The Dynamic Evolution of Industrial Electricity Consumption Linkages and Flow Path in China

Abstract

An in-depth investigation into the evolutionary characteristics, transmission mechanisms, and optimization pathways of electricity consumption linkages across China’s industrial sectors highlights their substantial theoretical and practical significance in achieving the “dual carbon” goals and advancing high-quality economic development. This study investigates the structural characteristics and developmental trends of electricity consumption linkages across China’s industrial sectors using an enhanced hypothetical extraction method. The analysis draws on national input–output tables and sector-specific electricity consumption data during the period from 2002 to 2020. Key transmission routes between industrial sectors are identified through path analysis and average path length calculations. The findings reveal that China’s industrial electricity consumption structure is marked by notable scale expansion and differentiation. The magnitude of inter-sectoral electricity flows continues to grow steadily. The evolution of these linkages exhibits clear phase-specific patterns, while the intensity of electricity consumption connections across sectors shows pronounced heterogeneity. Furthermore, the transmission path analysis revealed differentiated characteristics of electricity influence transmission, with generally shorter internal paths within sectors, significant cross-sectoral transmission differences, and manufacturing demonstrating good transmission accessibility with moderate path distances to major sectors. These insights provide a robust foundation for designing differentiated energy conservation policies, as well as for optimizing the overall structure of industrial electricity consumption.

1. Introduction

Electricity serves as the core energy carrier in modern society and has become an increasingly prominent focus of international concern amid global climate change and ongoing energy transitions. As the globe’s second-largest economic power and the leading consumer of energy resources, China holds a central position in influencing worldwide energy trends and developments. Its electricity utilization efficiency and structural composition are not only essential for the sustainable advancement of its national economy and social systems but also exert significant impacts on global energy patterns and climate governance frameworks [1,2]. According to the National Energy Administration, total social electricity consumption reached 8.710 trillion kilowatt-hours in 2023 [3]. Over the past decade, the average annual growth rate of electricity consumption has remained consistently above 5%, markedly outpacing the country’s GDP growth during the same period [4]. This persistent growth pattern underscores the pivotal role of electricity in driving China’s economic development and emphasizes the necessity of in-depth research into the structure of its electricity consumption [5].

China’s structural economic transformation has fundamentally reshaped electricity consumption patterns. Over the course of its structural transition, the tertiary industry’s GDP contribution increased from 46.9% to 54.3% during 2013–2023, while its share of total electricity consumption rose by 3.2 percentage points, reaching 17.9% in 2023. Meanwhile, industrial electricity consumption remained dominant at 68.4%, reflecting heterogeneous linkages across sectors [6]. In response to these structural shifts, the Chinese government established a comprehensive policy framework through key documents such as the “Carbon Peak Action Plan Before 2030” and the “14th Five-Year Plan for Modern Service Industry Development.” These measures are specifically designed to promote the shift toward low-carbon practices in energy-demanding industrial sectors and enhance energy efficiency within modern service sectors [4]. This dual-track governance approach not only directly influences electricity utilization characteristics across industries but also triggers indirect topological reconfigurations in cross-sectoral electricity consumption networks through policy intervention effects, thereby establishing an institutional foundation for analyzing the evolutionary dynamics of industrial electricity linkage systems.

The research on industrial electricity consumption linkages in developed economies has reached relative maturity from the international comparison perspective. Europe focused on policy optimization of inter-sectoral electricity dependency relationships under the background of renewable energy transition, emphasizing the reduction in industrial sector electricity intensity through energy efficiency policies and electricity market integration [7]. The United States concentrated on the structural changes in manufacturing energy consumption and the impact of technological progress on inter-sectoral energy linkages [8]. Japan emphasized the reshaping role of energy policy adjustments on industrial electricity consumption efficiency and inter-sectoral linkage intensity [9]. However, China, as the world’s largest developing economy and manufacturing powerhouse, exhibited significant uniqueness in its industrial electricity consumption linkages: the absolute dominance of manufacturing in electricity consumption, high electricity dependence in the construction industry driven by rapid urbanization, and policy-driven structural adjustments under the constraints of the “dual carbon” goals. These characteristics made China’s industrial electricity linkage analysis not only of important theoretical innovation value but also provided key empirical references for energy transitions in other developing countries. Under the “dual carbon” goals, enhancing energy utilization efficiency has emerged as a central priority for China’s economic development [10]. In 2023, Chinese energy consumption decreased by 10.5% compared to 2020. Nevertheless, a substantial gap remains when compared to developed countries. That year, Chinese electricity consumption reached 0.490 kWh/USD, significantly higher than that of Japan (0.280 kWh/USD) and Germany (0.250 kWh/USD) [11]. This gap underscores the considerable potential for improvement in China’s energy efficiency and emphasizes the necessity of investigating industrial electricity efficiency. Electricity consumption linkages among industries reflect the interdependent energy relationships embedded in production processes [12]. According to data from the National Bureau of Statistics, the intermediate input ratio of manufacturing to other industries reached 61.2% in 2022, representing an increase of 3.5 percentage points compared to 2012 [13]. This deepening of industrial linkages indicates that changes in electricity efficiency within a single industry can have widespread ripple effects across industrial chains. For example, technological advancements in energy conservation within the steel industry not only directly reduce its own electricity consumption but also indirectly influence the electricity efficiency of downstream manufacturing sectors through shifts in product pricing structures [14]. The development of new energy sources is driving transformative shifts in industrial electricity consumption patterns. In 2023, the share of renewable energy generation in China reached 32.7%, marking an increase of 15.2% percentage points compared to 2013. This transformation modifies the electricity supply structure and expands clean energy options for industrial sectors, thereby accelerating the green transition of industry. These structural changes also give rise to new dynamics in inter-industry electricity consumption linkages.

The in-depth examination of electricity consumption linkages across China’s industrial sectors holds significant relevance within this context [15]. Analyzing these linkages and their dynamic evolution enhances our understanding of how China’s economic structural transformation influences energy consumption patterns. This analysis provides a scientific foundation for formulating more targeted energy and industrial policies. This study aims to thoroughly investigate trends in inter-sectoral electricity consumption linkages by innovatively applying an improved hypothetical extraction method integrated with path analysis. The findings reveal the intricate relationship between China’s economic development and energy efficiency, offering valuable insights for advancing the “dual carbon” goals and fostering sustainable development.

