Exploring Carbon Emission Reduction Pathways: Analysis of Energy Conservation Potential in Yangtze River Economic Belt

Abstract

In response to the escalating global energy demands, the optimization of energy efficiency has emerged as a linchpin for sustainable development. This study considers the potential of energy conservation and emission reduction in one of the most economically vibrant and resource-intensive regions in China, the Yangtze River Economic Belt, encompassing 11 provinces and cities. The SBM-Undesirable model is used to measure the energy efficiency and analyze the temporal-spatial distribution. Moran’s I is employed to analyze the overall spatial pattern and local regional differences in energy efficiency. The systematic analysis shows that the temporal fluctuation exists in the development of energy efficiency, and the average of the Yangtze River Economic Belt exhibits a development pattern of “downstream > midstream > upstream” from the spatial perspective. The upstream region would require way more effort than others to decarbonize and improve efficiency. At the municipal level, the overall energy efficiency of 11 provinces and cities fails to reach an efficient state, and potential for improvement exists. Moreover, the potential model of energy conservation and emission reduction is constructed. We further explore the pathways of energy efficiency improvement for each region in the Yangtze River Economic Belt, including pathways of “High-Efficiency Type”, “High Emission Reduction Potential”, and “Extensive Development Type”.

1. Introduction

In an era defined by rapid industrialization and escalating global energy demands, the optimization of energy efficiency has emerged as a linchpin for sustainable development [1,2], as well as a foundational pillar in the comprehensive strategy aimed at decarbonizing industrial processes [3]. Evidences show that sustainable economic development is associated with increased energy efficiency [4]. As an important driving force for energy efficiency, the digital economy and artificial intelligence play an irreplaceable role to address the worsening global climate risks [5,6].

As the world’s largest energy-consuming and carbon-emitting country, China has promised that carbon dioxide emissions intensity will be reduced by 60–65% by 2030 compared with 2005 [7]. The Yangtze River Economic Belt in China is one of the most economically vibrant and resource-intensive regions, encompassing 11 provinces and cities, and plays a pivotal role in the nation’s economic growth, accounting for almost 20% of national GDP [8,9]. However, this prosperity has been accompanied by substantial energy consumption and environmental challenges, making the study of its energy efficiency not only a regional necessity but also a matter of national strategic importance [10,11]. Understanding and enhancing energy efficiency within this belt can serve as a model for other regions, contributing to broader goals of carbon neutrality and ecological conservation. As the total energy consumption of the economic belt continues to increase, its development focus would be on continuously improving energy efficiency as the global trend [12]. Combining with the continuous advancement of the industrialization process, a large number of high-energy-consuming industries are concentrated in the Yangtze River region, which results in a significant impact on the environmental quality of the Yangtze River region [13]. Under this situation, it is vital to understand how the energy efficiency performs from the spatial–temporal perspective, but there is a lack of the analysis in extant literature.

Previous research on energy efficiency has covered a wide spectrum, from theoretical models to empirical analyses. Improving energy efficiency is an effective approach for the economic transformation and green development of the Yangtze River Economic Belt. Existing literature has conducted analyses of various perspectives on the indicator selection, evaluation methods, and influencing factors of energy efficiency [14]. In terms of methods for measuring energy efficiency, Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA) methods [15,16,17] have been widely applied with different perspectives of advantages. More specifically, the SFA method was used to measure the energy consumption and efficiency of households among high/low-income areas in China in [18]. Lundgren used the SFA method to estimate energy demand and energy efficiency in different industries in Sweden [19]. Moreover, the super-efficiency DEA model and the Malmquist index model are used to comparatively measure the energy efficiency of the Pearl River Delta in China in [20]. Perspectives of geographic region and industrial subdivision are considered for analyzing the path to improve green energy efficiency in literature [21,22,23], and then key factors that affect energy efficiency are identified including economics, endowments, technology, policies, etc. [24,25]. The popular methods for measuring spatial correlation include Moran’s Index (Moran’s I) [26], Geary’s C index, and G index [27]. Particularly, Moran’s I is an important spatial statistical measure used to determine the presence or absence of spatial autocorrelation [28], and this approach has been commonly used to uncover spatial correlations across dimensions, such as enhancing resource and energy optimization [29]. With increasing emphasis on the environment, many scholars have begun to incorporate various environmental factors into the indicator system; meanwhile, various environmental pollutants are considered as undesirable outputs [30,31]. However, there remains a gap in comprehensively examining both the temporal dynamics and spatial disparities of energy efficiency across the entire belt. Additionally, while some research has explored energy conservation and emission reduction strategies, few studies have systematically modeled the potential for improvement and specific pathways for each province and city within this region. This divergence in research focus has led to a fragmented understanding of the region’s energy efficiency landscape, creating a need for a more holistic investigation.

Against this backdrop, the primary objective of this study is to conduct a comprehensive assessment of the energy efficiency of the 11 provinces and cities in the Yangtze River Economic Belt. Compared to existing literature, this study integrates a multi-dimensional analytical framework that encompasses a wide range of variables, especially including “Dual Carbon” policy-driven factors. By doing so, it not only provides a more complete picture of the energy efficiency status across the entire economic belt but also uncovers the potential pathways that drive energy efficiency improvement among the 11 provinces and cities. Therefore, we aim to achieve three key goals: first, to provide a detailed and accurate evaluation of energy efficiency; second, to uncover the temporal trends and spatial patterns of energy efficiency through dynamic and spatial econometric analyses; and third, to identify the potential for energy efficiency improvement and propose tailored strategies for each region.

2. Materials and Modeling

2.1. The Indicator System of Energy Efficiency

In literature, various perspectives of analytics have been conducted on the indicator selection, evaluation methods, and influencing factors of energy efficiency [2,32,33]. In literature, the most direct method for calculating energy efficiency only focuses on a single factor, that is, the ratio between energy input and economic output, which is usually characterized by the energy consumption per unit of Gross Domestic Product (GDP). With the rapid development of the Yangtze River Economic Belt, during the process of energy consumption mainly composed of non-clean energy sources, a large amount of environmental pollution has been caused [8]. Therefore, in the calculation of energy efficiency, not only does the desired output need to be considered, but also the undesired output that restricts green development should be taken into account.

Existing studies mainly select labor, capital, and energy as input indicators, take regional economic growth as the desired output indicator [34,35], and there is no unified standard for the undesired output indicators oriented toward environmental protection goals. These mainly include the emissions of environmental pollutants such as carbon dioxide, sulfur dioxide, and nitrogen oxides [36]. Moreover, the regulatory policy is also considered as the source of indicators to investigate the energy efficiency differences for discussing the policy effectiveness and potential improvement [37,38,39]. Therefore, the “Dual Carbon” strategy that was proposed by the Chinese government in 2020 includes a series of related policies toward carbon emissions [40,41,42], which was also referred to for the design of the indicator system. More specifically, the index system of energy efficiency evaluation for the Yangtze River Economic Belt is constructed as shown in

Table 1. The input-output indicators of the energy efficiency evaluation.

