The Digital Economy and the Energy “Internal Circulation”: Evidence from China’s Interprovincial Energy Trade

Abstract

In the context of the increasing instability and uncertainty of the world economy, China’s energy security is threatened. It is important to study how to build the energy “internal circulation” for China’s sustainable development. This paper measures the level of interprovincial energy trade and digital economy development in China through the gravity model estimation method and the entropy value method and examines the impact and mechanism of action of the digital economy on energy circulation. The findings are as follows: (1) Accelerating the development of the digital economy can significantly improve the level of energy “internal circulation” in China, which remains robust after endogeneity treatment and a series of robustness tests. (2) Digital economy has promoted the development of energy “internal circulation” from the production, distribution, exchange, and consumption of energy. (3) The higher the level of digital economy development, the more energy-rich regions and the central regions with a high degree of coupling and coordination between digital and energy systems, the more the digital economy contributes to the energy “internal circulation”. Therefore, it is proposed to accelerate the development of the digital economy, to pay attention to the in-depth integration of digital and energy, and to focus on the coordinated development of regions as effective ways to jointly promote the energy “internal circulation”.

1. Introduction and Review of the Literature

With the continued spread of the COVID-19 epidemic, the operation of the global industrial chain is unbalanced, the manufacturing industry is shrinking, and protectionism and trade barriers are rising. China faces an increasingly complex external environment. Therefore, on 14 May 2020, China first proposed to “build a new development pattern featuring domestic and international double circulation and mutual promotion”. Since then, the Fifth Plenary Session of the 19th CPC Central Committee has included “accelerating the building of a new development pattern” into the Proposal of the CPC Central Committee on formulating the 14th Five-Year Plan for National Economic and Social Development and the Long-Term Goal of 2035. Smoothing the “internal circulation” is to rely on the strong domestic market to connect all links of production, distribution, exchange, and consumption, to break the industrial monopoly and local protection, and to form a virtuous circulation of the national economy [1]. As an important factor of production and living materials, how can energy participate in the “internal circulation”? In April 2022, the State Council issued the Opinions on Accelerating the Construction of a Unified National Market, which pointed out the construction of a unified national energy market to support economic development and to ensure energy security. The energy security strategy based on the energy “internal circulation” has become the basic support and inherent requirement for promoting China’s economic development [2]. Most scholars believe that the segmentation of the energy market significantly inhibits the free flow and efficient agglomeration of energy, leading to information asymmetry and unreasonable interregional resource allocation [3]. In addition, some scholars have found that energy market segmentation will inhibit energy efficiency, green development, and industrial upgrading [4,5,6]. Therefore, as early as 2015, the EU officially launched the process of building an energy union to strengthen the interconnection and interaction of the internal energy market and to promote intra-energy circulation.

So, how do we understand the core implications of the energy “internal circulation”? The energy “internal circulation” is a virtuous circulation that takes the “internal circulation” as the carrier and support, connects the four links of production, distribution, circulation, and consumption, and follows each other in time and space with mutual synergy and orderly transformation. At the same time, the energy “internal circulation” can form a dynamic balance of demand pulling supply and supply creating demand, enhancing the flexibility and resilience of the energy system, and improving the ability to cope with short-term shocks. Its goal is to build an energy market with unified rules of market infrastructure, high standards connectivity of market facilities, fair and unified supervision, and further regulation of inappropriate market competition and market intervention, effectively smoothing the circulation barriers of all links and building a national unified energy market.

However, at the present stage, China suffers from market segmentation at the regional level of energy [7] and has been relying on interregional transmission to ensure a stable supply. Meanwhile, state-owned capital monopoly and government energy pricing will lead to sectoral segmentation of the energy market [8]. In addition, the popularization of renewable energy has led to a shift from centralized to distributed energy development [9] and the intelligent linkage and real-time management of energy flows have become imperative issues to be addressed. In this regard, scholar Lin Boqiang et al. pointed out that China’s main resources and energies are distributed in the inland areas, while the manufacturing industries are concentrated in coastal areas. This reverse distribution pattern of energy resources and demand has led to the formation of a trans-regional fossil energy transportation system and emphasizes that, as China enters the stage of high-quality development, the energy system with fossil energy supply and transportation as the core needs to be reshaped [10]. Huang Ying et al. analyzed that China is still in the period of planning and market coexistence, that there are still interprovincial barriers in the energy sector to maintain their interests and plan means to balance the interests of all parties, and that market prices have not been formed in the true sense yet. Moreover, market price failures will not only lead to the mismatch of energy products and reduce the efficiency of energy use but will also distort the regional distribution and investment of energy industries, causing adverse effects on the long-term development of the energy system [11]. Further, scholar Hong Yong et al. introduced technology innovation to explore the impact of market segmentation and technology innovation on energy efficiency. The research shows that technology innovation can significantly improve energy efficiency, while market segmentation cannot discourage energy efficiency but may also weaken the promoting effect of technology innovation on energy efficiency by inhibiting technology innovation [12]. Duan Wei et al. further concluded that the wide application of emerging energy technologies is an important means to reshape the energy system and to achieve green transformation [13].

