Sustainability-Driven Energy Efficiency Assessment: Divergent Policy Impacts of Single Factor Limits Versus Total Factor Coordination

Abstract

While China’s current energy policies predominantly adopt single-factor energy efficiency (SFEE) as the benchmark, academic research increasingly advocates total-factor energy efficiency (TFEE) assessments. This study examines the differences between these two energy efficiency evaluation paradigms in the context of sustainable development goals, particularly exploring the extent of such divergences. Guided by the “energy input minimization” principle, we construct a time-series dynamic analytical framework to systematically compare the impact of SFEE and TFEE on regional energy efficiency rankings from a sustainable development perspective. Specifically, this paper innovatively incorporates “new driving forces” into the production function, establishing a green development-oriented evaluation system that reveals the measurement bias of traditional production frameworks on energy efficiency and its influence on regional rankings. The results demonstrate: (1) China’s regional energy efficiency rankings remain largely consistent under both evaluation systems, with only minor adjustments for individual provinces, confirming the feasibility of adopting SFEE in policy formulation as an effective method for evaluating and comparing regional energy efficiency; (2) For most provinces under the “new normal” economic development context, continued use of traditional production frameworks would lead to underestimation of TFEE. After introducing factors such as human capital, intangible capital, technological innovation, and business environments, China’s energy efficiency polarization gap widens. The evaluation of efficiency indicators provides theoretical foundations and micro-level evidence for energy policy formulation under the “dual-carbon” goals.

1. Introduction

In September 2020, Chinese President Xi Jinping pledged at the general debate of the 75th session of the United Nations General Assembly to strive for having carbon dioxide emissions peak before 2030 and achieving carbon neutrality before 2060. Enhancing energy efficiency, recognized as the most cost-effective decarbonization pathway, serves as a strategic cornerstone for achieving China’s dual-carbon goals. The International Energy Agency’s Global Energy Review 2019 indicates that global energy efficiency improvements reduced CO₂ emissions by 200 million tons, underscoring the critical role of precise measurement in climate governance. Scientifically constructed energy efficiency metrics, as fundamental instruments for evaluating energy input-output relationships, directly determine the efficacy of policy design—only through accurate identification of regional efficiency gaps can targeted resource allocation be optimized to systematically advance carbon neutrality objectives. Currently, energy efficiency is mainly divided into single-factor energy efficiency (SFEE) and total-factor energy efficiency (TFEE) [1] (pp. 235–244).

SFEE is defined as the ratio of energy input to output, which reflects the energy utilization efficiency of economic entities. Among them, energy consumption per unit of GDP is regarded as an excellent substitute indicator for SFEE in practice [2] (pp. 113–138). International organizations such as the World Energy Council, the European Union, and the UK energy industry have incorporated “GDP per unit of energy use” into the energy consumption evaluation system. The Chinese government has also gradually strengthened the assessment of the indicator of “energy consumption per unit of GDP”. In 2006, the Outline of the 11th Five-Year Plan for National Economic and Social Development first proposed the goal of reducing energy consumption per unit of GDP by 20% in five years, marking that China had begun to use the reduction of energy consumption per unit of GDP as a benchmark for improving energy efficiency. Since 2014, reports on indicators such as the reduction rate of 10,000 yuan regional GDP energy consumption have been released in provinces (autonomous regions, and municipalities). In 2019, the State Council of China issued the Program for Improving “Dual Control” System of Energy Consumption and Intensity, further improving the dual control system of total energy consumption and energy intensity.

TFEE is defined as the ratio of the minimum amount of energy use to the actual amount of energy input [3] (pp. 3206–3217), which reflects the configuration status of energy factors in the production process of a single economic entity and the relative efficiency among different economic entities in the same production process [4] (pp. 110–121). On the one hand, analyzing the changes in TFEE over time sequence can explain the changes in total factor productivity in economic growth [5] (pp. 240–250). On the other hand, by comparing the relative values of TFEE in spatial distribution, the optimal energy input, energy saving potential, and efficiency improvement path of economic entities can be calculated. Therefore, research on TFEE is highly favored by economists [6,7]. The scientific validity and policy effectiveness of this measurement system fundamentally hinge on the factor selection and structural specification within production frameworks—namely, the construction of input-output models that incorporate foundational elements (capital, labor, energy) while strategically integrating emerging production factors. Prevailing research overlooks the imperative for dynamic framework adjustments to maintain profound correspondence with evolving economic development phases. Taking China as an example, Premier Li Keqiang pointed out in 2015 that China’s economic development had entered a period of new normal and was in the process of transforming old and new driving forces. Such structural transformations render conventional tri-factor models (capital-labor-energy) increasingly inadequate in capturing the authentic evolutionary trajectories of TFEE within the digital economy era, constituting a fundamental theoretical predicament in contemporary efficiency evaluation’s policy adaptability.

New driving forces refer to novel production factors and their innovative combinations that propel economies onto renewed expansion trajectories. Future economic dynamism will primarily emanate from the pervasive integration and transformative restructuring of socioeconomic systems through new-generation information technologies. New driving forces featured by “talents, technology, knowledge, and data resources” are the expansion of conventional driving forces such as capital, labor, and energy [8] (pp. 53–57). It is a factor input that cannot be ignored in the context of high-quality economic development [9] (pp. 1–11). Undoubtedly, new driving forces are an important initiative to spawn a technological revolution and thus improve the efficiency of factor use. Therefore, the introduction of “new driving forces” significantly affects the TFEE measurement.

Based on the above discussion, it is clear that SFEE and TFEE have the same objectives, but different analysis perspectives. SFEE aims to analyze the extent of energy utilization by economic entities after the event, while TFEE attempts to reflect the degree of energy allocation by economic entities in the production process. The Chinese government uses energy intensity (single-factor energy efficiency) to assess the energy efficiency of various regions, while scholars tend to use TFEE to evaluate the energy efficiency of various regions. On the other hand, China is in a period of vigorous development of “new driving forces”, but few scholars have incorporated the factors of “new driving forces” into the production framework. The present study focuses on the following questions: (1) Will SFEE and TFEE lead to differences in energy efficiency rankings among different regions? How can the differences between the two types of energy efficiency be quantified? (2) Is the TFEE based on the new normal period of the Chinese economy different from that under the conventional driving forces? Is the gap between this new total factor energy efficiency and SFEE more obvious? To address these problems, this paper first evaluates the indicators of SFEE and TFEE of each region in China from 2006 to 2019 in a rigorous way and discusses the differences in energy efficiency rankings of economic entities under the two energy efficiency evaluation modes. Based on this, data from 2015 to 2019 are selected, and the “new driving forces” factors are incorporated into the production framework to re-calculate the TFEE of each region and make a comparison between the TFEE of “new driving forces”, the existing single-factor energy efficiency, and conventional TFEE. This study hopes to contribute to a better understanding of the status quo of China’s energy efficiency, which has significant practical implications. Investigating these issues advances energy efficiency evaluation theories by enhancing measurement accuracy and analytical comprehensiveness, while providing deeper insights into China’s current energy efficiency landscape. The findings offer government agencies precise data infrastructure and evidence-based decision-making references with significant practical implications. Furthermore, the empirical framework establishes transferable methodological paradigms for international policymakers to formulate context-specific efficiency assessment protocols, leveraging China’s operational experiences to accelerate global progress toward sustainable development agendas.

