Study on the measurement and driving mechanisms for coordinated development level of digital economy and low-carbon economy: evidence from China

Abstract

The progress of the digital economy and low-carbon economy (hereinafter “both economies”) in China currently shows a digitalization trend and decarbonization urgency, and their intrinsic connections are becoming increasingly evident. Study on coordination of both economies is of crucial importance for China. In this study, an index system was constructed to measure the development level of both economies, and the entropy method was employed to calculate the index based on the data panel of China’s 282 prefecture-level cities during 2012–2021. Then, the coordinated development level of both economies (CDL) was assessed by the coupling coordination degree model. In addition, we conduct a sensitivity test on the weight setting of the composite index in the coupling coordination degree model to verify the robustness of CDL measurement to alternative weight specifications. Results indicate that CDL has steadily improved annually and is generally in the basic coordination stage. Cities in the basic maladjustment stage and the basic coordination stage still show considerable room for further improvement. The CDL across China’s four regions exhibits a pattern of “Eastern > Northeastern > Central > Western”. Next, the regional differences of CDL were decomposed by the Dagum Gini coefficient, revealing a decreasing trend in overall differences. Within-regional differences is the primary source of regional differences. More specifically, within-regional disparities rank from high to low as Western, Eastern, Northeastern, and Central; and the between-regional disparities, from high to low, are Western–Northeastern, Eastern–Western, Central–Western, Eastern–Central, Eastern–Northeastern, and Central–Northeastern. Accordingly, further empirical analysis using the spatial Durbin model identified several driving mechanisms for CDL. In the spatial econometric setting, we further perform robustness checks with alternative spatial weight matrices, including a geographic adjacency matrix, an economic distance matrix, and a gravity-model nested matrix, and we also report the decomposition of spatial effects. Moreover, we introduce a one-period lag of the dependent variable to estimate a dynamic spatial Durbin model, capturing path dependence in CDL and testing the robustness of the conclusions. Results demonstrate that government guidance, market regulation, technological innovation, and structural optimization mechanisms all promote CDL significantly, while the openness mechanism has a significant inhibiting effect. The direct effects are consistent with the above results, while the indirect effects indicate positive spatial spillovers from government guidance and market regulation, and a significant negative spatial spillover from openness; these findings remain broadly stable after replacing spatial weight matrices and after introducing the dynamic term, suggesting strong robustness in identifying the driving mechanisms. These findings can be interpreted through the technology–institution–structure analysis framework. Specifically, technological integration highlights digital technology empowering low-carbon technology innovation; institutional guarantee emphasizes synergy between government policies and market mechanisms; and structural transformation underscores the synergistic upgrading of industrial digitalization and low-carbonization.

Introduction

As highlighted in the 14th Five-Year Plan for Developing Digital Economy issued in January 2022 by the State Council that “as a key manifestation of the nation’s comprehensive strength in the digital era, digital economy serves as a crucial impetus in building of a modern economic system.” With the rapid development and wide application of information technology, data has become a new factor of production, driving the optimization of economic structure and the transformation of growth mode. The digital economy, through efficient information processing and transmission, promotes the optimization of resource allocation, the improvement of production efficiency, and the continuous emergence of new business forms and models. It is an inevitable product that meets the requirements of the information age. Simultaneously, the Guiding Opinions of the State Council on Accelerating the Establishment and Improvement of the Green, Low-Carbon, and Circular Development Economic System emphasize the importance of “fostering development of high quality and protection of high level, constructing and enhancing an economic system that is green, low-carbon, and circular, realizing carbon peak and carbon neutrality goals, as well as propelling China’s green progress towards a higher plane.” Facing both global climatic change and increasing constraints on resources and environment, the low-carbon economy necessitates minimizing fossil energy utilization and greenhouse gas emissions during economic activities, while promoting energy structure adjustments and industrial upgrades. As an economic pattern featured by low energy consumption, minimal pollution, and reduced emissions, the low-carbon economy becomes an essential choice for achieving sustainable development. Clearly, the digital economy is a specific form at a certain phase of socio-economic development, while the low-carbon economy is an inevitable requirement to a certain extent of socio-economic development. Currently, both the digital economy and low-carbon economy (hereinafter referred to as “both economies”) in China are advancing rapidly; economic digitization is a prevailing trend, and economic decarbonization is an unavoidable direction. The intrinsic connections and mutual reinforcement between these two spheres are becoming increasingly prominent. On one hand, both economies develop with heavy reliance on innovation of technology. Digital technology’s being extensively applied provides substantial support for researching, developing and implementing low-carbon technologies, while the demand for low-carbon solutions further drives innovation in digital technology. On the other hand, the integration of both economies has led to new emerging industries and business models, such as green-energy networks, intelligent transportation systems, and smart cities. These developments facilitate a shift in the pattern of economic growth, as well as contribute to optimization and upgrading of economic structures. Therefore, exploring the integration of both economies not only aligns with the objective requirements of contemporary development but also represents a crucial path for advancing China’s economy steps towards a phase of high-quality development, having attracted sustained attention and wide discussion from academic circles.