Classification	Indicator	Indicator Definition	Unit
Input	Labor input	Tear-end employment population	Tens of thousands of people
	Energy input	Total energy consumption	Tens of thousands Tons of standard coal
	Capital input	Capital stock	100 million RMB
Desired output	Economic benefits	Gross regional production	100 million RMB
	Environmental benefits	Green coverage rate of built-up area	%
Undesirable output	Environmental pollution	Carbon dioxide emissions	Tens of thousands Tons
		Sulfur dioxide emission	Tons

.

2.1.1. Input Indicators

The energy use is accompanied by the labor and capital inputs related to production and industrial resources. Therefore, this paper includes corresponding labor input, energy input, and capital input in the input indicators of the energy efficiency evaluation. Specifically, this paper uses the year-end employment population of each province or city to represent the labor input, and the total energy consumption of each province or city to represent the energy input [43]. Since the capital stock cannot be directly obtained, this paper takes the fixed asset investment of each province or city in 2000 as the base period, and draws on the approach in [44] to further estimate the capital stock of 11 provinces and cities during 2011–2020 using the perpetual inventory method.

2.1.2. Desired Output Indicators

The output indicators that are desired to be obtained in energy efficiency measurement are called desired output indicators. Referring to the related literature, the regional GDP is used to represent the economic benefits [35]. Regarding the environmental benefits, the social sustainability requires achieving a balance between the economic, energy, and environmental systems based on the energy–economy–environment theory. Aiming at the “Dual Carbon” targets, increasing the green coverage rate or forest coverage rate has strategic significance. Therefore, referring to [45,46], the green coverage rate of built-up areas in various provinces and cities is chosen as the environmental benefit indicator.

2.1.3. Undesirable Output Indicators

The selection of undesirable output indicators mainly focuses on the substances that have a polluting impact on the environment. During the energy consumption, a large amount of carbon dioxide is emitted, which has a huge negative impact on the ecological environment [47]. Similarly, sulfur dioxide emissions also cause serious pollution to the environment, which is the focus of emission reduction. Therefore, this paper selects carbon dioxide emissions and sulfur dioxide emissions as the environmental pollutant indicators [48,49].

2.2. The SBM-Undesirable Model for Measuring Energy Efficiency

On the basis of the energy efficiency indicator system, the novel SBM-Undesirable model [50] is then used to measure the energy efficiency of the Yangtze River Economic Belt from 2011 to 2020. In reality, with the rapid economic development of the Yangtze River Economic Belt, the consumption of energy resources dominated by non-clean energy has gradually increased, which leads to environmental pollution during the process of energy consumption. This indicates that it is necessary to consider not only the desired outputs but also the undesirable outputs that constrain social development in the case of measuring energy efficiency. Therefore, compared to the classic SBM method, the SBM-Undesirable model is chosen to perform the energy efficiency analytics since undesirable output indicators are involved [51].

Assuming there are N decision-making units, each unit has m types of input indicators, s types of desired output indicators, and k types of undesirable output indicators. The SBM-Undesirable model can be defined as follows:

$$\rho =\min {\displaystyle \frac{1-\frac{1}{m}{\sum }_{i=1}^m\left(\frac{S_i^{-}}{x_{\mathrm{i}0}}\right)}{1+\frac{1}{s+k}\left({\sum }_{r=1}^s\frac{S_r^{+}}{y_{r0}}+{\sum }_{t=1}^k\frac{S_t^{-}}{z_{t0}}\right)}}$$

(1)

$$\mathrm{St}. {\displaystyle {\sum }_{n=1}^N{\lambda }_n{x}_{in}+S_i^{-}=x_{i0},i=1,2,3\dots ,m}$$

$${\displaystyle {\sum }_{n=1}^N{\lambda }_n{y}_{\mathrm{r}n}-S_r^{+}=y_{r0},r=1,2,3\dots ,s}$$

$${\displaystyle {\sum }_{n=1}^N{\lambda }_n{z}_{\mathrm{t}n}+S_t^{-}=z_{t0},t=1,2,3\dots ,k}$$

$$S_i^{-},S_r^{+},S_k^{-},{\lambda }_n\ge 0$$

In the formula, ρ represents the objective function, which stands for the total factor energy efficiency value, and it satisfies the condition of 0 ≤ ρ ≤ 1. Therein, $S_i^{-}{,S}_r^{+}$, and $S_t^{-}$ are the slack variables for the indicators of input, desired output, and undesirable output. If a decision-making unit satisfies $S_i^{-}$ = 0${,S}_r^{+}$ = 0 and $S_t^{-}$ = 0, as well as the objective function ρ = 1, this indicates that this decision-making unit is efficient, known as DEA efficient [52]. Conversely, if it suggests that the decision-making unit is inefficient, certain improvements would be required in its inputs or outputs.

2.3. Exploratory Spatial Data Analysis

Exploratory Spatial Data Analysis (ESDA) has been commonly used in the research area of spatial aggregation and anomalies in order to explain the spatial location relationships, by describing and visualizing the spatial distribution patterns of objects or phenomena [53]. Among them, spatial correlation analysis is a key component of ESDA. It could detect the correlation degree of the same observational data among regions in the geographical space, which can be divided into positive correlation, negative correlation, and non-correlation. Moreover, the analysis of spatial correlation could be specifically divided into two types: global spatial correlation and local spatial correlation. Global spatial correlation reflects the overall distribution of a certain attribute value in a region that is used to determine whether the attribute value of this region has spatial aggregation characteristics. In turn, local spatial correlation mainly analyzes the spatial correlation among various regions that reflects the specific regions of spatial aggregation. Therefore, this study first uses global spatial correlation to determine whether there is an obvious spatial correlation in the region as a whole. If so, then local spatial correlation analysis is used to determine the specific location within the region. Considering the data features and related calculation, Moran’s I [54] is applied to represent the similarity of unit attribute values among adjacent regions in order to explore the energy utilization rate in the Yangtze River Economic Belt region.

