Impact of Digital Economy on Dual Circulation: An Empirical Analysis in China

Abstract

With the growth of Internet technologies, including 5G, blockchain, and big data, the digital economy has emerged as an important driving force of economic growth, offering a fresh viewpoint on the sustainability of dual circulation. Thus, this study analyzes the impact of the digital economy on dual circulation and the underlying influencing mechanism. The influence of the digital economy on the degree of dual circulation sustainability was measured using panel data of 30 provinces in China from 2011 to 2020, and the influence was found to be significant and statistically positive. The analysis of the mechanism indicates that the increasing technological innovation capacity can amplify the digital economy’s effect on dual circulation. The promotion effect of the digital economy has a spatial spillover effect. The function of the digital economy in fostering dual circulation is heterogenous, with a greater impact on central and eastern China. Therefore, this study proposes that increasing communication infrastructure investment and implementing differentiated policies supporting the digital economy should be considered by policymakers to boost dual circulation.

1. Introduction

As China’s economy has entered the stage of high-quality development, ensuring sustainable economic growth has emerged as the core challenge of China’s current economic development. The digital economy, which is the third type of economy after agricultural and industrial economies, has driven the sustainability of economic growth in recent years [1]. The digital economy provides a new impetus and perspective for the long-term economic growth. It is one of China’s most dynamic areas of economic development. Additionally, numerous fields, including industrial structure and science technology innovation, are integrated with the digital economy. It is a key growth point for global economic growth and innovation capacity enhancement. The Chinese government has particularly emphasized the growth of the digital economy over the years. At the G20 Summit in Hangzhou in 2016, China initiated and joined the G20 Initiative on Digital Economy Development and Cooperation, which marked the beginning of a series of strategies and actions for the digital economy by the international community. At the 19th National Congress of the Communist Party of China, it was advocated that the Internet, big data, and artificial intelligence would be heavily integrated into the economy through promotion. At the 2019 Central Economic Work Conference, the basic task of “vigorously developing the digital economy” was proposed, setting the direction of China’s economic transformation. The 14th Five-Year Plan outlined the tasks and particular objectives of digital economic development. Several digital economy policies have been implemented smoothly under the active guidance of central and local governments. China’s digital economy, in 2016, which was valued at 22.58 trillion RMB, reached 45.5 trillion RMB in 2021, according to the data from the China Academy of Information and Communications Technology (CAICT). Additionally, the digital economy’s expansion was much larger than that of the gross domestic product (GDP) during the same period. The digital economy’s position as the center of the national economy continues to solidify, and the role of the “gas pedal” is prominent. Data from CAICT suggest that the scale of China’s digital economy ranks first worldwide, and the pace of growth is 9.6%, second only to the United States’ digital economy. Simultaneously, against the backdrop of slowing global economic growth, prevalent unilateralism, and reverse globalization, President Xi, on 14 May 2020, proposed the Dual Circulation Development Paradigm, in which domestic economic circulation plays a leading role, while international economic circulation acts as in a supporting role. The primary goal of this paradigm is to highlight the potential of the domestic market as the major driving force for China’s sustainable development and to emphasize the importance of a mutual benefit relationship between self-reliance and opening up. As a new production factor, the digital economy can supply new growth points and poles for high-quality economic development by being combined with traditional industries, which is a significant driving force in establishing a dual circulation development pattern [2]. The most important issue in constructing a smooth economic circulation under the new pattern is achieving sustainable economic growth.

Academic research on the interaction between the digital economy and the dual circulation has shown that facilitation is multidimensional. From a micro perspective, the organic integration of the digital economy and traditional sectors can foster synergy among elements and increase the effectiveness of inter-production collaboration. From a macro viewpoint, the popularity of the Internet infrastructure is non-exclusive and non-competitive, which may promote high-quality economic development through higher resource allocation efficiency and total factor productivity (TFP). The relevant literature focuses mainly on two aspects. The first comprises the study of the mechanism and realization approach of the digital economy that contributes to the new development pattern of dual circulation at the theoretical level [3,4,5]. However, empirical studies on the relationship between the two are extremely rare. The second aspect includes a study on how the digital economy affects areas associated with dual circulation, such as enhancing consumption, export trade, boosting TFP, and enhancing high-quality economic development [6,7,8,9,10,11,12,13,14]. However, previous studies have not explored the path of the digital economy for dual circulation, which presents a possibility for contribution through this research.