The existing literature primarily examines how the digital economy supports regional development within a low-carbon context and how it facilitates both transformation and development of a low-carbon economy. Research has shown that, as a new economic paradigm with digitization as the focus, the digital economy has become a crucial driver for development in a low-carbon manner [1]. With the advancement of the green-development concept, by introducing the objective of “carbon peak and carbon neutrality”, the digital economy could leverage digital technologies to boost green total factor productivity [2], support the strategy for rural revitalization [3], and promote distinctive industries in ethnic regions to develop in a high-quality manner [4]. Additionally, the digital economy can effectively restrain regional carbon emission intensity and significantly promote regional low-carbon transformation [5, 6] through mechanisms such as green technology innovation, optimization of resource allocation efficiency, industrial structure upgrading, local government economic competition, and increased public environmental awareness [7–9]. However, the existing studies lack a thorough exploration of the coordination between both economies and fail to uncover the driving mechanisms behind this coordination. In light of this, a comprehensive measurement index system is developed for evaluation of both economies in this study. Methods including the entropy method, the coupling coordination degree model, and the Dagum Gini coefficient are employed for measuring the coordinated development level of both economies (CDL), and further analyzing the overall and regional characteristics of CDL and decomposing the magnitude and sources of regional differences. It finally addresses the driving mechanisms behind the CDL, and empirical testing of these driving mechanisms is conducted using the spatial Durbin model (SDM). The aim is to enrich studies on both measurement characteristics and driving mechanisms for CDL, offering valuable references and insights for future studies.

Based on an in-depth study on the CDL, three significant contributions were made in this paper. Firstly, an index system was developed for measurement of the respective development level of both economies, drawing from existing literature. This system encompasses the core components of both economies, providing a robust theoretical framework and quantitative tools for accurately assessing their coordinated development. Secondly, various methods were used, including the entropy method, the coupling coordination degree model, and the Dagum Gini coefficient, to perform a multi-dimensional and multi-level empirical analysis of the CDL, enhancing study results’ accuracy and reliability. In particular, by decomposing regional disparities using the Dagum Gini coefficient, this study reveals the regional heterogeneity in the CDL, thereby providing a scientific basis for formulating differentiated regional development policies. Finally, the driving mechanisms behind the CDL were explored in a deep way, utilizing the SDM for empirical testing. This study not only addresses a gap in existing research concerning the driving mechanisms of CDL but also highlights the crucial role played by factors like policy environment, technological innovation, as well as industrial structure in fostering their alignment, which is significant for not only understanding the internal dynamics of both economies but also guiding practical policy-making.

Literature review

Digital economy

With rapid advancement in the digital economy, scholars have made significant progress in researching its connotation, measurement, driving effects, and pathways for realization. First, in China, the term “digital economy” was first mentioned by President Xi Jinping at the G20 Hangzhou Summit in 2016 and was formally incorporated into the Government Work Report in 2017 as a concept. Pan et al. [10] defined this economy as a novel economic paradigm characterized by digital information as the core productive element, supported by IT, and primarily carried by modern information networks. Zhang et al. [11] categorized the digital economy’s connotation into three aspects: digital foundational facilities, digital industry, and digital convergence. Based on this, Zhao et al. [12] assessed the city-level comprehensive development of the digital economy, focusing on progress of the Internet and digital financial inclusion. Additionally, the profound integration between digital and real economies is crucial for building a modern industrial system and transforming economic development methods [13]. Existing studies predominantly lay emphasis on how the digital economy drives effects on industrial structure’s upgrading [14, 15], high-quality development of foreign trade [16, 17], agricultural and rural modernization [18, 19], and common prosperity [20, 21] through mechanisms such as regional innovation capability, government governance effectiveness, information infrastructure construction, human capital accumulation, and consumption upgrades under the “digital-real integration”. Lastly, the digital economy has evolved from digital technology to digital industry and then to digital economic activities, gradually approaching maturity [22]. However, there are still issues of imbalance and polarization between regions [23]. It is required to accelerate the construction of new development advantages such as digital technology, industrial digitalization, and digital industrialization [24], and formulate targeted strategies for digital economy development [25] to better drive the economy in China to develop in a high-quality manner.