2. Literature Review

The theoretical evolution of the research on the correlation of electricity consumption reflects the continuous deepening of the academic community’s understanding of energy flow in the economic system [16]. Early studies mainly focused on the quantitative analysis of direct electricity consumption. The input–output analysis proposed by Leontief in 1936 laid the methodological foundation for this field [17]. Subsequently, some scholars incorporated energy input into the input–output model, pioneering the input–output analysis of energy and providing theoretical support for subsequent research [18]. These pioneering studies have provided a preliminary analytical framework for the study of electricity consumption correlations in industrial sectors, but they are still limited to the consideration of direct electricity consumption.

With the deepening of research, the academic community gradually realized that focusing only on direct electricity consumption cannot fully reflect the complex energy dependence among industries. To solve such problems, electricity could be incorporated as an independent factor into the input–output analysis framework, and the importance of indirect electricity consumption relationships among industries was systematically expounded for the first time [19]. This theoretical breakthrough significantly expanded the connotation of the research on the correlation of electricity consumption, providing a new theoretical perspective for subsequent scholars [16]. The concept of “total energy intensity” was proposed, integrating direct and indirect energy consumption into a unified analytical framework, offering a foundation for an in-depth grasp of energy interdependencies across industrial sectors [20]. The concept of “embedded energy” was proposed, marking an important milestone in this transformation in research [21]. Their research highlights the implicit indirect energy consumption in the product manufacturing process, providing a new theoretical tool for understanding the complex electricity transmission mechanism among industries. On this basis, some scholars have proposed the energy footprint analysis method from the “consumption perspective”, further expanding the theoretical boundary of industrial electricity consumption correlation research, enabling researchers to grasp the energy flow in the global value chain more comprehensively [22]. The continuous improvement of the theoretical framework has gradually shifted the research focus to the structural characteristics and transmission mechanisms of electricity consumption correlations among industrial sectors. By constructing a multi-level energy flow model, the mechanism of indirect electricity transmission among industries was further clarified, providing important insights for an in-depth understanding of the complexity of industrial electricity correlation [23]. Applying this method reveals the energy flow pattern in international trade and opens up a new field for the study of industrial electricity consumption correlation from a global perspective [24]. The application of game theory methods in power system optimization and policy design provided new theoretical perspectives for understanding the strategic interactions of electricity consumption among industrial sectors. Analyzing strategic behaviors in electricity markets through the game theory framework helped to deeply understand the complex interactive relationships among sectors in a large-scale power system [25].

The energy policies on the electricity consumption correlation of industrial sectors have subsequently become the focus of academic attention. By constructing a dynamic energy price–industrial electricity consumption model, the impact of energy price changes on the industrial electricity consumption structure in Japan was deeply analyzed, and it was found that there were significant differences in the sensitivity of different industries to energy price changes [26]. Meanwhile, through the multi-regional input–output model, the impact of carbon emission policies on the energy usage patterns in the global industrial chain was systematically studied, revealing the cross-border energy consumption transfer effect that policy intervention might bring [27]. Policy effect analysis based on electricity price reduction policies had differentiated impacts on the sustainable development of different industries, with heavily polluting industries showing significantly higher response levels than clean industries, which provided important empirical evidence for understanding the industry heterogeneity effects of electricity price policies [28]. A cost–benefit analysis of new power system construction under the dual carbon goals revealed the reshaping mechanism of large-scale renewable energy integration on industrial electricity price structures, with continuous increases in system costs creating upward pressure on future electricity prices [29]. These studies provide important theoretical support for understanding the complex impacts of energy policies.

Under this research background, the hypothesis extraction method, as an effective tool for industrial correlation analysis, has gradually been introduced into related research fields. This method was first proposed by Paelinck [30]. Its core idea is to assume the “extraction” of a certain industry, thereby accurately assessing the importance of that industry [30,31]. The hypothesis extraction method has undergone systematic theoretical summary and methodological expansion, and multiple extraction schemes have been proposed, significantly improving the applicability and flexibility of this method [32]. This method was used to identify key industrial sectors, providing an important theoretical basis for the formulation of industrial policies [33]. In regional economic research, the hypothesis extraction method is used to identify regional pillar industries. Further, the hypothesis extraction method is combined with the multi-regional input–output model to construct a cross-regional industrial correlation analysis framework, providing theoretical support for studying the deep mechanism of regional economic integration [34]. With the deepening of research, the hypothesis extraction method is increasingly widely applied in the energy usage analysis of industrial sectors. Based on this method, a decomposition model of industrial energy consumption was constructed, and the direct and indirect contributions of different industrial sectors to energy consumption were systematically analyzed [35]. This research not only deepens the understanding of the industrial energy usage structure but also provides a new analytical tool for assessing the impact of industrial policies on energy consumption [36]. The hypothesis simulation model became successful in analyzing the individual and systematic effects of different sectors on national carbon emission intensity by integrating multi-regional input–output analysis and hypothesis extraction techniques [37]. Temporal and spatial analysis at the prefecture-level city level systematically revealed the spatial distribution characteristics and evolution patterns of the relationship between industrial electricity consumption and economic growth [38].

The application scope of the hypothesis extraction method in industrial energy research was further expanded. This method was innovatively applied to assess the impact of changes in energy usage in manufacturing on the macroeconomy, revealing the complex relationship between industrial energy efficiency and economic development [39]. The research further revealed the existence of significant nonlinear characteristics between Chinese electricity consumption and industrial growth, and this relationship showed obvious heterogeneity across different development stages and regions [40]. The dynamic evolution of energy consumption patterns during the process of industrial structure transformation was analyzed through the hypothesis extraction method, providing a new perspective for understanding the relationship between economic transformation and energy consumption efficiency [41]. With the expansion of its application scope, the hypothesis extraction method itself is also constantly developing and improving. The partial extraction method effectively improves the adaptability of the model to the actual economic situation by introducing a variable extraction ratio [42]. The nonlinear hypothesis extraction rule further enhances the ability of this method to capture complex industrial energy correlations, providing a theoretical basis for analyzing nonlinear energy flows in the industrial chain [43]. The diversified development of analytical methods provided strong support for an in-depth understanding of electricity consumption mechanisms. Research based on logarithmic mean Divisia index decomposition systematically analyzed the multiple driving factors of electricity consumption changes, which provided a scientific basis for formulating differentiated energy-saving policies [44]. The relationship between electricity consumption and economic transformation provided effective analytical frameworks for capturing dynamic interdependent relationships among variables, which revealed the temporal characteristics of China’s economic transformation [45]. Human capital investment was incorporated into the industrial electricity consumption analysis framework as a new influencing factor. This enriched the theoretical understanding of the economic effects of electricity consumption [46]. City-level electricity consumption research provided important supplements for understanding micro-mechanisms. The socioeconomic factors, such as per capita disposable income, had significant impacts on electricity consumption growth, which provided a scientific basis for urban energy planning and policy formulation [47]. The innovations of these methodologies not only enhance the applicability of hypothesis extraction in industrial energy research but also provide a more advanced analytical framework for deepening the understanding of the complexity of industrial energy systems.