2. Literature Review

This section systematically examines the literature on single-factor and TFEE measurement systems, aiming to clarify three critical propositions through analyzing the methodological evolution and policy applicability of both models: (1) the complementary functions and inherent limitations of SFEE versus total-factor efficiency metrics in policy evaluation; (2) the sensitivity of total factor efficiency measurements to divergent DEA model specifications; (3) the adaptability crisis of traditional measurement frameworks amid the old-new growth driver transition.

2.1. Research on Single-Factor Energy Efficiency

The evaluation indicators for SFEE mainly include energy productivity [10] (pp. 46–57) and energy consumption per unit of GDP [11,12,13]. However, energy productivity cannot reflect the efficiency of civil energy use [14] (pp. 128–137). Therefore, on the practical level, energy consumption per unit of GDP is used more often to represent the SFEE of countries (regions). According to the World Bank’s calculation based on purchasing power parity, China’s energy consumption in 2019 was 2.29 times that of Germany and 1.89 times that of Japan, indicating that China’s SFEE is relatively low and there is still significant room for improvement. With the goal of energy conservation and efficiency enhancement, scholars have used various methods to decompose the factors affecting China’s energy intensity at different periods, such as the LMDI index [15] (pp. 52–62), the extended PDA model [16] (pp. 69–87), the multi-dimensional PDA model [17] (pp. 281), and econometric regression analysis [18] (pp. 258–265). Empirical studies have shown that technological progress is the main factor in reducing China’s energy intensity and improving energy efficiency.

2.2. Research on TFEE

The evaluation methods for TFEE mainly include Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA). However, SFA is unable to handle scenarios involving multiple inputs and outputs, as it requires the output variable to be singular—a condition often inconsistent with real-world applications. Additionally, the predefined production function in SFA may introduce specification bias due to subjective assumptions. Consequently, many scholars opt for Data Envelopment Analysis (DEA), which eliminates the need for explicit functional form specification, thereby circumventing biases arising from arbitrary parametric assumptions. DEA has emerged as the most widely adopted non-parametric frontier efficiency analysis method [19] (pp. 429–444).

The key to using the DEA method to measure TFEE lies in calculating the minimum amount of energy input. Currently, many scholars have calculated the minimum amount of energy use based on the perspective of technical efficiency, that is, reducing all input factors under the condition of constant output, and thus obtaining the TFEE [3,20]. However, the efficiency obtained in this way is more like comprehensive production efficiency [21] (pp. 31–43), and there is a possibility of overestimating energy efficiency [2] (pp. 113–138).

As research continues over the past 40 years, the application of DEA for measuring efficiency has become relatively mature. According to the method of adjusting energy use by economic entities, the models can be categorized as follows: (1) The CCR and BCC models, in which all input factors are reduced in the same proportion [19,22]; (2) The DDF model, which optimizes input and output together [23] (pp. 229–240); (3) The QFI (Quasi-Fixed Inputs) model, in which some input factors are proportionately reduced [24,25]; (4) The SBM model, in which all inputs and outputs are improved according to the maximum redundancy [26] (pp. 498–509); (5) The directional SBM model, in which all inputs and outputs are improved in a predetermined direction of redundancy [27] (pp. 274–287).

The identification and refinement of production frontiers constitute another pivotal research dimension in TFEE analysis. Regarding frontier identification, Halkos et al. (2024) developed a hybrid window DEA model to simultaneously evaluate ecological efficiency under short-term and medium-term policy scenarios [28] (p. 1). Tsionas et al. (2023) innovatively integrated artificial neural networks (ANNs) with both DEA and SFA frameworks, constructing endogenous production frontiers that account for potential input correlations [29] (p. 105491). In frontier refinement, scholarly efforts have systematically expanded production frameworks by incorporating non-conventional factors including pollutants [23] (pp. 229–240), fixed asset depreciation [30] (pp. 103–112), factor price fluctuations [31] (pp. 2306–2319), and building floor space [32] (p. 118749).

2.3. Discussion on SFEE and TFEE

As of now, a consensus has not been reached yet about the consistency of the evaluation indicators of the two types of energy efficiency in China. In fact, both single factor energy efficiency and total factor energy efficiency indicators are proxy variables for energy efficiency, which cannot fully reveal the true energy efficiency level of economic entities [14] (pp. 128–137). However, there is a certain consistency between the two estimation values in reflecting the energy efficiency ranking of each region. Market-based environmental regulation policies can alleviate the distortion of factor markets and stimulate the energy-saving and efficiency-improving motivation of market entities, thereby improving both SFEE and TFEE [33] (pp. 165–181). In addition, the common influencing factors of the two measurement indicators include factor substitution [34] (pp. 361–367), energy structure [35] (pp. 65–81), and technological progress. In order to combine SFEE and TFEE, Tu and Shen first divided regions into groups based on their energy intensity, and then used the common frontier technology to calculate TFEE [36] (pp. 7–13). On the other hand, many scholars argue that there is a paradox between the two measurement indicators. With the implementation of energy intensity-constrained policies, China’s SFEE has indeed improved, but this command policy distorts the combination of enterprise production factors, leading to a decline in TFEE. Secondly, many researchers believe that the SFEE indicator deviates from the actual production technology and can exaggerate the efficiency of energy use [37] (p. 102758). Based on the technical efficiency perspective, using the CCR model to calculate the data from 1995 to 2004 in China, the study found that there is a significant difference between SFEE and TFEE [38] (pp. 66–76).