Low-carbon economy

Achieving climate-neutral and low-carbon transition is a major challenge in the 21 st century [26]. At present, scholars mainly focus on evaluating and studying the connotation measurement, development efficiency, influencing factors, and pathways of low-carbon economy. First of all, the “low-carbon economy” notion was initially proposed by the UK Department Trade, Energy White Paper: Our Energy Future – Creating a Low Carbon Economy [27]. Some scholars hold that a low-carbon economy, originating from clean energy utilization, industrial structure optimization and institutional concept innovation, means eliminating outdated production capacities with high energy consumption, pollution and emissions to achieve sustainable development in the economy and society [28]. Currently, there is still a lack of a unified measurement system for the low-carbon economy in the academic community. Zhang et al. [29] used the total factor carbon emission rate calculated through data-envelopment analysis as an indicator to assess the development level of the low-carbon economy. Wu [30] designed an index system for measuring low-carbon economic development level, with three main indices: social economy, resource flow, and energy flow. Besides, the economic growth in China exhibits a good decoupling pattern with carbon emissions, and low-carbon economy has entered a critical phase of stable development. However, because of natural endowments and past factors, its development level on the whole has a large gap between regional development levels [31]. Furthermore, studies regarding different factors’ impact on low-carbon economy mainly included green finance [32, 33], green technology innovation [34, 35], green service industry policy [36], and so on. Among them, Hu et al. [32] hold that green finance, as an emerging economic model, can effectively promote the transformation of low-carbon economy. He et al. [36] pointed out that the green service industry policy can significantly promote regional low-carbon economic growth, but the realization of this policy effect has a certain time lag. Lastly, Wang et al. [37] and Yang et al. [38] emphasized the importance of leveraging government macro-control, and suggested accelerating advancement of new energy sectors and construction of low-carbon city pilots, maximizing value of carbon assets and strengthening comprehensive, multi-level cooperation between regions to achieve sustainable socioeconomic progress.

Digital economy and low-carbon economy

The digital economy and low-carbon economy are both critical drivers for China’s high-quality along with sustainable progress under the new development concept. In the context of fully implementing the “dual carbon” strategic goals, it is essential to incorporate digital economy and Green and low-carbon development into the research framework [9]. Based on the above research, some scholars started to explore the integration of both economies. Existing literature mainly addresses how digital economy supports regional development within a low-carbon framework and how it facilitates transformation to the low-carbon economy. On one hand, with deepening of the concept of green development and introduction of the “dual carbon” strategic goals, some scholars argue that the digital economy could promote economic growth through enhancing both regional innovation and resource allocation efficiencies, thereby driving the increase in green total factor productivity [2]. Meanwhile, developing digital technology can aid the rural economy to develop in a green way [3], and provide a viable technological pathway for promoting green development of characteristic industries from ethnic areas [4]. On the other hand, digital economy’s growth can effectively empower regional low-carbon transformations. Firstly, Cao et al. [5] and Tan et al. [6] argued that digital economy development can restrain regional carbon emission intensity and promote regional low-carbon transformation significantly. However, due to differences in regional digital economy development potential and low-carbon industry development stages, digital economy’s implementation on low-carbon development shows a relatively small degree of regional heterogeneity [39]. Secondly, the digital economy facilitates regional low-carbon transformation and development through mechanisms such as optimization of resource allocation efficiency, industrial structure upgrading, local government economic competition, and increased public environmental awareness [7–9]. Nonetheless, current challenges such as limited technological innovation, inadequate talent supply, and urgent need for improving digital governance system are constraining these enabling roles of digital economy [40]. Lastly, regional carbon emission intensity exhibits spatial dependence. Scholars such as Liu et al. [41] and Zheng et al. [42] have noted that the digital economy has obvious characteristics of spatial dependence, which plays a significant spatial spill-over effect on intensity of carbon emission, though this impact being weaker compared with the local impact.

In conclusion, existing research primarily examines how the digital economy facilitates the transition to a low-carbon economy. It is widely acknowledged that the digital economy provides crucial support for this transition by optimizing resource allocation and fostering green innovation. However, existing studies still show notable limitations. Most studies focus on the one-way enabling role of the digital economy in low-carbon transition, while the internal logic of coordinated development between the digital economy and the low-carbon economy has not been systematically explained, and the driving mechanisms behind such coordination remain insufficiently explored. In particular, three gaps are prominent: an integrated theoretical framework clarifying the internal linkage and coordination mechanism between the two economies is still lacking; a measurement model that can quantify their development levels, coupling relationship, and coordination degree remains underdeveloped; and the internal driving mechanisms of their coordinated development have not been systematically revealed. Accordingly, this study aims to move beyond a one-way analytical framework by proposing a theoretical foundation for the coordinated development of both economies, measuring their development indices using the entropy method and assessing coordination with the coupling coordination degree model, while conducting a sensitivity test on the weight setting to verify robustness. Furthermore, we employ the spatial Durbin model to empirically identify the driving mechanisms of coordinated development, thereby enriching the evidence base for advancing “Digital China” and achieving the “dual carbon” goals.

As discussed above, the relationship between the digital economy and the low-carbon economy is not a simple tool–target relationship but features intrinsic coupling and coordinated development. The digital economy relies on continuous technological iteration and infrastructure construction and therefore faces issues of energy consumption and carbon emissions, which calls for low-carbon transition to support green and sustainable development. Meanwhile, low-carbon transition urgently requires digital and intelligent solutions. Their integrated development can form a virtuous cycle in which digital development drives low-carbon transition and low-carbon transition guides digital development. To systematically elucidate the internal mechanism of such coordination, drawing on technology spillover theory, synergy effect theory, institutional theory, and industrial structure theory, this study constructs an integrated technology–institution–structure analysis framework, as shown in Fig. 1.