2.3.1. Spatial Weight Matrix

By setting different weight matrices, the differences in spatial correlation could be represented, which are usually divided into three categories: the adjacency spatial weight matrix, the geographical distance spatial weight matrix, and the economic distance spatial weight matrix. Among them, the adjacency spatial weight matrix is represented by the adjacent distance between regions. It is assumed that the closer the distance between regions, the greater the interaction. Therefore, it is the most commonly used spatial weight matrix and is also the type of spatial weight matrix adopted in this study. The setting of the adjacency weight matrix is based on whether there is a common boundary line between two provinces to determine whether there is a common border between the two provinces. There are three situations: “Rook’s case”, “Bishop’s case”, and “Queen’s case”, which respectively mean that province or city i and province or city j have a common edge, a common vertex, or a common edge or vertex. In comparison, the “Queen’s case” is more comprehensive. Therefore, this paper selects the indicator of “Queen’s case” to measure whether two provinces are adjacent. The adjacency weight matrix is generally represented by “0–1”, where “1” represents an adjacent region and “0” represents a non-adjacent region. The specific formula is:

$$w_{\mathit{ij}}=\left\{\begin{array}{c}1,province or city i and province or city are jadjacent\\ 0,province or city i and province or city are NOT jadjacent\end{array}\right.$$

(2)

2.3.2. Global Moran’s I

The application of the global Moran’s I can test whether there is a spatial aggregation characteristic of the energy efficiency in the Yangtze River Economic Belt and can reflect its spatial correlation. The specific formula is shown in Formula (3). Among them, x_i and x_j are the energy efficiencies of province or city i and province or city j, respectively; w_ij is a binary adjacency weight matrix representing the spatial weight value. Here, I = [−1, 1]. When I = 0, it indicates that there is no spatial correlation in the entire region. When 0 < I ≤ 1, the spatial correlation is positive, and the closer it approaches 1, the stronger the positive correlation is. When −1 ≤ I< 0, the spatial correlation is negative, and the closer it approaches −1, the stronger the negative correlation is.

$$I={\displaystyle \frac{N}{{\sum }_{i=1}^N{\left(x_i-\overline{x}\right)}^2}}\times {\displaystyle \frac{{\sum }_{i=1}^N{\sum }_{i=1}^N{w}_{ij}\left(x_i-\overline{x}\right)\left(x_j-\overline{x}\right)}{{\sum }_{i=1}^N{\sum }_{i=1}^N{w}_{ij}}}$$

(3)

In the process of analyzing spatial correlation, it is necessary to use an index to conduct the null hypothesis test for the random distribution. The hypothesized variable of the global Moran’s I follows a normal distribution, and a Z-test could be used to determine whether there is a significant global spatial correlation. The specific formula is shown in Formula (4). Among them, E(I) and Var(I) represent the expectation and variance of the global Moran’s I, respectively. Through the above formula, the Z-value could be obtained. If the Z-value is greater than 1.96, that indicates that there is spatial correlation among regions, and the null hypothesis is rejected at the 5% significance level, further indicating that there is a significant spatial correlation in the energy efficiency of the Yangtze River Economic Belt.

$$Z={\displaystyle \frac{I-E\left(I\right)}{\sqrt{Var\left(I\right)}}}$$

(4)

2.3.3. Local Moran’s I

Since the global Moran’s I only identifies whether there is spatial correlation in the entire Yangtze River Economic Belt, we have to further locate the specific position of the spatial correlation phenomenon. The local Moran’s I is then used to reflect the degree of correlation between the energy efficiency of each province or city and that of its neighboring regions. The specific formula of the local Moran’s I is shown in Formula (5). When I_i is a non-zero value, provinces with similar energy efficiency are clustered together, and the clusters can be divided into four types: high–high clustering, high–low clustering, low–low clustering, and low–high clustering.

Similar to the global Moran’s I, the Z-test can also be used to determine whether there are obvious characteristics of local spatial correlation. The definition of Z_i is shown in Formula (5), where E(I_i) and Var(I_i) are the expectation and variance of the local Moran’s I, respectively. The local Moran’s I can be visually expressed through the Moran scatter plot and the LISA cluster map. In this study, the Moran scatter plot is mainly used to study the local spatial correlation of the Yangtze River Economic Belt.

$$Z_i={\displaystyle \frac{I_i-E\left(I_i\right)}{\sqrt{Var\left(I_i\right)}}}$$

(5)

2.4. The Potential of Energy Conservation and Emission Reduction

2.4.1. The Potential of Energy Conservation

Most provinces and cities in the Yangtze River Economic Belt exhibit excessive energy consumption as well as the emission of environmental pollutants. In order to further analyze these situations in different regions and find the pathway for effective improvement, the potential of energy efficiency is modeled based on the SBM model of undesirable outputs. Relevant definitions on energy efficiency (EE) are proposed as:

$$E{E}_{t,i}={\displaystyle \frac{TE{I}_{t,i}}{AE{I}_{t,i}}}={\displaystyle \frac{AE{I}_{t,i}-LE{I}_{t,i}}{AE{I}_{t,i}}}$$

(6)

Among them, EE_t,i represents the energy efficiency of the i-th provincial-level administrative region in period t; TEI_t,i (Target Energy Input) represents the energy input of the reference target unit on the optimal production frontier of the i-th provincial-level administrative region in period t; AEI_t,i is the actual energy input (actual energy input) of the i-th provincial-level administrative region in period t; LEI_t,i (Loss Energy Input) represents the input slack of the energy indicator of the i-th provincial-level administrative region in period t, which can also be regarded as the savable energy input. According to the above formula, the energy-saving potential SPE_t,i (Saving Potential of Energy) of the i-th provincial-level administrative region in period t can be obtained.

$$SP{E}_{t,i}={\displaystyle \frac{LE{I}_{t,i}}{AE{I}_{t,i}}}$$

(7)

Specifically, the higher the value of SPE_t,i, the greater the energy inefficiency of the i-th province or city in period t, and its energy-saving potential is also greater.

2.4.2. The Potential of Emission Reduction

Similarly, energy efficiency (EE for short) can also be defined as:

$$E{E}_{t,i}={\displaystyle \frac{TP{I}_{t,i}}{AP{I}_{t,i}}}={\displaystyle \frac{AP{I}_{t,i}-LP{I}_{t,i}}{AP{I}_{t,i}}}$$

(8)

Among them, EE_t,i represents the energy efficiency of the i-th province or city in period t; TPE_t,i (Target Pollutant Emission) represents the emissions of environmental pollutants of the reference target on the best production frontier of the i-th province or city in period t; APE_t,i is the actual emissions of environmental pollutants (Actual Pollutant Emission) of the i-th province or city in period t; LPE_t,i (Loss Pollutant Emission) represents the slack of the output index of environmental pollutants of the i-th province or city in period t, and it can also be regarded as the reducible emissions of environmental pollutants. According to the above formula, the energy-saving potential RPE_t,i (Reducing Potential of Emission) of the i-th province or city in period t can be obtained:

$$RP{E}_{t,i}={\displaystyle \frac{LP{E}_{t,i}}{AP{E}_{t,i}}}$$

(9)

Specifically, the higher the value of RPE_t,i, the greater that the energy inefficiency of the i-th province or city in period t, and its potential for emission reduction is also greater.

