Research on Technology Spillover of Digital Economy Affecting Energy Consumption Intensity in Beijing–Tianjin–Hebei Region

Abstract

As a new economic paradigm, the digital economy is critical to economic growth and environmental protection. This paper empirically explores the impact of the digital economy on regional energy consumption intensity in the Beijing–Tianjin–Hebei region from 2010 to 2018. It is found that the digital economy has a significant inhibitory effect on regional energy consumption intensity. This effect remains valid even after passing the endogeneity and robustness tests. The paper confirms that technological innovation is the primary means by which the digital economy affects energy consumption intensity. The analysis of spatial spillover effects shows that the digital economy promotes the improvement of energy consumption intensity in surrounding areas through technology spillover effects. A heterogeneity analysis demonstrates that the technology spillover effect has a significant inhibitory effect on the energy consumption intensity of the surrounding areas for economically developed cities. Currently, the digital economy is a significant driver for enhancing productivity and quality. The integration and application of digital technologies have enabled technological innovation in the real economy, effectively reducing regional energy consumption.

1. Introduction

Energy consumption is a major factor contributing to greenhouse gas emissions. With the sharp increase in energy consumption, global carbon dioxide emissions are projected to rise from 32.1 billion tons in 2015 to 36.8 billion tons by 2022. As the world’s most populous country, China’s total energy consumption has increased from 4.30 billion tons in 2015 to 5.41 billion tons in 2022. Energy conservation and emission reduction have become unavoidable issues for all countries. In 2021, the CPC Central Committee and the State Council issued the Opinions on Improving the Accurate and Comprehensive Implementation of the New Development Concept and Doing a Good Job in Carbon Peaking and Carbon Neutrality, which further clarified that “energy conservation should be integrated throughout the entire process and all sectors of economic and social development”. The report of the 20th National Congress of the Communist Party of China emphasized the need to accelerate the establishment of a digitalized China. The advent of modern information technology and data-driven approaches has led to the emergence of a new economic form, the digital economy. This economic form has the potential to drive social progress, enhance production efficiency, and facilitate industrial development. As a nascent technology, digital technology plays a pivotal role in enhancing energy efficiency and curbing pollution emissions, and it is, therefore, held to higher standards of environmental responsibility [1]. The development of the digital economy is enabling the acceleration of Chinese-style modernization. According to the research report on the development of China’s digital economy in 2023, the scale of China’s digital economy reached 50.2 trillion yuan in 2022, marking a nominal increase of 10.3% year-on-year. This is equivalent to the proportion of the secondary industry in the national economy. The rapid development of the digital economy provides a powerful driving force for economic and social development. Although the digital economy utilizes artificial intelligence and big data processing, it can optimize the operation and management of the energy system to a certain extent, thereby reducing energy consumption intensity [2]. However, the expansion of digital infrastructure and energy consumption for the development of the digital economy may also result in an increase in energy consumption intensity [3]. Therefore, there may be a very close correlation mechanism between the digital economy and energy consumption intensity. To clarify the internal relationship between the two, it is necessary to conduct in-depth analysis and systematic data validation.

The main reasons for selecting Beijing–Tianjin–Hebei as the research focus in this study are as follows: firstly, Beijing–Tianjin–Hebei and its surroundings is the most critical area of air pollution in China. One of the primary contributors to air pollution in this area is the energy consumption intensity. In 2021, the energy consumption intensity in the Beijing–Tianjin–Hebei region reached 0.25 tons of standard coal per 10,000 yuan of GDP, and it showed a year-on-year growth trend, which shows that the improvement in energy consumption intensity in the Beijing–Tianjin–Hebei region still has a long way to go. Second, Beijing–Tianjin–Hebei is the hub of China’s digital economy development. Beijing, Tianjin, and Hebei have established an ecological system for digital economy innovation and research, positioning themselves as leaders in the digital economy ecosystem and attracting top innovation resources. In 2021, the scale of the Beijing–Tianjin–Hebei digital economy exceeded 4 trillion, accounting for 10% of the country’s total. The digital economy is experiencing positive momentum. With the development of the digital economy, the application of digital technology can improve the efficiency of production, distribution, and consumption and reduce the waste of resources and energy. But the development of the digital economy also requires energy consumption to support it, which may lead to some contradictions between the two. Focusing on the current situation, in the face of severe environmental pollution in the Beijing–Tianjin–Hebei region, can the digital economy effectively reduce energy consumption intensity? What is the specific mechanism of action? Is there a spatial spillover effect? The above issues need to be scientifically verified and logically analyzed.

The remainder of this paper is organized as follows: Section 2 surveys the relevant literature; Section 3 presents a theoretical analysis and formulates the research hypotheses; Section 4 describes the models and data used in this study. Section 5 and Section 6 present the empirical analysis. Section 7 presents the conclusions and policy recommendations of the study, along with an analysis of its limitations and future prospects.

2. Literature Review

2.1. The Digital Economy

In The Fourth Industrial Revolution, Schwab proposed that the fourth industrial revolution is the convergence of technologies, with the digital technology representing one such convergence [4]. As a new, more advanced and sustainable type of economic form, the rapid evolution of the digital economy and its great impetus to the development of society are pivotal to the promotion of change in the technological field, acceleration of the upgrading of industrial structure, and the growth of regional economies [5,6,7]. The digital economy also marks the advent of the fourth industrial revolution and will serve as the primary trajectory for the country’s future development [8,9]. Currently, there is no uniform definition of the digital economy. The concept of the digital economy was initially proposed by Don Tapscott [10] and has since evolved in tandem with the development of the underlying technologies [11]. Information and Communication Technology (ICT) encompasses technologies related to information processing, communication, and the convergence of various information processing technologies [12]. The digital economy is also regarded as a novel paradigm propelled by digital technologies [13]. Zhang et al. define a true “digital economy” as one in which a portion of economic output is derived entirely or primarily from digital technologies [14].