This research proposes the following hypotheses based on the existing literature analysis: (1) The electricity consumption linkages among China’s industrial sectors presented significant structural differentiation characteristics, with different sectors playing differentiated functional roles in the electricity linkage network. (2) The evolution of industrial electricity consumption linkages followed clear phased patterns, which reflected the internal logic of China’s economic structural transformation.

3. Model and Data

3.1. Model

3.1.1. Vertical Integration Measurement Method

The economic system includes n industrial sectors. Each sector participates in production as both an intermediate input producer and consumer. Based on input–output theory, the basic balance relationship between sectors can be expressed as

$$\mathrm{x}=\mathrm{A}\mathrm{x}+\mathrm{y}$$

(1)

where x represents the total output vector, y represents the final demand vector, and A represents the direct consumption coefficient matrix. Equation (1) can be transformed into

$$\mathrm{x}={(\mathrm{I}-\mathrm{A})}^{-1}\ast \mathrm{y}$$

(2)

where I represents the n-dimensional identity matrix, and ${(\mathrm{I}-\mathrm{A})}^{-1}$ represents the Leontief inverse matrix.

Let ${\mathrm{w}}_{\mathrm{j}}$ represent the direct electricity consumption of sector j, and ${\mathrm{x}}_{\mathrm{j}}$ represent the total output of sector j. The direct electricity coefficient of sector j is

$${\mathrm{q}}_{\mathrm{j}}={\displaystyle \frac{{\mathrm{w}}_{\mathrm{j}}}{{\mathrm{x}}_{\mathrm{j}}}}$$

(3)

where ${\mathrm{q}}_{\mathrm{j}}$ represents the direct electricity coefficient of sector j, indicating the electricity consumption per unit of product in sector j.

Although the direct electricity coefficients provided intuitive measures of sectoral electricity consumption intensity, their analytical scope was limited to direct consumption within sectoral boundaries and did not cover indirect electricity demand embedded through industrial chains. This limitation was significant in practical analysis: the direct electricity coefficient of manufacturing was 0.0152 in 2020, but the actual electricity consumption intensity would be significantly higher than this direct measure when considering electricity-intensive intermediate inputs purchased from upstream sectors such as steel and chemicals. Therefore, accurately assessing sectoral electricity consumption characteristics required the introduction of more comprehensive analytical indicators. The complete electricity coefficient vector B is obtained by right multiplying the direct electricity coefficient row vector q with the inverse matrix:

$$\mathrm{B}=\mathrm{q}\ast {(\mathrm{I}-\mathrm{A})}^{-1}$$

(4)

where q = (${\mathrm{q}}_{\mathrm{j}}$) represents the direct electricity coefficient row vector; vector B = (${\mathrm{b}}_{\mathrm{j}}$), and its element b_j represents the complete electricity coefficient of sector j, indicating the direct and indirect electricity consumption in various sectors of the economic system caused by a per unit change in the final demand of sector j.

Pasinetti [48] constructed the vertical integration consumption (VIC) based on the Leontief inverse matrix:

$${\mathrm{V}\mathrm{I}\mathrm{C}}_{\mathrm{j}}=\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{q}}_{\mathrm{i}}{\mathsf{\alpha }}_{\mathrm{i}\mathrm{j}}{\mathrm{y}}_{\mathrm{j}}$$

(5)

where ${\mathrm{V}\mathrm{I}\mathrm{C}}_{\mathrm{j}}$ represents the vertical integration consumption of sector j; ${\mathrm{q}}_{\mathrm{i}}$ represents the direct electricity coefficient of sector i; ${\mathsf{\alpha }}_{\mathrm{i}\mathrm{j}}$ represents the element of Leontief inverse matrix ${(\mathrm{I}-\mathrm{A})}^{-1}$, namely the complete consumption coefficient; ${\mathrm{y}}_{\mathrm{j}}$ represents the final demand of sector j. VIC represents the quantity of direct and indirect electricity consumption required to meet the final demand of industrial sector j. It combines direct electricity coefficients, complete consumption coefficients, and final demand to express sectoral electricity characteristics through quantities rather than coefficients. The numerical differences between VIC and direct consumption intuitively demonstrated the complexity of electricity transmission through industrial chains. The direct electricity consumption of the construction industry was 10.111 billion kWh in 2020, while the vertically integrated consumption reached 185.0457 billion kWh, with the difference exceeding 18 times. This significant difference indicated that although the construction industry had relatively low electricity consumption itself, it indirectly drove large-scale electricity consumption in the entire economic system through substantial demand for electricity-intensive materials such as cement, steel, and glass, which fully reflected the important role of terminal demand sectors in the electricity linkage network.

3.1.2. Improved Hypothetical Extraction Method

Hypothetical Extraction Method

Strassert [49] applied the hypothetical extraction method to analyze economic impacts during industrial structure changes. Based on this, Cella [50] adopted the concept of the hypothetical extraction method to reveal relationships among net forward linkages (NFL), net backward linkages (NBL), and total linkages of industrial sectors. The method divides the economic system into two industrial groups, ${\mathrm{B}}_{\mathrm{s}}$ and ${\mathrm{B}}_{-\mathrm{s}}$. For each industrial group ${\mathrm{B}}_{\mathrm{s}}$, its economic structure can be described as