3. Empirical Analytics

3.1. Data Description

Data sources include the China Energy Statistical Yearbook, China Environment Statistical Yearbook, China Statistical Yearbook, China Carbon Accounting Database, statistical yearbooks of various provinces and cities, and statistical bulletins of provincial and municipal statistical bureaus. For data preprocessing, the interpolation method is mainly used to fill in missing data for some provinces. Due to the data availability, the energy efficiency of the Yangtze River Economic Belt during 2011–2020 is chosen as the research object.

3.2. Measurement of the Energy Efficiency in the Yangtze River Economic Belt

Based on the energy efficiency evaluation system constructed in Section 2.1, the SBM-Undesirable model is further employed to measure the energy efficiency of the Yangtze River Economic Belt, which is divided into three major regions: downstream, midstream, and upstream. The specific measurement result obtained through the MaxDEA (pro 8.3) software is shown in

Table 2. The measurement result of the energy efficiency in Yangtze River Economic Belt.

Region	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	Average
Shanghai	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000
Jiangsu	1.000	1.000	1.000	1.000	1.000	1.000	1.000	0.892	0.897	0.871	0.966
Zhejiang	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000
Anhui	1.000	1.000	0.718	0.701	0.675	0.659	0.659	0.695	0.701	0.711	0.752
Jiangxi	0.806	0.808	0.800	0.791	0.758	0.737	0.719	0.680	0.687	0.694	0.748
Hubei	0.699	0.708	0.740	0.728	0.732	0.714	0.705	0.697	0.706	0.663	0.709
Hunan	0.737	0.743	0.764	0.751	0.755	0.734	0.725	0.656	0.673	0.676	0.721
Chongqing	0.695	0.717	0.769	0.756	0.791	0.785	0.776	0.720	0.734	0.787	0.753
Sichuan	0.661	0.678	0.689	0.676	0.676	0.658	0.673	0.653	0.659	0.714	0.674
Guizhou	0.497	0.521	0.542	0.542	0.551	0.528	0.536	0.497	0.498	0.511	0.522
Yunnan	0.521	0.524	0.535	0.510	0.489	0.461	0.455	0.478	0.488	0.481	0.494
Downstream	1.000	1.000	0.929	0.925	0.919	0.915	0.915	0.897	0.899	0.896	0.929
Midstream	0.747	0.753	0.768	0.757	0.748	0.728	0.716	0.678	0.689	0.677	0.726
Upstream	0.594	0.610	0.634	0.621	0.627	0.608	0.610	0.587	0.595	0.623	0.611
Average	0.783	0.791	0.778	0.769	0.766	0.752	0.750	0.724	0.731	0.737	0.758

¹ In this study, Shanghai, Jiangsu, Zhejiang, and Anhui are classified as downstream regions; Jiangxi, Hubei, and Hunan are classified as midstream regions; and Chongqing, Sichuan, Yunnan, and Guizhou are classified as upstream regions.

.

3.2.1. From the Perspective of Overall Performance

From 2011 to 2020, the average energy efficiency of the 11 provinces and cities in the Yangtze River Economic Belt was 0.758, which reached only 75.8% of the optimal level, with a significant gap from the best frontier. In terms of temporal changes, the overall energy efficiency of the Yangtze River Economic Belt fluctuated and declined during the study period, with the energy efficiency value falling from 0.783 in 2011 to 0.737 in 2020, although the decline was not significant. Moreover, in 2012, the average energy efficiency of the Yangtze River Economic Belt rose to its highest, reaching 0.791. The main reason for this increase is the improvement in energy efficiency in various provinces and cities in the middle and upstream regions. In 2011, the National Development and Reform Commission announced the “Implementation Plan for Energy-saving and Low-carbon Actions for Ten Thousand Enterprises”, which implemented the green development concept of energy conservation and emission reduction in the 12th Five-Year Plan and clarified energy-saving management measures, thereby improving energy efficiency from 2011 to 2012. However, after that, the energy efficiency of the Yangtze River Economic Belt gradually declines, reaching its lowest point during the entire study period in 2018 with an efficiency value of 0.724. This was due to the unequal growth of economic development and environmental carrying capacity, which in turn leads to widespread haze weather in many regions and a continuous decline in energy efficiency. And in 2020, the energy efficiency slowly recovered to 0.737.

3.2.2. From the Perspective of Upstream, Midstream, and Downstream Regions

The average energy efficiency values during the study period are 0.929, 0.726, and 0.611, respectively, none of which reached the effective production frontier. This result indicates that there is still considerable space for energy conservation and emission reduction in the energy efficiency of the Yangtze River Economic Belt. On the whole, the energy efficiency of the downstream, midstream, and upstream regions of the Yangtze River Economic Belt exhibited a development pattern of “downstream > midstream > upstream” during the study period. Correspondingly, the average energy efficiency in the downstream region was the highest, followed by the midstream region, and the upstream region had the lowest average energy efficiency. The significant difference in energy efficiency between the downstream region and the other two regions is related to the development status of each province and city. The higher energy efficiency in the downstream region is mainly due to its superior geographical location and energy policies. The downstream region is located on the eastern coast with a developed economy, a high degree of openness, and a gradual shift toward a more environmentally friendly tertiary industry within its industrial structure. Also, the downstream region is one of the regions with the strongest technological strength, and policies there are at the forefront of economic system reform in China. Leveraging its industrial advantages, the downstream region has achieved significant breakthroughs in the process of industrialization and urbanization, which leads to the superior performance in emission reduction technology and energy utilization compared to the midstream and upstream regions. In contrast, the midstream and upstream regions are located inland at a slower development phase, where they also face issues such as unreasonable industrial structure distribution, limited technological capabilities for energy conservation and emission reduction, and severe environmental pollution, which in turn lead to lower energy efficiency compared to the downstream region.

3.2.3. From the Perspective of Provinces and Cities

Shanghai and Zhejiang have been in a completely effective state on energy efficiency during the research period. Meanwhile, there are high economic benefits in Shanghai and Zhejiang, as well as in the top three among the 11 provinces and cities regarding the per capita GDP. Also, the pollution emissions and pollution levels in Shanghai and Zhejiang are low. Jiangsu is on the frontier of production in energy efficiency from 2011 to 2017, while its energy efficiency is less than 1 from 2018 to 2020, and the value of energy efficiency gradually decreases to 0.871. Similarly, Anhui reached the frontier of energy efficiency in 2011–2012, and its energy efficiency gradually fell to 0.711 in 2013–2020. In the midstream regions, the energy efficiency in Jiangxi keeps declining from 2011 to 2020, while the energy efficiency in Hubei and Hunan declines with fluctuation. Meanwhile, the economic development is relatively rapid, but the control of environmental protection is ignored to some extent, resulting in the decline of energy efficiency during the research period. In the upstream regions, except for Chongqing, the average energy efficiency of Yunnan and Guizhou is lower than 0.6, which is relatively low. The late start of economic development and low technical level in the upstream region will lead to low energy efficiency. None of the provinces and cities in the middle and upstream regions reached the average energy efficiency during the research period. Hence, the energy efficiency of most provinces and cities during the sample research period fails to reach 1, which is an invalid state of DEA with a large space to improve.

