Impact of Environmental Regulation and Industrial Agglomeration on Carbon Emissions in the Yangtze River Economic Belt

Abstract

Carbon reduction is an important aspect of achieving high-quality environmental development; environmental regulation and industrial agglomeration are important ways to affect carbon emissions. Therefore, studying the relationship between industrial agglomeration, environmental regulation, and carbon emissions has important theoretical and practical significance. Firstly, this article adopts the method of location entropy to measure the indicators of industrial agglomeration. Secondly, it proposes an environmental regulation indicator system based on the experience of previous scholars and measures the environmental regulation indicators using the entropy method. Next, eight types of energy consumption are used to measure carbon emissions based on the carbon emission coefficient method provided by the IPCC. Finally, based on the Moran index, the spatial correlation of carbon emission levels in various regions of the Yangtze River Economic Belt (YREB) is tested. A spatial econometric model was introduced to explore the relationship between industrial agglomeration, environmental regulation, and carbon emissions at a deeper level, and the following conclusions were drawn: (1) The regression coefficient of the spatial term of industrial agglomeration on carbon emissions is 0.848, which is significantly positive at the 10% level, indicating that under the influence of spatial effects, industrial agglomeration has a significant promoting effect on carbon emissions. (2) The regression coefficient of the spatial term of environmental regulation on carbon emissions is −0.011, which is significantly negative at the 10% level, indicating that environmental regulation has an inhibitory effect on carbon emissions under the influence of spatial effects. Based on the above conclusions, useful suggestions have been provided for optimizing industrial structure, improving environmental regulation levels, and alleviating carbon emission issues.

1. Introduction

Since the reform and opening up, China’s rapid economic growth has led to a sharp increase in energy consumption and carbon emissions [1]. The excessive emission of greenhouse gases, led by carbon dioxide, is an important cause of global ecological environment deterioration, causing a significant burden on resources and environmental carrying capacity. This contradicts China’s goal of “accelerating ecological civilization system reform and building a beautiful China”, and the situation of strict control of carbon dioxide emissions is very severe [2]. In 2020, at the 75th United Nations General Assembly, the Chinese government proposed that carbon dioxide emissions should strive to reach a peak by 2030 and strive to achieve carbon neutrality by 2060 [3,4]. The proposal of the goal of carbon peaking and carbon neutrality has raised China’s green development path to a new height, and green will become the background of China’s social and economic development in the coming decades. As the world’s largest developing country and the largest emitter of carbon dioxide, China has actively carried out multiple carbon emission reduction practices and adopted measures such as building a national carbon emission trading market and piloting low-carbon cities to promote energy conservation, reduce emissions, and promote low-carbon economic transformation [5]. However, the issue of global warming is becoming increasingly severe, and the demand for energy in urbanization and industrialization is still increasing, making China face enormous pressure to reduce carbon emissions. For the context of China’s national conditions and dual carbon policy, it is particularly important to complete the green and low-carbon transformation task in the process of achieving industrialization and urbanization and embark on a low-carbon development path that considers the constantly expanding energy demand [6].

With the increasingly severe issues of global warming and carbon emissions, many scholars have also conducted in-depth research on the influencing factors of carbon emissions. Many studies have shown that the impact of industrial agglomeration and environmental regulations on carbon emissions is becoming increasingly close. Industrial agglomeration refers to a process in which the same industry is highly concentrated in a specific geographical area, and industrial capital elements continuously converge within a spatial range. However, during the process of industrial agglomeration, environmental problems may also arise to some extent [7]. For example, Li et al. demonstrated that industrial agglomeration has a significant positive effect on environmental pollution [8]. Chen et al. proposed that when industrial agglomeration reaches a certain scale, the spillover effects of infrastructure, such as knowledge and technology, gradually form, contributing to environmental pollution control. Excessive industrial agglomeration often leads to congestion effects, thereby increasing carbon emissions [9]. Some studies have also found that industrial agglomeration affects carbon productivity through scale, technology spillover, and competition effects. Environmental regulations are a green policy tool that, as an ecological constraint on resources, mainly play an important role in coordinating the allocation of production factors, adjusting the industrial structure, and innovating green economy transformation [10,11,12]. In terms of environmental regulation and carbon emissions, the “Industrial Carbon Peak Implementation Plan” proposes relevant implementation plans for environmental regulation, indicating that the impact of environmental regulation on carbon emissions cannot be underestimated. Wang et al. used Chinese cities as research samples to explore the direct and indirect effects of environmental regulations on carbon emissions. They believe that environmental regulations can effectively suppress carbon emissions by forcing industrial structure upgrading and technological innovation [13]. Smulders et al. believe that the transition period before the formal implementation of carbon emission taxes often increases fossil energy consumption, thereby increasing carbon emissions and creating a green paradox effect. According to the research of relevant scholars, both environmental regulation and industrial agglomeration have an important impact on carbon emissions [14]. Therefore, this study on the spatial effects of industrial agglomeration and environmental regulation on carbon emissions can explore the internal relationship between the three and meet the goals of the dual carbon policy.