In conclusion, is the digital economy promoting dual circulation development in China? What is the mechanism that affects its function? How does the digital economy’s impact on the development of dual circulation differ spatially? These are the topics that need investigation in relation to the digital economy. Studying the above issues provides theoretical and empirical evidence for dual circulation, and references and suggestions for people who develop and put forth policies of considerable lasting benefit. Consequently, this study assesses the degree of the digital economy and dual circulation of 30 Chinese provinces from 2011 to 2020 and then constructs a theoretical framework with an innovation capacity. Moreover, this research uses mediating effect to examine the effect mechanism and finds that the digital economy adds greatly to dual circulation development, with technological innovation capabilities serving as the most influential mechanism. Then, the spatial Durbin model is adopted, which indicates that the degree of dual circulation development is affected by the digital economy’s spatial spillover impact.

The following are the contributions of this study. First, based on the characteristics of the spatio–temporal evolution of the digital economy and dual circulation, a spatial model is used to explore the relationship between the two. Hence, the absence of the digital economy and dual circulation in the economic geography is compensated. Second, it provides policy insights for creating a new pattern of dual circulation development. The digital economy, “mass innovation”, “mass entrepreneurship”, and the domestic and international circulation are important areas of economic growth at this stage. Through empirical research, this study finds that innovation capability significantly helps the digital economy to grow the dual circulation economy, which is crucial to the realization of the new pattern of economic development.

The remainder of this paper consists of the following sections. The theoretical mechanism and research hypotheses are detailed in the next section. The third section presents the research design by introducing the model and describing the variables. The fourth section presents the results of the empirical tests, which include the benchmark test, spatial effect test, regional heterogeneity test, robustness test, and endogeneity test. Finally, the fifth section provides study findings and policy recommendations.

2. Theoretical Analysis and Research Hypothesis

2.1. Direct Effect of the Digital Economy on Dual Circulation

The digital economy improves the degree of dual circulation development by expanding domestic demand and consumption [5]. First, the digital economy can deeply explore the domestic market demand and create a large domestic market. E-commerce platforms rely on the mining and analysis of a large amount of data and accurate marketing to properly match different levels of consumer demand, provide consumers with preferred products and services, and enhance the efficiency of capital circulation. Then, the penetration of the digital economy into various industries creates abundant jobs and optimizes the employment structure [15]. For example, the improvement in the digital infrastructure, the development of smart agriculture, and the upgrading of logistics and transportation capacity create opportunities for publicity and sales channels for agricultural products in remote areas. This effectively raises the income and material consumption demand of local farmers and improves the nation’s living standard. Finally, the digital economy improves the domestic factor market and promotes domestic factors’ free flow. Integrating the digital economy with various industries produces an open market environment that facilitates efficient and free flow of various factors of production. Simultaneously, digital trade reduces transaction costs by reducing information asymmetry and facilitating the matching of supply and demand in factor markets, providing sufficient power for the endogenous market and guaranteeing the development of a new dual circulation economic pattern [16,17].

The digital economy contributes to dual circulation patterns by strengthening foreign trade ties. First, the emergence of the digital economy effectively reduces trade costs, including the costs of searching, transportation, and communication [18,19]. Digital technology has been popularized in daily life, enabling both buyers and sellers to be matched through the network, breaking down information barriers and thus reducing search costs. With a high degree of cooperation between digital technology and logistics and supply chain, digital trade effectively reduces transportation costs. The continuous maturity of augmented reality (AR) technology and the high popularity of 5G communication technology realize efficient interconnection at home and abroad, and at the supply and demand ends, providing a strong guarantee for reducing communication costs [20]. Then, the digital economy expands the trade market and enhances trade efficiency. Advances in communication technology and the rise of e-commerce have lowered the import and export trade threshold and broken the original geographical restrictions. Simultaneously, some services have been digitized by using the Internet to expand the trading space, thus promoting business development. Finally, the growth of digital trade platforms overseas promotes the transformation of the main consumer and even that of major international trade. This expansion of the scale of imports is accompanied by the realization of diversified and advanced domestic consumer demand and the stimulation of domestic industrial upgrading. The changes in the digital economy in trade costs, markets, and subjects can enable the synergistic development of the domestic market and import and export trade [21,22]. Moreover, the elevation of the digital economy can build a different economic pattern with the domestic market as the driving force, domestic industries as the pillar, and foreign trade as the link, facilitating the formation of domestic circulation and international circulation through mutual promotion [23]. This study proposes the following hypothesis based on the theoretical analysis listed above.

