Spatial Influence of Digital Economy on Carbon Emission Efficiency of the Logistics Industry across 30 Provinces in China

Abstract

The logistics industry (LI) is a key pillar of the global economy, and its carbon emission efficiency (CEE) is crucial for achieving carbon neutrality. The rapid development of the digital economy (DE) has had a profound impact on the LI, but the spatial impact on its CEE is currently unclear and requires further research. Firstly, based on the collection of relevant data, we use the entropy weight method and linear weighted sum method to measure the level of development of the DE. Secondly, the SBM model is used to measure the CEE level of the LI. Using Moran’s I index model and OLS and GWR models, we analyze the impact and spatial distribution characteristics of the DE on the CEE of the LI and propose development strategies. The article uses statistical data from 30 provinces in China from 2013 to 2022 as an example to demonstrate the implementation process of the method. The results show that the DE has a positive impact on the CEE of the LI, and there are spatial differences. Based on this, this article proposes policy recommendations for the development of green and low-carbon logistics and digital logistics that are tailored to local conditions, providing theoretical and methodological support for low-carbon research in the LI, and providing reference for other countries and regions to explore the path of green and low-carbon transformation.

1. Introduction

Global attention is now focused on carbon emissions and the related climate challenges [1]. The China Carbon Accounting Database states that in 2022, China’s total carbon emissions made up 28.87% of the world’s total. China, a significant source of worldwide carbon emissions, has made a commitment to actively reduce carbon emissions, aiming to reach carbon neutrality by 2060 and a carbon peak by 2030 [2,3]. It is crucial for China to advocate for a high-quality economic development model, create a new environmentally friendly and low-carbon growth pattern, and achieve the “dual carbon” goals [3,4,5]. Economic progress is greatly aided by the logistics sector. Its development influences direct energy consumption-related carbon emissions along with indirect operational carbon emissions [6,7]. Improving carbon emission efficiency (CEE) and achieving low-carbon transformation in the logistics industry (LI) are crucial components of China’s “dual carbon” goals, which encourage superior economic development [8]. The digital economy (DE) is a crucial force that drives economic transformation and a new type of economy. China’s DE is at a stage of rapid development. The Global DE White Paper 2023 indicated that China’s DE has a compound annual growth rate of 14.2%, which is 1.6-fold the overall compound annual growth rate of the DE of the United States, China, Germany, Japan, and South Korea during the same period; this rapid growth is gradually influencing various fields [9]. The DE serves as a crucial element in fostering the development of an ecological civilization, acting as a novel catalyst for economic expansion [10]. In the process of achieving the “dual carbon” goal, emphasizing the development of novel technologies and promoting the digitization and advancement of heritage industries is conducive to improving energy utilization efficiency [11,12]. The development of the DE has significantly improved resource utilization, thereby reducing energy consumption, which is advantageous for reducing carbon emissions and improving CEE [13,14,15]. Researching how the DE may encourage carbon reduction in the LI is important from a practical standpoint [11,16]. Therefore, elucidating the influence of the DE on LI’s CEE and determining whether there are any regional differences is crucial. The measurement of scientific and objective carbon emission data in LI provides strong support for the government to formulate policies. How the DE affects the CEE of China’s LI is worthy of in-depth exploration by various departments in their upcoming strategies. Studying the impact of the DE development on the LI’s CEE can help achieve carbon reduction and encourage the growth of an economy characterized by excellence and sustainable prosperity. Therefore, this endeavor has become a focus in academia and government departments.

First, research on LI’s CEE is rich in content and diverse from a research perspective. CEE is currently a popular research topic in LI. LI achieves CEE by reducing carbon emissions, which leads to higher economic output [8]. Therefore, scholars have proposed various evaluation indicators and methods for CEE using different models to calculate LI’s CEE [8,17,18,19]. Simultaneously, certain scholars have begun to consider the factors influencing LI’s CEE [20] and the specific regional LI’s CEE, including provincial and urban levels [21]. In addition, researchers have studied the geographical evolution features of the CEE in LI [22]. Based on a series of studies, we propose strategies to improve LI’s CEE [20,21,22].

Second, studies have been undertaken to explore the correlation existing between the DE and CEE. Han et al. (2022) indicated that the DE has promoting effects on CEE levels and carbon reduction [14,15,16]. According to other academics, it is unclear how DE and carbon emissions are related [23]. Research has been conducted on the role of the DE in CEE from different perspectives, such as the impact of the DE on CEE in Chinese cities [24], the marine carbon emission rate [25], industrial CEE [26], and carbon emission performance [27]. The primary research methods used include Tobit regression and the least squares regression model [23,24,25,26,27], and to learn more about how the DE affects the CEE of the provincial LI, more research is required.

Third, relevant scholars have also conducted research on the factors influencing CEE in LI. While clarifying the concept of carbon emission efficiency, scholars point out that population, technology, spatial agglomeration [7], industrial structure [9,15], environmental regulations [20], LI development level and economic development level [20,21], optimization of the DE and energy structure, technological innovation, and application [10,11,13,14,15,16,22,28,29] are the main factors affecting CEE of LI, emphasizing that the DE may reduce carbon emissions by improving CEE [22,28].