Based on current research, both industrial agglomeration and environmental regulation have a certain impact on environmental pollution and carbon emissions. Therefore, this study on the relationship between these three factors is of great significance for the current research system and the implementation of dual carbon policies. In addition, this article has the following innovative points in the research process: firstly, in terms of research content, the article studies the impact of industrial agglomeration and environmental regulation on carbon emissions from a dual perspective, expanding the analysis of factors affecting carbon emissions. Secondly, in terms of indicator selection, this article combines the experience of scholars and adopts comprehensive indicators for measurement, making the data more authentic and stable. Thirdly, in terms of research objects, the research object of the article is the important strategic development region of China—the Yangtze River Economic Belt (YREB) (Figure 1). The YREB spans the three major regions of east, west, and central China and is one of the three major strategies implemented by the central government. It is a globally influential inland economic belt, a coordinated development belt for interaction and cooperation between the east, center, and west, and a comprehensive opening belt for both domestic and foreign development along the coast, rivers, and borders. It is also a leading demonstration belt for ecological civilization construction [15,16]. Therefore, conducting research on it has certain representativeness and can provide certain guidance for carbon reduction in China.

This paper is structured as follows: Section 2 is a literature review; Section 3 is the research methods, models, and data; Section 4 is the experimental section; and Section 5 is the conclusions and recommendations.

2. Literature Review

2.1. Industrial Agglomeration and Carbon Emissions

Industrial agglomeration and the adjustment and upgrading of the industrial structure are important factors affecting carbon emissions and environmental pollution. Studies related to industrial agglomeration should be based on measuring industrial agglomeration levels. Currently, many methods exist to measure industrial agglomeration, such as location entropy, Gini coefficient, E-G index, etc. [17,18,19]. What is the impact of industrial agglomeration on carbon emissions? Different schools of thought have different interpretations, mainly including the following three categories. The first viewpoint is that industrial agglomeration has an inhibitory effect on carbon emissions. Lu and Zhu studied the impact of industrial agglomeration on carbon emissions under the background of widespread excessive government intervention in industrial agglomeration in China. They found that industrial agglomeration could reduce carbon emission intensity [20]. Han et al. believe that industrial agglomeration and economic agglomeration can promote technological upgrading, thereby improving the utilization of energy efficiency and reducing carbon emissions [21]. Ding found from a spatial perspective that industrial collaborative agglomeration can effectively promote carbon emission reduction and play a positive role in promoting regional green development [22]. The second viewpoint is that there is a positive correlation between industrial agglomeration and carbon emissions. Chen et al. believe that agglomeration will cause the expansion of the industrial scale, increase energy consumption, and further lead to an increase in carbon dioxide emissions [23]. Li and Zhang et al. also believe that the agglomeration of productive and service industries will cause environmental effects, especially the greenhouse effect, which is not conducive to carbon emission reduction [24,25]. The third viewpoint is that the relationship between industrial agglomeration and environmental pollution or carbon emissions is not a simple linear relationship, but a non-linear relationship. Wang proposed a nonlinear relationship between industrial agglomeration and sulfur dioxide emission intensity through empirical research [26]. Wu et al. analyzed the differences in the impact of industrial agglomeration on urban carbon dioxide emissions from three aspects: scale, composition, and technological effects [27]. When discussing the relationship between industrial agglomeration and carbon intensity in China, Li and Liu found an inverted U-shaped relationship between industrial agglomeration and carbon emissions, and industrial agglomeration promoted the reduction of carbon emissions through technological progress [28].

2.2. Environmental Regulation and Carbon Emissions

Environmental regulation aims to protect the environment and regulate various behaviors that pollute the public environment. Therefore, an important factor affecting carbon emissions is environmental regulation. There are three main viewpoints on research on environmental regulation and carbon emissions. The first viewpoint is that environmental regulation can effectively promote carbon reduction. Xu et al. proposed that environmental regulations are an important measure for the government to control environmental pollution and an important component of environmental regulations. Therefore, environmental regulations can effectively promote carbon reduction [29]. Hashmi and Alam proposed that the effective implementation of environmental regulations and the adoption of green technologies are the main catalysts for mitigating global warming trends and reshaping carbon reduction strategies [30]. Bu et al. believe that implementing environmental regulations has curbed pollution and improved environmental quality, so environmental regulations play a positive role in environmental protection [31]. The second viewpoint is that there is a green paradox phenomenon in the impact of environmental regulations on carbon emissions, which means that environmental regulations can lead to an increase in carbon emissions. For example, Lai et al. found through data from 30 provinces in China that under the influence of market segmentation, environmental regulations can actually lead to an increase in carbon emissions [32]. Huang et al. also found that there is a green paradox in the impact of environmental regulation on carbon emissions through the study of China’s provincial panel data [33]. The third viewpoint is that the impact mechanism of environmental regulation on carbon emissions is complex. Hu and Wang found that there is a threshold for the impact of environmental regulation on carbon productivity by studying the relationship between environmental regulation and carbon emissions [34]. Pei et al., using China’s provincial panel data from 2005 to 2015, found that for the entire energy-intensive industry group, environmental regulation can not only potentially directly reduce carbon emissions, but can also indirectly reduce carbon emissions through technical efficiency [35]. Some scholars have found that environmental regulations with different intensities have varying degrees of impact on carbon emissions. For example, Wang et al. found, based on the Dynamic Comprehensive Measure, that the relationship between concentrated manufacturing industries with high environmental regulation intensity and carbon emissions shows an inverted U-shaped trend; in contrast, agglomeration manufacturing industries with lower environmental regulation intensity are more likely to exhibit a positive U-shaped trend in their relationship with carbon emissions [36].