In summary, academics have deepened their understanding of energy efficiency by investigating Chicanas’ two types of energy efficiency from various perspectives. However, there are still some limitations. Firstly, existing literature mainly calculates the TFEE of various regions in China based on the “technical efficiency” perspective and compares it with energy intensity, with scattered focuses on analyzing the similarities and differences between the TFEE and SFEE of each region from the perspective of “minimizing energy input”. “Minimizing energy input” assumes that the non-energy input of economic entities is effectively utilized, so as to calculate the minimum amount of energy use in the direction of energy input by economic entities, and obtain a more accurate TFEE [2] (pp. 113–138), which is more conducive to analyzing the similarities and differences between the two energy efficiency indicators. Secondly, the research on China’s TFEE using the DEA method is relatively abundant, but lacks objective criteria for model selection. Additionally, existing production frameworks have not included “new driving forces” factors. Since 2015, the role of “new driving forces” in supporting China’s economy has been continuously consolidated, gradually becoming a core impetus for promoting high-quality economic development. Thirdly, most scholars directly compare the energy intensity of each region with the TFEE. However, energy intensity cannot reflect the relative efficiency between economic entities, and cannot be directly compared with the scalar form of TFEE. Therefore, this paper attempts to re-measure and compare the SFEE and the TFEE of China’s various regions. Specifically, this paper first calculates the energy consumption per unit of GDP of each region based on the prices in 2000. Based on the perspective of “minimizing energy input”, it introduces “sequential production technology” to calculate the TFEE of each region from 2006 to 2019, and qualitatively analyzes the differences between the two energy efficiency indicators in ranking and quantitatively discusses their numerical deviations. Furthermore, by taking the data from 2015 to 2019 and including the “new driving forces” factors in the production framework, this paper explores the relationship between TFEE and SFEE during the period of the new normal of China’s economic development.

The marginal contribution of this paper is twofold: (1) This study pioneers empirical exploration of the relationship between dual energy efficiency metrics, providing theoretical foundations and empirical evidence for China’s energy governance policies; (2) It innovatively integrates “new growth drivers” into production frameworks for total-factor efficiency measurement, advancing research relevance to China’s developmental realities. The research limitations of this paper include: (1) The sample period (2006–2019) partially overlaps with post-”dual-carbon” strategic implementation; (2) Directional distance function specifications require further refinement to account for factor price distortions. These limitations outline critical pathways for future research extensions.

The remaining structure is organized as follows: Section 3 details computational methods and theoretical models for single- and multi-factor energy efficiency evaluation; Section 4 specifies indicator selection and data sources, including variable processing procedures; Section 5 presents theoretical and empirical analytical results; The conclusion synthesizes policy recommendations and research prospects.

3. Research Methods

This study develops a dual-dimensional energy efficiency assessment framework to systematically reveal measurement divergences between single-factor and total-factor efficiency metrics alongside their policy implications. Methodologically, the framework operates through two coordinated approaches: (1) Establishing comparable single-factor efficiency indicators via constant-price GDP standardization procedures; (2) Constructing dynamic total-factor efficiency metrics by integrating sequential production technology and comparatively analyzing three DEA variants through their technical specifications and optimization mechanisms.

3.1. SFEE Measurement Method

This paper uses the energy consumption per unit of GDP to represent the efficiency SFEE of different regions, $SFEE={\displaystyle \frac{E}{GDP}}$. In order to facilitate comparison, the GDP calculated at constant prices in 2000 is used in this paper.

As mentioned in the previous section, the dimensioned SFEE cannot be directly compared with the overall energy efficiency. Therefore, this paper draws on the approach of Shi dan [39] (pp. 49–58) and constructs a single-factor energy efficiency Indicator (SFEEI) to quantitatively compare the difference between these two types of energy efficiency.

$$SFEE{I}_i={\displaystyle \frac{\mathit{min}(\frac{E_i}{GD{P}_i})\left|i=1\dots N\right.}{\frac{E_i}{GD{P}_i}}}$$

(1)

where GDP_i is the gross national product of the i-th region, calculated at comparable prices, and E_i is the energy consumption of that region.

Obviously, such a SFEE is a scalar between 0 and 1. When SFEEI = 1, it indicates that the energy consumption per unit of GDP of the region has reached its minimum. When 0 < SFEEI < 1, it indicates that there is still room for improvement in the energy consumption per unit of GDP of the region.

3.2. Discussion on Measurement Model of TFEE

Considering that the energy consumption per unit of GDP is an indicator of energy economic efficiency, the production framework in this study only considers desirable outputs. At the same time, to better align with actual production conditions, all models in this study introduce “sequential production technology” to satisfy the property of “no regression in technology”.

As previously explained, the energy distance function can minimize energy input and is more suitable for measuring TFEE. Therefore, the TFEE in this study can be expressed as “the ratio of the minimum energy use required to achieve the same output while keeping other input factors constant to the actual energy use”. Considering that the energy consumption per unit of GDP is an indicator of energy economic efficiency, the production framework in this study only considers desirable outputs. At the same time, to better align with actual production conditions, all models in this study introduce “sequential production technology” to satisfy the property of “no regression in technology”. According to the research of Zhang [6] (pp. 165–172), there are currently three mainstream TFEE measurement models, namely: Sequential Energy-Quasi-fixed Inputs (SE-QFI), Sequential Energy-SBM model (SE-SBM), and Sequential Energy-Directional SBM (SE-DSBM). The SE-QFI model assumes that quasi-fixed inputs, such as capital and labor, cannot be freely adjusted in the short term. While the SE-SBM model not only quantifies total energy redundancy but also identifies optimal energy input targets through non-radial efficiency measurement, it remains constrained by potential overestimation biases arising from piecewise-linear production frontiers. This methodological limitation is systematically addressed by the SE-DSBM, which incorporates directional distance vectors to mitigate efficiency inflation under nonparametric technological heterogeneity. The mathematical formulations of these three models are specified as follows:

1.

Sequential Energy-Quasi-fixed Inputs, SE-QFI

$$\begin{gathered} \min \mathrm{\theta }\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \\ st. \sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{X}_n^t\le X^o;\sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{E}_n^t\le \theta E^o;\sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{Y}_n^t\ge Y^o; \\ {\lambda }_n^t\ge 0;n=\mathrm{1,2}\dots N;t=\mathrm{1,2}\dots T \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \end{gathered}$$

(2)

2.

Sequential Energy-SBM model, SE-SBM

$$\begin{gathered} \min 1-{\displaystyle \frac{S_E}{E}}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \\ st. \sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{X}_n^t\le X^o;\sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{E}_n^t+{\mathrm{S}}_E=E^o;\sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{Y}_n^t\ge Y^o; \\ {\lambda }_n^t\ge 0; {\mathrm{S}}_E\ge 0; \mathrm{n}=\mathrm{1,2}\dots N;t=\mathrm{1,2}\dots T \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \end{gathered}$$

(3)

3.