Hypothesis 1.

The digital economy contributes positively to dual circulation.

2.2. Indirect Effect of the Digital Economy on Dual Circulation

Artificial intelligence, 5G, and cloud computing, representing the major applications of digital technology, provide a convenient ground for technology innovation. The digital economy can promote technological progress by influencing the market size, knowledge spillovers, and factor combinations. First, the digital economy promotes diversified consumer demand for products, and instant communication technologies have transformed the one-way product supply into an efficient exchange of opinions between supply and demand [24]. Cloud computing and big data provide optimized methods for matching information in the market. Therefore, they greatly enhance the richness of product and service markets and establish a stable basis for creating breakthrough technologies. Second, the proliferation of the network infrastructure can accelerate the cross-regional integration of production inputs such as money, skills, and technology. The network infrastructure can accelerate the cross-regional integration of production factors such as capital, talent, and technology. Meanwhile, manufacturing, research, and production technology diffuse within the system, creating a knowledge spillover effect. Wang et al. found that government digitalization initiatives could promote enterprise innovation [25].

Additionally, the advancing innovation capacity significantly improves TFP. Digital innovation is one of the most important drivers of TFP [26]. The consistent growth in TFP can fundamentally improve resource utilization and manufacturing efficiency in all industrial sectors. The following hypothesis is presented in this study.

Hypothesis 2.

The digital economy influences dual circulation development by strengthening technological innovation.

2.3. Spatial Spillover Mechanism of Digital Economy on Dual Circulation

The physical exchange of information is a key component of the digital economy, implying that efficient data transmission, knowledge accumulation, and technology diffusion break the temporal and spatial limitations of economic activities and enhance the breadth and width of regional economic cooperation with positive spillover effects [27]. Additionally, according to the growth pole theory, economic development does not occur simultaneously in all regions, but rather from one or several growth centers to other centers. Therefore, the diffusion effect is generated by one or several growth centers in other regions. Furthermore, diffusion efficiency is greatly enhanced by the spread of digital technologies such as the Internet. Yilmaz et al. argue that investments in telecommunications infrastructure had a negative spillover effect on production, which indicated that even when one region could earn profit from convenient communication conditions, the neighboring regions might still suffer from that [28]. Similar evidence from Spain and China suggests that the Internet has a spatial spillover effect on the gross output [29,30,31]. Moreover, the Internet has spatial spillover effects on export trade and digital finance [8,32,33,34,35]. Similarly, the digital economy, which is strongly tied to the Internet, also has spatial spillover effects, for which this research proposes the following hypothesis.

Hypothesis 3.

The digital economy can influence the growth of dual circulation in adjacent regions through the spatial spillover effect.

The conceptual framework is depicted in Figure 1.

3. Method and Materials

3.1. Model Construction

The direct transmission mechanism is modeled to test the hypotheses.

$$CC{D}_{i,t}={\alpha }_0+{\alpha }_{1}Dig{e}_{i,t}+{\alpha }_2{Z}_{i,t}+{\mu }_i+{\delta }_t+{\epsilon }_{i,t}$$

(1)

In Equation (1)$, CC{D}_{i,t}$ is the degree of dual circulation development of region $i$ during period $t$.$Dig{e}_{i,t}$ is an indication of the level of digital economy development of region $i$ during period $t$. Vector $Z_{i,t}$ is a collection of control variables. The variable ${\mu }_i$ denotes regional fixed effects and ${\delta }_t$ denotes time fixed effects. ${\epsilon }_{i,t}$ indicates the random disturbance term. ${\alpha }_0$ is the intercept term.