2.3. Literature Summary

Throughout existing research, most of the literature has conducted in-depth research on the impact of industrial agglomeration or environmental regulation on carbon emissions and has formed basic research directions and mainstream perspectives. However, there are still the following shortcomings. Firstly, the existing research on environmental regulations and quantitative indicators of carbon emissions lacks a certain comprehensive and standardized system. Secondly, there is very little literature on the impact of industrial agglomeration, environmental regulation, and carbon emissions from the dual perspectives of industrial agglomeration and environmental regulation on carbon emissions within the same research framework. Thirdly, the current research system lacks an in-depth analysis of the impact of industrial agglomeration and environmental regulations on the spatial effects of carbon emissions. In view of the above shortcomings, this paper takes the YREB as the research object, improves the measurement methods of relevant indicators, and deeply examines the internal correlation among carbon emissions, industrial agglomeration, and environmental regulation. This has important practical significance and practical value for improving inter-regional industrial agglomeration, reducing environmental pollution and carbon emission intensity, and providing useful policy enlightenment for realizing national carbon neutrality and peak goals.

3. Models, Methods, and Data

This section mainly introduces the measurement methods of various variables, as well as the models and data used in the article. This article measures the explanatory variable and the dependent variable based on the indicator measurement analysis method. It uses the measured data to conduct a significance analysis of the spatial econometric model. The specific plan is as follows: Section 3.1 introduces the use of location entropy to measure industrial agglomeration. Section 3.2 introduces the use of the entropy method to measure environmental regulation indicators. Section 3.3 introduces the calculation formula for the carbon emission coefficient. Section 3.4 introduces the spatial econometric model used in the article. Section 3.5 is an explanation of the data sources.

3.1. Industrial Agglomeration and Location Entropy

The indexing measurement of industrial agglomeration includes various methods, such as Elision-Glaeser (E-G) index, industry concentration, location entropy index, and spatial Gini coefficient. By comparing the advantages, disadvantages, and feasibility of each measurement index, this paper adopts the location entropy index to measure the agglomeration level [37,38]. Location entropy is a meaningful index in measuring the spatial distribution of factors in a certain region, reflecting the specialization degree of an industrial sector and the status and role of a certain region in a high-level region. The specific formula is as follows:

$$A{\mathrm{gg}}_{it}=\frac{q_{ij}/q_j}{q_i/q}$$

(1)

where ${A\mathrm{g}\mathrm{g}}_{it}$ represents the industrial agglomeration level of region j in the YREB, $q_{ij}$ is the number of employed people in the manufacturing industry in region j, $q_j$ is the total employed population in region j, $q_i$ is the number of employed people in the manufacturing industry in the YREB, and q is the total employed population in the YREB.

3.2. Entropy Method and Environmental Regulation

The entropy method is a mathematical method used to judge the degree of dispersion of an index. The greater the dispersion degree, the greater the influence of this index on the comprehensive evaluation. The entropy value can be used to judge the dispersion degree of a certain index, which provides a basis for a multi-index system [39,40]. The entropy method is generally divided into four steps.

Step 1: Construct the initial decision matrix. X is the initial decision matrix, m is the number of nodes for suitability evaluation, n is the number of factors, $x_{ij}$ is the value of each sample parameter, i = 0, 1, 2…, m, j = 0, 1…, n.

$$X=\left[\begin{array}{cccc}{x}_{11}& x_{12}& \dots & x_{1\mathrm{n}}\\ x_{21}& x_{22}& \dots & x_{2n}\\ \dots & \dots & \dots & \dots \\ x_{m1}& x_{m2}& \dots & x_{mn}\end{array}\right]$$

(2)

Step 2: Normalize the initial decision matrix. It is necessary to normalize the matrix because the dimension and measures of the data are not consistent. The normalized representation of the decision matrix is Y.

$$Y=\left[\begin{array}{cccc}{y}_{11}& y_{12}& \dots & y_{1\mathrm{n}}\\ y_{21}& y_{22}& \dots & y_{2n}\\ \dots & \dots & \dots & \dots \\ y_{m1}& y_{m2}& \dots & y_{mn}\end{array}\right]$$

(3)

$$y_i{}_j=\frac{(x_i{}_j-{(x_i{}_j)}^j{{}_m}_{in})}{({(x_i{}_j)}^j{{}_m}_{ax}-{(x_i{}_j)}^j{{}_m}_{in})}$$

(4)