Sequential Energy-Directional SBM model, SE-DSBM

$$\begin{gathered} \max {\displaystyle \frac{ {\mathrm{S}}_E}{E}}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \\ st. \sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{X}_n^t\le X^o;\sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{E}_n^t+{\mathrm{S}}_E=E^o;\sum _{\mathrm{t}=1}^T\sum _{\mathrm{n}=1}^N{\lambda }_n^t{Y}_n^t\ge Y^o; \\ {\lambda }_n^t\ge 0;n=\mathrm{1,2}\dots N;t=\mathrm{1,2}\dots T \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \end{gathered}$$

(4)

where X represents non-energy inputs, E signifies energy consumption, Y denotes desirable outputs, N means the number of reference units, the superscript o indicates the o-th evaluated unit, and T refers to the sample period.

As shown in Figure 1, this paper first constructs isoquant curves QQ’, and point A is used as an example to explain it. Among them, E represents energy inputs, X represents other inputs, and A, B, C, and D represent decision-making units. To calculate the TFEE using the energy distance function, other input factors must be kept constant. Therefore, no matter how the energy input is adjusted, point B is always the benchmark for point A. All other conditions remain the same, $O{E}_1$ is the energy input at point A, $O{E}_2$ is the energy input at point B, and $E_1{E}_2$ is the energy consumption difference between the two. The expression of the TFEE of point A is as follows:

$${TFEE}_A={\displaystyle \frac{optimal energy utilization}{actual energy input}}={\displaystyle \frac{{OE}_2}{{OE}_1}}$$

(5)

It can be seen that the essence of the three models is the same, with only differences in the expression method. The Energy-Directional SBM model means that, with other factors unchanged, the energy input of point A should be reduced by ${\displaystyle \frac{{OE}_2}{{OE}_1}}$ times. The Energy-SBM model means that, with other factors unchanged, the energy input of point A can be reduced by $E_1{E}_2$ 10 thousand tons of standard coal. The Energy-Directional SBM model means that, with other input factors unchanged, the energy usage is reduced by ${\displaystyle \frac{{OE}_2}{{OE}_1}}$ times according to the energy direction. Therefore, consistent TFEE estimates can be obtained based on these three mainstream measurement models. That is, the TFEE of this paper can be expressed as follows:

$$\begin{gathered} TFEE=TFEE\_SE-QFI={\displaystyle \frac{\theta E}{E}}=\theta \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}=TFEE\_SE-SBM=TFEE\_SE-DSBM=1-{\displaystyle \frac{S_E}{E}} \end{gathered}$$

(6)

4. Indicator Selection and Data Sources

As mentioned earlier, since 2006, China has allocated the target of reducing energy consumption per unit of GDP to each province. Therefore, this paper analyzes energy efficiency by using labor, fixed capital stock, and energy consumption as input factors in 30 provinces (autonomous regions and municipalities directly under the central government) on the Chinese mainland from 2006 to 2019, and GDP in provinces as outputs. The specific explanation is as follows:

Labor input factor L: The number of employees in each province’s total social workforce is used to represent the labor input factor, and its value is equivalent to (the number of employees at the end of the current year + the number of employees at the end of the previous year)/2. The basic data is derived from the statistical yearbooks of each province from 2007 to 2020 (the difference in labor quality is not considered for the time being). The unit is ten thousand people.

Fixed capital stock K: The perpetual inventory method is used to calculate the fixed capital stock in each province of China from 2006 to 2019, i.e., Kt = Kt−1(1−δt) + It. Specifically, the provincial capital stock data for the year 2000, which was calculated by Zhang et al. [20] (pp. 31–43), is used as the base-year capital stock data. Then, the actual investment amount for each year is calculated by dividing the nominal fixed capital formation of that year by the corresponding indicator of gross fixed capital formation (the nominal fixed capital formation of 2006–2017 is from the National Bureau of Statistics, and that of 2018–2019 is replaced by the total investment in fixed assets of each province in the same year). The depreciation rate is fixed at 9.6%. The unit is 100 million yuan.

New driving forces: According to the research results of Zheng and Xiong [9] (pp. 1–11), the new driving forces indicator for each province in China from 2015 to 2018 is directly obtained. At the same time, based on the growth rate of the new driving forces indicator published by the National Bureau of Statistics of China from 2015 to 2019, an extension is made on their research results.

Energy consumption E: The energy input factor is represented by the annual energy consumption of each province, which is sourced from the China Energy Statistical Yearbooks from 2007 to 2020. The unit is 10,000 tons of standard coal.

Desirable output (GDP): The actual GDP of each province is represented as the desirable output indicator and is adjusted using the regional GDP indicator in constant prices in 2000. The nominal GDP of each province in the base period and the price index of each province can be directly obtained from the National Bureau of Statistics. The unit is 100 million yuan.

5. Empirical Results and Discussions

Building upon the theoretical frameworks and variable selections established in preceding chapters, this section conducts a focused comparative analysis between single-factor and TFEE models to elucidate the evaluation divergence between industrial practices and academic research paradigms. Furthermore, through innovative integration of “new driving forces” into production frontiers, we develop a refined TFEE framework for subsequent comparative evaluation with single-factor metrics. These analytical advancements collectively enhance interpretative capacity regarding the operational mechanisms underlying governmental “energy consumption per 10,000 yuan GDP” policy instruments.

5.1. Comparison and Analysis of the SFEE and TFEE Rankings

It is considered that SFEE and TFEE are not directly comparable, and the rankings based on the two types of energy efficiency can objectively reflect the energy efficiency of each region. Therefore, this paper starts by discussing the results of energy efficiency rankings.

Table 1. Energy Consumption per Unit of GDP and TFEE Rankings by Region.

	2006		2008		2010		2015		2019
Province	The Rankings of Energy Consumption per Unit of GDP	The Rankings of TFEE	The Rankings of Energy Consumption per Unit of GDP	The Rankings of TFEE	The Rankings of Energy Consumption per Unit of GDP	The Rankings of TFEE	The Rankings of Energy Consumption per Unit of GDP	The Rankings of TFEE	The Rankings of Energy Consumption per Unit of GDP	The Rankings of TFEE
Anhui	9	9	9	9	9	9	8	8	8	9
Beijing	2	2	1	2	1	3	1	1	1	2
Fujian	4	4	4	4	4	4	4	4	4	4
Gansu	23	23	23	23	23	23	23	23	23	23
Guangdong	1	3	2	3	2	1	3	3	3	3
Guangxi	10	11	12	12	15	15	19	19	20	20
Guizhou	27	27	27	27	27	27	24	24	24	24
Hainan	5	5	6	7	7	7	9	9	13	14
Hebei	25	25	25	25	25	25	26	26	25	25
Henan	14	14	14	15	13	14	14	14	11	11
Heilongjiang	17	17	19	18	18	18	16	16	17	17
Hubei	16	16	15	14	14	13	12	12	12	12
Hunan	13	13	13	13	12	11	10	11	9	8
Jilin	22	22	22	22	22	22	18	18	18	18
Jiangsu	7	7	7	6	6	6	5	5	5	5
Jiangxi	8	8	8	8	8	8	11	10	10	10
Liaoning	20	20	20	20	21	21	21	21	22	22
Inner Mongolia	26	26	26	26	26	26	25	25	26	26
Ningxia	30	30	30	30	30	30	30	30	30	30
Qinghai	28	28	28	28	28	28	29	29	29	29
Shandong	15	15	16	17	17	17	17	17	16	16
Shanxi	29	29	29	29	29	29	28	28	27	27
Shaanxi	19	19	18	19	19	19	22	22	21	21
Shanghai	3	1	3	1	3	2	2	2	2	1
Sichuan	18	18	17	16	16	16	13	13	14	13
Tianjin	11	10	10	10	11	12	15	15	15	15
Xinjiang	24	24	24	24	24	24	27	27	28	28
Yunnan	21	21	21	21	20	20	20	20	19	19
Zhejiang	6	6	5	5	5	5	6	6	7	6
Chongqing	12	12	11	11	10	10	7	7	6	7