With the rapid development of the new technological revolution, the current digital revolution represented by digital and digital technology is becoming a key force to smooth energy circulation and to reshape the energy system. Scholar Zhang Lin et al. found that the digital economy development can have a positive impact on the technology market development by enhancing the marketization level and the regional innovation capacity [14]. Further, Li Tao et al. examined the mechanism and influence channels of the digital economy on energy efficiency. The results showed that the digital economy has a significant positive promoting effect on regional total factor energy efficiency, that market trade is an important intermediary transmission way, and that the promoting effect of the digital economy on energy efficiency shows a gradually decreasing trend in the east–west–central region. It also pointed out that the government should vigorously promote the deep integration and balanced development of the digital economy and energy industry in all regions [15]. Therefore, in the context of the digital economy, China is opening a “new stage” in the “great changes not seen in a century”. It is important to enhance technological innovation capacity, increase disposable income, and reduce energy distribution costs to accelerate the release of the dividends of the development of the digital economy, ensure efficient and unimpeded energy “internal circulation”, and accelerate the establishment of a unified energy market.

The possible marginal contributions of this paper are mainly as follows: (1) The paper innovatively starts from the perspective of “internal circulation” and takes “energy elements” as the practical starting point to explore the impact of the digital economy development on the energy “internal circulation”. It not only expands the research horizon of the new development pattern of “double circulation” but also provides an important basis and policy reference for the construction of the big energy market. (2) In combination with the substantial connotation of China’s “internal circulation” and the special reality of China’s energy sector, this paper analyzes the theoretical mechanism from four aspects—production, distribution, exchange, and consumption—and conducts empirical tests to make the research conclusions more scientific and instructive.

The rest of this paper consists of the following sections. The theoretical mechanisms and research hypotheses are detailed in the next section. Section 3 introduces the research design through the model introduction and variable description. Section 4 details the results of the empirical tests, including the benchmark test, robustness test, mechanism test, and heterogeneity test. Finally, the fifth part gives the research results and policy recommendations.

2. Theoretical Analysis of the Energy “Internal Circulation” through the Digital Economy

2.1. Basic Hypothesis

Digital economy, characterized by high technology and high energy consumption [16,17], can be used to break through the blockages in energy production, distribution, exchange, and consumption. High energy consumption can effectively release domestic energy demand, which is important to smooth out the energy “internal circulation” [18]. First, the digital economy has increased the interaction between distributed end-users such as smart buildings, home energy storage and electric vehicles, and energy production entities, breaking through the energy circulation blockage caused by information asymmetry and forming a two-way interconnection circulation between supply and demand [19]. Secondly, the application of the digital economy has improved the interaction of energy production management. Collecting energy data in the energy production and the supply and distribution chain and sharing it among energy sectors can curb oligopolistic behavior, break energy market segmentation caused by the local protection of energy, and promote the energy “internal circulation” [20]. Finally, the energy-intensive nature of the digital economy urgently requires clean and sustainable energy. Therefore, it increases the integration of renewable energy resources and fossil energy, which is important to smooth the “internal circulation” between energy sources [21]. Based on this, the following hypothesis is proposed:

Hypothesis 1 (H1).

If the energy sector is further digitalized, then the development of the energy “internal circulation” will be enhanced.

2.2. Hypothesis of Mechanism

The development of the digital economy, typified by data and digital technologies, can effectively improve all aspects of energy production, distribution, exchange, and consumption, shortening the time between the different stages of conversion and thus building a “great domestic circulation” of energy.

Digital economy development can improve energy production efficiency and promote energy “internal circulation”. Digital technology facilitates the integration of data and energy systems, thus improving the operational efficiency of the energy production chain. It also makes it possible to use a wide range of renewable energy sources and increase energy production [22,23,24]. Furthermore, the effective combination of big data and digital technology can accurately predict energy demand and thus customize energy production to avoid underproduction or overproduction and achieve a dynamic balance between energy supply and demand [25].

Hypothesis 2 (H2).

The digital economy can accelerate the energy “internal circulation” by empowering energy production.

The development of the digital economy can optimize the income distribution of workers and accelerate the energy “internal circulation”. In the flow of the energy economy, it is often difficult to accurately measure the contribution of labor due to the high cost of supervision. However, digital empowerment in the distribution chain makes it possible to intelligently measure the contribution of workers in the labor process, thereby increasing employee motivation to produce [26]. In addition, the development of the digital economy can increase the disposable income of residents, narrow the income distribution gap, and enhance people’s pursuit of a better life [27,28,29]. Therefore, residents spend more in energy-intensive areas such as transportation and communication, health care, and housing [30], thereby increasing indirect energy consumption.

Hypothesis 3 (H3).

The digital economy can optimize the distribution of labor income to promote the energy “internal circulation”.

Digital economy development and digital technology can improve energy transportation efficiency and can accelerate the energy “internal circulation”. Digital technology can integrate information on energy demand locations, energy demand quantities, traffic conditions, weather conditions, etc., in order to accelerate the rapid aggregation and matching of energy elements [31]. Moreover, by relying on big data and digital technology, the energy system can collaborate with energy transportation channels such as roads, railroads, sea transport, pipelines, and power grids to build an efficient and smooth energy transportation network, to reduce the time consumption of the transportation process, and to accelerate the energy “internal circulation” [32].