3.2.4. From the Perspective of the Trend over Time

The energy consumption of the Yangtze River Economic Belt constantly rises overall during the research period, which causes a great environmental burden on the ecological environment. Therefore, it is necessary to further explore the time change trend of the Yangtze River Economic Belt. The temporal differences in energy efficiency of the Yangtze River Economic Belt and its three major regions (i.e., the upstream, midstream, and downstream regions) from 2011 to 2020 are analyzed, as shown in Figure 1.

On the whole, the energy efficiency of the Yangtze River Economic Belt generally forms an “N-shaped” fluctuating downward trend during the research sample period. In terms of temporal changes, the energy efficiency of the lower and middle reaches of the Yangtze River shows a fluctuating downward trend, while the energy efficiency of the upper reaches shows a fluctuating upward trend. Especially, the energy efficiency of the lower reaches decreases rapidly from 2012 to 2013, and then decrease slowly from 2013 to 2020; The energy efficiency of the middle reaches rises slowly from 2011 to 2013, and then fluctuates downward from 2013 to 2020; The energy efficiency of the upper reaches fluctuates upward during the entire research period from 2011 to 2020, and reaches the highest efficiency value of 0.623 in 2020.

Particularly, the increase in energy efficiency of the Yangtze River Economic Belt during the period of 2011–2012 is largely due to a favorable policy environment. A series of policy measures were issued in 2011 in China that emphasized the need to improve the ecological environment and set specific targets for reducing environmental pollution emissions. From 2012 to 2018, the energy efficiency of the Yangtze River Economic Belt fluctuated slightly and generally showed a downward trend, with an average annual decline rate of 1%. During this period, the decrease in energy efficiency of the Yangtze River Economic Belt might be closely related to changes in the economic environment at that time. With the continuous advancement of industrialization and urbanization, the energy demand in China shows rigid growth. However, due to the limited resources and environmental capacity, the economy and resources as well as environment were in a state of disharmony, which led to a decline in energy efficiency during this period. In the subsequent development process of the upstream regions, not only economic outputs but also the emissions of environmental pollutants are taken into account. Therefore, from 2018 to 2020, the energy efficiency of the Yangtze River Economic Belt recovered.

3.3. The Spatiotemporal Analysis of Energy Efficiency

3.3.1. From the Perspective of Global Moran’s I

The global Moran’s I is applied to test the global spatial correlation of the energy efficiency of 11 provinces and cities in the Yangtze River Economic Belt from 2011 to 2020 in the Geoda (1.18) software, and the calculation results are shown in

Table 3. The global Moran’I ′ I index of the Yangtze River Economic Belt.

Year	Moran’s I	Sd	Z-Value	p-Value
2011	0.7646	0.2030	4.1705	0.01
2012	0.7530	0.2018	4.1380	0.01
2013	0.6207	0.1883	3.8152	0.01
2014	0.6198	0.1891	3.7923	0.01
2015	0.5571	0.1887	3.4902	0.01
2016	0.5568	0.1893	3.4810	0.01
2017	0.5477	0.1895	3.4277	0.01
2018	0.6154	0.1915	3.7290	0.01
2019	0.6088	0.1907	3.7143	0.01
2020	0.5074	0.1805	3.3627	0.01
The Yangtze Belt	0.6151	0.1912	3.7121	0.01

. Overall, the global Moran’s I is 0.6151, which passes the significance level test of 1%, indicating that the energy efficiency of the entire Yangtze River Economic Belt as a whole shows clustering during the study period. The positive spatial dependence shows significant, and the spatial radiation driving effect of provinces and cities with high energy efficiency on the surrounding provinces and cities with low energy efficiency is relatively good.

From the perspective of temporal changes, the global Moran’s I indicates that the spatial clustering characteristics of the energy efficiency are unstable in the Yangtze River Economic Belt. Specifically, from 2011 to 2017, the global Moran’s I of the energy efficiency shows a significant downward trend, with an average decline rate of 3.6%; from 2017 to 2018, the positive spatial dependence of the energy efficiency shows a gradually strengthening development trend, with an increase rate of 6.7%; from 2018 to 2020, the global Moran’s I dropped significantly, with an average decline rate of 5.4%. From above, regarding the entire study time interval, the energy efficiency of the Yangtze River Economic Belt shows that the positive spatial dependence becomes weaker gradually, but each region is still in an unbalanced development state.

3.3.2. From the Perspective of Local Moran’s I

To further explore the location and specific manifestation of spatial clustering among the provinces and cities within the Yangtze River Economic Belt, local Moran scatter plots are applied to conduct a local spatial correlation analysis of the energy efficiency, respectively. Due to the limited space here, four scatter plots in 2011, 2015, 2020, and throughout the study period are taken as examples to analyze the local spatial correlation, as shown in Figure 2. According to the energy efficiency levels among adjacent provinces and cities in the Yangtze River Economic Belt, each province and city is classified into four quadrants of the local Moran scatter plot graph.

• The first quadrant is the “H-H clustering area” (high–high clustering area), in which provinces and cities with a high energy efficiency level are surrounded by adjacent provinces and cities with a high energy efficiency level. There is a strong positive promoting effect among adjacent provinces and cities, and the spatial clustering characteristics are obvious. The energy efficiency values of the provinces and cities in this area are relatively high, and there are significant diffusion characteristics among the provinces and cities.
• The second quadrant is the “L-H clustering area” (low–high clustering area), in which provinces and cities with a low energy efficiency level are surrounded by adjacent provinces and cities with a high energy efficiency level. There is a large difference in energy efficiency among adjacent provinces and cities, and there is an obvious phenomenon of unbalanced spatial development.
• The third quadrant is the “L-L clustering area” (low–low clustering area), in which provinces and cities with a low energy efficiency level are surrounded by adjacent provinces and cities with a low energy efficiency level. The energy efficiency of the provinces and cities in this area is in a low state, and the growth of energy efficiency is slow.
• The fourth quadrant is the “H-L clustering area” (high–low clustering area), in which provinces and cities with a high energy efficiency level are surrounded by adjacent provinces and cities with a low energy efficiency level. There is the significant characteristics of unbalanced development between provinces and cities and their adjacent provinces and cities, and the radiation effect of high-energy-efficiency areas on low-energy-efficiency areas is not significant.

