, and the total effect was − 2.9732. These results show that the strengthening of environmental regulation was beneficial to the improvement of green energy efficiency in the region but leads to a decline in neighbouring places to a greater extent, resulting in a negative overall effect, which shows that the strengthening of local environmental constraints will force local governments and enterprises to strengthen regulations, forcing enterprises to produce in a cleaner way and leading to the improvement of green energy efficiency in the region. However, such changes also lead to the flow of local enterprises with high pollution and consumption to the surrounding areas, resulting in the decline of green energy efficiency in those areas. According to the regression results, the negative indirect effect (spillover effect) was much larger than the positive direct effect, which is worth the Chinese government's attention. For the Chinese government, future development requires better regional coordination and planning. ② The direct influence, indirect influence and total influence of the industrial structure (STR) were all 0. In fact, it can be seen from the regression results in Table 5 that the coefficient value of this variable was very small. Therefore, the direct and indirect regression results of the SDM model were nearly 0. This result shows that the proportion of the secondary industry had a negligible impact on green energy efficiency because in China, all regions with large or small proportions of industry use the same industrial production technology and process, which is the embodiment of the standardization of China's industrial production. ③ The direct effect (-0.1193), indirect effect (-0.0196) and total effect (-0.1389) of foreign direct investment (FDI) were all less than 0, indicating that FDI had a negative impact on green energy efficiency. China was still at a low position in the international industrial chain, and the industry structure still belongs to traditional modes, which were characterized by high consumption and pollution. Foreign direct investment mainly invests in these low value-added primary and extensive industries, which leads to an increase in foreign direct investment but does not help to improve green energy efficiency. ④ The direct effect of the industrial agglomeration degree (AGGLO) was negative (-5.7435), but the indirect effect was positive and has a large coefficient value (14.9218). Therefore, the total effect was positive (9.1783), which indicates that the regional industry concentration had a negative effect on the local green energy efficiency value but a positive impact on surrounding areas. This result shows that there was a phenomenon of “low-level agglomeration” in China's industrial layout. Currently, traditional industries account for too much of China's overall industrial layout, while the new manufacturing industry has not been developed sufficiently due to its late start. Most of the products of traditional industries had low technological content and added value. They were generally at the “middle and low section” of the value chain and were mainly engaged in original equipment manufacture or the production of spare parts and peripheral components. Without independent technology and management skills, they could only rely on resource consumption and ecological destruction to survive. In particular, the negative environmental externalities brought by the agglomeration of traditional manufacturing industries with high pollution, such as steel, building materials and the chemical industry, cover up the positive effect on ecological improvement, resulting in the status quo of “more agglomeration, more pollution” from industrial enterprises. ⑤ For manufacturing firm scale, the direct effect was positive (0.0584), the indirect effect was negative (-0.1321), and the total effect was negative (-0.0737). Generally, with an increase in company scale, the technical efficiency value can be improved by the economic scale effect to increase regional green energy efficiency but is not conducive to the surrounding areas, which also illustrates the urgent need for an overall regional plan in China.

Conclusions

(1) From 2010 to 2019, although China's green energy efficiency performance improved significantly, the overall situation was still relatively poor. The efficiency value has a great correlation with regional economic development. (2) Eco-environmental constraints, foreign direct investment and industrial agglomeration are not conducive to regional green energy efficiency improvement. Industrial structure and manufacturing enterprise scale have positive effects on efficiency. (3) The effect of government regulation caused by environmental constraints on green energy efficiency in local regions was positive, but the spillover effect on adjacent land was negative, and the total effect was negative. The three kinds of influence are 0 of industrial structure and less than 0 of foreign direct investment. The direct effect of industrial agglomeration was negative, but the indirect and total effects were positive. At the enterprise scale, the direct effect was positive, and the indirect effect and the total effect were negative.

Therefore, we make the following suggestions.

More effort should be made to develop energy conservation and emission reduction technologies to improve energy consumption efficiency. According to the calculation results of the DEA model, China's green energy efficiency value was generally low, and waste was a serious problem because China’s energy industry technological level was less developed.⁵ Therefore, the Chinese government needs to encourage energy companies to optimize their structure through mergers and acquisitions and close outdated production capacity. That is, it must not only increase the development of related technologies, innovation and new products to strengthen the transformation and application of technological achievements but also weed out outdated production facilities such as power companies through mergers and reorganization. The limited energy resources can be used more scientifically and rationally.

China must establish reasonable environmental regulation methods to promote enterprise production in a cleaner way. Intensifying the reversed transmission mechanism of the environment on energy efficiency, the economy and energy control will force enterprises to upgrade technology and transform the economic development mode. China should strengthen regional evaluation and assessment and make energy conservation and emission reduction targets more binding. In addition, environmental regulation policy formulation should pay more attention to applicability, take measures according to local conditions, and take full consideration of the local environmental carrying capacity and energy resource endowment to further improve policy effectiveness.