Hypothesis 4 (H4).

The digital economy can improve energy transportation efficiency to drive the energy “internal circulation”.

The digital economy can be developed to stimulate energy consumption and to pull the energy “internal circulation” from the demand side. The growth of the digital economy has accelerated the spread of digital infrastructure, which directly strengthens the reliance on energy demand [33]. In addition, the development of the digital economy has expanded the scale of economic activity and production, increasing the demand for energy for social production [34,35]. Furthermore, the development of the integration of the digital economy and energy systems in China is focused on improving the efficiency of energy access, leading to a rebound effect of energy, stimulating the consumption of energy consumers, and increasing energy consumption [36].

Hypothesis 5 (H5).

The digital economy can increase energy consumption and accelerate the energy “internal circulation”.

Therefore, the high-tech and energy-intensive attributes of the digital economy will directly contribute to the energy “internal circulation” and will also accelerate the development of the energy “internal circulation” by enabling energy production and by optimizing the labor income distribution, energy transportation efficiency, and energy consumption.

3. Theoretical Model

3.1. Model Construction

The direct transmission mechanism is modeled to test the hypotheses.

$${\mathit{EIC}}_{\mathit{ijt}}={\beta }_0+{\beta }_1{\mathit{Digital}}_{\mathit{ijt}}+{\beta }_2{\mathit{X}}_{\mathit{ijt}}+{\mu }_i+{\nu }_t+{\epsilon }_{\mathit{ijt}}$$

(1)

In Equation (1), ${\mathit{EIC}}_{\mathit{ijt}}$ is the logarithm of the number of energy trade in province i and province j in year t, indicating the degree of development of the energy “internal circulation” in each province; ${\mathit{Digital}}_{\mathit{ijt}}$ is the level of digital economic development in province i and province j in year t; ${X\prime }_{\mathit{ijt}}$ is the logarithm of the product of control variables in province i and province j in year t, which are the value added of secondary industry (SI), residential consumption (RC), fuel retail prices (RU), logistics infrastructure transportation network density (TD), and geographical distance between provinces (Lndist). In addition, ${\mu }_i$ is a region-fixed effect, controlling for regional non-time-varying heterogeneity. ${\upsilon }_t$ is a time-fixed effect, with the variation generated for all regions over the sample time absorbed by this term. ${\epsilon }_{\mathit{ijt}}$ is the error term.

To further investigate whether the digital economy affects the energy “internal circulation” by linking energy production, distribution, exchange, and consumption, a mediating effects model is built to test the mechanism.

$$M_{\mathit{ijt}}={\lambda }_0+{\lambda }_1{\mathit{Digital}}_{\mathit{ijt}}+{\lambda }_2{X}_{\mathit{ijt}}+{\mu }_i+{\nu }_t+{\epsilon }_{\mathit{ijt}}$$

(2)

$${\mathit{EIC}}_{\mathit{ijt}}={\alpha }_0+{\alpha }_1{\mathit{Digital}}_{\mathit{ijt}}+{\alpha }_2{M}_{\mathit{ijt}}+{\alpha }_3{X}_{\mathit{ijt}}+{\mu }_{\mathit{ij}}+{\nu }_t+{\epsilon }_{\mathit{ijt}}$$

(3)

where $M_{\mathit{ijt}}$ is the mechanism variable representing energy production, distribution, exchange, and consumption in terms of energy production efficiency, energy sector workers’ wages, energy transportation efficiency, and energy consumption, respectively.

3.2. Variable Explanation and Description

3.2.1. Explanatory Variable: Energy “Internal Circulation” (EIC)

The volume of interprovincial energy trade is used to represent the degree of development of energy “internal circulation” in China [37]. The volume of interprovincial energy trade is measured as follows: First, China’s coal consumption accounts for 67% of the overall energy consumption [38]. To characterize the interprovincial energy flow more comprehensively, the interprovincial energy flow matrix of rail transportation is obtained by dividing the rail coal transport volume by the coal consumption ratio. Secondly, since energy transportation includes railroads, roads, pipelines, and water transportation, the friction coefficient of energy flow is calculated by referring to Yu Yang’s method of estimating interprovincial trade flows [39], using the ratio of rail freight to total railroads, roads, pipelines, and water transportation as the amplification factor. The formula is as follows.

$$F^{\mathit{rs}}=\frac{H^{\mathit{rs}}}{\frac{H^{\mathit{ro}}{H}^{\mathit{os}}}{H^{\mathit{oo}}}}$$

(4)

$F^{\mathit{rs}}$ is the friction factor of energy in province r and province s. $H^{\mathit{rs}}$ is the amount of energy transported from province r to province s in million tons of standard coal. $H^{\mathit{ro}}$ is the total energy output from province r, $H^{\mathit{os}}$ is the total energy input from province s, and $H^{\mathit{oo}}$ is the total energy output from all provinces.