2.2. The Digital Economy and Energy Consumption

There are two opposing views on the impact of the digital economy on energy consumption. One perspective posits that the digital economy exerts a positive influence on reducing energy consumption [15,16]. In contrast, the opposing view suggests that the digital economy increases energy consumption [17,18]. The digital economy can decrease energy consumption intensity by encouraging intelligent and efficient energy use. Information and communication technologies (ICT) and services have made an important contribution to economic development, resulting in significant improvements in energy efficiency [17]. Henryk Dzwigol et al. found a negative correlation between digitization and energy consumption by exploring the intricate relationship between digitization and energy in EU countries [19]. Erdmann and Hilty demonstrated that the digital economy has the potential to reduce traffic flows and energy demand, drive technological efficiency gains and structural change, and reduce greenhouse gas emissions [20]. Ishida [21], Kumar and Manas [2], and Lange et al. [22] found that increased investment in information and communication technology (ICT) has improved energy use efficiency and reduced energy consumption intensity. These findings were based on examples from Japan, India, and Germany, respectively. Although scholars have already explored the relationship between the digital economy and energy consumption, the academic community has not reached a unanimous point of view; there are still some differences. Some scholars are skeptical that the digital economy can inhibit the development of energy consumption. Wang and Zhu [23] pointed out that the digital economy, facilitated by information and communication technologies, has significant potential for energy conservation. However, there is uncertainty about whether this potential can be realized. Moyer et al. [24] studied the dynamic impacts of digital information technology development on economic development, energy systems, and carbon emissions. They concluded that the impact of digital information and communication technologies on carbon emissions is limited.

2.3. Spatial Effects of the Digital Economy

The digital economy usually has positive spatial spillover effects, and the development of the digital economy promotes the formation of innovation ecosystems, which is conducive to promoting the improvement of regional energy consumption intensity and green development. Fang [25] analyzed the impact mechanism of the digital economy on urban energy efficiency from the perspectives of technological empowerment and technological spillover and found that the digital economy can significantly promote the improvement of energy efficiency in neighboring cities through technological advancement. Tang et al. identified spatial spillover effects in the digital economy development among Chinese provinces [26]. The characteristics of data elements, including non-exclusivity of use, non-competitiveness, non-scarcity of sources, and non-exhaustiveness of value, result in data elements with considerable spatial spillover effects [27]. Jie Wu et al. further confirmed the positive spatial spillover effect of the digital economy on urban energy efficiency [28]. Keller adds a discussion of spillover distances from the perspective of knowledge and technology diffusion [29]. However, there is also some disagreement in the academic field about the spatial effects of the digital economy. Xue et al. [1] conducted a spatial analysis and determined that over time, the digital economy exerts a negative spatial spillover effect on energy intensity. Deng and Zhang [30] similarly found that the development of the digital economy has negative spatial spillover effects on environmental pollutants in neighboring cities. Nevertheless, the digital economy may also result in the concentration of high-quality production factors in the central city [31,32], which may give rise to a “siphon effect” and is not conducive to energy conservation and emission reduction in the surrounding areas [26].

A review of the literature reveals that while the academic community has conducted some research on the relationship between the digital economy and energy consumption, there are still certain shortcomings. First, most scholars’ discussion of the impact of the digital economy on energy development is particularly one-sided [33]. Few scholars have discussed the impact of the digital economy on energy intensity. Secondly, although some scholars have studied the impact of the digital economy on energy consumption intensity based on spatial econometric modeling, the majority of scholars consider the direct spillover effects of the digital economy and do not analyze the mechanism by which the digital economy affects energy intensity. Thirdly, the regional heterogeneity of the impact of the digital economy on energy intensity is also insufficient, and scholars have not conducted research on the spatial heterogeneity due to the economic development gap between regions. In light of the aforementioned considerations, this study aims to elucidate the mechanism through which the digital economy affects energy consumption intensity, with a particular focus on technology spillovers. To this end, spatial econometric models are employed to analyze the spatial effects and heterogeneity of the digital economy on energy consumption intensity in the Beijing–Tianjin–Hebei region. The findings of this study will serve as a reference for the promotion of the digital economy and the alleviation of the challenges posed by energy consumption intensity and environmental pollution in the Beijing–Tianjin–Hebei region.

Compared with previous studies, the innovations of this study are as follows: firstly, in terms of research content, the study considers the spatial differences in the impact of the digital economy on energy consumption intensity. This study aims to integrate the digital economy and technological innovation into the framework of energy consumption research. It examines the correlation between the digital economy and energy consumption intensity by considering spatial spillover effects, transitioning from individual variances to the regional level as a whole. This approach helps bridge the existing research gaps in this field. Secondly, although most studies in this area focus on the provincial level, conducting research at the Beijing–Tianjin–Hebei regional level allows for a more detailed examination from a micro perspective. This region provides an opportunity to delve deeper into the spatial and temporal evolution characteristics and the relationship of influence between the digital economy and energy consumption intensity. Finally, the analysis examines the impact of digital technology on energy consumption intensity in various regions of Beijing–Tianjin–Hebei from a perspective of heterogeneous effects. The goal is to provide insights into how different regions of Beijing–Tianjin–Hebei can implement digital transformation to mitigate the dividend effect of energy consumption intensity.

3. Hypotheses Development

3.1. The Impact of the Digital Economy on Energy Consumption Intensity

From a macro perspective, the development of the digital economy can facilitate the transformation and upgrading of the industrial structure, influencing energy consumption intensity [34,35]. On the one hand, from the demand-side perspective, the digital economy, with its fundamental innovation, interconnectivity, and breakthrough of space–time boundaries, has promoted profound changes in the primary, secondary, and tertiary industries. The transformation of the industrial structure, optimization, and upgrading has led to a low-carbon transformation of the energy demand structure. This transformation has facilitated the shift from traditional fossil energy and “double-high” (high energy consumption and emissions) production to clean energy and green production methods. On the other hand, from the supply-side perspective, the industrial structure has a positive impact on the development of the digital economy, and the structural optimization of the digital economy improves on the disadvantages of the traditional energy factors and reduces energy consumption [36].

At the micro level, the development of the digital economy promotes the transformation of business models towards intelligence, efficiency, and convenience, thereby reducing energy consumption. The advantage of transcending time and space allows for the digital economy to realize online services and intelligent commerce transactions, which, in turn, reduces the energy consumption generated by the circulation of people. This also enhances the possibility of improving the supply chain [37]. On the other hand, the digital economy’s new transaction mode will reshape consumers’ awareness and concept of consumption. Modern consumption concepts, such as “green consumption” and “low-carbon consumption”, will be disseminated to guide consumers towards more sustainable consumption patterns, reducing energy consumption.

Hypothesis 1 (H1).

The digital economy has a dampening effect on energy consumption intensity.