Finally, research on the logistics industry’s CEE and the construction of a DE indicator system should be conducted. The majority of academics hold the view that capital, labor force, and utilization of energy are the most crucial input production factors in efficiency measurement [12,30,31] and that each province is aiming to maximize its output and minimize carbon emissions. Therefore, the total output value of LI and carbon dioxide emissions were selected as the expected and unexpected output indicators [21,28]. Regarding the measurement index system of the DE, Liu and Wang (2022) measured the evolution of the rate of internet penetration, the number of workers, production connected to the internet, and the advancement of digital inclusive finance [24]. Three factors were used by Liao et al. to create indicators: digital economic innovation, digital economic development, and digital economic infrastructure [31]. Four factors were taken into consideration when Wei et al. (2022) created indicators: data value, industrial digitization, digital governance, and digital industrialization [32]. Yu et al. (2023) constructed indicator systems from four aspects: digital infrastructure, digital industrialization, digital technology innovation, and industrial digitization [28,33,34]. This study provides a reference and guidance for the establishment of a CEE and DE indicator system for LI.

Research on CEE and DE by domestic and foreign scholars has provided crucial references for exploring the effect of the DE on LI’s CEE. However, certain shortcomings remain. Specifically, although the role of the DE in CEE has been studied, most research has only considered the impact of the DE on urban and provincial carbon emissions or efficiency, as well as its impact on industrial CEE. Overall, further exploration is required to elucidate the impact of the DE on LI’s CEE and spatial distribution characteristics.

• CEE of LI and DE are two intricate industrial systems that encompass various dimensions, multiple source elements, and multiple indicators. Therefore, accurately measuring and evaluating the CEE of LI and DE systems is a crucial prerequisite for conducting research on the effect of the DE on LI’s CEE.
• Several factors must be considered when studying the effect of the DE on LI’s CEE. Therefore, it is necessary to scientifically perform data processing and model construction to provide a quantitative basis for proposing targeted policy recommendations on LI’s CEE and the development of the DE.
• Considering regional resource endowments, LI, and the level of DE development, it is necessary to formulate tailored policies for the digital and low-carbon transformation development of LI, as well as DE development strategies.

To overcome these difficulties, this study draws and conducts research based on the research ideas of Wang et al. [24,27,28,30,33]. This study is more systematic and comprehensive in the construction of indicator systems than existing studies [9,21,28,33]. We measure from the perspective of all factors, comprehensively considering the expected capital output, manpower, energy consumption, output value, and unexpected output of CO₂ emissions in LI. In addition, we comprehensively consider the infrastructure construction of DE, digital industries, industrial digitization, and innovative development to construct a DE indicator system. We put forward a novel methodology intended to address measuring, evaluating, and analyzing the impact of the DE on LI’s CEE. This method is applied to 30 provinces in China to verify its effectiveness. Then, depending on the evaluation results, we make recommendations and targeted countermeasures. To fortify professionals and policymakers within the logistics and DE industries, our methodology provides a numerical basis for policy recommendations.

Therefore, the framework structure of this article is as follows: Section 2 is Materials and Methods, which introduces data collection and processing, data models, and applications. Section 3 is a case study, taking 30 provinces in China (due to the lack of data in Xizang, Taiwan, Hong Kong, and Macao, the sample is not considered) as an example, to measure, evaluate, and analyze the impact of the digital economy on the carbon emission efficiency of the logistics industry. Section 4 is discussion. Section 5 conclusion, policy implications and directions for future research.

2. Materials and Methods

2.1. Research Framework

This study collects and processes data related to CEE in the DE system and LI, embeds them into appropriate models for calculation and evaluation, applies the results to real-life scenarios, and proposes targeted policy recommendations.

This research implements the slacks-based measure (SBM) model with unexpected outputs to measure LI’s CEE to better reflect objective facts. The entropy weight and linear weighted sum methods were used to assign weight and measure the level of DE development, eliminating errors caused by inconsistent indicator dimensions. We created geographically weighted regression (GWR) and ordinary least squares (OLS) models to simulate regression and compared them after assessing the spatial correlation of CEE in the logistics sector using Moran’s I model. We visualized the spatial distribution characteristics of the regression coefficients for the impact of DE on CEE in LI. Based on this, countermeasures and suggestions were proposed to promote digital and low-carbon transformation and the development of LI. The process of this method is illustrated in Figure 1.

2.2. Index System

This study was based on research results on CEE in LI [6,7,8,9,10,20,21,22] and on the connotation and measurement of the DE [28,29,32,33,34]. The indicators were selected based on scientific and quantifiable principles. We constructed LI’s CEE and DE indicator system which includes two systems and twenty-six indicators and conducted research to gauge the system’s development status.

CEE system indicators for LI were built from the input and output perspectives. In contrast to earlier research indicators [6,7,8,9,20,21,22], in the input indicators of this study, capital represents the fixed assets investment amount of LI, labor indicates LI employees, and energy reflects the energy consumption of LI; in the output indicators, the expected output represents the output value of LI, and the expected output is the measurement of carbon dioxide emissions. This indicator system is more scientific compared to those of the other studies. The indicator system of the DE system integrates existing research [28,29,32,33,34,35,36]. The approach is made more comprehensive by adding a digital inclusive finance index and taking into account the building of infrastructure, the digital industry, digital innovation, and the industrial digitalization level (

Table 1. Measurement index system for CEE in the LI and DE.