$$\begin{gathered} \left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_{\mathrm{s}}}{{\mathrm{x}}_{-\mathrm{s}}}}\right)=\left(\begin{array}{cc}{\mathrm{A}}_{\mathrm{s},\mathrm{s}}& {\mathrm{A}}_{\mathrm{s},-\mathrm{s}}\\ {\mathrm{A}}_{-\mathrm{s},\mathrm{s}}& {\mathrm{A}}_{-\mathrm{s},-\mathrm{s}}\end{array}\right)\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_{\mathrm{s}}}{{\mathrm{x}}_{-\mathrm{s}}}}\right)+\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{y}}_1}{{\mathrm{y}}_2}}\right)\iff \left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_{\mathrm{s}}}{{\mathrm{x}}_{-\mathrm{s}}}}\right)=\left(\begin{array}{cc}{\mathsf{\Delta }}_{\mathrm{s},\mathrm{s}}& {\mathsf{\Delta }}_{\mathrm{s},-\mathrm{s}}\\ {\mathsf{\Delta }}_{-\mathrm{s},\mathrm{s}}& {\mathsf{\Delta }}_{-\mathrm{s},-\mathrm{s}}\end{array}\right)\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{y}}_{\mathrm{s}}}{{\mathrm{y}}_{-\mathrm{s}}}}\right) \\ {(\mathrm{I}-\mathrm{A})}^{-1}=\left(\begin{array}{cc}{\mathsf{\Delta }}_{\mathrm{s},\mathrm{s}}& {\mathsf{\Delta }}_{\mathrm{s},-\mathrm{s}}\\ {\mathsf{\Delta }}_{-\mathrm{s},\mathrm{s}}& {\mathsf{\Delta }}_{-\mathrm{s},-\mathrm{s}}\end{array}\right) \end{gathered}$$

(6)

In addition, it is assumed that an industrial group ${\mathrm{B}}_{\mathrm{s}}^{*}$, ${\mathrm{B}}_{\mathrm{s}}^{*}$ does not trade with other sectors in the virtual economic system, and its assumed production association is given by the following formula:

$$\begin{gathered} \left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_s^{*}}{{\mathrm{x}}_{-\mathrm{s}}^{*}}}\right)=\left(\begin{array}{cc}{\mathrm{A}}_{\mathrm{s}\mathrm{s}}& 0\\ 0& {\mathrm{A}}_{-\mathrm{s},-\mathrm{s}}\end{array}\right)\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_{\mathrm{s}}^{*}}{{\mathrm{x}}_{-\mathrm{s}}^{*}}}\right)+\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{y}}_{\mathrm{s}}}{{\mathrm{y}}_{-\mathrm{s}}}}\right)\iff \left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_{\mathrm{s}}^{*}}{{{\mathrm{x}}_{-}^{*}}_{-\mathrm{s}}}}\right) \\ =\left(\begin{array}{cc}{\left(\mathrm{I}-{\mathrm{A}}_{\mathrm{s},\mathrm{s}}\right)}^{-1}& 0\\ 0& {\left(\mathrm{I}-{\mathrm{A}}_{-\mathrm{s},-\mathrm{s}}\right)}^{-1}\end{array}\right)\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{y}}_{\mathrm{s}}}{{\mathrm{y}}_{-\mathrm{s}}}}\right) \end{gathered}$$

(7)

A comparison of these two economic systems yields the following equation for measuring the influence degree of industrial groups through production changes:

$$\begin{gathered} \mathrm{x}-{\mathrm{x}}^{\mathrm{*}}=\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{x}}_{\mathrm{s}}-{\mathrm{x}}_{\mathrm{s}}^{\mathrm{*}}}{{\mathrm{x}}_{-\mathrm{s}}-{\mathrm{x}}_{-\mathrm{s}}^{\mathrm{*}}}}\right) \\ =\left(\begin{array}{ll}{\mathsf{\Delta }}_{\mathrm{s},\mathrm{s}}-{\left(\mathrm{I}-{\mathrm{A}}_{\mathrm{s}.\mathrm{s}}\right)}^{-1}& {\mathsf{\Delta }}_{\mathrm{s},-\mathrm{s}}\\ {\mathsf{\Delta }}_{-\mathrm{s},\mathrm{s}}& {\mathsf{\Delta }}_{\mathrm{s},-\mathrm{s}}-{\left(\mathrm{I}-{\mathrm{A}}_{-\mathrm{s},-\mathrm{s}}\right)}^{-1}\end{array}\right)\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{y}}_{\mathrm{s}}}{{\mathrm{y}}_{-\mathrm{s}}}}\right) \\ =\left(\begin{array}{cc}{\mathrm{C}}_{\mathrm{s},\mathrm{s}}& {\mathrm{C}}_{\mathrm{s},-\mathrm{s}}\\ {\mathrm{C}}_{-\mathrm{s},\mathrm{s}}& {\mathrm{C}}_{-\mathrm{s},-\mathrm{s}}\end{array}\right)\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{y}}_{\mathrm{s}}}{{\mathrm{y}}_{-\mathrm{s}}}}\right) \end{gathered}$$

(8)

This yields total, backward, and NFL of industrial group ${\mathrm{B}}_{\mathrm{s}}$. Total linkage:

$${\mathrm{T}\mathrm{L}}_{\mathrm{s}}={\mathrm{u}}^{\prime }(\mathrm{x}-{\mathrm{x}}^{*})$$

$${\mathrm{B}\mathrm{L}}_{\mathrm{s}}={\mathrm{u}}^{\prime }\left({\mathrm{C}}_{\mathrm{s},\mathrm{s}}{\mathrm{C}}_{-\mathrm{s},\mathrm{s}}\right){\mathrm{y}}_{\mathrm{s}}$$

$${\mathrm{F}\mathrm{L}}_{\mathrm{s}}={\mathrm{u}}^{\prime }\left({\displaystyle \genfrac{}{}{0pt}{}{{\mathrm{C}}_{-\mathrm{s},\mathrm{s}}}{{\mathrm{C}}_{-\mathrm{s},-\mathrm{s}}}}\right){\mathrm{y}}_{-\mathrm{s}}$$

(9)

where u′ represents the unit vector (1,…, 1); TL represents the total linkage; BL is the backward linkage; FL is the forward linkage.

Improvement in Linkage Degree Decomposition

Sanchez Choliz and Duarte [51] improved the HEM method based on this theoretical foundation. This study applies this method to analyze industrial electricity linkage characteristics. The improved HEM method decomposes linkages in the hypothetical extraction method into four independent factors in VIC form:

$${\mathrm{I}\mathrm{E}}_{\mathrm{s}}={\mathrm{q}}_{\mathrm{s}}{\left(\mathrm{I}-{\mathrm{A}}_{\mathrm{s},\mathrm{s}}\right)}^{-1}{\mathrm{y}}_{\mathrm{s}}$$

$${\mathrm{M}\mathrm{E}}_{\mathrm{s}}={\mathrm{q}}_{\mathrm{s}}^{\prime }{\mathrm{C}}_{\mathrm{s},\mathrm{s}}{\mathrm{y}}_{\mathrm{s}}={\mathrm{q}}_{\mathrm{s}}^{\prime }({\mathsf{\Delta }}_{\mathrm{s},\mathrm{s}}-{\left(\mathrm{I}-{\mathrm{A}}_{\mathrm{s},\mathrm{s}}\right)}^{-1}){\mathrm{y}}_{\mathrm{s}}$$