Finally, the interprovincial energy trade volume is calculated according to the gravitational model transport volume estimation method [40].

$$T^{\mathit{rs}}=\frac{Y^r{D}^s}{{\displaystyle \sum Y^r}}{F}^{\mathit{rs}}$$

(5)

$T^{\mathit{rs}}$ is the trade volume of energy from r to s province, $Y^{\mathit{rs}}$ is the total energy production in r province, and $D^s$ is the total energy demand in s province.

The data come from the China Transportation Yearbook and the China Energy Statistical Yearbook.

3.2.2. Core Explanatory Variable: Digital Economy’s Development (Digital)

The digital economy is a new economic form with information as the core production factor, supported by information technology, with modern information networks as the carrier and with information technology to provide products or services [41]. Referring to the research on digital economy measurement, this paper divides the digital economy into three dimensions: digital users, information and communication platforms, and digital transactions. In addition, it selects 13 indicators to measure the development level of the digital economy [42,43]. This is shown in

Table 1. Digital economy’s development and the weights of indicators.

Categories	Variables	Weight	Index Attribute
Digital Users	Cellphone subscribers	0.039	+
	Mobile Internet users	0.038	+
	Broadband Internet users	0.047	+
Digital Platform	Number of domains	0.096	+
	Length of fiber optic cable	0.044	+
	Information transmission, software and technology services’ employees	0.086	+
	Software business revenue	0.134	+
	Total telecom business	0.087	+
Digital Trading	Number of websites owned by enterprises	0.097	+
	Enterprises using computers	0.075	+
	Percentage of e-commerce	0.022	+
	E-commerce sales	0.099	+
	Online retail sales	0.136	+

. First, the continuous development of the digital economy has changed people’s lifestyles in three ways: social networks, search engines, and mobile Internet. The exponential growth in the amount of information people generate has laid the foundation for the development of the digital economy. Therefore, cell phone subscribers, mobile Internet subscribers, and broadband Internet subscribers are chosen as the basis for measuring the development of the digital economy [44]. Secondly, information and communication platforms drive the digital transformation of industries. The wide application of ICT can improve the efficiency of enterprises, change the industrial chain structure of manufacturing and service industries, gradually lead enterprises to digitalization, form a virtuous cycle of information product supply and enterprise industries, and inject power into the digital economy [45]. Therefore, the length of fiber optic cable, the number of domain names, information transmission, software, and technical services’ employees and software business income, and the total telecommunication businesses are chosen as the measurement indicators for the basis of information technology and information development [46]. Finally, the massive development of mobile payments and Internet finance has made digital transactions an important part of consumption in society [47]. Digital transactions cannot be made without the use of portals and computers established by enterprises, and the percentage of e-commerce companies can reflect the breadth of digital transactions. Meanwhile, e-commerce sales and online retail sales can represent the scale of digital transactions in the province, with the larger scale presenting a higher level of digital economy development. Therefore, the number of websites owned by enterprises, the computer usage rate of firms, and the share of e-commerce, e-commerce sales, and online retail sales are selected as the digital transaction indicators.

Since there are many evaluation indicators of the level of digital economy development, the extreme difference method is used to standardize the indicators [48]. Considering the existence of zero values, infinitesimal shifts are required after the standardization process, i.e., positive indicator $X_{\mathit{ij}}{}^{\prime }=\frac{X_{\mathit{ij}}-{\mathit{minX}}_{\mathit{ij}}}{{\mathit{maxX}}_{\mathit{ij}}-{\mathit{minX}}_{\mathit{ij}}}+0.00001$, negative indicator $X_{\mathit{ij}}{}^{\prime }=\frac{{\mathit{maxX}}_{\mathit{ij}}-X_{\mathit{ij}}}{{\mathit{maxX}}_{\mathit{ij}}-{\mathit{minX}}_{\mathit{ij}}}+0.00001$, and standardized value $S_{\mathit{ij}}=\frac{X_{\mathit{ij}}{}^{\prime }}{{\displaystyle \sum _{j=1}^m{X}_{\mathit{ij}}{}^{\prime }}}$. To avoid the bias caused by subjective factors, the entropy value method is used to determine the indicator weights [49].

3.2.3. Mediating Variables

Based on sorting out the theoretical mechanisms and referring to existing studies, energy production is selected as the mediating variable, containing the production of coke, raw coal, crude oil, gasoline, kerosene, diesel oil, fuel oil, and natural gas. The total energy production is then summed according to the standard coal conversion coefficient [50]. Wage data for workers in the energy sector are directly obtained from the China Statistical Yearbook [51]. The energy transportation efficiency is measured by the total factor efficiency method, which can reflect the real-time interaction of various factors more closely to the actual situation than the single factor efficiency. The input factors are transportation routes, energy industry workers, and energy fixed capital investment, while the output factors are energy transportation volume and turnover. The energy consumption quantity data are directly from the China Energy Statistical Yearbook [52]. These four mechanisms correspond to the production, distribution, exchange, and consumption stages of the energy “internal circulation” [53].

3.2.4. Control Variables

The level of economic development and the share of industry in each province directly reflects their energy demand; the retail price of fuel in each province directly affects energy production; the density of railroads, inland waterways, roads, and the distance between provinces reflect the speed of logistics and can affect the domestic circulation of energy. Therefore, the value added of secondary industry (SI), the level of population consumption (RC), the retail price of fuel in each province (RU), the density of transportation infrastructure (TD), and the distance between provincial capitals (Lndist) are selected as control variables [54,55].