3.2. Indirect Effects of the Digital Economy on Energy Consumption Intensity

The digital economy relies on information technologies such as blockchain, big data, the internet, cloud computing, and artificial intelligence to integrate and innovate with the real economy, reducing energy consumption. The digital economy is considered a novel economic form that facilitates the modernization of traditional industries and the rapid growth of emerging industries, exerting a profound influence on enterprises’ digital transformation [38]. Interconnectivity and data-driven and factor innovations are among its advantages [39]. Empowered by digital elements, technological innovation can improve the efficiency of energy utilization to a certain extent, thereby realizing energy saving and emission reduction [40].

First, the digital economy can facilitate the advancement of technological innovation [41,42], which, in turn, results in a reduction in energy consumption. Technological innovation is a crucial factor in the digital economy’s ability to address energy consumption [34]. Schumpeter’s theory of innovation states that innovation is the recombination of production factors. Information and communication technology is a production factor with technological attributes that can effectively integrate factors of production such as labor, capital, and technology, thus promoting technological innovation [43]. In the digital era, enterprises will rely on digital technology to improve their product innovation ability and efficiency. Technological innovation promotes the transformation of the energy consumption structure to low-carbon, clean, and green, reducing energy consumption, according to Davido’s law [44].

Secondly, the development of digital technology has the potential to enhance the efficiency of the green economy and help businesses achieve their energy control and emission reduction objectives. From the perspective of the substitution effect, information and communication technology (ICT) has the potential to enhance the proportion of low-carbon industries and the utilization of energy resources [45]. From the spatial dimension, digital technology crosses spatial and temporal boundaries, improves the efficiency of spatial resource allocation, supports regional synergy in advancing the development of the green economy, and broadens the space for the promotion and utilization of clean energy [17], which, in turn, reduces the transitional consumption of traditional energy.

Hypothesis 2 (H2).

The digital economy can act as a disincentive to energy consumption intensity through technological innovation.

3.3. Spillover Effects of the Digital Economy on the Impact of Energy Consumption Intensity

The digital economy has significant externalities, and it has a significant spillover effect on energy consumption intensity in surrounding areas [36,46]. First of all, the digital economy follows Metcalfe’s law, which states that the value of a network is proportional to the square of the number of nodes in the network. When the digital economy reaches a certain critical point, its external economy will experience explosive growth [47]. The channels of spillover effects of the digital economy on energy consumption are mainly through the diffusion of digital technologies, the sharing of digital resources, and the penetration of economic and social technologies [36]. With the continuous development of the digital economy, their external economy will become increasingly significant. Secondly, one important aspect of the data economy is its ability to compress spatiotemporal distance by improving the efficiency of information transmission [25]. This, in turn, enhances the speed and reduces the cost of information dissemination and acquisition, leading to increased breadth and depth of inter-regional economic activity linkage. Additionally, it accelerates the spillover effect of network externality, improves spatiotemporal correlation, and enhances energy utilization efficiency. Finally, the digital economy is highly permeable. The technical and data elements introduced by the digital economy can overcome geographical limitations; improve the efficiency of resource allocation in factor markets; accelerate the exchange of production factors such as technology, talent, and knowledge across space and time; enable innovation in neighboring regions; and, thus, enhance inter-regional cooperation. The enhancement of technological innovation across various production sectors improves production processes and methods, facilitates the adoption of energy-saving production methods, and reduces regional energy consumption. Simultaneously, the spatial and temporal flow of factors can create a new location advantage, resulting in the rapid integration of production factors across different regions. This leads to a new spatial combination of modes and the reconfiguration of economic space. The digital economy can achieve higher-quality and more efficient outputs with fewer factor inputs, thus realizing innovative, coordinated, and shared development and promoting harmonious coexistence between human beings and nature [48].

Hypothesis 3 (H3).

The spatial spillover effects of the digital economy on energy consumption intensity are significant.

4. Research Design

4.1. Model Construction

According to the above theoretical assumptions, in order to test hypothesis H1, this paper constructs the following econometric model to analyze the impact of the digital economy on energy consumption intensity:

$$\mathit{ln}{EI}_{it}={\beta }_0+{\beta }_1{Dige}_{it}+{\beta }_c{Control}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}$$

(1)

In Formula (1), i denotes the city, t denotes the time in years; $EI$ is urban energy consumption intensity; $Dige$ indicates the development level of the city’s digital economy; Control means a series of control variables affecting urban energy consumption intensity; ${\beta }_1$ represents the direct influence coefficient of the digital economy on urban energy consumption intensity, and β_c is the indirect influence coefficient of control variables on urban energy consumption intensity; ${\mu }_i$ and ${\delta }_t$ stand for the city and time fixed effects; ${\epsilon }_{it}$ is the random disturbance term.

To further explore the potential indirect mechanism of the digital economy’s impact on energy consumption intensity, this paper proposes a specific model to examine whether technological progress serves as an intermediary variable between the digital economy and energy consumption intensity:

$$\mathit{ln}{EI}_{it}={\alpha }_0+{\alpha }_1{Dige}_{it}+{\alpha }_c{Control}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}$$

(2)

$${TI}_{it}={\beta }_0+{\beta }_1{Dige}_{it}{+\beta }_c{Control}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}$$

(3)

In Formulas (2) and (3), TI is technological innovation; ${\alpha }_0$ and ${\beta }_0$ represent constant terms. ${\alpha }_1$,${\alpha }_c$, ${\beta }_1$, ${\beta }_c$ are coefficients to be estimated.

Finally, we further research the potential spatial spillover effect of the digital economy on energy consumption intensity through its associated technological innovations. In Formula (1), the spatial interaction terms of digital economy, energy consumption intensity, and other control variables are introduced, and it is further expanded into a spatial panel measurement model:

$$\ln {EI}_{it}={\beta }_0+\rho W\ln {EI}_{it}+{\phi }_{1}W{TI}_{it}+{\beta }_1{TI}_{it}+{\phi }_{c}W{Control}_{it}+{\beta }_c{Control}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}$$

(4)

In Formula (4), $\rho$ represents the spatial autoregressive coefficient; W is the spatial weight matrix; ${\phi }_1$ and ${\phi }_c$ are the elastic coefficients of the interaction terms in the space of explanatory variables and control variables. Equation (4) includes the explained variable and the spatial interaction of the explained variable, which is called the spatial Durbin model (SDM).