More specifically, from 2011 to 2012, Shanghai, Jiangsu, Zhejiang, Anhui, and Jiangxi were all in the “H-H agglomeration area” of the first quadrant. The energy efficiency of the five provinces and cities reached the production frontier and gathered together; only Hubei was in the “L-H agglomeration area” of the second quadrant; Hunan, Chongqing, Guizhou, Yunnan, and Sichuan were in the “L-L agglomeration area” of the third quadrant. From 2013 to 2014, Shanghai, Jiangsu, Zhejiang and Anhui still fell into the “H-H agglomeration area” of the first quadrant, while Jiangxi fell from the first quadrant to the second quadrant, changing from the “H-H agglomeration area” to the “L-H agglomeration area”, indicating that there was an obvious polarization phenomenon in the energy efficiency between Anhui and its surrounding provinces and cities during this period. The provinces and cities in the “L-L agglomeration area” of the third quadrant were not much different from the previous period. In addition to Hunan, Chongqing, Guizhou, Yunnan, and Sichuan, Hubei also entered the third quadrant. From 2015 to 2017, the agglomeration of provinces and cities was the same as that from 2019 to 2020: the “H-H agglomeration area” in the first quadrant included Shanghai, Jiangsu and Zhejiang, while Anhui entered the “L-H agglomeration area” in the second quadrant, indicating that the energy efficiency of Anhui decreased significantly during this period. The “L-L agglomeration area” in the third quadrant included Hunan, Hubei, Guizhou, and Yunnan; the “H-L agglomeration area” in the fourth quadrant only included Chongqing, indicating that Chongqing’s energy efficiency was improved during this period. In 2018, Shanghai, Jiangsu, and Zhejiang were in the “H-H agglomeration area” of the first quadrant, while Anhui and Jiangxi were in the “L-H agglomeration area” of the second quadrant. Compared with the period from 2015 to 2017, Chongqing fell back from the fourth quadrant to the “L-L agglomeration area” of the third quadrant, and the polarization of energy efficiency around it was weakened, indicating that the radiation driving effect of surrounding provinces and cities on Chongqing was enhanced during this period. The local agglomeration of the Yangtze River Economic Belt in 2018 was the same. Most provinces and cities in the economic belt were agglomerated in the “H-H agglomeration area” and “L-L agglomeration area”, and the high–high agglomeration and low–low agglomeration were obvious.

In general, the provinces and cities in the Yangtze River Economic Belt were mainly distributed in the first and third quadrants during the study period, which indicated that the energy efficiency of the Yangtze River Economic Belt had an obvious positive spatial correlation and an obvious spatial competition relationship among provinces and cities. The “H-H agglomeration area” was mainly in the downstream area, and the “L-L agglomeration area” was mainly in the middle and upper reaches. The spatial agglomeration of energy efficiency among provinces and cities was obvious, and the spatial non-equilibrium characteristics were significant.

Above all, the energy efficiency of the regions located in the first and third quadrants shows positive spatial correlation in space, and there are significant spatial clustering distribution characteristics. However, the energy efficiency of the regions located in the second and fourth quadrants shows negative spatial correlation in space, with discrete distribution characteristics.

3.4. The Potential Analysis of Energy Efficiency Improvement

According to the model in Section 2.4, we further analyze the potential of energy efficiency improvement to further explain the phenomenon that the energy efficiency has not yet reached the optimal production level in most provinces and cities along the Yangtze River Economic Belt.

3.4.1. Analyzing the Potential of Energy Conservation

In this case, the potential refers to the possibility of a certain province or city to improve its energy efficiency. When the external environment changes, this possibility has the ability to develop into reality. The results of the potential ratio and amount (coal equivalent) are depicted in

Table 4. The measurement result of energy-saving potential in the Yangtze River Economic Belt.

Region	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	Average	Rank (Ratio)
Shanghai	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	10
Jiangsu	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	6.3% 1989	6.1% 1985	9.8% 3212	2.2% 719	9
Zhejiang	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	0.0% 0	10
Anhui	0.0% 0	0.0% 0	19.4% 2266	20.0% 2404	22.4% 2756	23.1% 2925	23.3% 3037	18.6% 2477	18.0% 2493	24.5% 3604	16.9% 2196	7
Jiangxi	5.7% 395	5.1% 370	6.9% 522	10.0% 804	13.9% 1171	15.2% 1328	17.1% 1538	22.2% 2061	21.5% 2076	24.9% 2440	14.2% 1271	8
Hubei	32.1%	31.0%	22.8%	21.1%	16.9%	18.0%	18.5%	19.9%	18.9%	23.3%	22.3%	4
	5314	5481	3586	3442	2620	2865	2998	3316	3268	3791	3668
Hunan	32.5%	30.1%	18.6%	17.9%	13.4%	15.2%	17.1%	25.7%	23.8%	26.4%	22.1%	5
	5257	5043	2777	2739	1942	2257	2603	3989	3816	4291	3471
Chongqing	33.9%	32.1%	22.0%	21.9%	11.7%	11.3%	12.8%	19.8%	18.6%	5.9%	19.0%	6
	2976	2980	1773	1885	909	904	1062	1690	1653	451	1628
Sichuan	39.2%	38.1%	32.9%	34.1%	28.6%	29.9%	28.5%	31.5%	31.3%	14.8%	30.9%	3
	7726	7834	6313	6786	5230	5617	5489	6270	6502	2417	6018
Guizhou	64.1%	62.7%	57.5%	56.2%	51.1%	51.1%	48.9%	51.3%	50.7%	51.9%	54.6%	1
	5817	6192	5346	5458	4774	4908	4815	5153	5283	5509	5325
Yunnan	47.6%	47.3%	42.6%	43.8%	43.2%	45.0%	45.5%	42.7%	41.4%	45.8%	44.5%	2
	4544	4932	4289	4576	4499	4826	5079	4949	5039	5949	4868
Downstream	0.0%	0.0%	4.8%	5.0%	5.6%	5.8%	5.8%	6.2%	6.0%	8.6%	4.8%
	0	0	566	601	689	731	759	1117	1120	1704	729
Midstream	23.4%	22.1%	16.1%	16.3%	14.7%	16.1%	17.6%	22.6%	21.4%	24.9%	19.5%
	3656	3631	2295	2328	1911	2150	2380	3122	3053	3508	2803
Upstream	46.2%	45.0%	38.7%	39.0%	33.6%	34.3%	33.9%	36.3%	35.5%	29.6%	37.2%
	5266	5484	4430	4676	3853	4064	4111	4516	4619	3581	4460
Average	23.2%	22.4%	20.2%	20.5%	18.3%	19.0%	19.3%	21.6%	20.9%	20.7%	20.6%
	2912	2985	2443	2554	2173	2330	2420	2900	2919	2879	2651

¹ In this study, the potential amount is calculated as coal equivalent per ten thousand tons. Among them, for the provinces and cities where the slack of the energy input index is 0, it does not mean that there is no room for energy conservation in that province or city, but rather that during the current sample under investigation, the amount of energy saved is 0.

.