shows the rankings of energy consumption per unit of GDP and TFEE in each region. It can be seen from the table that most regions in China maintain the same ranking under the two types of energy efficiency, and only a few provinces show changes in the ranking position. Among them, the provinces with differences in ranking positions in 2008 are the most: Heilongjiang and Shaanxi, Henan and Hubei, Shandong and Sichuan, Jiangsu and Hainan showed changes in ranking position. In other years, only two or three regions may show differences in ranking positions.

Furthermore, it can be observed that regardless of whether it is the ranking of SFEE or TFEE, the positions of most provinces remain unchanged during the sample period. Chongqing, Hubei, Hunan, Sichuan, and Henan have significantly risen in the rankings, while Guangxi, Hainan, Tianjin, and Xinjiang have shown a significant downward trend. Taking Guangxi and Chongqing as examples, a specific analysis is conducted. Looking at the rankings of single-factor energy efficiency, there was little difference between Guangxi and Chongqing in 2006, with only two places apart (Guangxi ranked 10th and Chongqing ranked 12th). However, the governments of the two regions issued different level-constrained targets for reducing their respective energy consumption per unit of GDP, and adopted different attitudes in terms of implementation. In order to achieve the energy consumption target of the 11th Five-Year Plan, the Chongqing government proposed to reduce energy consumption by 4% in 2006. However, data in 2007 showed that the target was not reached as expected. The Chongqing government immediately made a corresponding review and implemented the Implementation Plan of Statistic Indicators for Energy Consumption per Unit of GDP and Implementation Plan for Energy Consumption per Unit of GDP Monitoring System in 2008 to indicate Chongqing’s determination to save energy and reduce energy consumption. Moreover, Chongqing is located in western China, with abundant coal, natural gas, and hydropower resources, and it has persisted in implementing a new industrialization strategy. Therefore, the energy efficiency of Chongqing can continue to improve, ranking sixth in the country by 2019 (seventh in TFEE rankings). On the other hand, in 2006, only a 3% reduction in energy consumption per unit of GDP was required by the Guangxi government, and the region’s industrial foundation was weak, with relatively extensive economic growth. As of the end of the 13th Five-Year Plan period, the output value of the six high energy-consuming industries in Guangxi still accounted for more than 45% of the total output value of the industry, according to Barometer of the Completion of Dual Control Targets for Energy Consumption by region in the First Half of 2021. Therefore, it can be seen that the economic growth of Guangxi Province is mainly driven by heavy industry, and its scope for energy saving and efficiency improvement is relatively limited. As a result, its energy efficiency ranking dropped to 20th in China by 2019.

In summary, although the industry and academia use different efficiency evaluation models, the deviation between practice and theory does not cause contradictions in the rankings of regional energy efficiency. At the same time, the energy consumption per unit of GDP is clear and can be widely accepted by the public. Its calculation is simple and easy to operate, facilitating the industry to update policy dynamics and adjust work plans in a timely manner. On the contrary, TFEE measurement is based on production functions, which is a complex and abstract concept. Even after decades of research, scholars have not been able to provide a definite form of the production function. In addition, the measurement of TFEE involves many variables, which pose certain difficulties to statistical work. Therefore, the author believes that the industry’s use of energy consumption per unit of GDP to assess energy efficiency in various regions is a manifestation of industry wisdom, reflecting the principle of “simpler is better”. This is an advantage that TFEE cannot match.

5.2. SFEE Indicators and Total Factor Energy

Efficiency Gap Analysis

To quantify the specific differences between the industry and academia, this paper analyzes SFEE indicators and TFEE.

Table 2. Comparison of the SFEEIs and TFEE by region.

	2006		2010		2015		2019
Province	Single-Factor Energy Efficiency Indicators	Total Factor Energy Efficiency	Single-Factor Energy Efficiency Indicators	Total Factor Energy Efficiency	Single-Factor Energy Efficiency Indicators	Total Factor Energy Efficiency	Single-Factor Energy Efficiency Indicators	Total Factor Energy Efficiency
Beijing	0.9652	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
Tianjin	0.5969	0.6078	0.5301	0.5301	0.4714	0.4714	0.4731	0.4731
Hebei	0.3199	0.3199	0.2851	0.2851	0.2536	0.2536	0.2616	0.2616
Shanxi	0.2164	0.2164	0.2097	0.2097	0.1780	0.1780	0.1709	0.1709
Inner Mongolia	0.2872	0.2872	0.2809	0.2809	0.2713	0.2713	0.2073	0.2073
Liaoning	0.4835	0.4835	0.4243	0.4243	0.4008	0.4008	0.3517	0.3517
Jilin	0.4296	0.4296	0.3752	0.3752	0.4437	0.4437	0.4405	0.4405
Heilongjiang	0.5054	0.5054	0.4714	0.4714	0.4666	0.4666	0.4491	0.4491
Shanghai	0.9316	1.0000	0.8981	1.0000	0.9003	1.0000	0.9120	1.0000
Jiangsu	0.8225	0.8225	0.8033	0.8155	0.7330	0.7587	0.7482	0.7769
Zhejiang	0.8474	0.8474	0.8230	0.8283	0.7139	0.7270	0.6937	0.7115
Anhui	0.7113	0.7113	0.6933	0.7143	0.6138	0.6324	0.6239	0.6306
Fujian	0.9176	0.9176	0.8555	0.8700	0.7996	0.7996	0.7897	0.7897
Jiangxi	0.7311	0.7311	0.7150	0.7150	0.6027	0.6033	0.6115	0.6161
Shandong	0.5165	0.5165	0.4761	0.4761	0.4450	0.4450	0.4565	0.4565
Henan	0.5272	0.5272	0.5151	0.5151	0.5329	0.5329	0.5991	0.5991
Hubei	0.5139	0.5139	0.5122	0.5155	0.5671	0.5671	0.5719	0.5719
Hunan	0.5361	0.5361	0.5266	0.5378	0.6030	0.6030	0.6195	0.6318
Guangdong	1.0000	1.0000	0.9397	1.0000	0.8601	0.9474	0.8324	0.8842
Guangxi	0.5989	0.5989	0.5099	0.5099	0.4416	0.4416	0.4172	0.4172
Hainan	0.8856	0.8856	0.7996	0.7996	0.6070	0.6070	0.5536	0.5536
Chongqing	0.5647	0.5647	0.5576	0.5605	0.7024	0.7251	0.6986	0.7100
Sichuan	0.5049	0.5049	0.4947	0.5028	0.5435	0.5592	0.5414	0.5593
Guizhou	0.2670	0.2670	0.2608	0.2608	0.2788	0.2788	0.3010	0.3010
Yunnan	0.4549	0.4549	0.4371	0.4371	0.4150	0.4150	0.4173	0.4173
Shaanxi	0.4981	0.4981	0.4641	0.4641	0.3947	0.3947	0.3823	0.3823
Gansu	0.3611	0.3611	0.3550	0.3550	0.3170	0.3170	0.3189	0.3189
Qinghai	0.2333	0.2333	0.2200	0.2200	0.1524	0.1524	0.1632	0.1632
Ningxia	0.1667	0.1667	0.1606	0.1606	0.1172	0.1172	0.0914	0.0914
Xinjiang	0.3525	0.3525	0.3085	0.3085	0.1837	0.1837	0.1698	0.1698