$${\mathrm{N}\mathrm{B}\mathrm{L}}_{\mathrm{s}}={\mathrm{q}}_{-\mathrm{s}}^{\prime }{\mathrm{C}}_{\mathrm{s},-\mathrm{s}}{\mathrm{y}}_{\mathrm{s}}={\mathrm{q}}_{-\mathrm{s}}^{\prime }{\mathsf{\Delta }}_{-\mathrm{s},\mathrm{s}}{\mathrm{y}}_{\mathrm{s}}$$

$${\mathrm{N}\mathrm{F}\mathrm{L}}_{\mathrm{s}}={\mathrm{q}}_{-\mathrm{s}}^{\prime }{\mathrm{C}}_{\mathrm{s},-\mathrm{s}}{\mathrm{y}}_{-\mathrm{s}}={\mathrm{q}}_{-\mathrm{s}}^{\prime }{\mathsf{\Delta }}_{\mathrm{s},-\mathrm{s}}{\mathrm{y}}_{-\mathrm{s}}$$

(10)

The improved HEM method enables a quantitative analysis of linkage influence effects in electricity resource consumption between industrial sectors and determines internal transfer flows during electricity consumption by sectors. Additionally, this study introduces the Net Effect (NT) indicator to better evaluate the comprehensive positioning of industrial sectors in electricity linkage networks. NT represents the difference between NBL and NFL. A positive NT value indicates net input characteristics of the sector in electricity linkage networks, meaning that the electricity effects received from other sectors exceed those transmitted to other sectors; conversely, a negative value indicates net output characteristics.

In traditional input–output analysis, the common approach is to analyze and identify the electricity-related linkages and characteristics of economic sectors by means of direct or total electricity coefficients and degrees of correlation. This method, compared to traditional methods, quantified the systemic importance by hypothetically “extracting” specific sectors, and this method could simultaneously capture both direct and indirect effects. More importantly, the four-factor decomposition mechanism (IE, ME, NBL, and NFL) provided a more detailed decomposition framework than traditional input–output multipliers, which could accurately identify the specific transmission channels of electricity effects.

3.1.3. Average Propagation Path Length (APL)

This research introduced the APL based on the improved hypothesis extraction method to comprehensively understand the transmission mechanisms and network characteristics of electricity consumption linkages among industrial sectors. The APL measured the average number of steps required for electricity consumption changes in one sector to propagate through the industrial chain and affect another sector, which provided an important tool for deeply understanding the complexity and transmission efficiency of electricity influence propagation. The APL was based on the structural path analysis method and utilized the Leontief inverse matrix L to quantify the transmission distances between sectors. An auxiliary matrix H was defined to quantify the influence propagation among sectors, as shown in Equation (11).

$$\mathrm{H}=\mathrm{L}\left(\mathrm{L}-\mathrm{I}\right)$$

(11)

The calculation of the APL needed to distinguish between diagonal elements and non-diagonal elements to accurately reflect different propagation mechanisms. The APL measurement between the path target industry i and path origin industry j for each sector was calculated as Equation (12).

$${\mathrm{q}}_{\mathrm{i}\mathrm{j}}=\left\{\begin{array}{c}{\displaystyle \frac{{\mathrm{h}}_{\mathrm{i}\mathrm{j}}}{{\mathrm{l}}_{\mathrm{i}\mathrm{j}}}},\mathrm{i}\ne \mathrm{j}\\ {\displaystyle \frac{{\mathrm{h}}_{\mathrm{i}\mathrm{j}}}{\left({\mathrm{l}}_{\mathrm{i}\mathrm{j}}-1\right)}},\mathrm{i}=\mathrm{j}\end{array}\right.$$

(12)

where ${\mathrm{q}}_{\mathrm{i}\mathrm{j}}$ represented the average propagation path length between sector i and sector j, ${\mathrm{l}}_{\mathrm{i}\mathrm{j}}$ represented the elements of the Leontief inverse matrix L, and ${\mathrm{h}}_{\mathrm{i}\mathrm{j}}$ represented the elements of the auxiliary matrix H. The network characteristics and transmission path differences in electricity influence propagation among industrial sectors could be identified through the APL analysis. The combination of the APL analysis and the improved hypothesis extraction method provided a more complete analytical perspective for understanding the dynamic evolution of China’s industrial electricity consumption structure and provided a scientific basis for formulating differentiated energy-saving measures based on path analysis.

3.2. Data

This study primarily uses data from China’s input–output tables from 2002 to 2020 published by the National Bureau of Statistics. The relevant data could be obtained through the statistical data section of the official website of the National Bureau of Statistics (http://www.stats.gov.cn/). China compiles benchmark input–output tables every five years, with extension tables for other years. This study selects input–output tables from 2002, 2005, 2007, 2010, 2012, 2015, 2017, 2018, and 2020 as basic data. Sectoral electricity consumption data comes from the corresponding years of the China Energy Statistical Yearbook. Considering industrial hierarchical characteristics and electricity linkage degrees, input–output table sectors merge into eight industrial sector groups while maintaining inter-sectoral differences. The specific classification includes the following: (1) Agriculture, Forestry, Animal Husbandry, and Fishery (AFAHF); (2) Mining Industry (MI); (3) Manufacturing (MF); (4) Production and Supply of Electricity, Heat, Gas, and Water (PSEHGW); (5) Construction (GT); (6) Transportation, Storage, Postal and Hotel Communication Services (TSPHCS); (7) Wholesale and Retail Trade (WRT); (8) Other Industries (OTH).

However, several limitations also existed in data applications. First, the compilation of input–output tables had obvious time lags, with benchmark tables compiled every five years. Although extension tables filled the time gap, their accuracy was relatively low. Second, the consistency problem of sectoral classification might affect the accuracy of intertemporal comparisons, particularly during periods of rapid economic structural change. Third, input–output tables reflected average technological relationships and could not capture firm-level heterogeneity and the dynamic effects of technological progress.

This grouping scheme further considered three key factors based on industrial hierarchy characteristics and electricity linkage degrees. The consideration for selecting eight groups rather than more or fewer was to balance analytical precision with statistical significance: too many groups would result in too few industries within some categories, making it difficult to reflect the representativeness of groupings; too few groups would blur the electricity consumption differences among different types of industries, reducing the policy relevance of the analysis. Compared with similar international research, developed countries usually focus more on service industry subdivision, while developing countries emphasize manufacturing differentiation. China’s eight-group scheme fully reflects the structural characteristics of the current economic development stage, where manufacturing dominance coexists with rapid service industry development.