Target Layer	Primary Indicators	Secondary Indicators	Symbol	Direction
CEE of the LI (LCE)	Input indicators	(Capital investment) fixed assets investment (CNY 100 million)	$x_{1k}$	+
		(Labor input) LI employees (10,000 people)	$x_{2k}$	+
		(Energy input) energy consumption of the LI (10,000 tons of standard coal)	$x_{3k}$	−
	Output indicators	(Expected output) total output value of the LI (CNY 100 million)	$y_{rk}^g$	+
		(Unexpected output) CO2 emissions from the LI (10,000 tons)	$y_{rk}^b$	−
Development level of DE (DE)	Digital infrastructure	Mobile phone penetration rate (unit/100 individuals)	${\mathrm{a}}_{11}$	+
		Fiber-optic cable line density (km/km2)	${\mathrm{a}}_{12}$	+
		No. of domain names per person (pieces)	${\mathrm{a}}_{13}$	+
		No. of web pages per person (pieces)	${\mathrm{a}}_{14}$	+
	Digital industrialization	Proportion of software business revenue to the GDP	${\mathrm{a}}_{15}$	+
		Per capita total telecommunications business (CNY 100 million)	${\mathrm{a}}_{16}$	+
		Fixed assets investment in the information service industry (CNY 100 million)	${\mathrm{a}}_{17}$	+
		No. of employees in the information transmission, software, and information technology service industries (10,000 individuals)	${\mathrm{a}}_{18}$	+
	Digital innovation	Employment in scientific research and technology services industries (10,000 individuals)	${\mathrm{a}}_{19}$	+
		Research and experimental development expenditure (CNY 100 million)	${\mathrm{a}}_{20}$	+
		Total No. of undergraduate talents (individuals)	${\mathrm{a}}_{21}$	+
		No. of patent applications per 10,000 people (pieces/10,000 individuals)	${\mathrm{a}}_{22}$	+
	Industrial digitization	No. of websites owned by each hundred enterprises (pieces)	${\mathrm{a}}_{23}$	+
		E-commerce transaction volume (sales revenue: CNY 100 million)	${\mathrm{a}}_{24}$	+
		No. of enterprise e-commerce situations (including e-commerce enterprises/total No. of enterprises)	${\mathrm{a}}_{25}$	+
		Digital Inclusive Finance Index	${\mathrm{a}}_{26}$	+

).

Digital infrastructure, the industrialization of digital technology, digital innovation, and the level of digitalization within industries are direct influences on the development of the DE. The integration of digital technology with LI intervenes in the industry’s carbon emissions and economic output. The application of digital technology and platforms, while predicated on a certain energy consumption [36], provides essential support for intelligent distribution systems and supply chain management within LI. The widespread adoption and application of digital technology can reduce LI’s empty driving rate and carbon emissions, achieving cost reduction and efficiency enhancement [37]. It can break through the constraints of time and space, match supply with demand, and promote the intelligent and green, low-carbon development of LI, thereby reducing carbon emissions [20,38,39]. The economic output of the LI is increased by the popularization and application of digital technology. For instance, the higher the level of digital infrastructure such as mobile phone penetration, the more accurate the prediction of logistics demand and supply, leading to higher service satisfaction and, consequently, higher economic benefits. Therefore, Hypothesis 1 is proposed: the DE has a promoting effect on CEE of LI. Additionally, considering the different development levels of the DE in various provinces [40,41,42], the impact on CEE of the LI may also vary. Hence, Hypothesis 2 is proposed: the promoting effect of the DE on the CEE of LI exhibits spatial heterogeneity.

Drawing on existing research findings [8,22,31,32,33,34], we constructed a variable indicator system that includes the impact of the DE on LI’s CEE, as shown in

Table 2. Variable indicator system.

Variable Type	Variable Name	Measurement Indicators	Symbol
Dependent variable	CEE of the LI	SBM calculation result	LCE
Explanatory variables	Development level of DE	Linear weighting method composite index	DE
Control variables	Development level of the LI	Proportion of the LI’s output value to GDP	LDL
	Economic development level	Per capita GDP	RJGDP
	Integrity with the outside world level	Percentage of GDP that comes from all imports and exports	OPEN
	Industrial structure	Percentage of GDP attributable to the secondary industry’s output value	IS
	Rules pertaining to the environment	Percentage of industrial added value from completed investments in pollution control	ER

. The LI’s CEE was the dependent variable, and the level of DE development was the explanatory variable. First, LI’s level of development directly impacts its CEE. Second, economic development both raises the carbon emissions of LI and offers institutional, financial, and technical support for reducing carbon emissions. Third, optimizing and upgrading industrial structures can promote the development of green, environmentally friendly, and high-tech industries. Industrial linkages can promote the green and environmentally friendly transformation of LI and reduce carbon emissions. The fourth is to increase environmental supervision efforts, promote green and clean energy to replace fossil fuels, and force the logistics sector to shift towards green and environmentally friendly development. Finally, integrating with the outside world can expand the boundaries of the logistics market; it is advantageous for LI, optimizing resource allocation and reducing carbon emissions within a larger market scope.

2.3. Data Source and Processing

We selected 30 provinces nationwide as samples (Xizang, Hong Kong, Macao, and Taiwan were not involved in this study due to data availability and other reasons). Divided by economy, the analysis was conducted in four regions: western, eastern, central, and northeastern regions. The data on CEE in LI and DE systems were obtained from the China Statistical Yearbook, China Internet Development Report, China Information Industry Yearbook, and China Energy Yearbook from 2014 to 2023. Data for individual indicators were unavailable for 2020, and the missing value processing method such as interpolation and averaging were adopted to ensure completion [43]. To avoid the price impact caused by inflation, the GDP and other indicators were adjusted to be at constant prices based on 2012 data. We searched for fossil fuel conversion and carbon emission coefficients using the International Panel on Climate Change (IPCC) National Greenhouse Gas Emission Inventory Guidelines (IPCC 2018) [43], and to calculate the energy consumption [7] and carbon emissions of the LI (Equation (1)).