4.2. Variable Definition and Measurement

4.2.1. Description of Variables

The measurement of the digital economy is still imperfect and primarily relies on provincial panels. While these panels can highlight differences between provinces, they cannot accurately reflect differences between cities within a province. This paper focuses on research objects selected for this study and locates them at the city level. Based on the Beijing–Tianjin–Hebei regional city panel data, this paper relies on the methodology of Xu et al. [49], Jing et al. [50], and Zhao Tao et al. [51] to construct the digital economy at the city level in the Beijing–Tianjin–Hebei region by selecting four indicators: Internet penetration, related practitioners, related outputs, and mobile phone penetration. The raw data for these indicators are available from the China Urban Statistical Yearbook. The system of indicators of the level of development of the digital economy is displayed in

Table 1. Indicator system for the level of urban digital economy development.

Primary Indicators	Secondary Indicators	Tertiary Indicators
Level of development of the digital economy	Internet penetration rate	Number of Internet broadband access users per 100 people
	Relevant practitioners	Percentage of employees in computer services and software industry in urban units
	Relevant output	Total telecommunication services per capita
	Cell phone penetration rate	Number of cell phone subscribers per 100 people

.

Based on a comprehensive measurement system of the indicators mentioned above, the initial data in the indicator system are standardized. The entropy value method is then applied to reduce the dimensionality, resulting in the core explanatory variable of this paper: Dige, a comprehensive index of the level of digital economy development.

Energy consumption intensity is the energy consumption per unit of GDP, which is interpreted as follows:

$$\mathrm{E}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y} (EI)=\frac{\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} (\mathrm{t}\mathrm{o}\mathrm{n}\mathrm{s} \mathrm{o}\mathrm{f} \mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d} \mathrm{c}\mathrm{o}\mathrm{a}\mathrm{l})}{\mathrm{G}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{s} \mathrm{r}\mathrm{e}\mathrm{g}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{a}\mathrm{l} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t} (\mathrm{m}\mathrm{i}\mathrm{l}\mathrm{l}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{y}\mathrm{u}\mathrm{a}\mathrm{n})}$$

Technological innovation TI: there are more methods for the measurement of technological innovation, and most of them use the amount of patents granted as a measure. This paper uses the amount of invention patents granted per 10,000 people as a measure of technological innovation [52].

In addition to choosing the above core variables, five control variables are introduced into the model to control the factors affecting the change in energy consumption intensity. Among them, foreign direct investment $lnFDI$ [53] is measured as the logarithm of the ratio of the amount of actual foreign investment utilized by each city to GDP in the year. Science expenditure $lnSS$ [54] is measured as the logarithm of the ratio of regional science expenditure to GDP. Highway Passenger Transportation $lnHPT$ [34] is measured as the logarithm of the total amount of regional highway passenger transportation. Population size $lnPS$ [55] is measured as the logarithm of the year-end resident population. Government intervention GOV [56,57] is expressed as the ratio of local general public budget expenditures to GDP.

In general, the regression results may be affected by the different units of measurement of the data. In order to increase the accuracy of the empirical results as well as to enhance the stability of the data, reduce the fluctuation of the data, and improve the effectiveness of the statistical analysis, the variables are standardized.

4.2.2. Data Source and Description

Due to the accelerated growth of the digital economy and the heightened energy intensity in the Beijing–Tianjin–Hebei region, this paper selects the Beijing–Tianjin–Hebei region as the subject of study. Concurrently, the lack of data and the unavailability of certain data necessitate the inclusion of data from 13 prefecture-level cities and above in the Beijing–Tianjin–Hebei region between 2010 and 2018 in the sample of this paper. This approach allows for the formation of a balanced panel observation of 117 large samples, which reduces the error and enables the drawing of more comprehensive and generalized conclusions. The data used in this paper were sourced from various publications, including the China Statistical Yearbook, China Urban Statistical Yearbook, China Statistical Report on Internet Development, Statistical Annual Reports of Some Prefectural Cities, CNRDS database, and CEIC China Economic Database. To ensure data stability and reduce fluctuations, logarithmic treatment was applied to the accepted variables, mediator variables, and some control variables due to the different units between variables. The specific variable definitions are displayed in

Table 2. Variable definition.

Symbol	Variable	Definition
$lnEI$	Energy consumption intensity	Logarithmic total energy consumption to GDP ratio
Dige	Level of development of the digital economy	indicators on Internet penetration, related workforce, related outputs, and cell phone penetration
TI	Technological innovation	The ratio of Patent Granted to Year-End Resident Population
$lnFDI$	Foreign direct investment	The ratio of real FDI to GDP in logarithmic terms
$lnSS$	Science expenditures	The ratio of science expenditure to GDP taken in logarithms
$lnHPT$	Road passenger traffic	Regional road passenger traffic taken in logarithm
$lnPS$	Population size	Year-end resident population in logarithm
GOV	Government intervention	The ratio of local general public budget expenditure to GDP

.

4.2.3. Data Processing and Model Selection

Prior to conducting empirical analysis, it is necessary to perform a multicollinearity test. The objective of this test is to ascertain whether the explanatory variables are highly correlated with one another. In the event that the correlation between the variables is particularly high, it can result in distorted regression results and a loss of credibility. The variance inflation factor (VIF) is a commonly utilized indicator of the severity of multicollinearity. The VIF represents the ratio between the estimated variance of the variables in the presence of multicollinearity and the estimated variance of the variables in the absence of multicollinearity. A higher VIF value indicates a more severe degree of multicollinearity. In general, a VIF value below 10 is considered to pass the multicollinearity test. The results of the multicollinearity test are presented in

Table 3. Multicollinearity test.

Variable	VIF	1/VIF
Dige	3.56	0.28
$lnFDI$	1.54	0.65
$lnSS$	3.94	0.25
$lnHPT$	4.16	0.24
$lnPS$	2.66	0.38
GOV	1.52	0.66
Mean VIF	2.90

. All variables exhibit VIF values below 5, thus passing the multicollinearity test.

When static regression analysis was selected for this study, individual fixed effects were initially tested, with a p-value of 0.000, which is significant at the 1% level. This result demonstrates that the selection of a fixed-effects model is preferable to a mixed-effects model. Finally, considering the individual and temporal differences that may exist simultaneously in data, we ultimately selected a two-way fixed effects model for regression analysis.

5. An Empirical Test of the Impact of the Digital Economy on Urban Energy Consumption Intensity under Static Panel Modeling

5.1. Descriptive Statistical Analysis

This section begins with a descriptive analysis of each variable.

Table 4. The statistical description of variables.