In addition to the direct influence represented in Equation (1), it is determined whether the regional innovation capability is a mediating variable between the two to analyze the possible mechanism of the effect of the digital economy on the degree of dual circulation development. The particular test steps are as described below. Based on the significant coefficient ${\alpha }_1$ of model (1), we first create the regression model of $Dige$ on the mediating variable $Ino.$ Then, the linear regression equation of $Dige$ and $Ino$ on $CCD$ is constructed. Finally, we determine the existence of the mediating effect by the significance of the regression coefficients of ${\beta }_1, {\gamma }_1,\mathrm{and}{\gamma }_2$. The following are the specific equations of the aforementioned regression model.

$$In{o}_{i,t}={\beta }_0+{\beta }_{1}Dig{e}_{i,t}+{\beta }_2{Z}_{i,t}+{\mu }_i+{\delta }_t+{\epsilon }_{i,t}$$

(2)

$$CC{D}_{i.t}={\gamma }_0+{\gamma }_{1}Dig{e}_{i.t}+{\gamma }_{2}In{o}_{i.t}+{\gamma }_3{Z}_{i.t}+{\mu }_i+{\delta }_t+{\epsilon }_{i.t}$$

(3)

A spatial interaction term is added to the basic regression model, which is expanded to a spatial econometric model to explore the spatial spillover effect of the digital economy on the degree of dual circulation development.

$$CC{D}_{i,t}={\alpha }_0+\rho WCC{D}_{i,t}+{\theta }_{1}WDig{e}_{i,t}+{\alpha }_{1}Dig{e}_{i,t}+{\theta }_{2}W{Z}_{i,t}+{\alpha }_2{Z}_{i,t}+{\mu }_i+{\delta }_t+{\epsilon }_{i,t}$$

(4)

where $\rho$ represents the spatial autoregressive coefficient, $W$ represents the spatial weight matrix, and ${\theta }_1$ and ${\theta }_2$ represent the spatial interaction terms of the explanatory and control variables, respectively. Therefore, to investigate the impact of the digital economy on dual circulation thoroughly, an adjacency weight matrix, a geographic weight matrix, and an economic weight matrix are created to quantify the geographical importance of the digital economy.

(1)

Adjacency weight matrix, which is written as W₁:

$$W_{ij}=\left\{\begin{array}{l}1,\mathrm{when}\mathrm{province}i\mathrm{and}j\mathrm{are}\mathrm{adjacent}\hfill \\ 0,\mathrm{when}\mathrm{province}i\mathrm{and}j\mathrm{are}\mathrm{not}\mathrm{adjacent}\hfill \end{array}\right.$$

(5)

(2)

Geographic weight matrix, expressed as W₂:

$$W_{ij}=\left\{\begin{array}{l}1/d_{ij}，i\ne j\hfill \\ 0，i=j\hfill \end{array}\right.$$

(6)

d_ij represents the direct distance between the capitals of region i and j in Equation (6).

(3)

Economic weight matrix, indicated as W₃:

$$W_{ij}=\left\{\begin{array}{l}1/\left|Y_i-Y_j\right|，i\ne j\hfill \\ 0，i=j\hfill \end{array}\right.$$

(7)

In Equation (7), $Y_i$ represents the GDP per capita in region $i$ of 2011.

3.2. Measure and Description of Variables

3.2.1. Explained Variables

The coupling coordination degree of dual circulation (CCD): The coupling coordination degree is measured by two primary indicators: domestic circulation and international circulation. Based on the presented studies, this study developed the secondary indicators representatively connected with the two primary indicators. The secondary indicators of domestic circulation is the fundament of consumption, the intention of consuming, the structure of consumption, the scale of production, pattern of production, and production efficiency [36]. The secondary indicator of international circulation is outward direct investment import trade, export trade, and technology import [37,38]. Ultimately, considering the disparity and inadequacy of domestic and international circulation, the degree of dual circulation’s sustainability is measured by coupling the coordination degree model, which evaluates the level of coordinated development between domestic and international circulation [39,40]. The coupling coordination degree reflects the scale of the degree of benign coupling of interactions between systems or components, indicating a condition of outstanding coordination between systems or elements, which is a widely accepted method for assessing the connection and coordinated growth of economic subsystems [41]. Hence, CCD is considered as an index to evaluate the coupling coordination degree of dual circulation in order to measure sustainable growth.