Fig. 1.

“technology–institution–structure” theoretical framework

Specifically, the framework is developed from three dimensions, namely technological integration, digital technology empowering low-carbon technology innovation; institutional guarantee, synergy between government policies and market mechanisms; and structural transformation, synergistic upgrading of industrial digitalization and low-carbonization, and it further clarifies the interactions among these dimensions.This framework aims to reveal the internal mechanism underlying the coordinated development of the digital economy and the low-carbon economy from three interrelated dimensions, namely technological integration, institutional guarantee, and structural transformation. Specifically, the first dimension is technological integration, which refers to the synergistic mechanism through which digital technology empowers low-carbon technology innovation. Drawing on technology spillover theory and synergy effect theory, digital technology exhibits strong pervasiveness and complementarity. Prior research indicates that firms with a higher level of digitalization tend to possess more advanced production technologies and management systems. Accordingly, these firms are more willing and more capable of implementing carbon reduction measures, thereby exerting a more pronounced impact on green technology innovation [43]. Related studies also suggest that the development of network infrastructure contributes to improvements in low-carbon development [44]. The integration of the digital economy and the low-carbon economy generates significant synergy effects, which mainly operate through two channels. First, efficiency enhancement: digital technology enables precise monitoring, substantially improves resource utilization efficiency, and reduces uncertainty in low-carbon technology research and development. Second, system optimization: digital technology facilitates intelligent dispatching of energy and manufacturing systems, thereby unlocking decarbonization potential at the system level. The second dimension is institutional guarantee, which represents a regulatory and incentive framework characterized by synergy between government policies and market mechanisms. According to institutional theory, effective institutional arrangements can guide the behavior of market participants and align technological innovation directions. Such an institutional guarantee is reflected in two aspects. First, guidance and regulation: governments employ policy instruments such as carbon taxes and data governance rules to clarify social value orientations, set development directions for the integration of the two economies, and delineate necessary normative boundaries. Second, incentives and coordination: institutional design reduces the risks of integrated innovation through measures such as research and development subsidies and tax incentives. At the same time, the establishment of a carbon emissions trading market can stimulate firms’ endogenous motivation for voluntary emissions reduction. Moreover, cross-departmental coordination mechanisms help address silo problems during transition, thereby promoting knowledge diffusion and collaboration. The third dimension is structural transformation, which refers to the synergistic upgrading of industrial digitalization and low-carbonization. Based on industrial structure theory and path dependence theory, structural transformation is a process featuring both lock-in effects and opportunities for unlocking. This dimension is embodied in three aspects. First, industrial chain restructuring: digitalization drives intelligent production processes, while low-carbonization requires cleaner energy use, and their synergy jointly facilitates industrial chain restructuring. Second, structural reshaping of the energy system: digital technology improves supply–demand matching, mitigates the intermittency of renewable energy, and promotes the transition of the energy system toward a greener and smarter direction, thereby enabling structural reshaping of the energy system. Third, orderly factor flows and optimal allocation: the transition process is accompanied by the orderly flow and reallocation of factors such as capital and data from traditional high-carbon sectors toward greener sectors, improving overall allocation efficiency.

The interactions among the three dimensions can be summarized as follows. First, the depth and breadth of technological integration depend on the standards and incentives provided by institutional guarantee, and technological integration is also the most dynamic force driving industrial transformation. Second, institutional guarantee not only provides application scenarios for technological integration and sets rules and pathways for structural transformation, but is also reshaped by governance challenges arising from technological change and by shifts in the interest structure during industrial transformation. Third, structural transformation is the concrete manifestation of the outcomes of technological integration, and its transformation needs in turn stimulate further technological integration. Meanwhile, the depth of structural transformation relies on the safeguarding mechanisms provided by institutional guarantee. Therefore, the technology–institution–structure theoretical framework constitutes a dynamic model. Overall, institutional guarantee provides incentives and rules for technological integration, thereby fostering integrated innovation. These emerging technologies are applied under institutional guidance and further drive industrial restructuring. The new industries and emissions reduction benefits generated by successful transformation, in turn, strengthen societal support for institutional reform and technological investment, forming a virtuous cycle. In this sense, the coordinated development of the digital economy and the low-carbon economy is jointly shaped by technological integration, institutional guarantee, and structural transformation. This theoretical framework emphasizes promoting the creative integration of digital technology and low-carbon technology, building an incentive-compatible system of institutional guarantee, and guiding the synergistic upgrading of industrial digitalization and low-carbonization, thereby ultimately realizing the coordinated development of the two economies. It also provides a theoretical foundation for the subsequent construction of the measurement system, empirical identification of driving mechanisms, and the design of differentiated policy implications in this study.

Materials and methods

Construction of index system for measuring coordinated development of both economies

Digital economy

Given the rich connotations of the digital economy that are difficult to measure with a single indicator, this study refers to the digital economy evaluation index system developed by Zhao et al. [12]. Considering data availability at the prefecture-level city scale, the digital economy is evaluated from four dimensions, namely digital infrastructure, digital talent development, digital product output, and the development of digital inclusive finance. The particular index system is presented in Table 1, and the digital economy development level index (DE) is obtained through the entropy method.