4. Results and Discussion

4.1. Overall Characteristics of Industrial Electricity Consumption Linkages

The analysis of China’s industrial electricity consumption linkages from 2002 to 2020 based on the improved hypothetical extraction method indicates significant characteristics of scale expansion and structural differentiation in industrial electricity consumption structure. As shown in Figure 1, the total direct electricity consumption of all sectors increased from 1469.403 billion kWh in 2002 to 6622.367 billion kWh in 2020, with an average annual growth rate of 8.95%. MF has consistently maintained its position as the largest electricity consumption sector. Its proportion of electricity consumption increased from 56.93% in 2002 to 60.18% in 2020, which reflects the continuously strengthening dominant position of MF in China’s industrialization process. During the same period, the PSEHGW functioned as the second-largest electricity consumption sector. Although its electricity consumption scale increased from 285.921 billion kWh to 993.408 billion kWh, its proportion decreased from 19.46% to 15.00%. This decline reflects the improvement in electricity consumption efficiency of the energy supply sector.

Significant electricity transfer effects exist between industrial sectors from the comparison of direct electricity consumption and VIC. The VIC of MF in 2020 (2690.394 billion kWh) remained significantly lower than its direct electricity consumption (3985.338 billion kWh). This indicates substantial electricity transfer to other sectors through industrial chains. In contrast, the GT industry shows strong electricity integration effects. Its VIC in 2020 (1850.457 billion kWh) substantially exceeded its direct electricity consumption (101.110 billion kWh). This characteristic fully reflects the electricity aggregation effect of the GT industry in industrial chains.

MF functions as a significant net electricity output sector through the analysis of net electricity flow directions between industrial sectors. Its net output scale significantly expanded from 217.063 billion kWh in 2002 to 1294.944 billion kWh in 2020. This continuously expanding net output characteristic indicates the increasingly strengthening influence of MF on the downstream electricity consumption industry as an important intermediate goods supplier. Its hub role in industrial chains becomes increasingly prominent. The electricity supply industry and MI also show net output characteristics. Their net output scales increased from 234.583 billion kWh and 123.090 billion kWh in 2002 to 739.023 billion kWh and 519.233 billion kWh in 2020, respectively. This reflects the further strengthening supporting role of these basic industries in industrial chains.

In contrast, the GT industry shows significant net electricity input characteristics. Its net input scale substantially increased from 308.334 billion kWh in 2002 to 1749.347 billion kWh in 2020. This characteristic reflects the continuously deepening dependence of the GT industry on upstream industry electricity input as a typical end-consumption sector. It also reflects the electricity consumption characteristics of the GT industry in China’s rapid urbanization process.

The industrial electricity consumption linkages in China from 2002 to 2020 demonstrate characteristics of continuous scale expansion, deep structural differentiation, and deepening linkages. This evolution trend represents an inevitable result of China’s economic structural transformation and reflects deep interactive relationships between industrial development and energy consumption. The industrial electricity structure transformation comprises the continuously strengthening electricity influence of MF as an intermediate goods supplier, the continuously deepening electricity dependence of GT as an end-consumption sector, and the gradual emergence of service sector electricity characteristics. These findings help deepen the understanding of relationships between China’s industrial development and energy consumption. They also provide important empirical evidence for formulating precise energy consumption and emission reduction policies.

4.2. Dynamic Evolution of Industrial Electricity Consumption Linkages

The industrial electricity consumption linkages in China exhibit significant dynamic evolutionary characteristics from 2002 to 2020. As shown in Figure 2, the backward linkage (NBL) of MF increased from 181.580 billion kWh in 2002 to 682.267 billion kWh in 2020. The forward linkage (NFL) expanded from 398.643 billion kWh to 1977.211 billion kWh. These changes reflect the continuous strengthening of the electricity linkage effects of MF in industrial chains. This significant enhancement in linkage intensity closely relates to the elevation of China’s MF position in global value chains. It also reflects structural changes in energy consumption during the deepening process of industrial chain division.

The evolution of industrial electricity linkages demonstrates distinct phase characteristics throughout the research period. The period from 2002 to 2007 represents a rapid expansion phase. The net output scale of MF expanded from 217.063 billion kWh to 274.803 billion kWh, achieving an average annual growth rate of 4.840%. This phase coincides with China’s industrial expansion period after joining the WTO. The development of numerous export-oriented MF companies strengthened their electricity influence on downstream industries. The period from 2008 to 2012 functions as a structural adjustment phase. The growth of MF net output scale relatively slowed down from 274.803 billion kWh to 418.915 billion kWh, with the average annual growth rate decreasing to 3.450%. This change reflects the policy effects of China’s industrial structure optimization and adjustment after the financial crisis. The period from 2013 to 2020 exhibits a deep transformation phase. The net output scale of MF rapidly climbed to 1294.944 billion kWh, which reflects the further strengthening of MF’s role as an intermediate goods supplier during industrial upgrading.

The evolution of electricity linkages in the GT industry shows a significantly strengthening trend of net input characteristics. Its net input scale increased from 308.334 billion kWh in 2002 to 1749.347 billion kWh in 2020, achieving a high average annual growth rate of 10.520%. This continuously strengthening net input characteristic closely relates to the acceleration of China’s urbanization process. It reflects the continuously deepening dependence of the GT industry on upstream electricity input. The electricity net input growth became more significant, particularly after 2012. The average annual growth rate of net input scale reached 12.360% during this phase, which highlights the profound impact of new urbanization strategy implementation on GT industry electricity demand.

The evolution of electricity linkages in the service sector demonstrates apparent structural upgrade characteristics. The wholesale and retail commerce, hospitality, and food service sectors experienced a transformation from net output to net input in electricity flow direction, reaching a net output scale of 22.971 billion kWh in 2020. This transformation reflects the strengthening influence of the service industry on downstream industries during industrial upgrading. It demonstrates the elevation of the modern service industry position in industrial chains. During the same period, transportation, storage, and postal services also show similar transformation trends in electricity linkages. Their net output scale increased from 5.177 billion kWh in 2002 to 17.869 billion kWh in 2020, which highlights the strengthening electricity influence of modern service industries in other sectors.