3.3. Data Sources and Descriptive Statistics

As data from Hong Kong, Macau, Taiwan, and Tibet were inaccessible, this study selects a trade matrix of 30 Chinese provinces from 2006 to 2020, forming a balanced panel observation of 13,500 province–years. The data are mainly obtained from the China Statistical Yearbook and the China Energy Statistical Yearbook.

Table 2. Descriptive statistics.

Category	Variables	Obs	Mean	Std	Min	Max	Source
Explained variables	EIC	13,500	9.256	1.238	4.739	14.152	Calculated by author
Explanatory variables	Digital	13,500	1.832	0.812	1.005	6.183	Calculated by author
Mediating variables	EP	13,500	8.640	1.029	5.763	11.491	China Energy Statistical Yearbook
	CI	13,500	9.693	0.573	8.298	11.188	China Statistical Yearbook
	TE	13,500	1.138	0.764	0.252	7.102	Calculated by author
	EC	13,500	9.325	0.695	6.824	10.631	China Energy Statistical Yearbook
Control variables	RC	13,500	19.117	1.117	16.484	22.244	China Statistical Yearbook
	SI	13,500	17.227	1.563	10.860	21.412	China Statistical Yearbook
	RU	13,500	9.229	0.238	4.796	9.538	China Statistical Yearbook
	TD	13,500	−0.564	1.142	−5.346	1.851	China Statistical Yearbook
	Lndist	13,500	7.049	0.607	4.701	8.147	Calculated by author

EIC—energy “internal circulation”; Digital—digital economy’s development; EP—energy production; CI—the workers’ wages in the energy sector; TE—energy transportation efficiency; EC—energy consumption; RC—the level of population consumption; SI—the value added of secondary industry; RU—the retail price of the fuel in each province; TD—the density of transportation infrastructure.

shows the descriptive statistics of the main variables in this study.

4. Empirical Results and Analysis

4.1. Results of Benchmark Regression

Table 3. Baseline regression and robustness test.

	(1)	(2)	(3)	(4)	(5)	(6)
Digital	0.6678 ***	0.8015 ***	1.0875 ***	0.8622 ***	0.1641 ***	1.1218 ***
	(7.3040)	(15.5678)	(12.0626)	(6.1741)	(6.0775)	(8.8187)
SI		0.3274 ***	0.5645 ***	0.0472 ***	0.4040 ***	0.7209 ***
		(9.9846)	(13.3177)	(7.8106)	(9.7551)	(13.1042)
RC		0.1872 ***	0.5555 ***	0.0358 ***	0.5164 ***	0.8074 ***
		(4.2437)	(8.0600)	(4.5987)	(7.3031)	(8.8203)
RU		0.1383 ***	0.0744 ***	0.0433 ***	0.0940 ***	0.0929 ***
		(9.1946)	(2.9726)	(3.6914)	(3.7508)	(3.8274)
TD		0.1061 **	0.1166 **	0.0359 ***	0.0797	0.1735 ***
		(2.3293)	(2.4300)	(6.5225)	(1.6058)	(2.6109)
lndist		−0.2695 ***	−0.2529 ***	−0.0143 ***	−0.3017 ***	−0.5032 ***
		(−3.8219)	(−3.6771)	(−2.8680)	(−4.1878)	(−5.7131)
L.EIC				0.9614 ***
				(21.0401)
Constant	9.1849 ***	2.2279 ***	11.6447 ***	0.0491	9.5592 ***	24.6766 ***
	(156.7618)	(4.0121)	(8.9977)	(0.2125)	(7.3078)	(10.2003)
Time Fixed	Yes	Yes	Yes	Yes	Yes	Yes
Individual Fixed	Yes	Yes	Yes	Yes	Yes	Yes
Observation	13,320	12,847	12,847	11,919	12,847	12,818
R-sq	0.122	0.184	0.107	0.410	0.215	0.134

Note. *** p < 0.01, ** p < 0.05.

reports the results of benchmark regression and the results of the endogenous treatment of instrumental variables. Based on controlling time- and region-fixed effects, no control variables are added in column (1), while control variables are added in column (2). The regression results show that the effect of the digital economy on the energy “internal circulation” is significantly at the level of 1%, indicating that the development of the digital economy is conducive to accelerating the energy “internal circulation”. The development of the digital economy is highly technological and energy intensive. High technology enables the digital economy to empower the energy system, connects the energy industry chain with the industrial ecosystem, and provides important technical support for energy circulation. High energy consumption has increased the energy demand, promoted the reformation and upgrading of energy demand, activated the energy demand of residents, and injected vitality into the energy “internal circulation” [56]. Hypothesis 1 is verified.

4.2. Robustness Test

To test the robustness of the model regression, three methods are taken: replacing the explained variable, replacing one period lag of the explained variable, and replacing the core explanatory variable. In column (3) of Table 3, the interprovincial energy flows are used to estimate the explained variables without the gravity model; the regression results are still robust. Column (4) lags the interprovincial energy trade data by one period, and the energy interaction level between the two provinces in the last period will affect the energy interaction in the next period. In the regression results, the digital economy and the energy interaction level in the last period are both significant at 1%. In column (5), principal component analysis is adopted to re-measure the development level of the digital economy. The regression results of the model are still robust and the regression results of this paper are credible.