3.2.2. Explanatory Variables

Digital economy’s development (Dige): Referring to existing studies, the digital economy’s level of development is measured on the basis of Internet development and digital financial inclusion [13]. Referring to Huang et al., this research selects data on four dimensions to measure Internet development, including broadband penetration rate, Internet-related employment status, Internet-related production, and the penetration rate of smartphones [42]. The China Digital Inclusive Finance Index is also employed to quantify digital financial inclusion. Furthermore, the level of digital economy development is computed using principal component analysis with the standardized data mentioned above.

3.2.3. Mediating Variables

Technological innovation (Ino): Referring to Li and Cui, the innovation input factor perspective is chosen to quantify the level of technological innovation, denoted as Ino [43].

3.2.4. Control Variables

Foreign direct investment (FDI): The region’s attractiveness to investment from abroad may be a powerful contributor to global commerce and have a bigger influence on dual circulation. The actual use of foreign investment to regional GDP is used to measure the reliance on foreign investment. Urbanization level (Urban): This study measures the urban resident population compared with the regional resident population. The marketization index (Market): The marketization level represents the market’s capacity to distribute production factors, calculated by the Report on China’s Marketization by Province. Involvement of government (Gov): This indicator measures the extent to which the government intervenes in a competitive market. This study uses the ratio of a region’s fiscal expenditure to its GDP to represent the degree of government involvement.

3.3. Data Sources and Descriptive Statistics

As data from Hong Kong, Macau, Taiwan, and Tibet were inaccessible, data from 30 Chinese provinces from 2011 to 2020 were chosen for this study, forming a balanced panel of 300 province–years of observations. This study is based on data that were obtained mainly from the Statistical Yearbook of China, the China Science and Technology Statistical Yearbook, Major Science and Technology Indicators Database of China, annual statistic reports of some provinces, and Wind Database.

Table 1. Statistical descriptions of the main variables.

	Variables	Obs	Mean	Std	Min	Max
Explained variables	CCD	300	0.386	0.145	0.108	0.756
Explanatory variables	Dige	300	0.314	0.104	0.0773	0.782
Mediating variables	Ino	300	0.134	0.128	0.00895	0.842
Control variables	FDI	300	0.0217	0.0203	0.000107	0.131
	Urban	300	0.577	0.126	0.350	0.943
	Market	300	6.525	1.900	2.330	10.856
	Gov	300	0.265	0.117	0.121	0.758

CCD: The degree of dual circulation; Dige: The level of digital economy’s development; Ino: technological innovation; FDI: foreign direct investment; Urban: urbanization level; Market: marketization index; Gov: involvement of government.

shows the descriptive statistics for the primary variables of this research.

4. Analysis and Empirical Results

4.1. Results of Benchmark Regression

Table 2. Results of benchmark regression and mediation regression.

Variables	Benchmark Regression	Mediating Effects
	CCD (1)	CCD (2)	Ino (3)	CCD (4)
Dige	0.621 ***	0.575 ***	0.634 ***	0.278 **
	(3.942)	(3.591)	(2.763)	(2.002)
Ino				0.312 ***
				(7.423)
FDI		0.658 ***	0.687 **	0.586 ***
		(3.542)	(2.651)	(3.762)
Urban		−0.156	−0.450 ***	−0.075
		(−1.280)	(−2.522)	(0.205)
Market		0.016 ***	0.015 **	0.011 ***
		(3.583)	(2.271)	(2.746)
Gov		0.071	−0.088	0.136 *
		(1.031)	(−0.852)	(1.811)
N	300	300	300	300
R-squared	0.975	0.980	0.944	0.984

t-statistics in parentheses; ***, **, and * indicate p < 0.01, p < 0.05, and p < 0.1, respectively; Dige: The level of digital economy’s development; Ino: technological innovation; FDI: foreign direct investment; Urban: urbanization level; Market: marketization index; Gov: involvement of government.

depicts the influence of the digital economy on the growth of dual circulation. As can be seen in Table 2, the coefficient of Dige is strongly positive, demonstrating that the digital economy has played a role in advancing dual circulation; this supports Hypothesis 1. For the control variables, the estimated coefficient of FDI is considerably positive, indicating that a 1% rise in foreign capital investment facilitates the growth of dual circulation by 0.658%. The coefficient of the Market is positive, passing the significance test at the 99% level. This shows that a greater degree of marketization helps distribute market resources, influencing the growth of dual circulation. The redistribution of resources is driven by the promotion of marketization, which encourages the expansion of dual circulation.