Step 3: Calculate the entropy $e_j$ of each factor.

$$p_i{}_j=y_i{}_j/{\displaystyle \sum _{i=1}^m{y}_{ij}}$$

(5)

$$e_j=-\frac{1}{\ln m}{\displaystyle \sum _{i-1}^m{p}_{ij}\ln }{p}_{ij}$$

(6)

Step 4: The weight $w_j$ of each factor can be calculated as follows:

$$w_j=(1-e_j)/{\displaystyle \sum _{j=1}^m(1-e_j)}$$

(7)

In order to profoundly reflect the discriminative ability of indicators and determine a better weight, the entropy weight method is used to measure environmental regulation in this paper. According to the study by Yang (2020) and Du (2020) et al., environmental regulation indicators are mostly expressed by the amount of industrial wastewater discharge, industrial waste gas discharge, and industrial solid waste production [41,42]. In addition, Xie (2022) and Zhao (2020) et al. also proposed the important role of environmental governance investment costs in environmental regulation [43,44]. Based on this, this paper uses the ratio of operation cost of industrial wastewater treatment facilities to industrial wastewater discharge, the ratio of operation cost of industrial waste gas treatment facilities to industrial waste gas discharge, the comprehensive utilization rate of industrial solid waste, and the ratio of environmental pollution control cost to gross industrial product to determine the comprehensive index of environmental regulation. Specific indicators are shown in

Table 1. Composite Index of environmental regulation.

Variable	Level Indicators	The Secondary Indicators
Composite Index of environmental regulation (${Env}_{it}$)	Environmental type index	Operating costs of industrial wastewater treatment facilities/Industrial wastewater discharge
		Operating costs of industrial waste gas treatment facilities/Industrial exhaust emission
		Comprehensive utilization rate of industrial solid waste
	Market type indicator	Cost of environmental pollution control/Gross industrial product

.

3.3. Carbon Emissions

Carbon emissions mainly come from the combustion of fossil fuels. Due to the lack of direct monitoring and statistical data, most studies use the carbon emission coefficient provided by IPCC to estimate the total carbon emissions [45,46]. Considering data availability, this paper selects eight energy sources, namely raw coal, coke, crude oil, gasoline, kerosene, diesel oil, fuel oil, and natural gas, to calculate carbon emissions. It uses per capita carbon emissions CE = C/population as the research variable, and the specific formula is as follows.

$$C={\displaystyle \sum _{i=1}^8{E}_i}\times CE{F}_i$$

(8)

$$CE{F}_i=H_i\times C{H}_i\times CO{R}_i\times \frac{44}{12}\times {10}^{(-6)}$$

(9)

where $E_i$ represents the total consumption of fossil energy i, and the data are from China Energy Statistical Yearbook. ${CEF}_i$ is the carbon emission coefficient of energy i. $H_i$ is the average low calorific value, ${CH}_i$ is the carbon content per unit calorific value, and ${COR}_i$ is the carbon oxidation rate. The final carbon emission coefficient ${CEF}_i$ is shown in

Table 2. Carbon emission coefficients of eight major fossil energy sources.

$\mathbf{Energy} \mathbf{Types} \left({\mathit{E}}_{\mathit{i}}\right)$	$\mathbf{Average} \mathbf{Low} \mathbf{Calorific} \mathbf{Value} \left({\mathit{H}}_{\mathit{i}}\right)$	Carbon per Unit Calorific Value ${\mathit{C}\mathit{H}}_{\mathit{i}}$ (TC/TJ)	$\mathbf{Carbon} \mathbf{Oxidation} \mathbf{Rate} \left({\mathit{C}\mathit{O}\mathit{R}}_{\mathit{i}}\right)$	$\mathbf{Carbon} \mathbf{Emission} \mathbf{Factor} \left({\mathit{C}\mathit{E}\mathit{F}}_{\mathit{i}}\right)$
coal	20,908 kJ/kg	26.37	0.94	1.9003 kgCO2/kg
coke	28,435 kJ/kg	29.5	0.93	2.8604 kgCO2/kg
crude oil	41,816 kJ/kg	20.1	0.98	3.0202 kgCO2/kg
gasoline	43,070 kJ/kg	18.9	0.98	2.9251 kgCO2/kg
kerosene	43,070 kJ/kg	19.5	0.98	3.0179 kgCO2/kg
diesel	42,652 kJ/kg	20.2	0.98	3.0959 kgCO2/kg
fuel oil	41,816 kJ/kg	21.1	0.98	3.1705 kgCO2/kg
natural gas	38,931 kJ/m3	15.3	0.99	2.1622 kgCO2/kg

.

3.4. Spatial Econometric Model

Spatial econometrics is a branch of econometrics that deals with the analysis of spatial interactions (spatial autocorrelation) and spatial structure (spatial inhomogeneity) in regression models of cross-sectional and panel data. Spatial econometric models include the spatial lag model (SAR), spatial error model (SEM), and spatial Dubin model (SDM), which uses the Moran index to test the existence of spatial correlation. The spatial econometric model also introduces the spatial weight matrix to express the spatial strength relationship between regional individuals more intuitively [47,48,49].