Table 1.

Index system for measuring coordinated development level of both economies

Target level	Functional level	Indicator level	Unit	Direction
Digital economy (DE)	Digital infrastructure	Internet users per 100 people	per 100 people	Positive
Digital economy (DE)	Digital infrastructure	Mobile phone users per 100 people	per 100 people	Positive
Digital economy (DE)	Digital infrastructure	Artificial intelligence development level	index	Positive
Digital economy (DE)	Digital talent development	Share of employees in computer services and software	percent	Positive
Digital economy (DE)	Digital product output	Number of digital economy patents	count	Positive
Digital economy (DE)	Digital product output	Total telecom business per capita	CNY per person	Positive
Digital economy (DE)	Inclusive development of digital finance	Digital Financial Inclusion Index	index	Positive
Low-carbon economy (LE)	Social economy	GDP per capita	CNY per person	Positive
Low-carbon economy (LE)	Social economy	Disposable income per capita	CNY per person	Positive
Low-carbon economy (LE)	Social economy	Urbanization rate	percent	Positive
Low-carbon economy (LE)	Social economy	Share of tertiary industry in GDP	percent	Positive
Low-carbon economy (LE)	Ecological environment	Green coverage rate of built-up areas	percent	Positive
Low-carbon economy (LE)	Ecological environment	Park green area per capita	square meters per person	Positive
Low-carbon economy (LE)	Ecological environment	Harmless treatment rate of household waste	percent	Positive
Low-carbon economy (LE)	Ecological environment	Wastewater treatment plant capacity	10,000 cubic meters per day	Positive
Low-carbon economy (LE)	Consumption and emissions	Energy consumption per unit industrial value-added	tons of standard coal per 10,000 CNY	Negative
Low-carbon economy (LE)	Consumption and emissions	Carbon emission intensity	tons per 10,000 CNY	Negative
Low-carbon economy (LE)	Consumption and emissions	Industrial smoke and dust emissions	tons	Negative
Low-carbon economy (LE)	Consumption and emissions	Share of renewable energy	percent	Positive

Low-carbon economy

Making reference from the low-carbon economy development level measurement indicator systems developed by Gao [45] and Wu [30], and based on the accessibility of data at the prefecture-level city scale, a low-carbon economy evaluation indicator framework is established from three aspects: social economy, ecological environment, and consumption and emissions. As depicted on Table 1, the entropy method is employed to calculate the low-carbon economy development level index (LE).

Methods

Entropy method

As a common objective weighting method for comprehensive evaluation, the entropy method was used to determine the relative importance of multiple indicators. This method assesses the quantity of information conveyed by every specific indicator to ascertain its weight, thereby eliminating subjective errors introduced by human factors. This study uses the entropy method for comprehensively evaluating both economies’ development levels of China, and the specific calculation formulas are given below:

First, standardize the indicators.

If is a positive indicator:

1

If is a negative indicator:

2

In Eqs. (1) and (2), is the original data of item of the prefecture-level city, while is the standardized value for each indicator. To mitigate the influence of outliers and ensure the robustness of the results, all continuous variables are winsorized at the 1% and 99%.

Second, calculate the index weight.

3

4

5

6

In the above Eqs. (3)-(6), represents the proportion of indicators after standard, represents index quantity, represents indicator’s entropy, represents the redundancy of information entropy, and represents indicator’s weight.

Third, calculate the index comprehensive score.

7

8

In Eqs. (7) and (8), DE is the digital economy development level index, LE is the low-carbon economy development level index, with DE (LE) . A higher value of these indices respectively suggests a greater degree of development for both economies within the region.

Coupling coordination degree model

The level of coordination in the interaction between systems is referred to as the coupling coordination degree. This study follows the approaches of scholars like Li et al. [46] and Yang et al. [47], using the coupling coordination degree model to determine the CDL in China. This model can reflect the dynamic and balanced development state between both economies, representing the overall coordination of system interactions. The following is the calculation formulas:

9

10

11

In the above Eqs. (9)-(11), represents the all-round coordination index between both economies; and are coefficients of both economies () to be determined, measuring the relative importance of both economies to the whole system’s coupling synergies respectively. Considering that both economies are of equal importance, the undetermined coefficients are determined as 0.5. represents the coupling degree (), while an increased value of indicates a greater coordination relationship between both economies; CDL is the coupling coordination degree between both economies, and CDL . A higher CDL value implies a greater extent of coordination between both economies.