2.3.1. Calculation of Carbon Emissions

In most countries, 70% of carbon emissions originate from conventional energy consumption, and the increase in greenhouse gases is primarily due to the combustion of fossil fuels. The consumption of raw coal, kerosene, coke, crude oil, gasoline, diesel, and natural gas in LI was included in LI’s overall energy usage according to the standard [44], referring to the carbon emissions calculation method and parameters published by the IPCC in 2006. The equivalent carbon dioxide emissions of the logistics sector were calculated using the formula for computing carbon emissions from energy consumption. The equation used was as follows:

$$C_i={\sum }_{j=1}^8{X}_{ij}\times B_j$$

(1)

where $C_i$ is the carbon dioxide emissions from the LI in the i-th province (city) in 10,000 tons, j = 1, 2 … 8, is the type of energy, $X_{ij}$ symbolizes the energy source’s consumption (the j-th), and $B_j$ symbolizes carbon emission coefficient (the j-th energy source) in the LI.

2.3.2. Calculating the Weights of DE System Indicators

To increase the precision of the evaluation results for the development level of the DE and eliminate the bias of subjective weights and other factors, weights were assigned to each indicator using the entropy weighting approach [41]. The following were the precise stages in the calculation:

Step 1: We used the range normalization method to perform dimensionless standardization on the raw DE data. For the positive and negative indications, which have different interpretations, separate algorithms were used, as shown in (2) and (3):

$$\mathrm{Positive}\mathrm{index}:Y_{tj}={\displaystyle \frac{a_{tj}-{min}_j}{{max}_j-{min}_j}}$$

(2)

$$\mathrm{Negative}\mathrm{indicator}:Y_{tj}={\displaystyle \frac{{{max}_j-a}_{tj}}{{max}_j-{min}_j}}$$

(3)

where $a_{tj}$ an $Y_{tj}$ represent the values of the t-th year and j-th indicator data (t = 1, 2, 3 …, n; j = 1, 2, 3 …, m), respectively, of the DE subsystem before and after dimensionless standardization. The ${max}_j$ representative in this study is 1.01-fold the MAX value in the j-th indicator data of the DE system throughout the time spent studying. The ${min}_j$ symbolizes 0.99-fold the minimum value of the j-th indicator data in the DE system during the study period. This indicates that ${max}_j=1.01\ast max{a}_{tj};{min}_j=0.99\ast {mina}_{tj}$.

Step 2: we determined the j-th indicator’s proportion in year t, as shown in (4):

$${\mathrm{P}}_{tj}={\displaystyle \frac{Y_{tj}}{\sum Y_{tj}}}$$

(4)

Step 3: we calculated $e_j$ of index j, as shown in (5):

$$e_j=-k\sum P_{tj}\ln P_{tj}$$

(5)

In Equation (5), k$>0,k=\left({\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$\ln n$}\right.}\right)$, n refers to the year, $e_j\ge 0.$

Step 4: we calculated the information entropy redundancy $g_j$ of index j,

$$g_j=1-e_j$$

(6)

Step 5: we calculated the weight $w_j$ of indicator j, as shown in (7):

$$w_j=g_j/\sum g_j$$

(7)

where $w_j$ represents the weight of the j-th indicator; the larger the weight, the smaller the entropy value and the less information it carries.

2.4. Data Model

2.4.1. Model of Development Level

Utilizing a multi-objective linear weighting approach, the comprehensive development level of the DE was determined. The specific Equation (8) is as follows:

$$F_i=\sum _{j=1}^m{w}_j{Y}_{tj}$$

(8)

In Equation (8), the higher the value of $F_i$, the higher the level of comprehensive development of the DE. Among them, the weight of the j-th indicator in the DE indicator system calculated in formula (7), $Y_{tj}$ represents the standardized value of the jth DE system indicator in the t-th year.

2.4.2. SBM Model for Unexpected Output

Tone’s (2001) proposition introduced the SBM model to address the oversight of slack variables in radial efficiency models. Additionally, the model’s attributes of being non-radial, non-angular, and non-dimensional mitigate biases stemming from indicator dimensions and angle-related choices [45]. Therefore, we adopted an unexpected output SBM model to measure the LI’s CEE, with the basic formula as follows:

$$\left\{\begin{array}{c}min\theta ={\displaystyle \frac{1-\frac{1}{m}{\sum }_{i=1}^m\frac{s_i^{\_}}{x_{ik}}}{1+\frac{1}{q_1+q_2}({\sum }_{r=1}^{q_1}\frac{s_r^g}{y_{rk}^g}+{\sum }_{r=1}^{q_2}\frac{s_r^b}{y_{rk}^b})}}\\ {\displaystyle \frac{s.t.X\lambda +s^{-}=x_k}{Y^g\lambda -s^g=y_k^g}}\\ {\displaystyle \frac{Y^b\lambda +s^b=y_k^b}{s^{-},s^g,s^b,\lambda \ge 0}}\end{array}\right.$$

(9)

where $\theta$ is the LI’s CEE value. An efficiency value of one signifies that the optimal performance threshold has been achieved. The super-efficiency model, in contrast to traditional DEA, enables the identification of efficient decision-making units (DMUs) with efficiency scores exceeding unity. Efficiency ($\theta$) escalates as the parameters K, ranging from 1 to n—where n is the aggregate of DMUs—alongside m,$q_1$, and $q_2$ denote the quantities of inputs, expected, and unexpected outputs for each unit. $x_{ik}$, $y_{rk}^g$, and $y_{rk}^b$ indicate the input vector, expected output vector, and unexpected output vector of the decision-making unit; X, $Y^g$, and $Y^b$ indicate the input, expected output, and unexpected output matrices, respectively; $\lambda$ is a column vector; $s^{-},s^g,\mathrm{a}\mathrm{n}\mathrm{d}{s}^b$ are relaxation variables for the input, expected output, and unexpected output, respectively.