3.4.1. Moran Index

To check whether there is spatial dependence or clustering, some spatial autocorrelation tests are needed for analysis. In general, it can be divided into global dependencies and local dependencies. The global dependence is mainly expressed as the global spatial attributes of the observed variables. Spatial autocorrelation is usually expressed by the global Moran’s I index, whose expression is as follows:

$$Moran’sI=\frac{n{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{w}_{ij}(x_i-\overline{x})(x_j-\overline{x})}}}{{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{w}_{ij}{{\displaystyle \sum _{i=1}^n(x_i-\overline{x})}}^2}}}$$

(10)

where $x_i$ and $x_j$ are the values of the selected variables in countries i and j, respectively. $w_{ij}$ represents the elements in the spatial weight matrix.

3.4.2. Spatial Lag Model

Spatial lag model is a statistical model based on spatial dependence, which establishes the model by considering the interactions and dependencies between regions.

$$y_{it}={\mu }_i+{\delta }_t+\rho {\displaystyle \sum _{j=1}^n{w}_{ij}{y}_{jt}+}{\beta }_{xt}+{\epsilon }_{it}$$

(11)

3.4.3. Spatial Error Model

Spatial error model is a statistical model used to explain the statistical behavior of spatial data.

$$y_{it}={\mu }_i+{\delta }_t+{\beta }_{xt}+\lambda {\displaystyle \sum _{j=1}^n{w}_{ij}{\phi }_{jt}+}{\epsilon }_{it}$$

(12)

3.4.4. Spatial Dubin Model

Spatial Dubin model is a combination and expansion form of the spatial lag model and the spatial Error term model, which can be established by adding corresponding constraints to the spatial lag model and the spatial error model.

$$y_{it}={\mu }_i+{\delta }_t+\rho {\displaystyle \sum _{j=1}^n{w}_{ij}{y}_{jt}+\beta x_{it}+{\displaystyle \sum _{j=1}^n{w}_{ij}{y}_{jt}\gamma +}}{\epsilon }_{it}$$

(13)

where $y_{it}$ is the dependent variable, and in this paper, it is represented by carbon emissions per capita. $\rho$ is the spatial autoregressive coefficient, which implies the effect of adjacent regions. $x_{it}$ is an independent variable, including environmental regulation and industrial agglomeration, as well as four control variables, including industrial structure, degree of opening to the outside world, urbanization rate, and government intervention. $\beta$ is the coefficient of vector X and represents the direct effect of the independent variable on the dependent variable. $\gamma$ is the spatial autocorrelation coefficient. ${\phi }_{jt}$ is the error term in the SEM model. $\lambda$ is the spatial autocorrelation coefficient of the error term. $\epsilon$ is an error term, and the element $w_{ij}$ is the spatial weight matrix. W represents the spatial correlation between regions i and j. In general, we stipulate that the diagonal element $w_{ij}$ in the matrix is zero; that is, the region has no spatial correlation with itself.

3.5. Data

In addition to the above descriptions of independent and dependent variables, the paper also introduces control variables. According to the research and discussion of relevant scholars, it is found that the influence of industrial structure, degree of opening to the outside world, urbanization rate, and government intervention on carbon emissions cannot be ignored, so this paper selects these four variables as the control variables of the model (

Table 3. Control variables.

Control Variables	Explain	The Literature
The industrial structure (a1)	Gross product of tertiary industry/gross product of secondary industry	Zhang et al. (2022) [50]
Degree of openness (a2)	Total imports and exports/gross domestic product	Yu et al. (2022) [51]
Urbanization rate (a3)	Urban population/total population	Shi and Li (2018) [52]
Government intervention (a4)	Revenue/gross domestic product	Ahmed and Almukhtar [53]

).

The data for this article were obtained from the following sources: China Statistical Yearbook, Urban Statistical Yearbook, Provincial Statistical Yearbook, and Energy Statistical Yearbook from 2011 to 2021. The data used include the number of employed people in the manufacturing industry, the total employed people, the total amount of three waste emissions and pollution control, the carbon emission coefficient of eight energy sources, the gross product of the secondary industry and the tertiary industry, the total import and export volume, the urbanization rate, and the proportion of fiscal revenue in GDP. Some missing values are supplemented using the interpolation method.