shows the specific values of SFEE indicators and TFEE indicators for each region during the sample period.

As shown in Table 2, the SFEEIs for each region are less than or equal to the TFEE indicators. Based on the results, this paper constructs the total factor optimal production frontier using Beijing, Shanghai, and Guangzhou as three points, as shown in points A, B, and C in Figure 2. Among them, E represents energy input, Y represents other inputs, and the dots on the diagram represent the decision-making unit. Using points P and C as examples, this paper explains the results as follows:

P point

$$\begin{gathered} \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}={\displaystyle \frac{{OE}_1}{{OE}_3}} \\ \mathrm{s}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{l}\mathrm{e} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}={\displaystyle \frac{{OE}_2/E_{2}A}{{OE}_3/E_{3}P}}={\displaystyle \frac{{OE}_1/E_{1}Q}{{OE}_3/E_{3}P}}={\displaystyle \frac{{OE}_1}{{OE}_3}} \\ \therefore \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}=\mathrm{s}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{l}\mathrm{e} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y} \end{gathered}$$

C point:

$$\begin{gathered} \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}={\displaystyle \frac{{OE}_5}{{OE}_6}} \\ \mathrm{s}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{l}\mathrm{e} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}={\displaystyle \frac{{OE}_2/E_{2}A}{{OE}_6/E_{6}C}}={\displaystyle \frac{{OE}_4/E_{4}S}{{OE}_6/E_{6}C}}={\displaystyle \frac{{OE}_4}{{OE}_6}}<{\displaystyle \frac{{OE}_5}{{OE}_6}} \\ \therefore \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}<\mathrm{s}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{l}\mathrm{e} \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y} \end{gathered}$$

Based on this, it can be gleaned that the SFEE defined in this paper aims to compare with the point of minimum energy consumption in the sample, while TFEE has an additional binding condition that it needs to seek the smallest energy use under the conditions of other input factors and output being constant. Therefore, the production frontier constructed according to TFEE will be smaller than the frontier constructed according to single factor energy efficiency. In other words, the target under SFEE is more stringent, resulting in a smaller energy efficiency indicator.

From the table, it can be observed that when an economic entity is in Phase I, such as in China’s provinces of Heilongjiang, Inner Mongolia, Shaanxi, Guangxi, Yunnan, Tianjin, Shanxi, Jilin, Xinjiang, Guizhou, Gansu, Hainan, Ningxia, and Qinghai, the SFEE is consistent with the TFEE, indicating a correspondence between the practical and academic domains. When an economic entity is at point A, such as Beijing in China, the SFEE equals the TFEE, both at 1. When an economic entity is in Phase II, such as in China’s provinces of Guangdong, Jiangsu, Shandong, Zhejiang, Henan, Sichuan, Fujian, Shanghai, Hunan, Hubei, Anhui, Hebei, Liaoning, Chongqing, and Jiangxi, the SFEE is lower than the TFEE.

Therefore, the author believes that when evaluating high energy-consuming regions, SFEE can be used for simple research. However, for economically developed regions, using energy consumption per unit of GDP alone to evaluate energy efficiency may be misleading. For example, Beijing, Shanghai, and Guangdong Provinces in China should all be energy efficiency benchmarks from the perspective of TFEE. However, when considering single-factor energy efficiency, Guangdong and Shanghai still have certain gaps compared to Beijing. Through investigation, it was found that this is mainly due to the different industrial structures. At the end of 2010, SHOUGANG GROUP completely moved out of Beijing, and the Prohibited and Restricted Catalogue of Additional Industries in Beijing (2014 Edition) explicitly stated that Beijing will limit the development of various manufacturing industries and only retain the new energy vehicle industry. Therefore, Beijing’s energy consumption control has achieved excellent results. Beijing, as China’s political, cultural, and financial center, ranks first in the country in terms of scientific and technological innovation, so Beijing is an energy efficiency benchmark in both SFEE and TFEE. On the other hand, the steel industry in Shanghai and the petrochemical industry in Guangdong are still developing, which leads to higher energy consumption levels in Shanghai and Guangdong, resulting in higher levels of energy consumption per unit of GDP and lower single-factor energy efficiency. However, as a developed coastal region in China, Guangdong has maintained its position as the top GDP contributor for many years, with a high degree of marketization, large industrial scale, and high level of openness. Shanghai, as China’s financial center, integrates manufacturing, finance, trade, technology, and shipping, with highly developed land, sea, and air transportation, and has a deep commercial and cultural heritage. Therefore, although energy consumption in these two regions is slightly higher than in Beijing, it should be acknowledged that Guangdong and Shanghai can be considered as energy efficiency benchmarks. Guangdong’s green petrochemical development in recent years has set an example for all provinces in China and has become a world-class green petrochemical industry cluster base. China Baowu Steel Group Corporation Limited in Shanghai is now equipped with world-class fourth-generation high-temperature gas-cooled reactor nuclear power technology, making it a leader in China’s modern steel industry.