Variable	Observation	Mean	S.D.	Min	Max
$lnEI$	117	−0.078	0.458	−1.428	0.866
Dige	117	0.103	0.175	0.005	0.854
TI	117	9.230	16.777	0.353	89.621
$lnFDI$	117	−4.042	0.781	−7.127	−2.097
$lnSS$	117	−6.416	0.811	−7.733	−4.265
$lnHPT$	117	8.697	1.078	6.873	11.790
$lnPS$	117	6.531	0.467	5.664	7.227
GOV	117	0.171	0.053	0.074	0.365

displays the results of the analysis. The energy consumption intensity in the Beijing–Tianjin–Hebei region ranges from −1.428 to 0.866, with a standard deviation of 0.458. The digital economy development level ranges from 0.005 to 0.854, while the technological innovation level has a minimum value of 0. The Beijing–Tianjin–Hebei region exhibits significant development differences in terms of energy consumption intensity, digital economy development level, and technological innovation level. The minimum value for these indicators is 353, while the maximum value is 89.621, with a standard deviation of 16.777. The standard deviation is 16.777, indicating significant differences in energy consumption intensity, digital economy development, and technological innovation among cities in the Beijing–Tianjin–Hebei region.

5.2. Base Regression Analysis

Table 5. Benchmark regression results.

Variable		$\mathit{l}\mathit{n}\mathit{E}\mathit{I}$
	(1)	(2)	(3)	(4)	(5)	(6)
Dige	−1.019 ***	−1.024 ***	−0.919 ***	−0.919 ***	−0.941 ***	−0.927 ***
	(−3.00)	(−3.01)	(−2.76)	(−2.76)	(−2.80)	(−2.76)
lnFDI		0.017	0.011	0.006	0.008	0.002
		(0.62)	(0.42)	(0.24)	(0.31)	(0.08)
lnSS			0.127 **	0.138 ***	0.139 ***	0.134 ***
			(2.60)	(2.77)	(2.78)	(2.68)
lnHPT				−0.049	−0.054	−0.060
				(−1.11)	(−1.19)	(−1.31)
lnPS					0.404	0.563
					(0.60)	(0.82)
GOV						0.764
						(1.17)
cons	0.289 ***	0.362 ***	1.249 ***	1.791 ***	−0.996	−2.232
	(6.20)	(2.84)	(3.44)	(2.95)	(−0.21)	(−0.47)
Observations	117	117	117	117	117	117
N	13	13	13	13	13	13
City	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
adj-R2	0.943	0.942	0.945	0.946	0.945	0.945

Note: *** and **, indicate significance at the 1% and 5%, respectively, and values in parentheses indicate t-statistic values.

presents the benchmark regression analysis results, including the stepwise inclusion of control variables. The results demonstrate that the digital economy has a significant negative impact on energy consumption intensity at the 1% level, regardless of the gradual addition of control variables. The regression results in Column 6 show that the impact of the digital economy on the energy consumption intensity decreases by 0.9 percentage points for every percentage point increase in the digital economy. This finding suggests that the digital economy has a significant inhibitory effect on the energy consumption intensity. The digital economy has facilitated the high efficiency, intensification, and refinement of China’s energy supply. This has resulted in a high degree of matching between energy supply and production processes, enabling the efficient use of energy. Additionally, the digitalization of energy production and transmission has minimized energy loss in the supply chain and reduced energy consumption in all segments.

5.3. Endogeneity Test

To address the issue of endogeneity caused by omitted variables and bidirectional causality, this paper employs the historical data of the post and telecommunications of each city in 1984 as the instrumental variable for the core explanatory variable, digital economic development level. The two-stage least squares (2SLS) method is used for estimation and analysis, following the approach of Huang et al. [58]. The reason for selecting the post and telecommunications historical data as an instrumental variable is that they satisfy the relevance requirement. The historical post and telecommunications number is the continuation of traditional information and communication technology in terms of modern digital economic development. The post and telecommunications historical data will affect the contemporary level of digital economic development through the continuation and inheritance of the technology. However, the number of post and telecommunications offices, which make up the historical communication infrastructure, has significantly decreased in importance compared to the contemporary era. Therefore, the number of historical post offices does not have a direct and significant impact on the current energy consumption intensity, satisfying the exclusivity condition. However, the historical data on post and telecommunications from 1984 are cross-sectional and cannot be used to estimate and analyze panel data. Therefore, a time-varying variable needs to be introduced. This paper draws on the methodology of Zhao [43] and constructs an interaction term by multiplying the number of Internet users per 100 people by it. This serves as an instrumental variable (IV) for the development of the digital economy in this paper.

Table 6. Results of endogeneity test.

	Instrumental Variable Approach
Variable	Dige	lnEI
	2SLS Phase I	2SLS Phase Ⅱ
Dige		−1.0778 ***
		(−4.92)
IV	0.0004 ***
	(5.29)
Control variable	Yes	Yes
City	Yes	Yes
Year	Yes	Yes
Kleibergen-Paap rk LM		6.073 **
		[0.0137]
Kleibergen-Paap rk Wald F		27.940
		{16.38}
Observations	117	117
N	13	13

Note: *** and ** indicate significance at the 1% and 5% respectively; ( ) values are z-values, [ ] values are p-values, and { } values are critical values at the 10% level of the Stock–Yogo weak identification test.

presents the results of the endogeneity test conducted through two-stage least squares. The Kleibergen-Paap rk Wald F statistic of 27.940 exceeds the critical value of 16.38 at the 10% significance level of the Stock–Yogo weak identification test, indicating that the selected instrumental variables are reasonable. The regression results of instrumental variables show that the regression coefficient of the digital economy development is significantly negative at a 1% statistical level. This indicates that the development of the digital economy can significantly reduce urban energy consumption intensity.

5.4. Robustness Test

Based on the previous analysis, it is evident that the digital economy can decrease regional energy consumption intensity. To ensure the reliability and robustness of the analysis, a robustness test is conducted to examine the impact of the digital economy on energy consumption intensity.

Table 7. Robustness test results.

	(1)	(2)	(3)	(4)
	Replacement of Core Explanatory Variable	One-Period-Lagged Explanatory Variable	Shrinkage Treatment	Adding Control Variable
Dige	−0.101 *	−0.857 **	−0.910 ***	−0.980 ***
	(−1.86)	(−2.32)	(−5.24)	(−2.89)
cons	−2.279	−3.358	−2.553	−1.035
	(−0.06)	(−0.67)	(−0.60)	(−0.21)
Control variable	Yes	Yes	Yes	Yes
City	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Observations	117	117	117	117
adj-R2	0.943	0.949	0.958	0.946

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and values in parentheses indicate t-statistic values.

presents the robustness test results.