4.2. Analysis of the Mediating Effect

We found the mechanism through which the digital economy affects dual circulation using the estimation of models (2) and (3). The findings of the mediating mechanism regression are shown in Table 2. The coefficient value of Dige is positive and statistically significant, proving that the digital economy can enhance the potential for technological innovation. By adding the mediating factors to model (2), the coefficient of Dige is 0.278, and the significance level is over 5%. Additionally, the coefficient of Dige has declined markedly compared with that in column (2). In conclusion, technical innovation is a mechanism for the indirect influence of the digital economy on dual circulation; this validates Hypothesis 2.

4.3. Analysis of the Spatial Model

Before undertaking the spatial econometric estimation, the spatial autocorrelation test was conducted to assess the degree of digital economy and dual circulation development, and component analysis, and

Table 3. The Moran’s I index of variables.

Year	CCD		Dige
	Moran’s I	Z-Statistics	Moran’s I	Z-Statistics
2011	0.244 ***	2.915	0.230 ***	2.994
2012	0.245 ***	2.917	0.242 ***	3.161
2013	0.318 ***	3.675	0.206 ***	2.736
2014	0.335 ***	3.857	0.187 ***	2.573
2015	0.353 ***	4.047	0.163 ***	2.302
2016	0.311 ***	3.609	0.184 ***	2.585
2017	0.361 ***	4.131	0.153 ***	2.198
2018	0.347 ***	3.994	0.152 ***	2.146
2019	0.356 ***	3.999	0.151 ***	2.132
2020	0.359 ***	3.997	0.154 ***	2.153

*** indicate p < 0.01, respectively; CCD: The degree of dual circulation; Dige: The level of digital economy’s development.

shows the Moran’s I index of variables. As indicated in Table 3, Moran’s I index estimated the spatial effects of each year with matrix W₂. It reveals that Moran’s I indexes are greater than zero and attain 1% significance. The study demonstrates that each province’s digital economy and dual circulation development indexes exhibit a high spatial autocorrelation, which means clustering in the spatial distribution.

Figure 2 and Figure 3 show that the digital economy’s development level and dual circulation increased markedly from 2011 to 2020. The level of the digital economy in Beijing, Shanghai, Jiangsu, and Guangdong is consistently the highest in China. Figure 2 demonstrates that dual circulation development is further advanced in the eastern coastal zone. From 2011 to 2020, the level of dual circulation development in the central and western regions is also rising according to the overall development trend.

After referencing Elhorst [44], following the Lagrange Multiplier test, Spatial Durbin Model fixed effects, Hausman test, and SDM model simplified test, the SDM model with dual fixed effects in space and time was determined to be the best option. The further effect decomposition was then performed.

Table 4. Results of the spatial model.

Variables	W1	W2	W3
Dige	0.029 ***	0.044 ***	0.064 ***
	(2.595)	(3.792)	(10.410)
W×Dige	0.067 ***	0.193 **	0.145 ***
	(2.595)	(2.116)	(7.290)
$\rho$	0.449 ***	0.611 ***	0.593 ***
	(6.666)	(8.413)	(8.454)
FDI	0.816 ***	0.871 ***	0.347 **
	(5.240)	(5.452)	(2.161)
Urban	−0.170	0.042	−0.077
	(−1.587)	(0.361)	(−1.528)
Market	0.012 ***	0.009 **	0.010 **
	(3.307)	(2.512)	(2.379)
Gov	0.167 **	0.114	−0.289 ***
	(2.454)	(1.513)	(−6.736)
Direct	0.040 ***	0.053 ***	0.065 ***
	(3.153)	(3.579)	(9.855)
Indirect	0.135 ***	0.143 *	0.153 ***
	(2.939)	(1.653)	(5.024)
Total	0.176 ***	0.196 **	0.219 ***
	(3.282)	(2.018)	(6.323)
N	300	300	300
R2	0.279	0.882	0.868

Z-statistics in parentheses; ***, ** and * indicate p < 0.01, p < 0.05, and p < 0.1, respectively; Dige: The level of digital economy’s development; FDI: foreign direct investment; Urban: urbanization level; Market: marketization index; Gov: involvement of government.

details the effects of the digital economy on the degree of dual circulation development under W₁, W₂, and W₃. The spatial autoregressive coefficients of the dual circulation development index are significantly positive, as are the direct, indirect, and total effects of the digital economy on the intensity of dual circulation. This reveals that the digital economy contributes positively to the dual circulation development of the region and the surrounding regions. Hence, Hypothesis 3 holds.