5.3. Measurement of TFEE Under the Introduction of “New Driving Forces”

As China’s economy transitions into the new normal phase, the growth-driving capacity of traditional factors continues to diminish. Dynamic coupling systems formed by non-traditional elements—human capital, intangible assets, and technological innovation—break through the bottleneck of diminishing marginal returns inherent to conventional factors. Exemplified by skilled labor accelerating green technology adoption, R&D investments catalyzing digital energy-efficiency tool development, and intellectual property accumulation enabling integrated multi-energy optimization, these interactions underpin our pioneering integration of “new driving forces” as supplementary inputs into production frontiers.

Table 3. TFEE by region under the Introduction of “New Driving Forces”.

Time	2015		2016		2017		2018		2019
Province	New and Old Driving Forces	Conventional Driving Forces	New and Old Driving Forces	Conventional Driving Forces	New and Old Driving Forces	Conventional Driving Forces	New and Old Driving Forces	Conventional Driving Forces	New and Old Driving Forces	Conventional Driving Forces
Anhui	0.6686	0.6324	0.683	0.6376	0.6993	0.6440	0.7008	0.6473	0.6955	0.6306
Beijing	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000
Fujian	0.8875	0.7996	0.915	0.8126	0.9163	0.8080	0.9026	0.8043	0.8964	0.7897
Gansu	0.317	0.3170	0.3328	0.3328	0.3215	0.3215	0.3145	0.3145	0.3189	0.3189
Guangdong	1	0.9474	1	0.9403	1	0.9188	1	0.8941	1	0.8842
Guangxi	0.45	0.4416	0.4512	0.4359	0.4537	0.4331	0.4482	0.4296	0.4375	0.4172
Guizhou	0.2788	0.2788	0.285	0.2850	0.294	0.2940	0.3025	0.3025	0.301	0.3010
Hainan	0.607	0.6070	0.5993	0.5993	0.5869	0.5869	0.5722	0.5722	0.5536	0.5536
Hebei	0.2783	0.2536	0.2828	0.2539	0.2861	0.2546	0.289	0.2599	0.2941	0.2616
Henan	0.6027	0.5329	0.6306	0.5490	0.6609	0.5719	0.6676	0.5786	0.6989	0.5991
Heilongjiang	0.4889	0.4666	0.4926	0.4647	0.4983	0.4648	0.4874	0.4596	0.4784	0.4491
Hubei	0.625	0.5671	0.6357	0.5677	0.6507	0.5769	0.6477	0.5798	0.6461	0.5719
Hunan	0.6661	0.6030	0.68	0.6056	0.6938	0.6158	0.6977	0.6298	0.704	0.6318
Jilin	0.4437	0.4437	0.4582	0.4582	0.4629	0.4629	0.4566	0.4566	0.4405	0.4405
Jiangsu	1	0.7587	1	0.7653	1	0.7740	1	0.7883	1	0.7769
Jiangxi	0.6273	0.6033	0.6388	0.6051	0.6574	0.6165	0.6562	0.6233	0.6549	0.6161
Liaoning	0.4407	0.4008	0.4427	0.3925	0.4295	0.3829	0.4113	0.3721	0.3923	0.3517
InnerMongolia	0.2831	0.2713	0.2844	0.2686	0.2805	0.2619	0.2391	0.2269	0.2196	0.2073
Ningxia	0.1172	0.1172	0.1165	0.1165	0.1038	0.1038	0.0969	0.0969	0.0914	0.0914
Qinghai	0.1524	0.1524	0.1572	0.1572	0.1581	0.1581	0.1564	0.1564	0.1632	0.1632
Shandong	0.5114	0.4450	0.5195	0.4454	0.5381	0.4590	0.5443	0.4634	0.5417	0.4565
Shanxi	0.178	0.1780	0.1768	0.1768	0.1755	0.1755	0.1747	0.1744	0.1713	0.1709
Shaanxi	0.4024	0.3947	0.4042	0.3903	0.4114	0.3907	0.4136	0.3950	0.4025	0.3823
Shanghai	1	1	1	1	1	1	1	1	1	1
Sichuan	0.6067	0.5592	0.6167	0.5632	0.6287	0.5688	0.6267	0.5685	0.6222	0.5593
Tianjin	0.4714	0.4714	0.4895	0.4895	0.5009	0.5009	0.489	0.4890	0.4731	0.4731
Xinjiang	0.1837	0.1837	0.1805	0.1805	0.1747	0.1747	0.1751	0.1751	0.1698	0.1698
Yunnan	0.4229	0.4150	0.4338	0.4170	0.4452	0.4209	0.4466	0.4244	0.4429	0.4173
Zhejiang	0.8075	0.7270	0.8097	0.7223	0.8122	0.7219	0.8079	0.7200	0.8057	0.7115
Chongqing	0.7349	0.7251	0.7648	0.7473	0.7819	0.7523	0.7576	0.7311	0.7449	0.7100

reflects the TFEE of various regions under “new driving forces”. After the introduction of new driving force factors, Guangdong and Jiangsu have entered the “frontier”, and the energy efficiency of most regions has been significantly improved. This likely stems from their abundant innovation resources, which facilitate modern industrial system construction and emerging cluster cultivation. By leveraging technological innovation to lead industrial upgrading and supply chain collaborations, these provinces accelerate new quality productive forces, thereby enhancing productivity and resource utilization efficiency while reducing energy consumption and emissions. Concurrently, most provinces are aggressively deploying strategic and future industries. Zhejiang’s “one chain, one policy” initiative exemplifies efforts to expand emerging industries, while Liaoning and Yunnan prioritize AI sector development. Supported by targeted policies and optimized business ecosystems, new driving forces are effectively embedded into production processes, demonstrating systematic alignment between regional industrial strategies and national energy transition objectives. However, even with the introduction of new driving forces, the TFEE of regions that are relatively backward in efficiency, such as Guizhou, Ningxia, Qinghai, Tianjin, and Xinjiang, has not improved. Therefore, this paper concludes that under the new normal of economic development, the gap in energy efficiency among regions in China is expanding based on the value of TFEE.

Meanwhile, we discussed the TFEE and energy consumption per unit of GDP rankings of the “new driving forces” in various regions, and the results are shown in

Table 4. Rankings of TFEE and Energy Consumption per Unit of GDP for “New Driving Forces”.