2.4.3. Moran’s Index Model

The spatial dependence of LI’s CEE was analyzed using the global Moran’s I index [46], and the calculation formula is as follows:

$$I={\displaystyle \frac{{\sum }_{i=1}^n{\sum }_{j=1}^n{w}_{ij}\left({\theta }_i-\overline{\theta }\right)\left({\theta }_j-\overline{\theta }\right)}{S^2{\sum }_{i=1}^n{{\sum }_{j=1}^{n}w}_{ij}}}$$

(10)

where $S^2={\sum }_{i=1}^n\left(e_i-\overline{e}\right)$/n is the variance, and $w_{ij}$ is the weight correlation coefficient of the LI’s CEE between the i-th and j-th provinces ($i\ne j$); the range of values $I$ is [−1,1], and when it is greater than zero, it indicates that the CEE of the inter-provincial LI is positively correlated in space; which is to say, the high-value areas of the inter-provincial LI’s CEE are adjacent to the high-value areas, or the low-value areas are adjacent to the low-value areas. When the value of $I$ is zero, it indicates that the CEE of the inter-provincial LI is randomly distributed in space and has no correlation. When it is less than zero, it indicates that the CEE of the inter-provincial LI is negatively correlated in space; that is, high-value areas of the inter-provincial LI’s CEE are adjacent to low-value areas.

2.4.4. OLS Model

The OLS model is the most fundamental form of regression analysis that requires the least number of model conditions to minimize the sum of the squared distances from all observations on the scatter plot to the regression line [47].

$$y={\beta }_0+{\beta }_1{x}_1+{\beta }_2{x}_2+\cdots \cdots {\beta }_n{x}_n+\epsilon$$

(11)

where y is the dependent variable, ${\beta }_1,{\beta }_2{\cdots \cdots \beta }_n$ is the regression coefficient, $x_1{,x}_2{\cdots \cdots x}_n$ is the explanatory variable, and $\epsilon$ is the random error/residual term.

2.4.5. GWR Model

We explored spatial heterogeneity by constructing a GWR surface model. This local regression model utilizes the different spatial positions of each element for calculations, which can solve the problem of spatial non-stationarity and express the spatial dependence between variables [48]. Its model is

$$y_i={\beta }_0\left({\mu }_i,v_i\right)+{\sum }_{k=1}^n{\beta }_k\left({\mu }_i,v_i\right)x_{ik}+{\epsilon }_i$$

(12)

where $y_i$ is the dependent variable, $x$ is the explanatory variable, ${\beta }_0\left({\mu }_i,v_i\right)$ is the spatial coordinate of data $i$, and ${\epsilon }_i$ is the random error/residual term.

2.5. Data Application

DE has gradually become a crucial engine for high-quality economic development, and digital transformation will also become an inevitable trend for digital transformation and high-quality development in various industries. However, in the DE era, the digitalization, low-carbon transformation, and high-quality development of the LI still face challenges. Therefore, studying the effect of the DE on LI’s CEE has crucial practical significance. Based on a comprehensive consideration of the CEE of the DE and LI, this study constructed an indicator system consisting of two subsystems. Because the two systems have multiple sources of data, involve a wide range, and contain a large amount of data, this study adopted a data-driven approach. This research leveraged data-centric techniques to precisely gauge, assess, and pinpoint the influence of the DE on LI’s carbon efficiency and emissions (CEE). Drawing from the empirical evaluation, strategic policy insights were formulated to furnish a foundation for decision-making among professionals and executives in the realms of the DE and LI. Specific applications are shown in Figure 2.

3. Case Study

3.1. Background of the Case Study

Utilizing the methodological models, this paper selects 30 provinces (excluding Tibet and Hong Kong, Macau, and Taiwan due to data availability and other reasons) as the sample. By 2023, the value added by China’s core DE industries has already reached 10% of the Gross Domestic Product (GDP), achieving the target set in the “14th Five-Year Plan for DE Development” for the year 2025 ahead of schedule. China’s logistics market has been the largest in the world for seven consecutive years, with the total social logistics value reaching CNY 347.6 trillion in 2022, and the total revenue of the LI being CNY 12.7 trillion. However, behind the rapid growth of China’s LI lies a significant amount of energy consumption, and the emissions of CO₂ and pollutants have been increasing annually. The development of the DE can promote the digital transformation of LI, breaking the traditional economic growth model characterized by high energy consumption and high pollution through the input of multiple factors into the development of LI, and advancing the green development of the economy. Therefore, this paper measures the development level of China’s DE and CEE of LI, analyzes the spatial impact of China’s DE on CEE of LI, and puts forward countermeasures and suggestions tailored to local conditions. This is of great significance for promoting the sustainable development of China’s and the world’s DE and LI.

3.2. Results

3.2.1. Analysis of the Development Level of DE

From 2013 to 2022, the development level of China’s DE has generally shown an upward trend, with an accelerated growth rate after 2017. This is related to the 2016 Hangzhou Summit, which proposed the development of DE as the primary path for China’s innovative growth, as well as the strategic plan proposed in 2017 to accelerate the construction of a digital China. This study analyzed four regions based on economic division: western, eastern, central, and northeastern regions. From a regional perspective, there were significant differences between the eastern, central, and western regions. With the central region being stronger than the northeast region and the western region being the least, the eastern region had a significantly higher level of digital economic development than the other regions. At the provincial level, the 30 provinces in China exhibited a spatial development pattern of “high in the southeast, low in the northwest, and balanced in the central region”. During this period, the development of China’s DE has shown significant regional differences, with the eastern coastal areas standing out. Provinces and cities such as Beijing, Guangdong, Jiangsu, Shanghai, and Zhejiang, due to their strong economic foundation, robust innovation capabilities, favorable policy environment, and vast market potential, have ranked at the forefront of the country in terms of digital economic development, as shown in Figure 3.