The internal effect (IE) of MF significantly strengthened from 380.020 billion kWh in 2002 to 1766.409 billion kWh in 2020. This strengthening of IE reflects both the improvement in MF’s electricity efficiency and the impact of industrial technological progress on energy consumption. In contrast, the IE growth of PSEHGW remained relatively moderate, increasing from 41.917 billion kWh to 199.636 billion kWh. This indicates positive achievements in improving electricity efficiency by energy supply sectors.

The evolution of mixed effects (MEs) reflects a deepening trend of electricity interaction between industries. The mixed effect of MF increased from 57.866 billion kWh in 2002 to 241.717 billion kWh in 2020, showing significantly higher growth than other sectors. This indicates increasingly close electricity interactions between MF and other sectors. This strengthening of interactive relationships reflects both the deepening of industrial chain division and industrial synergy effects driven by energy conservation and emission reduction policies.

AFAHF demonstrates unique cyclical characteristics in its electricity linkage evolution. Its flow direction experienced a structural change from net input dominance (2002–2012) to net output transformation (2013–2020). This transformation profoundly reflects the fundamental change in industrial positioning during agricultural modernization. The IE of these sectors reached 46.658 billion kWh in 2020, with ME reaching 1.491 billion kWh. This indicates that the transformation of agricultural production methods has gradually changed these sectors from traditional electricity-dependent sectors to sectors with electricity influence on other sectors, which reflects the profound impact of agricultural modernization on industrial electricity linkage structures.

The evolution of China’s industrial electricity consumption linkages from 2002 to 2020 reflects both the general laws of macroeconomic transformation and special trajectories of industrial structure optimization. The basic characteristics of China’s industrial electricity linkage evolution comprise the continuous strengthening of MF’s electricity influence on downstream industries, a significant enhancement of GT industry electricity dependence, the gradual emergence of service industry electricity spillover effects, and the cyclical evolution of agricultural electricity structure. These evolutionary characteristics not only deepen our understanding of relationships between industrial development and energy consumption but also provide an important basis for formulating differentiated energy conservation policies.

4.3. Key Sector Identification and Characteristic Analysis

The analysis based on relative effect indicators reveals significant structural characteristics in China’s industrial electricity consumption linkages from 2002 to 2020. As shown in Figure 3, MF demonstrates the strongest electricity self-sufficiency capacity from an IE perspective. Its relative effect value increased from 5.570 in 2002 to 5.520 in 2020. This value remains at a high level despite slight fluctuations. Such consistently high-intensity IE reflects the core position of MF in the industrial system and indicates significant room for improvement in MF electricity efficiency. The IE relative value of PSEHGW increased from 0.61 to 0.62. This moderate change indicates steady progress in electricity efficiency within energy supply sectors.

Mixed-effect analysis reveals the complexity of electricity interactions between industries. The relative value of MF’s mixed effect consistently remains above 7.000, reaching a peak of 7.740 in 2010. Although it slightly declined afterward, it still maintains a high level of 6.920. This characteristic indicates close electricity linkages between MF and other sectors. Changes in its production activities generate significant impacts on the electricity structure of the entire economic system through industrial chains. GT shows low ME, reaching only 0.008 in 2020. This indicates that its electricity consumption primarily depends on direct linkages with specific sectors, while indirect industrial effects remain relatively weak.

The comparative analysis of forward and NBL further highlights electricity interdependence relationships between sectors. The relative effect of MF’s NFL increased from 3.71 in 2002 to 4.180 in 2020, while NBL decreased from 1.690 to 1.440. This expanding “scissors gap” reflects MF’s gradual downstream extension trend in industrial chains. In contrast, the GT industry exhibits significant backward linking characteristics. Its backward linkage relative effect increased from 2.880 to 3.700, while NFL consistently maintained low levels (only 0.003 in 2020). This demonstrates the typical characteristics of end-consumption sectors.

The linkage effects of service sectors show structural optimization trends. The forward linkage relative effect of WRT, accommodation, and catering industries increased from 0.250 to 0.430. This indicates the gradually strengthening driving role of service industries in industrial chains. Meanwhile, their backward linkage relative effect increased from 0.63 to 0.380, reflecting increasingly close interactive relationships between service industries and upstream industries. This balanced development characteristic of bidirectional linkages aligns with the upgrading and transformation trends of modern service industries.

AFAHF demonstrates unique evolutionary characteristics in linkage effects. Their forward linkage relative effect decreased from 0.270 to 0.200, while NBL decreased from 0.560 to 0.210. This synchronous weakening of bidirectional linkages reflects the modernization transformation of agricultural production methods. Particularly after 2015, their IE relative value began to rise from 0.090 to 0.150 in 2020, indicating positive progress in agricultural sector electricity efficiency.

The comprehensive assessment based on relative effects enables the classification of China’s industrial sectors into three categories. The first category comprises high-linkage sectors represented by MF, characterized by strong IE and ME while maintaining high bidirectional linkages. The second category includes unidirectional linkage sectors represented by GT, exhibiting significant NBL but weak NFL. The third category consists of balanced development sectors represented by service industries, demonstrating coordination between forward and NBL. This classification reflects both the positioning of different sectors in industrial chains and reveals directions for industrial electricity structure optimization.

The MI shows negative IE and ME while maintaining strong NFL (0.880 in 2020). This special effect structure reflects its unique position in industrial chains. This sector influences the electricity structure of the entire industrial system through downstream transmission, which aligns with the fundamental position of resource-based industries.

The analysis based on relative effects deepens our understanding of industrial electricity linkages. The differentiated linkage characteristics demonstrated by different sectors reflect both the deep logic of China’s industrial structure transformation and provide an important basis for formulating precise energy conservation and emission reduction policies. The MF’s dominant position, GT sector’s unidirectional linkage characteristics, and service sector’s balanced development trends collectively constitute the basic pattern of China’s industrial electricity linkages. These findings carry important policy implications for optimizing the industrial structure and improving electricity efficiency.

4.4. Path Analysis of Industrial Electricity Consumption Structure

The analysis based on the average path length indicators revealed the transmission network characteristics of electricity influence among China’s industrial sectors, which reflected the transmission efficiency and complexity of electricity influence propagation in the industrial network. As shown in Figure 4, MF demonstrated good transmission accessibility in the electricity influence transmission network. The path length between MF and AFAHF was 2.649, and the path length with the GT industry was 2.656. The moderate transmission paths indicated that changes in MF’s electricity consumption could affect the electricity consumption of these downstream sectors through relatively direct routes. The path length between MF and mining was 2.785, which reflected that electricity influence transmission between the two productive sectors was relatively direct. However, the path length between MF and the power supply industry reached 3.360. The longer transmission path reflected that the impact of electricity consumption changes in productive sectors on energy supply decisions needed to go through complex market and policy intermediary links.