Province	2015		2016		2017		2018		2019
	Total Factor Ranking Under New Driving Forces	Energy Consumption per Unit of GDP Ranking	Total Factor Ranking Under New Driving Forces	Energy Consumption per Unit of GDP Ranking	Total Factor Ranking Under New Driving Forces	Energy Consumption per Unit of GDP Ranking	Total Factor Ranking Under New Driving Forces	Energy Consumption per Unit of GDP Ranking	Total Factor Ranking Under New Driving Forces	Energy Consumption per Unit of GDP Ranking
Anhui	8	8	8	8	8	8	8	8	10	8
Beijing	1	1	1	1	1	1	1	1	1	1
Fujian	5	4	5	4	5	4	5	4	5	4
Gansu	23	23	23	23	23	23	23	23	23	23
Guangdong	2	3	2	3	2	3	2	3	2	3
Guangxi	18	19	19	19	19	19	19	19	20	20
Guizhou	25	24	24	24	24	24	24	24	24	24
Hainan	12	9	14	11	14	11	14	13	14	13
Hebei	26	26	26	26	25	26	25	25	25	25
Henan	14	14	12	13	10	13	10	12	9	11
Heilongjiang	16	16	16	16	17	16	17	17	16	17
Hubei	11	12	11	12	12	12	12	11	12	12
Hunan	9	10	9	9	9	9	9	9	8	9
Jilin	19	18	18	17	18	17	18	18	19	18
Jiangsu	3	5	3	5	3	5	3	5	3	5
Jiangxi	10	11	10	10	11	10	11	10	11	10
Liaoning	20	21	20	21	21	22	22	22	22	22
Inner Mongolia	24	25	25	25	26	25	26	26	26	26
Ningxia	30	30	30	30	30	30	30	30	30	30
Qinghai	29	29	29	29	29	29	29	29	29	29
Shandong	15	17	15	18	15	18	15	16	15	16
Shanxi	28	28	28	28	27	27	28	28	27	27
Shaanxi	22	22	22	22	22	21	21	21	21	21
Shanghai	4	2	4	2	4	2	4	2	4	2
Sichuan	13	13	13	14	13	14	13	14	13	14
Tianjin	17	15	17	15	16	15	16	15	17	15
Xinjiang	27	27	27	27	28	28	27	27	28	28
Yunnan	21	20	21	20	20	20	20	20	18	19
Zhejiang	6	6	6	7	6	7	6	7	6	7
Chongqing	7	7	7	6	7	6	7	6	7	6

. As mentioned earlier, the energy efficiency rankings of most provinces under the two evaluation methods are consistent, while a few provinces have experienced changes in their rankings. The new energy indicator in this paper draws on Zheng’s (2021) data [9] (pp. 1–11). In his research, the development of new driving forces in Beijing, Shanghai, Guangdong, Jiangsu, and Zhejiang is active, which directly drives the improvement of TFEE. Due to the short time span, the energy efficiency rankings of various regions from 2015 to 2019 remain basically consistent, but it can still be seen that Guangxi Province’s energy efficiency ranking has been declining year by year.

6. Conclusions and Recommendations

6.1. Conclusions

This study aims to reveal the impact mechanisms of different energy efficiency evaluation paradigms on sustainable development policy formulation. Constructing an “energy input minimization”-oriented intertemporal dynamic analytical framework bridges the theoretical gap between academic research and governmental decision-making systems. The main conclusions are: (1) SFEE and TFEE systems exhibit significant policy compatibility, with regional energy efficiency rankings remaining largely consistent across China. Minor adjustments observed in individual provinces validate the effectiveness of current SFEE-based policy benchmarks, providing empirical support for establishing gradual energy efficiency improvement mechanisms under sustainable development frameworks; (2) Traditional production frameworks lead to systematic underestimation of TFEE. Under the “new normal” of economic development, continued reliance on conventional frameworks exacerbates this underestimation for most provinces. Introducing new production factors—including human capital, intangible capital, technological innovation, and business ecosystems—amplifies polarization in national energy efficiency, highlighting the critical role of green productivity configuration in sustainable energy efficiency assessments.

6.2. Policy Recommendations

The conclusions of the above study have important policy implications for energy efficiency evaluation, promotion of the “new driving forces” indicators, and improvement of energy efficiency in China. This paper provides the following suggestions:

Different energy efficiency methods are used according to the development stage. In the short term, given the intuitive operationalization and implementation simplicity inherent in single-factor efficiency metrics, along with their demonstrated minimal deviation from TFEE assessments in policy impact, China can maintain its current energy governance framework anchored in single-factor metrics. This approach sustains the dual-control mechanism (total energy consumption and intensity regulation) for provincial energy management, balancing methodological accessibility with practical governance efficacy. In the medium and long term, the government needs to establish a multi-factor energy efficiency evaluation system, improve the talent management system, and provide training for relevant professionals to achieve the conversion of single-factor and multi-factor energy efficiency, in order to provide support for meeting the higher requirements for energy conservation and emission reduction under the “new driving forces”.

Construct a mechanism for resource circulation between Eastern and Western regions. Establish a mechanism for the mutual flow of human capital. In the Western region, organizations can conduct on-the-job training and directional cultivation between enterprises on a regular basis, host-related skills competitions, establish a special fund for talents in the Western region, provide service guarantees such as regular health checkups for high-level talents, as well as financial subsidies for talents to work and settle in the Western region. At the same time, the Eastern region can provide talent assistance to the Western region by sending experts to conduct seminars, technical consultations, and other activities, building a cross-regional employment information platform, expanding the channels for releasing demands for technical personnel, and promoting the flow of talents between the Eastern and Western regions. A mechanism for mutual access to the business environment should be constructed, and the energy grid pricing through market competition should be formed. At the same time, coordinate and give full play to the service and guidance roles of various provincial and municipal governments to improve the legislative system of the cross-regional business environment, so as to promote local governments to formulate market-related local laws and guidelines and strengthen the joint supervision and management capabilities of the market. A mechanism of mutual benefit for intangible capital can be constructed. A multi-level platform system and an energy industry meta-universe can be established. Government departments should be empowered with a unified data exchange platform based on open data and technology from various provincial and municipal governments, realizing real-time interaction and the flexible response of information flow and value flow.

The technological advancement capacity related to energy efficiency in the central and western regions can be enhanced. On the one hand, the departments concerned can introduce a series of incentive policies for enterprise technology introduction, further increasing the tax incentives, loan support, financial subsidies, and other measures for high-tech enterprises. The local governments can also establish technology innovation funds or to apply for the establishment of a national commercialization of scientific and technological achievement transformation guidance funds. Besides, a collaborative system for IUR (Industry-University-Research) related to energy in the western region can be established to accelerate the improvement of the energy market trading mechanism to stimulate fair competition among enterprises, thus closing down the backward enterprises and enterprises with low energy efficiency. On the other hand, enterprises should strengthen investment and training for knowledge talents, technical talents, and practical talents, increase funding for low-carbon technology research and development and the transformation of scientific and technological achievements, actively engage with high-level science and technology professionals from universities, and promote the green transformation and further development of low-carbon energy technologies in the enterprises.