Overall, the energy-saving potential exists in the Yangtze River Economic Belt. During the period from 2011 to 2020, the average energy-saving potential of the Yangtze River Economic Belt was 20.6% with the amount of 265.1 million tons of coal equivalent. The energy-saving potential of the Yangtze River Economic Belt has shown a fluctuating downward trend year by year, and it decreased to 20.7% in 2020, which indicated that energy efficiency made continual improvements.

From the perspective of the three major regions, the upstream region of the Yangtze River Economic Belt has the highest energy-saving potential, which is 37.2% with the amount of 44.6 million tons of coal equivalent, nearly 17 percentage points higher than the average level. However, the energy-saving potential of the downstream region is relatively small, at only 4.8%. The upstream region has the largest energy-saving potential, which decreased from 46.2% in 2011 to 29.6% in 2020, showing a favorable downward trend year by year.

At the provincial and municipal level, Guizhou has the largest average energy-saving potential with the amount of 53.25 million tons of coal equivalent, and Yunnan ranks second, both cities with a value more than twice the average. In addition, the average energy-saving potential of Sichuan, Hubei, and Hunan is greater than the average. Among the 11 provinces and cities, Shanghai and Zhejiang have the smallest energy-saving potential, with the amount of 0, indicating that their energy utilization levels have reached the optimal state.

3.4.2. Analyzing the Potential of Emission Reduction

Similar to the analysis of energy-saving potential, the reducible amounts of carbon dioxide and sulfur dioxide, respectively, represent the slack amounts of the undesired output indicators of emission reduction, and the results are depicted in

Table 5. The measurement result of emission-reducing potential in the Yangtze River Economic Belt.

Region	CO2 Ratio	CO2 Amount	Ranking (Ratio)	SO2 Ratio	SO2 Amount	Ranking (Ratio)	Average Ratio	Average Amount
Shanghai	0.0%	0	10	0.0%	0	10	0.0%	10
Jiangsu	11.1%	8861	9	27.1%	65135	9	19.1%	9
Zhejiang	0.0%	0	10	0.0%	0	10	0.0%	10
Anhui	38.4%	14,885	3	56.3%	141,491	8	47.3%	6
Jiangxi	31.2%	7118	4	77.4%	292,511	4	54.3%	3
Hubei	30.2%	10,315	5	68.8%	214,982	7	49.5%	5
Hunan	23.0%	7081	7	69.6%	251,344	6	46.3%	8
Chongqing	23.5%	3828	6	81.0%	242,027	3	52.3%	4
Sichuan	21.0%	6770	8	72.8%	319,879	5	46.9%	7
Guizhou	65.1%	15,929	1	93.0%	576,441	1	79.1%	1
Yunnan	40.5%	8050	2	85.5%	372,539	2	63.0%	2
Downstream	12.4%	5936		20.9%	51,657		16.6%
Midstream	28.1%	8171		72.0%	252,946		50.0%
Upstream	37.5%	8644		83.1%	377,722		60.3%
Average	25.8%	7531		57.4%	225,123		41.6%

1 The reducible amount of carbon dioxide (in ten thousand tons) is much larger than the reducible amount of sulfur dioxide (in tons). Similarly, for provinces and cities where the reducible amounts are 0, it means that the amount of pollutant emissions that can be reduced in the currently investigated research samples is 0.

.

Overall, there is potential for reducing the emissions of both carbon dioxide and sulfur dioxide in the Yangtze River Economic Belt. The average emission reduction potential of carbon dioxide and sulfur dioxide in the Yangtze River Economic Belt is 25.8% and 57.4% respectively, and the average emission-reducing potential of both is 41.6%. From the perspective of the three major regions, similar to the results of energy-saving potential, the upstream region has the largest potential for reducing carbon dioxide emissions. From the perspective of provinces and cities, Guizhou has the largest average potential for reducing carbon dioxide emissions, with a value as high as 65.1%, and Yunnan ranks second, with a value of 40.5%. Guizhou contributes the most to the potential for reducing carbon dioxide emissions in the entire Yangtze River Economic Belt. Also, Guizhou has the largest average potential for reducing sulfur dioxide emissions, with an average value of 93.0%. Yunnan ranks second, with an average value of 85.5%. Except for Jiangxi, all provinces and cities with a sulfur dioxide emission reduction potential greater than the average emission reduction potential of the Yangtze River Economic Belt are concentrated in the upstream region. Except for Shanghai and Zhejiang, most provinces and cities in the Yangtze River Economic Belt still have a large space for emission reduction potential and reducible amounts, and the average emission reduction potential is greater than the average energy-saving potential. This shows that the emission reduction tasks of each province and city in the Yangtze River Economic Belt are more significant and urgent compared to the energy-saving tasks.

Comparing the emission-reducing potential of carbon dioxide and sulfur dioxide, the reducible amount of carbon dioxide (in ten thousand tons) is much larger than the reducible amount of sulfur dioxide (in tons), indicating that the task of reducing carbon dioxide emissions in the entire Yangtze River Economic Belt remains urgent and severe. The overall reducible amounts of carbon dioxide and sulfur dioxide in the Yangtze River Economic Belt are 75.31 million tons and 225,123 tons, respectively. Although the emission reduction potential of sulfur dioxide is greater than that of carbon dioxide, the reducible amount of sulfur dioxide is much smaller than that of carbon dioxide. Among them, Guizhou Province has the largest reducible amounts of both carbon dioxide and sulfur dioxide, which are 159.29 million tons and 576,441 tons, respectively.

From the perspective of temporal changes, the carbon dioxide emission reduction potential of the Yangtze River Economic Belt shows a fluctuating upward trend in a “V” shape. The emission reduction potential declined from 2011 to 2013 and then began to rise from 2013 to 2020, continuing to grow. During the research period, it increased from 21.0% in 2011 to 33.1% in 2020. In contrast, the change range of the sulfur dioxide emission reduction potential in the Yangtze River Economic Belt is larger, and there are significant differences in the emission reduction potential between 2011 and 2020. The overall change trend of the sulfur dioxide emission reduction potential in the Yangtze River Economic Belt is upward, with an average annual growth rate of 4.2%. Specifically, the sulfur dioxide emission reduction potential increased from 40.1% in 2011 to 77.7% in 2020. Overall, the sulfur dioxide emission reduction potential has shown a significant growth during the research period.

3.5. The Pathway Analysis of Energy Efficiency Improvement

The improvement potential of 11 provinces in the Yangtze River Economic Belt is analyzed from the two aspects of energy conservation and emission reduction. These regions are classified into four different regional combinations according to the average value of energy conservation potential and emission reduction potential, as a two-dimensional matrix as depicted in Figure 3. Each combination represents different energy conservation and emission reduction performances. Combined with the current status of energy conservation and emission reduction potential in the Yangtze River Economic Belt, we define regions with an energy conservation potential average value of less than 25% as those with high energy utilization efficiency, and regions with a comprehensive emission reduction potential average value of less than 40% for the two environmental constraints (carbon dioxide and sulfur dioxide) as low emission reduction regions.