Considering that the weight setting of the composite index T may affect the calculation of the coupling coordination degree, this study conducts a weight-sensitivity test based on the baseline specification following the expert suggestions. Specifically, the composite index in Eq. (9) is still expressed as T = αDE + βLE. Keeping other settings unchanged, we adjust the weights to α = 0.4 and β = 0.6 and recalculate the coupling coordination degree to obtain a new index, D46. The results show that the correlation coefficient between the baseline index D55 and the reweighted index D46 is 0.997 and is significant at the 1% level. The mean difference between the two indices is close to zero, and their temporal evolution trends and spatial distribution patterns are highly consistent. Moreover, after replacing the baseline index with D46 as the dependent variable and re-estimating the driving-mechanism model, the signs and significance levels of the core explanatory variables remain largely stable. These results indicate that the coupling coordination degree between the digital economy and the low-carbon economy is not sensitive to weight settings, supporting the robustness of our conclusions.

According to coupling coordination levels classified by Deng et al. [48] and Song et al. [49], the CDL is divided into five types (as indicated in Table 2).

Table 2.

Criteria for degree of coupling coordination

Degree	Type
	Severe disorder
	Basic disorder
	Basic coordination
	Moderate coordination
	High coordination

Dagum Gini coefficient

As a method for measuring the evolutionary trend of regional differences, the Dagum Gini coefficient is capable of avoiding data overlap and better analyze the main sources for regional differences. Specifically, the Gini coefficient (G) consists of three parts: within-region Gini coefficient (G_w), between-region Gini coefficient (G_b), and super-variable density (G_t), where G_t refers to the influence of inter-group interaction differences on the overall differences. The following is the calculation formulas:

12

13

14

15

In Eqs. (12)-(15), means the number of divided regions, means the city quantity at prefecture level, () refers to the coordination degree between both economies of () prefecture cities in region (), while () refers to the total of prefecture cities of region (). refers to the mean in all cities at prefecture level. refers to the Gini coefficient of region , refers to the Gini coefficient between regions and . () refers to the mean of prefecture-level cities in region ().

The Gini coefficient has a value between 0 and 1, and when the Gini coefficient is smaller, the lesser the difference there is. The relationship of G with G_w, G_b and G_t satisfies:

16

Spatial autocorrelation model

As a key prerequisite for spatial econometric analysis, there shall be spatial autocorrelation among variables. Conduct a global correlation analysis on the CDL to verify whether there is spatial autocorrelation from an overall perspective. This study uses the global Moran’s I index to measure this kind of spatial correlation between regions. The specific computing formula is given below:

17

In Eq. (17), is the average CDL in cities at prefecture level, is the sample variance, () means the CDL of the prefecture city (), and is the matrix of spatial weight. The Moran index I is . As the absolute value of I increases, the spatial correlation becomes stronger.

Spatial econometric model

Spatial econometric models are mainly categorized into three types: spatial lag model (SLM), spatial error model (SEM) and spatial Durbin model (SDM). The SDM is chosen to be established for empirical test based on Wald and LR tests. In order to examine the spatial spill-over effect of relevant variables on CDL in China, it is necessary to take into account not only the impact of explanatory variables within regions, but also the impact of explanatory variables within neighboring regions. And the specific formula for SDM is as follows:

18

In Eq. (18), means a city and means the year, means a dependent variable, while means an explanatory variable. means the spatial autoregressive coefficient, mainly reflecting the spatial spillover effect. And means the estimated coefficient of explanatory variable, means spatial lag coefficient, represents spatial fixed effect, while represents random error term. Specifically:

(1) Dependent variables. The dependent variable of this paper refers to the coupling coordination degree (CDL), calculated by the coupling coordination degree model and reflects the coordinated development level of both economies (CDL) in China.

(2) Explanatory variables. Explanatory variables in this paper include: Government Size (), calculated using the ratio of fiscal expenditure to regional GDP; Marketization Degree (), calculated using the marketization index based on the index system established by Fan et al. [50]; Regional Innovation Level (), calculated using the proportion of science and technology expenditure to fiscal expenditure; Openness Level (), calculated using the ratio of actual utilized foreign investment to urban area; Industrial Structure Upgrading (), calculated using the ratio of the value-added of the tertiary industry to that of the secondary industry.

(3) Spatial weight matrix. A binary matrix of spatial weight () was constructed in accordance with spatial econometrics theory, considering whether regions have a common geographic border (being neighboring regions). If regions and are neighboring regions, the element of matrix , indicating a spatial effect between neighboring regions; otherwise, , indicating no spatial effect between neighboring regions.Considering that cities are connected not only by geographical proximity but also by heterogeneous economic linkages, this study further constructs an economic distance matrix and a gravity-model nested matrix based on the baseline geographic adjacency matrix to characterize multidimensional spatial associations.

First, the economic distance matrix, . To reflect tighter economic linkages between regions with similar levels of economic development, is constructed using comprehensive indicators of economic scale such as GDP per capita at the prefecture-level city. Specifically, we first compute the distance between any two cities in terms of the economic indicator, then take its inverse and apply row standardization. Cities with closer economic development levels have larger values of , whereas cities with larger economic differences receive smaller weights. This matrix emphasizes economic proximity and complements geographical proximity.