3.2.2. Analysis of the Logistic Industry’s CEE

CEE of China’s LI in 30 provinces was calculated using the Slacks-Based Measure (SBM) model based on the collected data, as shown in

Table 3. Years 2013–2022 CEE of the LI.

	Province	2013	2014	2015	2016	2017	2018	2019	2020	2021	2022	Mean
1	Beijing	0.3219	0.3064	0.3340	0.3443	0.2934	0.1656	0.2511	0.2407	0.2938	0.3150	0.2866
2	Tianjin	0.4569	0.3743	0.3992	0.4006	0.5516	1.0498	0.4592	0.4279	0.4889	0.6339	0.5242
3	Hebei	0.5972	0.6342	0.6194	0.5750	0.6697	0.6493	0.6705	0.7791	1.0398	1.0167	0.7251
4	Shanxi	0.3085	0.2976	0.3143	0.3291	0.8991	0.6602	0.4686	0.4897	0.5681	0.4942	0.4829
5	Inner Mongolia	0.3846	0.3837	0.3272	0.3429	0.3455	0.5799	0.4780	0.5255	0.8741	0.8428	0.5084
6	Liaoning	0.2822	0.2806	0.4029	0.4317	0.5422	0.6880	1.0306	0.9331	1.0225	0.6181	0.6232
7	Jilin	0.2483	0.2285	0.2082	0.2038	0.2177	0.2883	0.2322	0.2742	0.3297	0.3214	0.2552
8	Heilongjiang	0.2814	0.2606	0.2189	0.2175	0.2232	0.2462	0.1674	0.1634	0.1871	0.1847	0.2150
9	Shanghai	0.3435	0.4503	0.3112	0.3007	0.3216	0.2922	0.4054	0.5159	0.7422	1.0159	0.4699
10	Jiangsu	0.5337	0.4231	0.4149	0.4232	0.4359	0.4030	0.4168	0.4678	0.6469	0.5706	0.4736
11	Zhejiang	0.3163	0.3315	0.3109	0.3253	0.3337	0.2843	0.3123	0.2894	0.3404	0.3902	0.3234
12	Anhui Province	0.2681	0.2532	0.2294	0.2174	0.2213	0.3493	0.4995	0.5108	0.5770	0.5621	0.3688
13	Fujian	0.3288	0.3468	0.3711	0.3914	0.4168	0.2474	0.3157	0.3498	0.4491	0.5079	0.3725
14	Jiangxi	0.3965	0.3270	0.3035	0.3034	0.3801	0.3599	0.4155	0.3961	0.4728	0.5209	0.3876
15	Shandong	0.4754	0.3766	0.3640	0.3782	0.4027	0.4367	0.4225	0.4553	0.6239	1.0043	0.4940
16	Henan	0.3246	0.3850	0.3418	0.3636	0.3653	0.3882	0.4572	0.4379	0.5456	0.5935	0.4203
17	Hubei	0.2304	0.2365	0.2259	0.2031	0.2190	0.2482	0.3392	0.2965	0.3703	0.4104	0.2780
18	Hunan	0.3320	0.3248	0.3048	0.3028	0.3275	0.2930	0.3003	0.3022	0.3345	0.3303	0.3152
19	Guangdong	0.3172	0.3183	0.3116	0.1630	0.3321	0.2680	0.3011	0.2877	0.3669	0.4114	0.3077
20	Guangxi	0.2398	0.2306	0.2364	0.2355	0.2533	0.1843	0.2125	0.2183	0.2557	0.2662	0.2333
21	Hainan	0.1705	0.2052	0.1794	0.1786	0.2217	0.1984	0.3121	0.2795	0.4046	0.4452	0.2595
22	Chongqing	0.1929	0.2199	0.2084	0.2176	0.2210	0.1959	0.2390	0.2289	0.2886	0.3205	0.2333
23	Sichuan	0.1669	0.1837	0.2082	0.2103	0.2161	0.1592	0.1972	0.1922	0.2230	0.2230	0.1980
24	Guizhou	0.3670	0.3575	0.3608	0.3618	0.3820	0.1844	0.2227	0.2235	0.2861	0.2908	0.3037
25	Yunnan	0.0920	0.0836	0.0834	0.0825	0.0801	0.1617	0.2368	0.2361	0.2866	0.3204	0.1663
26	Shanxi	0.2373	0.2250	0.1982	0.2052	0.2036	0.2187	0.2469	0.2778	0.3933	0.4128	0.2619
27	Gansu	0.2374	0.1398	0.1366	0.1197	0.1341	0.1956	0.2048	0.1842	0.2287	0.2769	0.1858
28	Qinghai	0.1100	0.1084	0.1165	0.1064	0.1025	0.1069	0.1375	0.1272	0.1619	0.1737	0.1251
29	Ningxia	0.4174	0.3552	0.3211	0.2875	0.2830	0.2628	0.3395	0.3617	0.4137	0.4134	0.3455
30	Xinjiang	0.2172	0.2069	0.1954	0.2258	0.1821	0.2061	0.3318	0.1977	0.2259	0.2912	0.2280

. It is observed that the CEE of LI in most provinces and cities has been growing amidst fluctuations, with the overall trend being relatively stable.