Reasonable regional development plans should be formulated to promote the rational distribution and flow of the energy industry across regions. From the regression results of the SDM model, we concluded that most economic variables have obvious spillover effects, which requires the government to establish a reasonable cross-regional development plan. However, this is not a simple task. The cross-regional development mechanism of industrial synergy has an extremely strict logic mechanism and system coordination. A good, harmonious mechanism needs the government, enterprises and society to work closely together. However, the government is bound to play a leading role in formulating scientific regional and industrial development policies. Additionally, enterprises should strive to break regional barriers, exploit the market, seek cooperation, and enhance competition in cross-regional flow. Society should strengthen the construction of such a cross-regional development mechanism, reinforce road traffic and information exchange construction between regions, and create a good environment for enterprise flow.

Of course, there is further space for improvement in this article. Optimization and breakthrough can still be made in the index system design, so as to make the measurement and evaluation of green energy efficiency more scientific and reasonable. Although the index selected in this paper is representative, but the measured efficiency numbers cannot entirely represent the development status of green energy. In addition, in order to reflect the differences among regions and provinces of the whole country, the index with the greatest influence on each province was not differentiated, which may affect the objectivity of the results. Therefore, the design of the index system can be optimized in the future research.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (Grant No. 72171124; Grant No. 71771126), the General research projects of philosophy and Social Sciences in Colleges and Universities of Jiangsu Province (Grant No. 2021SJA0107) and Social Science Research Foundation of Nanjing University of Posts and Telecommunications (Grant No. NYY220006).

ORCID iD

Mingran Wu

Mingran Wu is an Assistant Professor of School of Management, Nanjing University of Posts and Telecommunications. His main research interests are sustainable development.

References

Yang

Chen

. Opportunities and challenges of energy industry under the background of ‘carbon emission peak’and ‘carbon neutrality’. Natural Resource Economics of China 2021; 12: 1–10.

Poon

JPH

Casas

. The impact of energy, transport and trade on Air pollution in China. Eurasian Geography and Economics 2006; 47: 568–584.

Chuai

Huang

Wang

, et al. Spatial econometric analysis of carbon emissions from energy consumption in China. J Geog Sci 2012; 22: 630–642.

Arouri

MEH

Youssef

M’ henna

, et al. Energy consumption, economic growth and CO₂ emissions in Middle East and north african countries. Energy Policy 2012; 45: 342–349.

Zhao

. The coordination and dynamic relationship of the industrial energy consumption, economic growth and SO2 emission—analysis of time series data of China from 1990 to 2016. Environmental Processes 2018; 5: 789–806.

Song

Zhou

, et al. Does environmental regulation affect energy efficiency in China’s thermal power generation? Empirical evidence from a slacks---based DEA model. Energy Policy 2014; 66: 537–546.

Measurement and spatial statistical analysis of green science and technology innovation efficiency among Chinese provinces. Environ Ecol Stat 2021; 28: 423–444.

Janssens

Crippa

Guizzardi

, et al. Fossil CO₂ & GHG emissions of All world countries [M]. Luxembourg: Publications Office of the European Union, 2017.

Ren

. Coupling analysis of urbanization and ecological total factor energy efficiency — A case study from hebei province in China. Sustainable Cities and Society 2021; 74: 103183.

10.

Zhao

. Evaluation of development level and economic contribution ratio of science and technology innovation in eastern China. Technol Soc 2019; 11: 101194.

11.

. Experience and inspiration of Japanese low-carbon technology innovation. Enterprise Management 2021; 6: 15–19.

12.

Yang

Chen

. ‘The 14th five-year plan’ promotes energy transformation and achieves the peak of carbon emissions. Yuejiang Academic Journal 2021; 13: 73–85 + 124.

13.

Ernst & Young Carbon Neutralization Research Group. Read a book to understand carbon neutrality. Beijing: China Machine Press, 2021, pp.51-52.

14.

Zhao

. An evaluation and variation analysis of sustainable development capacity in different regions of China. Int J Environ Technol Manage 2020; 23: 397–413.

15.

Porter

. America’s green strategy. Sci Am 1991; 164: 168–170.

16.

Porter

Vander Linde

. Toward a new conception of the environment-competitiveness relationship. J Econ Perspect 1995; 9: 97–118.

17.

Lin

Polenske

. Input-Output anatomy of China’s energy Use changes in the 1980’s. Economic System Research 1995; 7: 32–45.

18.

Fisher-Vanden

Jefferson

Liu

, et al.

What is driving China’s decline in energy improvement?

Resource and Energy Economics 2004; 26: 2287–2295.

19.

Richard

Edward

. Environmental regulations and innovation activity in UK manufacturing industries. Resource and Energy Economics 2012; 34: 211–235.

20.

Herrerias

Cuadros

Luo

. Foreign versus indigenous innovation and energy intensity: further research across Chinese regions. Appl Energy 2016; 162: 1374–1384.

21.

Cagno

Ramirez-Portilla

Trianni

. Linking energy elficiency and innovation practices: empirical evidence from the foundry sector. Energy Policy 2015; 83: 240–256.