Second, the gravity-model nested matrix, . Based on geographical distance and economic scale, a comprehensive weight matrix is constructed following the gravity-model idea. The weight is proportional to the product of the economic scales of two cities, such as GDP, and inversely proportional to the square of their geographical distance, and the resulting matrix is row-standardized. This specification retains the distance-decay effect of geographical proximity while capturing the strength of economic linkages, thereby providing a more comprehensive characterization of the nested geographic–economic spatial association structure.

Spatial-matrix robustness design. Given that spatial econometric results may depend on the specification of the spatial weight matrix, the geographic adjacency matrix is used as the baseline matrix in the spatial regression. We then replace it with and to estimate the spatial Durbin model and compare the direction and significance of the key coefficients across the three matrices. The empirical results show that is significantly positive at the 1% level under the geographic matrix, the economic distance matrix, and the gravity-model nested matrix. Moreover, the estimates of the main explanatory variables and their spatial lags are highly consistent across the three specifications, indicating that our conclusions regarding the spatial effects of coordinated development between the digital economy and the low-carbon economy do not rely on any specific spatial matrix setting and are robust to alternative spatial weight matrices.

Data sources

This study selects 282 prefecture-level cities (including sub-provincial cities) in China from 2012 to 2021 as the research sample, excluding Tibet, Hong Kong, Macao, and Taiwan, resulting in a total of 2,820 observations. The data primarily come from the China Statistical Yearbook, the China Energy Statistical Yearbook, the China City Statistical Yearbook, as well as the statistical yearbooks and bulletins of respective cities, the website of the National Bureau of Statistics, and official municipal government websites. Data for the Digital Financial Inclusion Index at the city level are sourced from the Peking University Digital Financial Inclusion Index. To mitigate the influence of extreme values, all continuous variables at the city level were winsorized at the 1 st and 99th percentiles. Missing data were supplemented using linear interpolation.

Measurement analysis for coordinated development level of both economies

Overall evolution of coordinated development level of both economies

The evolution trends of DE, LE, and DL in China over the span of 2012–2021 are depicted in Fig. 2. It is evident that the LE demonstrates a consistently more stable growth trajectory in comparison to the DE. The DE overall increased, leaping from 0.211 to 0.331 in 2012–2014, with an increase rate reaching 56.5%. Such a phenomenon is caused by the wide-ranging implementation of big data technologies in diverse industries including finance, healthcare, retail, and logistics around 2013. The emergence of big data analysis platforms and the establishment of big data companies led many media to call 2013 the “Year of Big Data”. Big data technology was a crucial foundation for the digital economy’s development leap during this period. Additionally, the CDL climbed from 0.333 in 2012 to 0.441 in 2021, reaching a basic coordinated state by 2021. And by 2017, the CDL transitioned from a state of basic disorder to basic coordination on the whole, showing a positive development trend. Therefore, vigorously promoting the positive interaction of both economies is crucial for attaining a high coordination level for them.

Fig. 2.

Evolution trends of both economies and their coupling coordination degree in China

Figure 3 illustrates the quantitative development of the coupling coordination type between both economies in China from 2012 to 2021. The x-axis denotes the year, and the y-axis denotes the number of cities. Evidently, a highly coordinated city, Shanghai, began to emerge since 2015 through statistical analysis. By 2021, there were three highly coordinated cities: Shanghai, Guangzhou, and Shenzhen. Meanwhile, although the number of moderately coordinated cities has been steadily increasing during the sample period, it has not exceeded 20 by 2021. Additionally, since 2018, cities like Qujing, Zhaotong, and Pu’er in Yunnan Province have reached the basic coordination level, and all sample cities have moved out of the severely disordered state. Moreover, in 2012, the number of basically disordered cities was the highest, accounting for 81.2% of the sample cities. From 2012 to 2021, the number decreased from 229 to 90, with the proportion falling to 31.9%. Conversely, the number of basically coordinated cities increased from 34 to 180, accounting for 63.8% of the sample by 2021 and surpassing half. It can be seen that China’s basically disordered and basically coordinated cities have great potential for improvement in CDL. They are expected to enter the basically coordinated and moderately coordinated states on a large scale respectively in the future, and the number of highly coordinated cities will increase steadily (Fig. 3).

Fig. 3.

Evolution of the number of coupling coordination types between the digital economy and the low-carbon economy in China’s prefecture-level cities

Regional evolution of coordinated development level of both economies

From 2012 to 2021, the CDL in China has demonstrated a continuous upward tendency. However, there are considerable area-based variations in their coupling coordination due to differences in the advancement levels and speeds of both economies. According to the categorization of the National Bureau of Statistics, the 282 selected prefecture-level cities are divided into four major regions: Eastern, Central, Western, and Northeastern.