4. Spatial Econometric Analysis

4.1. Weight Matrix and Spatial Autocorrelation Test

The spatial weight matrix contains an adjacency matrix, a geographical distance matrix, a distance matrix, etc. Considering the geographic distance is an important index of spatial pattern, the adjacent provinces and cities also exist between spatial spillover effects. Based on the geographic distance weighting matrix as a model for the empirical analysis of the weighting matrix, the geographically weighted matrix formula is as follows (14), where $d_{ij}$ represents the geographical distance between province i and province j, and $w_{ij}$ is the element of the spatial weight matrix, with the diagonal element being 0.

$$w_{ij}=\frac{1}{d_{ij}}$$

(14)

To determine the spatial weight matrix, it is necessary to test the spatial correlation. The purpose of spatial autocorrelation analysis is to determine whether a variable is spatially correlated and to what extent. The spatial autocorrelation coefficient is often used to quantitatively describe the spatial dependence of things. Specifically, the spatial autocorrelation coefficient measures the spatial distribution characteristics of physical or ecological variables and their impact on the domain. If the values of a variable become more similar with the decrease in the measurement distance, the variable is spatially positively correlated. If the measured value is more different with the decrease in distance, it is spatially negatively correlated. If the measured values do not show any spatial dependence, then the variable shows spatial irrelevance or spatial randomness. Spatial autocorrelation is usually tested using the Moran index. Considering the volatility of data, this paper uses the logarithm of carbon emissions per capita to test the Moran index.

As can be seen from

Table 4. Spatial autocorrelation test.

Year	2010	2011	2012	2013	2014	2015
Moran index	0.232 *	0.181 *	0.139 *	0.219 *	0.259 *	0.236 *
Year	2016	2017	2018	2019	2020
Moran index	0.195 *	0.276 *	0.319 **	0.342 **	0.347 **

Note: * and ** denote significance at the 10% and 5% levels, respectively.

, the Moran index of the carbon emission level of the Yangtze River Economic Belt from 2010 to 2020 is positive and significant at the level of 10%, indicating that the carbon emissions in the Yangtze River Economic Belt are not independent of each other, but affected by the surrounding areas, and there is a positive correlation. The Moran index has been on the rise since 2010, which indicates that the spatial correlation is gradually increasing. In order to observe the spatial autocorrelation of carbon emissions more intuitively, scatter plots of 2010, 2014, 2017, and 2020 are depicted in Figure 2.

4.2. Descriptive Statistics and Spatio-Temporal Evolution Chart

From the data in

Table 5. Descriptive statistics of variables.

Variable	Observations	Mean	Std. Dev.	Min	Max
ln CE	121	1.845	0.308	1.314	2.466
Agg	121	0.976	0.324	0.414	1.622
Env	121	9.947	2.756	5.209	18.218
a1	121	1.169	0.379	0.592	2.751
a2	121	0.143	0.216	0.001	0.944
a3	121	0.573	0.137	0338	0.896
a4	121	0.163	0.512	0.089	0.327

, it can be seen that the mean and variance of the sample data in the article are stable within a certain range. Industrial agglomeration is measured using the location entropy method, environmental regulation is measured with the entropy method, and various control variables are also measured using the formula in Table 3.

According to the final data collection results, this paper uses ARCIS 10.2 to draw spatial–temporal evolution maps for the industrial agglomeration, environmental regulation, and carbon emission levels in 2010 and 2020 (Figure 3, Figure 4 and Figure 5), respectively, and makes an intuitive analysis.

According to the spatial–temporal evolution chart, it can be intuitively seen that Jiangsu and Zhejiang are the two provinces with higher levels of industrial agglomeration in the YREB region, while Guizhou and Yunnan have lower levels of industrial agglomeration. Moreover, the industrial agglomeration levels of each province in the Yangtze River Economic Belt were in a dynamic and stable state from 2010 to 2020. In terms of environmental regulation, the intensity of environmental regulation has gradually increased from 2010 to 2020, indicating that China has begun to pay more attention to environmental control. In terms of carbon emissions, Jiangsu, Zhejiang, and Shanghai have relatively high levels of carbon emissions in the Yangtze River Economic Belt. In contrast, Sichuan and Hunan have relatively low levels of carbon emissions. From the spatiotemporal evolution chart, it can also be seen that the level of carbon emissions is distributed in a block-like manner. That is, the carbon emissions in the surrounding areas of provinces with high carbon emissions are relatively high, while the carbon emissions in the surrounding areas of provinces with low carbon emissions are also relatively low. This indicates that the carbon emissions of each province and city are in a dynamic and stable state over time and exhibit spatial agglomeration and spillover effects. This is also a basis for using spatial econometric models in this article.

4.3. Result

4.3.1. Establishment of Spatial Dubin Model

Due to the significant spatial diffusion effect of carbon emissions and industrial agglomeration, spatial factors should be considered. Through comprehensive analysis, this paper constructs a dynamic spatial panel model that includes both dynamic and spatial factors and specifically adopts the spatial Dubin model with both time and space fixed.

From the spatial Dubin model regression results (

Table 6. Results of the spatial Dubin model.

ln CE	Coefficient	z	p > |z|
Agg	0.306 ***	2.81	0.005
Env	0.004 *	1.73	0.084
a1	−0.168 *	−1.93	0.054
a2	−0.945 ***	−5.1	0.000
a3	2.707 ***	3.8	0.000
a4	2.545 ***	4.67	0.000
W Agg	0.848 *	2.58	0.010
W Env	−0.011 *	−1.78	0.075
W a1	−0.622 ***	−3.94	0.000
W a2	−1.207 ***	−3.55	0.000
W a3	−1.753	−1.09	0.274
W a4	3.916 ***	3.05	0.002
Log-likelihood	193.8527
R2	0.3655

Note: * and *** denote significance at the 10% and 1% levels, respectively.