We have converted the CEE of the two-time series in Table 3, 2013 and 2022, into a map graph, as shown in Figure 4. The spatial development trend of CEE in the LI is characterized by “high efficiency changing and stabilizing, medium efficiency gradually improving and stabilizing, and low efficiency relatively stable”.

Analyzing from the perspective of the four major economic regions, as can be seen in Figure 5, the CEE of China’s LI has shown a fluctuating upward trend from 2013 to 2022, reaching its peak at 0.4726 in 2022. Spatially, there is a general pattern of “the east being higher than the west”, indicating significant regional disparities, with the eastern region having a relatively better economic foundation and higher CEE in the LI. Integrating

Table 4. Moran value test results.

Year	Moran’s I	Z-Score	p-Value
2013	0.3522	3.1925	0.0014 ***
2014	0.3796	3.5019	0.0005 ***
2015	0.4115	3.7767	0.0002 ***
2016	0.3925	3.5143	0.0004 ***
2017	0.2502	2.4264	0.0153 **
2018	0.2599	2.5250	0.0116 **
2019	0.3014	3.0966	0.0020 ***
2020	0.3951	3.6766	0.0002 ***
2021	0.4077	3.6732	0.0002 ***
2022	0.3429	3.1296	0.0017 ***

Note: ** p ˂ 0.05, *** p ˂ 0.01.

and Figure 5, specifically, Xinjiang, Gansu, Qinghai, Guangxi, Sichuan, Yunnan, and Heilongjiang have consistently been in the low-efficiency area. The high-efficiency areas have evolved from Tianjin, Hebei, Jiangsu, Shanghai, and Shandong in 2013 to Tianjin, Hebei, Inner Mongolia, Liaoning, Jiangsu, and Shanghai in 2022.

3.2.3. Spatial Autocorrelation Analysis of the Logistic Industry’s CEE

According to Equation (10), we examined the regional dependence and autocorrelation of the CEE in the LI, and the results are listed in Table 4. A significant positive spatial correlation between the LI’s CEE during the sample period was observed, indicating that the LI’s overall CEE is relatively clustered. From a temporal perspective, the Moran’s I value for the CEE of the LI showed an increasing trend amidst fluctuations from 2013 to 2021, but in 2022, they were lower than in the previous two years. When considering geographical factors, there is spatial instability in the CEE of the LI, further GWR analysis is required.

3.2.4. Analysis of Regression Results

To learn more about how DE affects CEE in the logistics sector, we conducted an estimation analysis using OLS and GWR models on sample data from 2013, 2016, 2019, and 2022. To eliminate the interference of collinearity, collinearity tests were conducted on each explanatory variable. The variance inflation factor values were less than 7.5 [35] (

Table 5. VIF values of OLS results from 2013, 2016, 2019, and 2022.

Variable	DE2013	LGL2013	IS2013	OPEN2013	ER2013	RJGDP2013
VIF value	6.6393	1.5699	1.7035	7.1516	1.9478	2.7402
Variable	DE2016	LGL2016	IS2016	OPEN2016	ER2016	RJGDP2016
VIF value	4.2537	1.1427	1.3231	3.9486	1.0879	2.9018
Variable	DE2019	LGL2019	IS2019	OPEN2019	ER2019	RJGDP2019
VIF value	5.6651	1.8129	1.1551	6.2353	1.6608	6.7596
Variable	DE2022	LGL2022	IS2022	OPEN2022	ER2022	RJGDP2022
VIF value	4.9126	1.3509	1.4769	7.3928	1.10153	5.0314

), indicating an absence of multicollinearity.

The fitting index is shown in

Table 6. Comparison of GWR and OLS model parameters.

Model Parameters	2013		2016		2019		2022
	GWR	OLS	GWR	OLS	GWR	OLS	GWR	OLS
R	0.6374	0.6373	0.6620	0.6618	0.3913	0.3909	0.7071	0.7069

based on the calculation of the OLS and GWR models. The comparison of the fitting degrees between the two models showed that the R values were slightly higher in the GWR: the higher the fitting degree, the better the model performance.

The OLS model only estimates the global regression situation and cannot explore the spatial regression conditions by region. Considering this, the article employs the Geographically Weighted Regression (GWR) to further investigate the spatial distribution characteristics of the impact of the DE on the CEE of the LI. From 2013 to 2022, due to the impact of factors such as the epidemic in the middle years, we only show the spatial distribution changes in the impact of the DE on the CEE of the LI from the beginning to the end of the period. The regression coefficients for the beginning year of the study period, 2013, and the end year, 2022, are presented here, as shown in

Table 7. GWR regression coefficient descriptive statistics.

Variable	DE		LGL		IS		OPEN		ER		RJGDP
Year	2013	2022	2013	2022	2013	2022	2013	2022	2013	2022	2013	2022
Max	1.2565	0.6116	5.1633	18.4101	0.4468	0.1732	−0.0741	−0.0337	−3.3099	11.7346	0.0148	0.0250
Median	1.2562	0.6109	5.1623	18.4083	0.4466	0.1728	−0.0741	−0.0338	−3.3193	11.7017	0.0148	0.0250
Min	1.2557	0.6107	5.1614	18.4065	0.4465	0.1721	−0.0742	−0.0340	−3.3276	11.6218	0.0148	0.0249

. The DE has played a role in promoting the improvement of CEE in the LI during these two years, suggesting the possibility of the validity of Hypothesis 1. Additionally, it can be observed that the development level of the LI, the industrial structure, and the economic development level are also positively correlated with the CEE of the LI during these two years.