5.4.1. Replacement of Core Explanatory Variables

The regression results are based on the entropy method used to measure the digital economy development index for analysis. In the robustness test section of this paper, the principal component analysis method is used to recalculate the level of digital economy development in Beijing–Tianjin–Hebei cities, and further regression tests are conducted. Table 7, Column 1, shows that the principal component analysis indicates a significant impact of the digital economy on the energy consumption intensity at the 10% level. The regression coefficient of −0.101 suggests that the digital economy has a significant inhibitory effect on the impact of energy consumption intensity.

5.4.2. One-Period Lagged Explanatory Variables

To prevent reverse causality among the core variables, this paper adds the core explanatory variables to the regression analysis model after lagging them by one period. Table 7, Column 2, shows a significant negative impact of the digital economy on energy consumption intensity with a regression coefficient of −0.857 at the 5% level. This conclusion aligns with the benchmark regression.

5.4.3. 1% Two-Way Shrinkage of the Full Sample Size

To avoid the influence of extreme values on the data regression results, this paper carries out the 1% two-way shrinkage treatment for the whole sample indicators and continues the regression analysis. Table 7, Column 3, indicates that the regression coefficient of the digital economy on energy consumption intensity is −0.910, which is significant at the 1% level. This conclusion is consistent with the previous benchmark regression results, further supporting their robustness.

5.4.4. Adding Control Variables

A review of existing studies has demonstrated that the level of openness to the outside world can have a significant impact on energy consumption [59]. Trade activities may influence the industrial structure and resource allocation of a country or region. The development of specific industries may necessitate a greater consumption of energy, which, in turn, affects the overall energy consumption level. In this paper, the ratio of total regional import and export trade to GDP is taken in logarithm to indicate the degree of regional openness to the outside world. Furthermore, a regression analysis of the relationship between the digital economy and the intensity of energy consumption is continued by adding control variables. Table 7, Column 4, demonstrates that by incorporating the control variable representing the level of openness to the outside world, the regression coefficient of the effect of the digital economy on the intensity of energy consumption is −0.980 and is statistically significant at the 1% level. This further corroborates the reliability of the benchmark regression results.

5.5. Analysis of Impact Mechanisms

Based on the analysis above, it is evident that the digital economy primarily reduces energy consumption intensity by promoting technological innovation. Therefore, a mechanism analysis is necessary for further verification. Currently, the step-by-step method is commonly used for testing mediation effects. However, the traditional three-step method may produce endogenous bias, as it is difficult to find an exogenous mediating variable, which leads to unreliable results. This paper draws on the improved method proposed by Jiang [60] to verify the relationship between the digital economy, mediating variables, and energy consumption intensity. The results presented in

Table 8. Mediating effect test results.

Variable	Model 1	Model 2
	TI	TI
Dige	10.301 ***	9.999 ***
	(16.76)	(18.75)
cons	−0.671 ***	−20.959 ***
	(−7.97)	(−2.76)
Control variable	No	Yes
City	Yes	Yes
Year	Yes	Yes
Observations	117	117
adj-R2	0.960	0.971

Note: *** indicates significance at the 1% levels, and values in parentheses indicate t-statistic values.

indicate that the regression coefficient of the digital economy’s effect on technological innovation is significantly positive at the 1% level, regardless of whether the control variables are included in the analysis. This suggests that the digital economy promotes technological innovation. Additionally, as discussed in the previous section, the digital economy significantly reduces the energy consumption intensity, and the core explanatory variables have a significant inhibitory effect on the explanatory variables. Furthermore, Tao [61], Sun [62], and Jules-Daniel Wurlod et al. [63] have consistently shown that technological innovation can significantly reduce energy consumption intensity. The digital economy can indirectly reduce energy consumption intensity through technological innovation

6. Analysis of the Spatial Effects of Digital Economy Development on Energy Consumption Intensity under Technological Spillovers

6.1. Spatial Correlation Test

To begin spatial econometric analysis, it is essential to test the spatial autocorrelation of the research variables using Moran’s I index.

Table 9. Global Moran’s Index of EI and TI.

Year	$\mathit{l}\mathit{n}\mathit{E}\mathit{I}$		TI
	Moran’s I	Z	Moran’s I	Z
2010	0.032	0.625	0.387 ***	4.301
2011	0.132	1.206	0.399 ***	4.320
2012	0.071	0.852	0.453 ***	4.394
2013	0.264 **	2.021	0.477 ***	4.404
2014	0.328 **	2.486	0.431 ***	4.418
2015	0.323 **	2.461	0.471 ***	4.328
2016	0.301 **	2.276	0.460 ***	4.313
2017	0.342 **	2.524	0.455 ***	4.342
2018	0.389 ***	2.910	0.500 ***	4.398

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

shows that from 2010 to 2018, Moran’s I index indicates a positive correlation between energy intensity and technological innovation, passing the 5% significance test and reaching the 1% significance level in some years. This suggests that the two variables are spatially clustered and have a significant positive correlation. Therefore, a spatial econometric model is suitable for analysis. While Moran’s I index of energy intensity in the years 2010–2012 is not statistically significant, it does not necessarily prove the absence of spatial correlation [64]. Therefore, further empirical analysis using a spatial econometric model is necessary.

6.2. Empirical Analysis of Spatial Effects

To choose the right spatial measurement model for spatial analysis, this study followed Elhorst’s [65] approach. Firstly, a simplification test of the spatial measurement model (LM-test, Wald test) was conducted. The analysis concluded that the spatial model can be either a spatial lag model or a spatial error model. Secondly, the Likelihood Ratio (LR) test was utilized to determine whether the SDM will degenerate to SAR and SEM models. The LR test indicates that the spatial Durbin model does not degenerate into the spatial error model or the spatial lag model. This suggests that urban technological innovation has a double spillover effect on both independent and dependent variables. Therefore, the spatial Durbin model is the optimal choice for analysis. Finally, the Hausman test was conducted to determine whether to use a random effect or fixed effect model. The results indicated that a time–space double fixed-effect model was more appropriate for analysis. In summary, this paper employs the spatial Durbin model (SDM) with time–space double fixed-effects for spatial econometric analysis.