4.4. Heterogeneity Test

There is variation in the degree of digital economy development and dual circulation development across regions due to the clear variances in resource endowments. Consequently, the effect of the digital economy on dual circulation development varies between regions. Thirty provincial administrative units are classified into eastern, central, and western regions in this study, and

Table 5. Results of the heterogeneity test.

Variables	Eastern Regions	Central Regions	Western Regions
	(1)	(2)	(3)
Dige	0.662 ***	0.891 ***	0.091
	(2.870)	(3.381)	(0.308)
FDI	0.988 ***	2.687 ***	0.335
	(5.194)	(3.678)	(0.416)
Urban	0.093	1.406 ***	0.594
	(0.588)	(5.191)	(1.422)
Market	0.018 ***	0.026 ***	0.004
	(3.099)	(3.922)	(0.485)
Gov	0.052	0.228 ***	−0.238
	(0.586)	(2.859)	(−1.347)
N	110	80	110
R2	0.982	0.977	0.946

t-statistics in parentheses; *** indicate p < 0.01, respectively; Dige: The level of digital economy’s development; FDI: foreign direct investment; Urban: urbanization level; Market: marketization index; Gov: involvement of government.

displays the results of the heterogeneity analysis. The coefficients of Dige in the eastern and central regions are prominent. The higher level of the digital economy and regional economic development in the eastern and central areas of China, and the greater distribution of the digital economy dividend, may account for this outcome.

4.5. Robustness Tests

4.5.1. Change of Explanatory Variables

Indicators of the digital economy comprise mainly coverage (Cover) and the usage of digital finance (Usage). Coverage evaluates the digital financial environment of a region. In contrast, the digital financial usage reflects mainly the region’s digital financial business service capability. Therefore, Cover and Usage are substituted for the primary explanatory variables to verify the validity of the conclusions.

Table 6. Results of changing the explanatory variables.

Variables	CCD (1)	CCD (2)
Cover	4.668 ***
	(5.249)
Usage		5.167 ***
		(3.573)
FDI	0.890 ***	0.882 ***
	(5.883)	(5.630)
Urban	−0.134	−0.144
	(−1.212)	(−1.228)
Market	0.017 ***	0.013 ***
	(3.947)	(2.925)
Gov	0.086	0.086
	(1.104)	(1.071)
N	300	300
R2	0.981	0.980

t-statistics in parentheses; *** indicate p < 0.01, respectively; FDI: foreign direct investment; Urban: urbanization level; Market: marketization index; Gov: involvement of government.

depicts the findings of the regression. After substituting the core explanatory variables, the positive and negative signs and significance of the values of the explanatory variable coefficients are typically similar to those obtained in the previous tests, demonstrating the robustness of this study.

4.5.2. Use of Instrumental Variable

Referring to Huang et al. and Nunn and Qian, the number of fixed telephones per 100 persons in each province in 1984 was used as an instrumental variable of the digital economy [42,45]. Before the widespread adoption of Internet technology, the telephone was the primary real-time communication method. Therefore, the level of Internet development would be influenced by the historical level of communication infrastructure construction and local communication habits, thus meeting the relevant requirement of the instrumental variables. Moreover, as Internet technology advances rapidly, people’s demand for relatively inefficient communication methods such as landline telephones gradually declines, and the number of landline telephones has little effect on import-export trade and consumption volume. This satisfies the exclusion requirement of the instrumental variables. For endogeneity testing, the two-stage least squares approach of the fixed effects model was adopted, and the findings are shown in

Table 7. Results of the instrumental variable.