22.

Otsuka

Goto

Sueyoshinergy

. Efficiency and agglomeration economics: the case of Japanese manufacturing industries. Regional Science Policy & Practice 2014; 6: 195–212.

23.

Ramirez-Portilla

Cagno

Trianni

. Is innovation an enabler of energy elficiency? An exploratory study of the foundry sector. Energy Procedia 2014; 61: 1191–1195.

24.

. Can internet development improve the energy efficiency of firms: empirical evidence from China. Energy 2021; 237: 121590.

25.

Chen

Lin

. Understanding the green total factor energy efficiency gap between regional manufacturing—insight from infrastructure development. Energy 2021; 237: 121553.

26.

Duttan

. Impact of environmental regulation on technical efficiency: a study of chemical industry in and around mumbai, science. Technology & Society 2011; 16: 333–350.

27.

Brunnermeier

Cohen

. Determinants of environmental innovation in US manufacturing industries. J Environ Econ Manage 2003; 45: 278–293.

28.

Hancevic

. Environmental regulation and productivity：The case of electricity generation under the CAAA-1990. Energy Economics 2016; 60: 131–143.

29.

Tan

Liu

, et al. Property rights protection, environmental regulation and corporate financial performance: revisiting the porter hypothesis. J Cleaner Prod 2020; 264: 121615.

30.

Sabuj Kumar

Bhandari

. Does Win-Win opportunity always result in pollution abatement? An application of directional distance function to Indian cement industry. The Journal of Developing Areas 2020; 54: 161–174.

31.

Hazarika

Zhang

. Factors that drive and sustain eco-innovation in the construction industry: the case of Hong Kong. J Cleaner Prod 2019; 238: 117816.

32.

Kyle

Xiang

. Induced innovation in clean energy technologies from foreign environmental policy stringency? Technol Forecast Soc Change 2019; 147: 198–207.

33.

Wang

Sun

Guo

. Environmental regulation and green productivity growth: empirical evidence on the porter hypothesis from OECD industrial sectors. Energy Policy 2019; 132: 611–619.

34.

Andrea

Guarini

Meliciani

. Green patents, regulatory policies and research network policies. Res Policy 2018; 47: 1537–1542.

35.

Zodrow

. Grandfather rules and the theory of optimal tax reform, North-Holland, 1992; 49: 163–190.

36.

Abdelfeteh

. Environmental regulation and eco-innovation: the porter hypothesis refined. Eurasian Business Review 2018; 8: 299–321.

37.

Qiao

Shi

. Has China's war on pollution slowed down the growth of its manufacturing and by how much? Evidence from the clean Air action. China Economic Review 2018; 53: 271–289.

38.

Grazi

van den Bergh

Waisman

. A simple model of agglomeration economies with environmental externalities, AFD Research Papers, No. 2016-18, French Agency for Development. 2016.

39.

Rodriguez

Dupont-Courtade

Oueslat

. Air pollution and urban structure linkage evidence from european cities. Renewable Sustainable Energy Rev 2016; 53: 1–9.

40.

Umut

Kemal

. Is foreign direct investment an engine for energy consumption? An empirical investigation for Turkey. Environmental Science and Pollution Research 2019; 26: 28092–28105.

41.

Lin

Wang

. How to boost energy productivity in China’s industrial sector: an integrated decomposition framework based on multi-dimensional factors. J Cleaner Prod 2020; 259: 120902.

42.

Agnihotri

Kumar

Dinesh

, et al. Foamy slag practice to enhance the energy efficiency of electric arc furnace: an industrial scale validation. Materials Today: Proceedings 2021; 46: 1537–1542.

43.

Sun

Zhang

Wang

. Evaluation of regional carbon emissions performance based on SE-SBM model: taking shandong province as An example. Ecological Economy 2016; 32: 68–73.

44.

Tone

. A slacks-based measure of efficiency in data envelopment analysis. Eur J Oper Res 2002; 143: 32–41.

45.

Peng

. China’s total factor energy efficiency evaluation: based on three stage global undesirable-hybrid-slack-based –model. Economic Problems 2020; 1: 11–19.

46.

Zhao

Sun

Zheng

. Water resource utilization efficiency and its spatial spillover effects measure in China. Acta Geographica Sinica 2014; 69: 121–133.

47.

Lesage

Pace

. Introduction to spatial econometrics. New York: CRC Press, 2009, pp. 27-41.

The impact of eco-environmental regulation on green energy efficiency in China - Based on spatial economic analysis

Abstract

Keywords

Introduction

Literature review

Calculation and dynamic analysis of green energy efficiency

Super-efficiency SBM model

Variable definition and data sorting

Calculation of super-efficiency of energy consumption

Dynamic evolution analysis of green energy efficiency

Global spatial autocorrelation analysis of green energy efficiency

Spatial economic analysis of the impact of eco-environmental regulation on energy efficiency in China

Research method selection

Empirical analysis

Conclusions

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

References