This paper presents the results of an analysis of three spatial models: SAR, SEM, and SDM. The objective is to facilitate analysis and comparison. In order to enhance the robustness of the spatial model, this study constructed the economic distance weight matrix W1 and the geographic distance weight matrix W2, respectively. The results of the analysis of the economic distance weight matrix W1 indicate that the log-likelihood values of the three models in

Table 10. Spatial econometric model regression results and effects decomposition.

Variable	SAR		SEM		SDM
	(1)	(2)	(3)	(4)	(5)	(6)
	W1	W2	W1	W2	W1	W2
ρ	−0.636 ***	−0.547 **			−0.643 ***	−1.016 ***
	(−4.44)	(−2.30)			(−7.16)	(−6.80)
λ			−0.687 ***	−0.942 ***
			(−5.23)	(−2.87)
TI	−0.117 ***	−0.084 ***	−0.090 ***	−0.098 ***	−0.154 ***	−0.100 **
	(−4.98)	(−4.13)	(−5.67)	(−3.48)	(−4.30)	(−2.07)
W × TI					0.063	−0.055
					(1.24)	(−0.46)
Log-likelihood	124.2	113.1	124.2	115.4	132.3	125.6
Direct effect	−0.134 ***	−0.086 ***			−0.187 ***	−0.103 **
	(−4.10)	(−3.74)			(−3.55)	(−2.12)
Indirect effect	0.063 ***	0.033 *			0.130 **	0.029
	(2.65)	(1.90)			(2.17)	(0.45)
Total effect	−0.071 ***	−0.054 ***			−0.057 ***	−0.074
	(−5.33)	(−5.64)			(−3.27)	(−0.91)
Fixed Effect	Yes	Yes	Yes	Yes	Yes	Yes
Observations	117	117	117	117	117	117
R2	0.147	0.405	0.258	0.389	0.010	0.010
Variance	0.006 ***	0.008 ***	0.006 ***	0.008 ***	0.005 **	0.000 ***
	(4.28)	(3.72)	(4.37)	(7.14)	(4.56)	(3.73)

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and values in parentheses indicate t-statistic values.

are 124.2, 124.2, and 132.3, which demonstrates that the SDM model exhibits a superior fitting effect and higher credibility than the SAR and SEM models. For SDM, the results calculated using the economic distance weight matrix W1 and the geographic distance weight matrix W2 are essentially identical. As an illustration, Column 5 of the economic distance weight matrix W1 reveals that the coefficient representing technological innovation in the SDM model is markedly negative. This suggests that technological innovation in the cities of the Beijing–Tianjin–Hebei region may result in a reduction in the intensity of energy consumption in the region. Furthermore, the spatial autoregressive coefficient of technological innovation is significantly negative, indicating that the development of the digital economy in the Beijing–Tianjin–Hebei region has a significant spatial spillover effect.

6.3. Decomposition of Spatial Effects

When a spatial effect is present, the regression coefficients of the spatial Durbin model cannot be fully explained. This is because the model takes into account both the spatial lag operator of the core explanatory variables and the explanatory variables themselves. As a result, changes in the core explanatory variables not only affect the explanatory variables in the region but also impact other regions. Therefore, the regression coefficients of the spatial interaction term cannot be used to discuss the marginal impact of technological innovation on energy consumption intensity [66]. LeSage et al. [67] proposed a partial differentiation method based on the scope and object of the spatial effect. This method decomposes the effect of the independent variable on the dependent variable into direct, indirect (spatial spillover), and total effects for spatial interpretation. Column 5 of Table 10 shows the regression effects decomposition for the spatial Durbin model. The results indicate that the direct effect of technological innovation on energy consumption intensity is negative and significant at the 1% level. Meanwhile, the indirect effect is positive and significant at the 5% level, suggesting that the technology spillover effect will increase energy consumption intensity in neighboring areas. The digital economy has a significant inhibitory effect on urban energy consumption intensity. Additionally, the development of the digital economy in a region has a significant inhibitory effect on energy consumption intensity in that region, but it has a positive spillover effect on energy consumption intensity in neighboring cities. The adoption and application of technological innovations in a region may lead to a loss of resources and increased competition, including the outflow of people and capital, as well as the shifting of supply chains. This can be attributed to the fact that technological innovations can disrupt traditional industries and create new ones, leading to changes in the economic landscape. This competition for resources may result in neighboring regions lacking the necessary support, limiting their development in terms of energy consumption intensity improvement. However, the application of technological innovations in one region may give it a competitive advantage in terms of technology and energy efficiency. Uneven technological development can cause neighboring regions to fall behind in terms of technology and energy efficiency, resulting in technological differences and disadvantages that may limit and impact neighboring regions’ ability to improve energy consumption intensity.

6.4. Further Tests: Spatial Heterogeneity Analysis

The benchmark regression, robustness test, and endogeneity test all indicate that the digital economy can significantly reduce regional energy consumption intensity, promoting local environmental progress and sustainable development. However, there are significant disparities in economic development among cities in the Beijing–Tianjin–Hebei region. Regions with varying levels of economic development may have distinct impacts on regional energy consumption intensity, and their technological spillovers may differ accordingly. In order to simultaneously analyze the heterogeneity of the impacts and spatial effects of technological advances brought about by the digital economy in different regions on energy consumption intensity, this paper refers to the methodology of Jiang et al. [68] and compares the economic development levels of 13 cities in the Beijing–Tianjin–Hebei region. The study categorizes Beijing, Tianjin, Tangshan, Shijiazhuang, Cangzhou, and Handan as cities with a higher level of economic development, while the remaining seven prefecture-level cities are categorized as having lower levels of economic development. The study then analyzes the heterogeneity of the lower-level cities from a spatial perspective using the SDM model and the partial differential method.

Table 11. Results of the spatial heterogeneity test.

	Economically Developed Cities	Non-Economically Developed Cities
Direct effect	−0.064	−0.038
	(−0.94)	(−0.13)
Indirect effect	−0.128 ***	0.739
	(−2.33)	(0.84)
Total effect	−0.174 ***	0.701
	(−6.16)	(0.66)
Fixed Effect	Yes	Yes
Observations	54	63
adj-R2	0.315	0.236

Note: *** indicates significance at the 1% levels, and values in parentheses indicate t-statistic values.

shows the test results, indicating a negative direct effect with poor significance levels for both economically developed and non-economically developed cities. However, for economically developed cities, the indirect effect is significantly negative and passes the significance test at the 1% level. This suggests that the inhibition of energy consumption intensity is mainly achieved through technological spillover, which significantly reduces the energy consumption of neighboring areas. The data indicate that economically powerful cities primarily reduce energy consumption intensity through technological spillover. This spillover effect can significantly decrease energy consumption in neighboring areas and positions these cities as key providers of green technology. However, non-economically developed cities are currently the main consumers of energy. The development of the digital economy will exacerbate energy consumption in neighboring areas, as these cities are the main subjects of technological spillovers that increase regional energy consumption.