The GT demonstrated typical terminal consumption sector characteristics in electricity influence transmission. The internal path length of the GT industry was only 1.243, which indicated that the transmission of electricity influence within the GT industry was extremely direct, and electricity consumption changes among various GT sub-industries could rapidly influence each other. However, the path length between the GT and MF reached 3.788. This long-distance transmission path reflects how electricity consumption in the GT industry influences building materials demand, which affects MF electricity consumption, requiring passage through multiple market links.

In the electricity influence transmission network, the path length is different. The path length between WRT and most other sectors remains between 2 and 3. This indicates that the electricity consumption changes in WRT industries can affect other sectors through medium-length transmission paths. However, the internal path length of transportation, warehousing, postal services, accommodation, and catering services is 1.783. The electricity influence transmission within this industry is relatively direct, but the transmission paths to other sectors are generally longer. This suggests that the electricity consumption changes in industries need to go through more intermediate links before they can affect other sectors. AFAHF exhibits significant differentiation in the transmission of electricity impact. The internal path length within these sectors is 1.673, while the path length between them and the PSEHGW reaches 4.471, with a span of 2.8 units. The transmission of electricity among the sub-sectors of agriculture is relatively direct. However, the transmission path between them and the power supply industry is longer, which reflects the complexity of the impact of changes in agricultural electricity demand on power supply decisions. Changes in agricultural electricity consumption also affect the load of rural power grids and the consumption decisions of regional power supply industries.

The transmission path analysis of electricity influence revealed the network characteristics of electricity influence propagation in China’s industrial system. The analysis indicated that the transmission paths of internal electricity consumption changes within sectors were generally shorter, which showed that electricity influence transmission among internal links was relatively direct. Cross-sectoral transmission paths presented significant differences, which reflected different degrees of complexity in electricity influence propagation across different industrial chains. Linkages with shorter path lengths were conducive to the rapid influence of electricity consumption changes, while linkages with longer paths needed to go through more transmission links to produce influence. These transmission characteristics provided a scientific foundation based on path analysis for understanding the propagation mechanisms of electricity influence in industrial networks and formulating differentiated energy-saving measures.

5. Conclusions

This comprehensive research identified four key findings regarding China’s industrial electricity consumption linkages. First, the electricity consumption structure experienced scale expansion and increasingly intensified structural differentiation: MF consolidated its dominant position through enhanced NFL in the industrial chain, with the proportion of electricity consumption growing from 56.93% in 2002 to 60.18% in 2020, and NFL increasing from 39.8643 billion kWh to 197.7211 billion kWh, while the GT industry as a terminal consumption sector showed deepening upstream dependence, with its vertically integrated consumption (185.0457 billion kWh) substantially exceeding direct consumption (10.111 billion kWh); meanwhile, energy supply sectors improved efficiency, and the service industry gained increasingly growing influence. Second, the evolution patterns of electricity consumption revealed three distinct phases—rapid expansion following China’s WTO accession (2002–2007) with MF net output growing at an annual average rate of 4.840%, structural adjustment during 2008–2012 with the growth rate slowing to 3.450%, and deep transformation marked by the rise in the service industry (2013–2020) with MF net output reaching 129.4944 billion kWh—all of which were influenced by industrial policy reforms and broader economic transformation. Third, the sectoral analysis highlighted differentiated linkage intensities: MF’s enhanced NFL (197.7211 billion kWh) far exceeded NBL (68.2267 billion kWh), which reinforced its core role in the industrial chain; the GT industry’s dominance in NBL with net input reaching 174.9347 billion kWh reflected its terminal consumption nature; the service industry’s rising forward influence with WRT achieving net output of 2.2971 billion kWh indicated ongoing industrial upgrading; the MI maintained fundamental but resource-dependent characteristics. Fourth, the transmission path analysis revealed the differentiated characteristics of electricity influence transmission in the industrial network. IE within sectors was generally shorter, with GT having the shortest internal path. Cross-sectoral transmission presented significant differences, with AFAHF having the largest path span. MF demonstrated good transmission accessibility and maintained moderate path distances with major sectors. This differentiated transmission structure reflected the degree of propagation complexity of electricity influence in different industrial chains, which provided a path analysis basis for formulating differentiated energy-saving measures.

Based on these research conclusions, the following policy recommendations are proposed:

First, the customized industrial electricity management strategy should be established. For key MF, strict energy efficiency standards should be established, with overall electricity efficiency improvement through industrial chain transmission. The GT industry should strengthen collaborative innovation with upstream industries to enhance energy conservation across the entire supply chain. Modern service industries should be vigorously developed under improved energy conservation standards and enhanced electricity efficiency frameworks. Meanwhile, industrial electricity consumption monitoring and evaluation systems should be established to guide industrial structure development toward low-carbon and high-efficiency directions.

Second, a phased energy conservation and emission reduction roadmap should be conducted. Electricity efficiency improvement targets should be scientifically set, and regular assessment indicators should be provided. By increasing investment in energy-saving technology research and development and strengthening policies supporting industrial transformation and upgrading, an innovation incentive mechanism should be established to promote energy conservation in all links of the industrial chain. In addition, a long-term supervision mechanism and regular assessment system should be established to ensure the steady progress of energy conservation goals through a combination of policy guidance and market incentives.

Third, a precise industrial electricity supervision system should be constructed. For industries with high electricity demand, their driving effect in the industrial chain should be fully utilized. For industries with high electricity dependence, coordination and cooperation between upstream and downstream should be strengthened. A linkage mechanism should be established to achieve simultaneous improvement in electricity efficiency and resource allocation optimization. At the same time, a big industrial electricity data analysis platform should be built to achieve efficient resource allocation and fully explore potential energy-saving space.

Fourth, targeted electricity policies should be formulated based on the differentiated transmission path characteristics. For the characteristic that IE within sectors were generally shorter, internal electricity consumption coordination within sectors should be strengthened to improve the implementation efficiency of internal energy-saving measures. For MF’s good transmission accessibility characteristics, its connecting role in the electricity influence transmission network should be fully utilized to drive energy-saving effects in other sectors through MF. For the transmission complexity reflected by the largest path span of AFAHF, hierarchical management mechanisms should be established, and coordination of intermediate links should be strengthened for linkages with longer transmission paths to ensure that energy-saving measures could effectively cover all sectors.