Most scholars believe that the digital economy is energy intensive. The industrial infrastructure of the digital economy and “digital tsunami” directly increase energy consumption, while the “income effect” of the indirect impact of digital technology application on energy consumption is much larger than the “substitution effect”. Therefore, the smooth circulation of domestic energy will provide a sufficient energy supply to sustain the deep development of the digital economy. Considering the validity of the benchmark regression may be affected by the bi-directional causality of energy and digital, the endogeneity should be dealt with. According to the research idea of Huang Qunhui [57], the number of telephones per 10,000 people in 1984 is selected as the instrumental variable. However, since it is a cross-sectional data sample that does not change with time, the interaction terms for the number of telephones per 10,000 people and the Internet broadband access ports in the previous year are adopted as the instrumental variable to address the issue of individual characteristics not changing with time. Areas with historically high fixed-line penetration are likely to be those with faster digital penetration, and the improvement of their communication technology facilities will affect the rapid growth of the digital economy and meet their relevance. Furthermore, the penetration rate of fixed telephone and the number of broadband access ports are the basis for the digital construction of the energy system, that is, they only affect the energy “internal circulation” from the digital economy, to meet the exclusivity. The Kleibergen–Paap rk LM statistic is significant at the 1% level, rejecting the null hypothesis that the instrumental variables are not identifiable. Kleibergen–Paap rk Wald F statistics and Cragg–Donald Wald F statistics are greater than Stock–Yogo weak instrumental variable identification, and the F test is at the critical value of 10% significance level, rejecting the null hypothesis of weak instrumental variables; the selection of instrumental variables in this paper is reasonable and reliable. Column (6) indicates that for every one unit increase in the level of development of the digital economy, the energy “internal circulation” will be accelerated by 112%. The robustness of the results in this paper is verified.

4.3. Mechanism Test

According to the above internal mechanism analysis, the digital economy speeds up the energy “internal circulation” by affecting the four major links of energy production, distribution, exchange, and consumption. The regression results are shown in

Table 4. Mechanism inspection of production, distribution, exchange, and consumption.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Digital	0.7002 ***	0.4146 ***	0.8482 ***	1.1016 ***	0.8043 ***	1.0715 ***	0.3633 ***	1.0307 ***
	(8.5381)	(8.0110)	(13.4164)	(12.2743)	(6.5598)	(11.8948)	(11.9269)	(11.4874)
EP		0.9443 ***
		(61.3389)
CI				0.5434 ***
				(3.9217)
TE						0.2152 ***
						(5.0012)
EC								0.1633 ***
								(6.2595)
SI	0.4365 ***	0.1507 ***	0.0822 ***	0.5486 ***	0.0030	0.5628 ***	0.2466 ***	0.5302 ***
	(12.2460)	(5.8372)	(18.1430)	(12.9369)	(1.5835)	(13.2712)	(17.9961)	(12.4543)
RC	0.3126 ***	0.2512 ***	0.3563 ***	0.4550 ***	0.0022	0.5626 ***	0.0569 **	0.5252 ***
	(5.0589)	(6.2163)	(57.9886)	(5.7871)	(0.5513)	(8.2374)	(1.9622)	(7.9204)
RU	0.0111	0.0642 ***	0.0043 ***	0.0718 ***	0.0028	0.0736 ***	0.0363 ***	0.0684 ***
	(0.7662)	(2.9361)	(3.7701)	(2.8869)	(1.4022)	(2.9431)	(4.3389)	(2.7008)
TD	0.1604 ***	0.0388	0.0258 ***	0.1420 ***	0.0074 ***	0.1215 **	0.0433 **	0.1388 ***
	(4.1601)	(1.2645)	(2.7380)	(2.9454)	(4.2935)	(2.5464)	(2.0190)	(3.0353)
lndist	−0.0573	−0.2025 ***	−0.0726 ***	−0.2397 ***	−0.0034	−0.2512 ***	−0.0079	−0.2430 ***
	(−1.0515)	(−4.1640)	(−4.3772)	(−3.3349)	(−1.2030)	(−3.6474)	(−0.2413)	(−3.7108)
Term of constant	4.6212 ***	7.1238 ***	0.9190 ***	14.3624 ***	0.0444	11.7340 ***	4.0605 ***	11.9675 ***
	(4.1869)	(8.1703)	(8.4610)	(10.8815)	(0.4948)	(9.1264)	(7.6402)	(9.4536)
Time fixed	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Individual fixed	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	13,021	12,847	13,021	12,847	13,021	12,847	13,021	12,847
R-sq	0.683	0.575	0.563	0.377	0.214	0.208	0.408	0.205

Note. *** p < 0.01, ** p < 0.05.