Variables	First Stage	Second Stage
	(1)	(2)
Dige		0.323 ***
		(0.097)
Instrumental variable	9.35 × 10−6 ***
	(1.27 × 10−6)
FDI	−1.093 ***	0.178
	(0.307)	(0.272)
Urban	−0.004	0.133 **
	(0.086)	(0.061)
Market	0.035 ***	0.030 ***
	(0.006)	(0.006)
Gov	0.252 ***	−0.425 ***
	(0.078)	(0.067)
Kleibergen-Paap rk Lagrange Multiplier statistics		16.810
		[0.000]
Kleibergen-Paap rk Wald F statistics		46.103
		{16.380}
R2	0.596	0.831

Z-statistics in parentheses; ***, ** indicate p < 0.01, p < 0.05, respectively; Dige: The level of digital economy’s development; FDI: foreign direct investment; Urban: urbanization level; Market: marketization index; Gov: involvement of government.

. After accounting for endogeneity, the results reveal that the digital economy still considerably influences the degree of dual circulation development. Meanwhile, Kleibergen-LM Paaprk’s statistic and Wald F statistic reject the unidentified hypothesis and the weak instrumental variable hypothesis, justifying the selection of instrumental variables.

5. Discussions and Conclusions

Achieving a balance between economic growth and the requirements of society as well as the environment is central to sustainable growth. The methods and impacts of the digital economy on dual circulation are analyzed in terms of direct, indirect, heterogeneity, transmission mechanisms, and spatial spillover effects using provincial panel data in China from 2011 to 2020. The conclusions are as follows. First, the digital economy has become a key factor in fostering dual circulation. The validity of the conclusion remains unchanged after the endogeneity test. Second, technological innovation is one of the important mechanisms for the digital economy to impact dual circulation, suggesting that technological innovation and the digital economy can be the primary drivers of dual circulation growth. Third, the spillover influence of the digital economy on dual circulation is significant. Additionally, the promotion impacts of the digital economy vary between regions—promotion effects are more pronounced in the eastern and central regions, but the influence of the digital economy in the western region is non-existent. The possible reason for this phenomenon is that the eastern and central regions enjoy more digital economy benefits from the digitized process than the western regions.

These findings demonstrate the importance of the digital economy in fostering dual circulation and sustainable growth, in which technological innovation can play a critical role. However, the digital economy’s contribution to the dual circulation of the eastern, central, and western areas varies, suggesting that a digital divide exists throughout China. The following are suggestions based on our findings.

(1)

The government must continue to enhance investments in telecommunication infrastructure and fully embrace its role as an enabler of the digital economy. Policymakers should comprehend the empirical rules of the spatial spillover effect, expedite the integration of 5G technology with traditional industries, enhance the service capacity of digital finance, and expand the scope of the digital economy to serve the real economy. Simultaneously, the structure of digital economy innovation policies should be altered to ensure its high-quality growth.

(2)

Promoting the degree of association of the digital economy should be emphasized to foster innovation in technology and dual economic circulation. The investment intensity in creative R&D in important areas such as software engineering and network information security should be increased, which can be foundational for the diffusion of the digital economy and breakthroughs. Then, related agencies must develop an inclusive and sensible regulatory policy and employ flexible management methods for emerging industries relevant to the digital economy. These measures can indirectly provide market support for the improvements in the digital economy. Additionally, boosting government public data disclosure and supply of technical standards, as well as improving rules and regulations about data ownership and usage rights are necessary to ensure a fair market for the digital economy.

(3)

Based on the development gap of each region, policymakers must implement a differentiated digital economy development strategy. According to the empirical results, a digital divide has been formed between regions in the process of information technology construction. This also demonstrates the need for adopting dynamic and diversified digital economy policies. While ensuring the continuous and stable growth of the digital economy in the eastern and central areas, appropriate policy tilts should be implemented in the western region, which is relatively undeveloped. Moreover, the diffusion of digital elements must be directed effectively to regions in the west to fill the digital economy gap between areas. Additionally, in less developed regions, we should promote inclusive digital finance, reduce transaction costs, and strengthen industrial ties with developed regions, reducing the differences and enhancing the synergies of digital economy governance among regions.

Although this study provides empirical evidence on the digital economy’s role in the growth of dual circulation, there is still room for improvement. First, this study examined provincial data, which are limited. Further research can analyze more microscopic data at the city level. Second, the mediating effect section is only from the perspective of the innovation capacity, and subsequent studies can further explore the multidimensional effects of different mechanisms on dual circulation, which can provide more practical policy implementation proposals. Finally, China’s digital economy is still dynamic and evolutive; therefore, other potential problems including the digital divide can be considered in subsequent studies.