Overall, the 11 regions are distributed in the matrix quadrants with quantities of 3, 0, 3, and 5, respectively. Most regions in the Yangtze River Economic Belt are located in the fourth quadrant, which indicates significant potential for emission reduction. Moreover, regions in the first quadrant include not only two regions at the forefront of energy efficiency production but also Jiangsu Province. These three regions have high energy utilization levels and relatively high control over environmental pollutants in the Yangtze River Economic Belt. Regions in the second quadrant exhibit the characteristics of “high energy conservation, low emission reduction”, though no regions currently fall into this category in the Yangtze River Economic Belt. Regions in the fourth quadrant represent “low energy conservation, high emission reduction”, including Anhui, Jiangxi, Hunan, Hubei, and Chongqing. These regions perform well in energy utilization but have large environmental pollutant emissions. Therefore, compared with energy conservation tasks, regions in this quadrant should prioritize emission reduction efforts. Sichuan, Yunnan, and Guizhou, which are all located in the upstream region, fall into the third quadrant, which indicates significant potential for both energy conservation and emission reduction. Currently, these regions face substantial energy waste and severe pollutant emissions in production, making them key targets for energy conservation and emission reduction work in the Yangtze River Economic Belt.

Regions in the Yangtze River Economic Belt are distributed across the first, third, and fourth quadrants, and each can adopt appropriate energy conservation and emission reduction approaches based on their regional characteristics. Regions in the first quadrant are “high-efficiency” in energy conservation and emission reduction, which achieve both effective resource utilization and low pollutant emissions, serving as benchmarks for other regions. Regions in the fourth quadrant, with severe pollutant emissions, could implement a “④ → ①” unilateral breakthrough approach that prioritizes environmental pollution reduction by strengthening emission reduction management to improve emission reduction levels. These regions should enhance pollution control, actively explore effective treatment and recycling methods, and improve energy efficiency. Regions in the third quadrant, with high energy conservation and emission reduction potential, represent extensive development models that face both low energy utilization and environmental pollution control issues. Their energy conservation and emission reduction improvement could adopt “③ → ② → ①” or “③ → ④ → ①” unilateral progressive approaches in order to formulate priority strategies for energy conservation or emission reduction based on current performance, then transitioning to the first quadrant. Additionally, a “③ → ①” leapfrog approach is feasible, which requires simultaneous focus on both energy conservation and emission reduction through measures such as introducing relevant policies, phasing out backward industries, optimizing industrial structures as well as introducing energy-saving technologies and talent, all of which jointly promote energy efficiency improvements in these regions.

4. Discussion and Conclusions

This paper firstly studied the energy efficiency of 11 provinces/cities in the Yangtze River Economic Belt and conducted an in-depth analysis of the measurement results. Secondly, it used time-varying trend charts to perform temporal differentiation analysis on the Yangtze River Economic Belt, and employs the global Moran’s I and local Moran’s I to analyze the overall spatial pattern and local regional differences in energy efficiency. Thirdly, the potential model of energy conservation and emission reduction is constructed, and we further explore the pathways of energy efficiency improvement for each region and city in the Yangtze River Economic Belt.

The temporal and spatial differentiation analysis of energy efficiency in the Yangtze River Economic Belt reveals that:

• Temporal Variation: Over the study period, the overall energy efficiency of the Yangtze River Economic Belt showed a gently declining trend. Regional differences were evident among the three major zones:

  ○

  Downstream region: Maintained a high overall energy efficiency level but trended downward.

  ○

  Midstream region: Experienced fluctuating declines.

  ○

  Upstream region: Had a lower overall energy efficiency but showed an upward trend.

• Spatial Pattern: The energy efficiency exhibited unbalanced development. The global Moran’s I passed the 1% significance test, indicating a significant positive correlation and strong spatial agglomeration characteristics. Local Moran scatter plots revealed that most downstream areas showed “H-H agglomeration” (high–high clustering), while most mid-upstream areas showed “L-L agglomeration” (low–low clustering). The energy efficiency presented a spatial distribution pattern of “significant differences between upstream, midstream, and downstream regions with severe polarization,” further confirming obvious spatial differentiation characteristics.

Analysis of energy conservation and emission reduction potential:

• Energy conservation potential: The energy conservation potential of the Yangtze River Economic Belt showed little variation and generally trended downward.
• Emission reduction potential: The Pollutant emission reduction potential fluctuated significantly. The reducing potentials for carbon dioxide (CO₂) and sulfur dioxide (SO₂) continuously increased. The overall emission reduction potential remained high, with the SO₂ reducing potential far exceeding the CO₂ reducing potential during the study period. However, the reducible amount of CO₂ was much larger than that of SO₂, highlighting the urgent need to address both CO₂ and SO₂ emission reduction tasks.

Region-specific emission reduction pathways based on potential:

• Regions in the first quadrant (high-efficiency type): Provinces/cities with relatively low energy conservation and emission reduction potentials serve as benchmarks for other regions, demonstrating efficient resource utilization and strict pollution control.
• Regions in the fourth quadrant (high emission reduction potential): Provinces/cities should adopt a “④ → ①” unilateral breakthrough pathway, prioritizing emission reduction tasks by strengthening emission management to improve emission reduction levels.
• Regions in the third quadrant (extensive development type): Provinces/cities may choose either a progressive pathway (“③ → ② → ①” or “③ → ④ → ①”) to gradually optimize energy efficiency and pollution control or a leapfrog pathway (“③ → ①”) to simultaneously advance energy conservation and emission reduction through industrial restructuring, technology adoption, and policy implementation.

Situating this research within the broader literature on regional energy efficiency, our observation of declining energy efficiency in downstream regions aligns with the existing studies on industrial overcapacity in coastal areas, while the upstream’s upward trend corroborates findings about technology spillovers from eastern China. Moreover, in order to further extend extant literature, our study demonstrates that the Yangtze River Economic Belt’s strong spatial clustering exacerbates regional disparities, highlighting the need for targeted policy interventions. These discrepancies underscore the importance of context-specific analysis in understanding energy efficiency dynamics across diverse economic regions, and potential pathways for energy efficiency improvement are indicated.

Meanwhile, the systematic framework for regional energy efficiency analysis applied in this research offers valuable references for other regions. The integrated framework proposed including an analysis of spatial–temporal dynamics and conservation potential that establishes a replicable methodology for regional energy governance, emphasizing that context-specific adjustments are essential for transferring its insights effectively. Finally, this study is constrained by its reliance on data up to 2020, potentially limiting generalizability to post-2020 trends. Future research could incorporate 2021–2025 data to assess recent developments, explore emerging variables, and validate findings against evolving contexts.