), it can be seen that the regression coefficient of industrial agglomeration on carbon emissions is significantly positive at the 1% level, indicating that industrial agglomeration promotes an increase in carbon emissions. This is consistent with the research conclusion of Chen et al. and the specific impact mechanism is that the expansion of the industrial scale increases energy consumption, further leading to an increase in carbon dioxide emissions [21]. In addition, the spatial coefficient of industrial agglomeration on carbon dioxide is significantly positive at the 10% level, indicating that the impact of industrial agglomeration on carbon emissions has a positive spatial spillover effect. The regression coefficient of environmental regulation on carbon emissions is significantly positive at the 1% level and approaches 0, indicating that the impact of environmental regulation on carbon emissions in the Yangtze River Economic Belt region has a weak green paradox characteristic, which also confirms the experimental conclusion of Huang et al. [33]. However, the spatial coefficient of environmental regulation on carbon emissions is significantly negative. This indicates that, considering the spatial effects between regions, environmental regulation still has the effect of promoting carbon emission reduction and has a negative spatial spillover effect.

The regression results of the control variables show that the spatial coefficient of industrial structure and opening up on carbon emissions is significantly negative at the 1% level. This indicates that under the influence of spatial effects, the optimization of industrial structure and the degree of opening up can effectively promote carbon emission reduction. The regression coefficient between the urbanization rate and carbon emissions is significant at the 1% level. Still, its spatial coefficient did not pass the significance test, indicating that the urbanization rate has a positive impact on carbon emissions but does not produce spatial spillover effects. The spatial coefficient of government intervention in carbon emissions is significantly positive, indicating that improper government intervention will increase the emission of carbon pollutants.

4.3.2. Model Checking and Effect Decomposition

To establish the spatial Dubin model well, it is necessary to test the effect of the model. In this paper, the temporal, spatial, and total effects of the spatial Dubin model were tested, and the test results are shown in

Table 7. The test results.

Variable	Spatial Fixed Effect	Time Fixed Effect	Double Fixed Effect
Agg	0.295 *	0.847 ***	0.306 ***
Env	0.008 **	0.012	0.004 *
a1	−0.014	0.037	−0.168 *
a2	−0.656 ***	−0.086	−0.945 ***
a3	1.806 *	−1.329 **	2.707 ***
a4	1.341 *	2.909 **	2.545 ***
W Agg	0.384	2.381 ***	0.848 *
W Env	−0.003	−0.013	−0.011 *
W a1	−0.135	−0.116	−0.622 ***
W a2	−0.359	−0.349	−1.207 ***
W a3	−1.736 *	−5.259 ***	−1.753
W a4	−1.934 *	10.29 ***	3.916 ***
$\beta$	0.310 **	−0.593 ***	−0.173
$\epsilon$	0.003 ***	0.022 ***	0.002 ***
R2	0.396	0.338	0.365

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

.

According to the above test results, it can be seen that the significant effect of the spatial Dubin model with dual fixation of time and space is better than that of the temporal fixation model and spatial fixation model. Moreover, according to the results of the LR test, the LR value of dual fixation is 193.85, that of temporal fixation is 54.21, and that of spatial fixation is 169.97. Therefore, in this paper, when choosing the spatial Dubin model, the time–space dual fixed model is better.

The above regression results are one-sided in analyzing the spatial spillover effect of explanatory variables. When studying the spatial spillover effect, we can measure and analyze the impact degree of the spillover effect by dividing it into direct, indirect, and total effects. Direct effect says a variable change region in the area has been influenced by the explained variable, an indirect effect says variable changes in the surrounding area region are the influence degree of the explanatory variables, the total effect is when variable changes in all areas of the region have been influenced by the explained variable, and the total effect is a direct effect and an indirect effect. The result of effect decomposition is shown in

Table 8. Spatial effect decomposition results.

Variable	Direct Effect	Indirect Effect	Total Effect
Agg	0.272 **	0.718 **	0.991 ***
Env	0.005 **	−0.011 **	−0.006
a1	−0.131	−0.52 ***	−0.651 ***
a2	−0.898 ***	−0.928 ***	−1.826 ***
a3	2.856 ***	−2.021	0.835
a4	2.415 ***	3.162 ***	5.577 ***

Note: **, and *** denote significance at the 5% and 1% levels, respectively.

.