. Columns (1) and (2) are regression results mediated by energy production. Column (1) indicates that the level of digital economy development has a significant facilitating effect on energy production and the premise of mediating mechanism is valid. When energy production is added into the regression equation, the level of digital economy development and energy production has a positive facilitating effect on the energy “internal circulation”, which means that the influence path mediated by energy production is established. Specifically, every one unit increase in the level of the digital economy development will unblock 0.4146 units of the energy “internal circulation” and increase energy production by 0.7002 units, thus increasing energy “internal circulation” by 0.6612 units. Hypothesis 2 is verified.

Columns (3) and (4) are the regression results for the wages of mediating workers in the energy system. The results show that a one unit increase in the level of digital economy development will have an impact of 1.1016 on energy “internal circulation”, will increase wages for energy system workers by 0.8482, and will have an indirect impact of 0.4609. The emergence of the new economic formation of the digital economy and its integration with other industries has significantly increased the wages of workers in the energy system, stimulated their enthusiasm for energy production, and increased the level of energy consumption per capita. Hypothesis 3 is verified.

Columns (5) and (6) are regression results mediated by energy transportation efficiency. The energy “internal circulation” caused by energy transportation efficiency at the development level of the digital economy is 0.1731. Higher efficiency of energy circulation will lead to a stronger facilitating effect of the digital economy on the energy “internal circulation” (mainly in the new infrastructure level), the construction of comprehensive transportation hubs, logistics hubs, and information hubs, the formation of modern transportation and logistics system, the speed of energy from the supply side to the demand side, and the improvement of energy supply security. Hypothesis 4 is verified.

Columns (7) and (8) are the mediating regression results of energy consumption. The results show that the impact of the digital economy on the energy “internal circulation” through energy consumption is 0.0614. The development of the digital economy makes Internet-based information access more transparent and extensive, improving the compatibility between the energy demand side and the supply side, realizing the real-time information matching between the supply and demand sides of energy, providing accurate energy consumption information, avoiding energy waste, reducing information barriers, and accelerating the energy circulation. Hypothesis 5 is verified.

4.4. Heterogeneity Analysis

To further study the influence of the level of digital economy development on the energy “internal circulation”, the level of digital economy development is divided into three settlements, namely the start-up stage, the development stage, and the maturity stage. FZ represents the initial stage, the development stage, and the maturity stage, respectively, with 1 for yes and 0 for no.

Table 5. Influences of different development stages of the digital economy on the energy “internal cycle”.

	(1)	(2)	(3)
Digital * FZ	0.0131	0.1028 ***	0.2215 ***
	(0.9227)	(5.6363)	(6.4954)
SI	0.5632 ***	0.5552 ***	0.5479 ***
	(13.3894)	(13.1219)	(12.8390)
RC	0.5682 ***	0.5727 ***	0.5632 ***
	(8.4793)	(8.6188)	(8.3392)
RU	0.0765 ***	0.0766 ***	0.0747 ***
	(3.0691)	(3.0779)	(2.9827)
TD	0.1259 ***	0.1339 ***	0.1348 ***
	(2.6532)	(2.8298)	(2.8335)
lndist	−0.2485 ***	−0.2440 ***	−0.2425 ***
	(−3.6185)	(−3.5597)	(−3.5440)
Constant	11.7270 ***	11.7832 ***	11.5190 ***
	(9.1829)	(9.2455)	(8.9598)
Time fixed effect	Yes	Yes	Yes
Region fixed effect	Yes	Yes	Yes
Observations	12,847	12,847	12,847
R-sq	0.106	0.108	0.112

Note. *** p < 0.01, ** p < 0.05, * p < 0.1.

shows that columns (1), (2), and (3) show that the higher level of the digital economy development will lead to a higher facilitating effect of the digital economy on the energy “internal circulation”. Specifically, the initial digital economy has no significant impact on the energy “internal circulation”, which is 10.28% in the development stage and increases to 22.15% in the maturity stage. This means that the higher development level of the digital economy will lead to stronger integration of its high-tech and energy-intensive attributes with the energy system and a more significant promotion of the energy “internal circulation”.

Due to the high-energy nature of the digital economy, its operation and development depend on energy supply. Different energy endowments may lead to different effects of the digital economy on the energy “internal circulation”. Therefore, the samples are divided into energy-rich and energy-poor regions based on the average value of total energy production. As shown in

Table 6. Influence of energy endowment on energy “internal circulation”.

	(1)	(2)
Digital * NY	0.9898 ***	0.8867 ***
	(20.0407)	(22.1250)
SI	0.2251 ***	0.4022 ***
	(5.9440)	(11.0813)
RC	0.3929 ***	0.4065 ***
	(6.6772)	(7.1619)
RU	0.0543 **	0.0344
	(2.0532)	(1.3262)
TD	0.0753 *	−0.0140
	(1.6912)	(−0.3262)
lndist	0.2816 ***	0.2062 ***
	(4.2568)	(3.1925)
Constant	−3.9587 ***	−5.9023 ***
	(−3.3525)	(−5.2227)
Time fixed effect	Yes	Yes
Region fixed effect	Yes	Yes
Observations	12,847	12,847
R-sq	0.166	0.192

Note. *** p < 0.01, ** p < 0.05, * p < 0.1.