We have chosen to visualize the regression coefficients of the DE for the years 2013 and 2022, which allows for an intuitive observation of the spatial impact evolution of the DE on the CEE of the LI, as depicted in Figure 6. The DE has a decreasing impact on the CEE of the LI, with a trend of decreasing from northwest to southeast in 2013 and gradually decreasing from west to east in 2022. This may be mainly related to the longer wintertime in the north, which leads to an increase in energy consumption input costs for the LI. Other conditions remaining unchanged, the DE has a greater impact on the CEE of the LI in the northwest and central regions with lower levels of digital economic development. It can also be seen that there is a certain spatial difference in the role of the DE in improving the CEE of the LI, and Hypothesis 2 is also valid. However, from Table 6 and Figure 6, it can be seen that the spatial differences in this impact are very small, indicating that there are spatial differences in the impact of the DE on the CEE of the LI in different regions in the same year, but they are very small. The DE is important for the low-carbon and digital transformation of the LI in various regions. The regression coefficient of the DE on the CEE of the LI in 2022 is smaller than that in 2013, possibly because by 2022, the development level and economic development level of the LI in the eastern region have already been relatively high, the industrial structure is more reasonable, and the agglomeration effect of the LI has become more widespread. The adoption and application of digital technology have become more widespread, and the impact of the DE on the eastern region is smaller than that on the northwest region. The pursuit of a better environment by people in developed regions or the situation of opening to the outside world may have a greater impact on the CEE of the LI.

4. Discussion

The rapid expansion of global digital economies (DEs) has brought the digital transformation of traditional industries [49], particularly the logistics industry (LI), to a critical juncture. As many nations intensify efforts to promote sustainable development through energy conservation and emission reduction initiatives, understanding the impact of the DE on the carbon emission efficiency (CEE) of the logistics sector becomes essential. Given its status as a key driver of economic growth and a significant source of carbon emissions, the LI’s transformation is fundamental to achieving sustainable regional development and long-term carbon-reduction targets [50].

The findings of this study demonstrate that the DE positively influences the CEE of the LI. However, there are notable regional disparities in its impact. Eastern regions, with higher levels of digital development, exhibit more significant improvements in carbon efficiency compared to the central and western regions. These spatial imbalances highlight the importance of targeted policy interventions that address regional gaps in digital infrastructure and economic development. The spatial analysis, which utilized Moran’s I model and geographically weighted regression (GWR), confirmed that regional proximity and the level of digital infrastructure are crucial in determining the effectiveness of DE on CEE in the logistics sector. Regions with more advanced digital capabilities show markedly higher carbon efficiencies.

5. Conclusions, Policy Implications, and Directions for Future Research

This study provides important contributions to the existing literature by integrating the CEE of DE and LI into a unified analytical framework. Through the application of a slacks-based measure (SBM) model and advanced spatial econometric methods, this study systematically evaluates the spatial dynamics between DE and the logistics industry’s carbon efficiency. By comparing ordinary least squares (OLS) and geographically weighted regression (GWR) models, the analysis identifies the most suitable approach for capturing regional variations, offering a scientific foundation for targeted policy recommendations.

Despite these valuable contributions, this study acknowledges certain limitations. The indicator system used, while encompassing multidimensional data, requires further refinement to enhance its comprehensiveness. Furthermore, data constraints, particularly the lack of complete data for 2023, indicate that future research should prioritize timeliness and accuracy in data collection. The mechanisms driving the spatial impact of the DE on the CEE of the logistics industry also warrant further exploration. Investigating the role of mediating variables and other region-specific factors could offer deeper insights into this relationship.

5.1. Policy Implications

Addressing regional disparities in the development of DE is critical for promoting a more balanced and sustainable transformation of the logistics industry. Targeted investments in digital infrastructure, particularly in the central and western regions, are necessary to bridge the development gap with more advanced eastern regions. Enhancing regional cooperation and facilitating the sharing of digital resources will improve the efficiency of digital infrastructure utilization and contribute to the overall enhancement of CEE in the logistics sector.

Promoting the digital transformation of logistics enterprises is also a key policy priority. Policymakers should support the adoption of digital technologies such as smart logistics and green logistics, while incentivizing logistics companies to embrace digitalization. Leading enterprises, such as China Ocean Shipping Group Co., Ltd., headquartered in Shanghai and SF Holding Co., Ltd., headquartered in Shenzhen, China. can play a critical role in setting industry standards and demonstrating the benefits of digitalization in improving operational efficiency and reducing carbon emissions. Additionally, the development of digital platforms to optimize logistics supply and demand matching will further enhance the sector’s carbon emission efficiency [51].

Finally, regional coordination for green development is vital. Policymakers must take into account the economic, industrial, and environmental specificities of different regions when formulating strategies to reduce carbon emissions in logistics activities. Integrating the logistics industry with other sectors and promoting the use of renewable energy will support high-quality, low-carbon development within the industry. Such efforts will not only help reduce carbon emissions but also contribute to broader national and global objectives of achieving carbon neutrality and sustainable economic growth.

5.2. Future Research Directions

This study opens several avenues for future research. First, the mechanisms underlying the spatial impact of DE on CEE deserve further investigation. The potential role of mediating variables and region-specific factors influencing the relationship between DE and CEE should be explored to provide a more comprehensive understanding of this dynamic. Additionally, improvements in data collection, particularly for more recent years, are essential to ensure the timeliness and accuracy of future analyses.

By continuing to investigate the spatial effects of DE on CEE, future research can offer more precise policy recommendations that address regional disparities and promote sustainable growth across all regions. This research is essential for advancing the green transformation of the logistics industry in a way that aligns with national and global sustainability goals.