7. Conclusions and Policy Implications

7.1. Conclusions and Discussion

This article examines the impact of the digital economy on energy consumption intensity in the Beijing–Tianjin–Hebei region. Using panel data from 117 samples across 13 cities from 2010 to 2018, the paper constructs an index of digital economy development based on the entropy method. Various empirical testing tools are applied, including the panel bidirectional fixed-effect model, instrumental variables, mediation effect model, and spatial Durbin model. The empirical study tests the influence, degree of effect, mechanism, spatial spillover effect, and role of regional heterogeneity of the digital economy intensity on regional energy consumption. The main findings are as follows.

Firstly, it can be observed that the digital economy has a significant impact on reducing regional energy consumption intensity, which is consistent with the findings of Guo et al. [36] and Huang et al. [34]. The direct effect of the digital economy on energy intensity is negative, indicating that the digital economy can significantly reduce energy intensity in the Beijing–Tianjin–Hebei region. Second, further analysis of the mechanisms through which the digital economy affects energy consumption intensity reveals that one of the primary channels is technological empowerment, which is consistent with the findings of Huang et al. [34] and Longo et al. [18]. The digital economy can significantly promote technological innovation, thereby reducing regional energy intensity. On the one hand, digital technologies can enhance energy efficiency [69,70], facilitate the transition to a low-carbon economy [29], and support the development of a green economy [71,72,73], thereby reducing overall energy consumption. On the other hand, the digital economy facilitates the optimization and upgrading of the industrial structure through technological innovation [74,75] while also playing an optimizing role in resource allocation [76] and promoting the growth of the regional economy. This ultimately results in decreased energy intensity. In any case, technological innovation represents the primary mechanism through which the digital economy affects energy intensity. Third, the analysis with a spatial econometric model indicates that the digital economy contributes to an increase in energy consumption intensity in neighboring areas through a technology spillover effect. This finding is consistent with the research results of Chen et al. [77]. However, it is notable that regional heterogeneity exists in this regard. For instance, in economically developed cities, the digital economy can have a significant dampening effect on the intensity of energy consumption in neighboring regions through technological spillovers. In contrast, for economically underdeveloped cities, technological spillovers have the opposite effect, increasing energy consumption in neighboring regions. Due to the significant disparities in economic development between different cities in Beijing–Tianjin–Hebei, similar to those observed in the eastern and western regions of China, the digital economy has the potential to facilitate regional energy conservation and emission reduction in the developed eastern regions [30]. However, in the underdeveloped regions in the central and western parts of the country, the energy conservation effect of the digital economy is even more limited [17], and the role of energy consumption is even more pronounced [1].

7.2. Policy Implications

In light of the aforementioned research findings, this paper puts forth the following recommendations for potential countermeasures:

(1)

The study reveals that there is significant potential for the development of the digital economy in the Beijing–Tianjin–Hebei region. To achieve this, it is necessary to enhance the digital ecosystem; invest in high-speed, stable, and secure network infrastructure; and promote the development of 5G technology. This will ensure the perfection of digital infrastructure and accelerate the transition from traditional economic growth. Simultaneously, it is necessary to accelerate the shift from the conventional economic growth path, meet the demands of modernization and high-quality development, spearhead the productivity surge, unleash the potential of data components, and strengthen the industrial foundation for cultivating new, high-quality productivity.

(2)

The positive role of the technology spillover effect of the digital economy can be utilized to reduce energy consumption. Additionally, it is important to increase the connection between the core digital economy region and neighboring regions and improve the spread of green technology to benefit a wider range of regions. Encouraging cooperation between different industries and regions, promoting a cross-border integration of the digital economy, and fostering innovation and resource sharing are also crucial. The aim is to promote a strong synergy and resonance in the field of digital economy in the Beijing, Tianjin, and Hebei region.

(3)

Technological innovation is a key driver for the digital economy to reduce energy consumption intensity. Therefore, to promote the development of a digitized, networked, and intelligent digital economy, it is crucial to focus on scientific and technological innovation, enhance the construction of digital infrastructure, and increase support for cutting-edge technologies such as artificial intelligence, big data, and the Internet of Things. It is also important to encourage collaboration between enterprises and research institutes to promote the industrialization of technological innovations and establish a network of research institutes. To promote the cross-application of technologies, establish an intellectual property protection system, and encourage open innovation, it is necessary to industrialize technological innovations.

(4)

The aim is to develop the inter-regional digital economy in a synergistic manner, taking into account regional differences in the level of digital economy development. This will allow for the positive effects of technological spillover from economically strong cities to be utilized in reducing energy consumption. Additionally, the development of the digital economy in non-economically strong cities should be prioritized, with zoning governance and rational planning based on local conditions for the progress of inter-regional digital economic infrastructure. The cities surrounding Beijing should leverage their digital ecosystem advantages and increase policy support for underdeveloped areas. This will allow for the “digital dividend” to provide a strong impetus for balanced and coordinated regional development, ultimately narrowing inter-regional development gaps.

7.3. Limitations and Future Prospects

Although this study employs spatial econometric modeling to investigate the impact of the digital economy on energy intensity in depth, it still has some limitations. Energy intensity is defined as the ratio of total regional energy consumption to gross regional product (GRP), which serves to reflect the status of regional economic development. Nevertheless, this study examines the relationship between the digital economy and energy intensity from the perspective of technological innovation, without focusing on the impact and mechanism of regional economic development on energy intensity. Moreover, the factors influencing energy intensity encompass a multitude of dimensions, including political, economic, social, and other aspects. Consequently, the potential for certain factors to be overlooked cannot be discounted. Concurrently, this study delineates the boundaries of its investigation as the Beijing–Tianjin–Hebei region, with a particular focus on the impact of the digital economy on energy intensity within the aforementioned region. However, it is important to note that the study lacks consideration of other regions. Therefore, in the future, it would be beneficial to consider the factors affecting regional energy intensity in a more comprehensive manner, to expand the scope of the study, and to explore the mechanism of digitalization affecting energy intensity in greater depth. This would enable the collection of more comprehensive and accurate evidence to support the formulation of policies.