From the above effect decomposition table, it can be found that the direct effect of industrial agglomeration on carbon emissions is 0.272, which is significantly positive at a 5% level. This indicates that industrial agglomeration in this region significantly promotes the increase in carbon emissions, and every increase in industrial agglomeration in this region will lead to a change in carbon emissions by 0.272 units. The indirect effect of industrial agglomeration on carbon emissions is 0.718, which is significantly positive at the 5% level. This indicates that the industrial agglomeration in the surrounding area promotes the increase in local carbon emissions, and every increase in industrial agglomeration in the surrounding area will lead to a change in regional carbon emissions of 0.718 units. The total effect is 0.991, which is significantly positive at the 1% level, indicating that industrial agglomeration will cause an increase in carbon emissions overall. The direct effect of environmental regulation on carbon emissions is 0.005, which is significantly positive at the 5% level. This shows that the effect of environmental regulation in the YREB significantly promotes the increase in carbon emissions in this region, and every increase in industrial agglomeration in this region will lead to a change in the carbon emissions of this region of 0.005 units. The indirect effect of environmental regulations on carbon emissions is −0.011, which is significantly negative at a 5% level, indicating that environmental regulations in surrounding areas promote the reduction of local carbon emissions; every unit increase in industrial agglomeration in surrounding areas will lead to a −0.011 unit change in regional carbon emissions. The total effect is −0.006, but it is not significant because the direct and indirect effects cancel each other, indicating that environmental regulation has a slight positive effect on carbon emissions when considering the spatial spillover effect.

5. Conclusions and Discussion

The Yangtze River Economic Belt region, as an important national development strategy, has economic significance in studying its carbon emissions. This article uses the spatial Dubin model to study the spatial effects of environmental regulation and industrial agglomeration on carbon emissions. The experimental results indicate that industrial agglomeration in the YREB has a significant positive impact on carbon emissions and that there is a spatial spillover effect. This finding also confirms Chen et al.’s experimental conclusion that excessive industrial agglomeration often leads to scale effects, thereby increasing carbon emissions [9]. This discovery can provide some inspiration for developing the industrial structure of the YREB. The impact of environmental regulations on carbon emissions exhibits a weak green paradox effect, which is slightly different from previous scholars’ conclusions. It can be seen that the legal system related to the environment in the YREB region still needs to be improved. However, considering the impact of spatial effects, environmental regulations have a significant inhibitory effect on carbon emissions under specific conditions. Therefore, the impact of industrial agglomeration and environmental regulation should not be underestimated to promote national carbon emission reduction and the dual carbon goals of carbon peaking and neutrality. In particular, the Yangtze River Economic Belt, an important development strategic area in China, should respond to the dual carbon goals and policies. Therefore, this article makes the following suggestions based on the experimental results.

Considering the positive impact of industrial agglomeration in the YREB on carbon emissions, relevant departments should raise the entry standards for enterprises in the agglomeration zone and promote the upgrading of industrial structures. Industrial agglomeration in the YREB has a positive impact on carbon emissions due to the unreasonable allocation of industrial structure and the non-standard entry of enterprises in the agglomeration area. Therefore, local governments cannot simply support enterprises to settle in when attracting investment. They should comprehensively consider factors such as enterprise production nature, energy consumption, pollution emissions, environmental protection technology status, etc., and improve the entry standards of enterprises in the agglomeration area. Especially in high carbon emissions areas such as Jiangsu and Shanghai, further investigation is needed to investigate the progress of transformation and upgrading of high energy consuming and polluting enterprises. Policy encouragement and support should be given to the transformation and upgrading of large and high energy-consuming enterprises and enterprises with environmental protection technology advantages to promote the improvement of independent research and development capabilities of enterprises in the cluster area and the upgrading of environmental protection technology, thereby reducing carbon emissions in the cluster area. As for the urban agglomeration of the YREB, each city should clarify its urban functional positioning based on its unique resource endowment advantages and develop differentiated leading industries. Moreover, to win a competitive advantage for the leading industries, it is necessary to support relevant industries and form a diversified collaborative agglomeration model [54,55].

Environmental regulations can effectively promote carbon emission reduction, but there may also be green paradoxes. Therefore, the government needs to formulate differentiated environmental regulation policies based on the actual situation of the province. There are significant differences in carbon emissions, economic development level, industrial structure, energy structure, and technological level among different provinces in the YREB. Carbon environmental regulation policy should avoid blanket policies, which require that the relevant departments, according to the actual circumstance of the provinces’ domain development in a scientific and reasonable environmental regulation policy, establish and perfect the corresponding system of environmental policy. The environmental policy system gives full play to the exemplary role of pilot projects, especially the market incentives planned to promote implementing the low carbon pilot and carbon trading pilot policy. At present, China’s environmental regulation policy is given priority to command controlling environmental regulation. Environmental regulation based on the market incentives mode is complementary, so the government should pay more attention to the market incentives, such as the carbon trading pilot policy model of environmental regulation tools, to advance the process and improve the extent to which the market incentives contribute to achieving carbon reduction [56,57].

Overall, this article delves into the spatial effects of industrial agglomeration and environmental regulation and proposes feasible suggestions. However, there are still the following shortcomings. Firstly, considering the availability of data, this article selected inter-provincial data from the Yangtze River Economic Belt for research without exploring data from prefecture-level cities. Secondly, although this paper studies the spatial effects of industrial agglomeration, environmental regulation, and carbon emissions, there is a lack of research on the mesomeric effect. Thirdly, the research object of the article is the Yangtze River Economic Belt, and whether the impact effect across the country is consistent or still needs to be investigated. The research group will conduct in-depth research on these aspects in the future.