, the impact of digital economy development on energy “internal circulation” is 98.98% in energy-rich areas and 88.67% in energy-poor areas. This is because China’s energy-poor regions, such as Beijing, Shanghai, and Guangdong, are the main regions of energy consumption. The development of the digital economy indirectly promotes the energy “internal circulation” from the demand side, but the impact is less than the direct promotion effect of energy-rich areas.

Because of the spatial mismatch between the centers of China’s digital economy and energy production, there are significant differences in the degree of digital economy development and energy endowment in different geographical regions and the influence of digital economy development on the energy “internal circulation” may be comprehensively influenced by regional characteristics. Therefore, China is divided into three regions: east, middle, and west (eastern region: Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Hainan; central region: Shanxi, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei, Hunan; western region: Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang).

Table 7. Regional heterogeneity.

	(1)	(2)	(3)
Digital	0.5821 ***	1.0086 ***	0.6680 ***
	(6.1425)	(6.2502)	(7.7933)
SI	0.5354 ***	0.1075	0.4386 ***
	(9.9386)	(1.1873)	(6.1410)
RC	0.1958 ***	0.9553 ***	0.4355 ***
	(2.6071)	(6.4050)	(3.3071)
RU	0.1416 ***	0.0852 *	−0.0175
	(6.6697)	(1.7023)	(−0.4824)
TD	0.0447	0.0861	0.1643 **
	(0.5839)	(0.7942)	(2.4286)
lndist	−0.3045 ***	−0.3945 ***	−0.0522
	(−2.9541)	(−2.8904)	(−0.4421)
Constant	6.1591 ***	12.4742 ***	4.7937 **
	(3.7071)	(4.3189)	(2.1051)
Time fixed effect	Yes	Yes	Yes
Region fixed effect	Yes	Yes	Yes
Observations	4714	3424	4709
R-sq	0.363	0.132	0.381

Note. *** p < 0.01, ** p < 0.05, * p < 0.1.

shows that the digital economy has the strongest impact on the energy “internal circulation” in the central region. The eastern region of China has a well-developed digital economy but is energy poor and a large number of energy resources need to be transferred from the western region, which is a long way away. The energy takes a long time in the exchange link, limiting the speed of energy circulation. The western region has a rich energy endowment, but the development of the digital economy is relatively weak. There are few digital applications in energy production, transmission, trading, consumption, and regulation, so it is impossible to effectively regulate the interconnection of information in all segments. As a result, the energy circulation in the central region is relatively slow, which is closer to the western energy center than in the east, and the energy exchange is smooth. Compared with the western region, the eastern region has a faster development of the digital economy with obvious spillover effects, and the digital and energy systems are better coupled and coordinated, effectively increasing the speed of energy “internal circulation”.

5. Conclusions and Recommendations

Based on the panel data on digital economy and energy trade in each province of China from 2006 to 2020, this paper studies the relationship between the digital economy and the energy “internal circulation” and explores the mechanism of the digital economy’s influence on energy circulation from four aspects: production, distribution, exchange, and consumption. The results show that: (1) The rapid development of the digital economy is conducive to making the circulation of energy in China unimpeded. (2) The digital economy accelerates the energy “internal circulation” by influencing energy production, distribution, exchange, and consumption. (3) The higher development level of the digital economy, the energy-rich area, and the central region with the strongest digital–energy coupling and coordination degree have a stronger promoting effect on the energy “internal circulation”. Therefore, the following policy recommendations are proposed:

First, based on China’s reality, we need to accelerate the development of the digital economy. China should make full use of its huge domestic market and relatively complete industrial system to accelerate the development of the digital economy, break the time and space constraints, effectively promote the rapid flow of energy, build a large energy “internal circulation”, and promote high-quality economic growth.

Second, we need to focus on the four major links to promote the integration of digital and energy. The production chain should strengthen the practical empowerment of digital technology in the modernized industrial system of energy, improve the supply of fossil energy and renewable energy, and rebuild the high-quality energy supply side. The distribution session needs to improve the efficiency of digital regulatory services, actively formulate countermeasures against monopoly, damage to consumer interest, and lack of labor rights protection, and build an efficient regulatory system. In the exchange session, we should also build region-wide digital transportation in the roadway, railway, shipping, and waterway sectors, upgrade logistics infrastructure and modernize warehousing, and minimize the time spent in the transportation of energy. In the consumption session, we should make full use of distributed energy to effectively coordinate and optimize the supply and energy use links on the user side, forming a dynamic balance of supply creating demand and demand pulling supply.

Third, we need to promote the coordinated development of the digital economy and energy system. The smooth operation of the energy “internal circulation” requires the interconnected development of the eastern, central, and western regions. The priority should be to combine the energy endowment advantages of the central and western regions with the technological, market, and digital industry advantages of the eastern region and to speed up the establishment of a supporting and coordinated development mechanism. In addition, the central and western regions need to further consolidate the foundation for the digital economy and to strengthen new infrastructure.

It should be noted that this paper has estimated the level of digital economy development and interprovincial energy trade for each province. However, the development of the digital economy has a more direct impact on electricity. In addition, the trade estimates using the gravity model are still subject to errors due to the lack of direct interprovincial energy trade data. How to accurately portray the impact of the digital economy on the interprovincial electricity cycle is a question worthy of further research in the future.