According to Fig. 4, from 2012 to 2021, the CDL varies significantly by region, generally showing a decreasing feature of “Eastern - Northeastern - Central - Western”, in which this level in the Eastern and Northeastern regions is much higher than that in the Central and Western regions. However, all four regions are showing an upward trend in coordination levels with similar patterns of change. Notably, although the CDL in the Central region is slightly better than that in the Western region, there is still a significant evolutionary difference in the coupling coordination types between these two regions (as shown in Fig. 5). In the Central region from 2012 to 2021, there were no severely disordered cities. However, in the initial phase of the sample, all severely disordered cities in China were concentrated in the Western region. It was not until 2018 that this region completely got out of the severely disordered state. Concurrently, the CDL in the Eastern and Northeastern regions are considerably higher than the country’s average. As for the Eastern region, by 2021, the only three highly coordinated cities were all located in this region, and 88.3% of cities were in a basic coordination state, with great potential for advancing to a moderate coordination state. And the Northeastern region experienced a shift of most cities from basic disorder to basic coordination from 2012 to 2021. However, several cities in this region have relatively unstable development, and their coupling types repeatedly change between basic coordination and moderate coordination.

Fig. 4.

Evolution trends of coupling coordination types between the digital economy and the low-carbon economy across China’s four major regions

Fig. 5.

Evolution of the number of coupling coordination types between the digital economy and the low-carbon economy in prefecture-level cities across China’s four major regions

Currently, many Central and Western cities are still experiencing a basic disorder phase, which need to strengthen the positive interaction between both economies to close the gap with the Eastern and Northeastern regions, reduce the quantity of cities in basic disorder, and reach a considerable degree of coordination. As for the Eastern and Northeastern regions, with remarkable capacity for high coordination levels, should pay more attention to improving the development levels of both economies and their synergetic interaction to achieve large-scale and high-level coordination as soon as possible.

Regional differences in coordinated development level of both economies

According to Section 4,  the CDL in China has obvious regional differences. To further study both extent and source of these differences across four regions, Dagum Gini coefficient is adopted for measuring the overall regional difference coefficient and the source for differences regarding CDL at the prefecture city level of China within 2012–2021. Table 3 shows the results.

Table 3.

Gini coefficient of coordinated development of both economies and its decomposition results

Year	Overall Gini coefficient (G)	Within-region Gini coefficient (Gw)				Between-region Gini coefficient (Gb)						Contribution rate(%)
		Eastern	Central	Western	Northeastern	Eastern-Central	Eastern-Western	Eastern-Northeastern	Central-Western	Central-Northeastern	Western-Northeastern	Within-region (Gw)	Between-region (Gb)	Super-variable density (Gt)
2012	0.129	0.090	0.073	0.183	0.091	0.10150	0.14846	0.09130	0.13412	0.09123	0.16447	23.471	47.721	28.808
2013	0.125	0.096	0.070	0.166	0.084	0.10351	0.14368	0.09348	0.12343	0.08947	0.15257	23.493	47.678	28.829
2014	0.117	0.093	0.070	0.144	0.086	0.10044	0.13005	0.09197	0.11152	0.09083	0.13691	23.511	45.765	30.724
2015	0.118	0.093	0.066	0.152	0.086	0.09785	0.13430	0.09178	0.11142	0.08714	0.14285	23.484	46.673	29.842
2016	0.118	0.095	0.068	0.148	0.082	0.10089	0.13237	0.09251	0.11287	0.08891	0.13880	23.649	45.618	30.733
2017	0.115	0.093	0.066	0.140	0.081	0.09799	0.12975	0.09074	0.10658	0.08636	0.13481	23.308	47.894	28.797
2018	0.114	0.093	0.067	0.138	0.082	0.09945	0.12782	0.09094	0.10621	0.08814	0.13274	23.383	46.418	30.199
2019	0.113	0.092	0.067	0.135	0.078	0.09876	0.12664	0.08876	0.10454	0.08601	0.12992	23.222	48.098	28.679
2020	0.112	0.090	0.068	0.133	0.080	0.09851	0.12467	0.08849	0.10396	0.08673	0.12849	23.319	47.121	29.559
2021	0.111	0.090	0.065	0.133	0.077	0.09664	0.12444	0.08764	0.10271	0.08575	0.12914	23.277	46.947	29.775
Average	0.117	0.093	0.068	0.147	0.083	0.09955	0.13222	0.09076	0.11174	0.08806	0.13907	23.412	46.993	29.595

Overall difference in coordinated development level of both economies and its sources

The overall difference of CDL is illustrated in Fig. 6. It can be seen that the G was relatively high from 2012 to 2021, but the trend of its evolution was generally decreasing. Specifically, G initially declined rapidly from 0.129 to 0.117, then increased to 0.118 and remained stable from 2014 to 2016, and finally experienced a gradual decrease to 0.111. Despite occasional slight increases, this difference has been steadily decreasing in the later samples years, indicating that regional CDL is strengthening with time passing by, and also is anticipated to have continuous improvement over the forthcoming years. Meanwhile, G_w, G_b and G_t can reflect the source of differences in CDL as contribution percentage (as shown in Fig. 7). Results show that G_b is the main source of regional difference in CDL, followed by G_t. G_w contributes the least. This suggests that narrowing the inter-regional coordination gap between digital and low-carbon economies is urgent and considering the impact of differences in interaction items on the overall difference.

Fig. 6.

Overall difference in coordinated development level of both economies and evolution trends of within-region Gini coefficient in China

Fig. 7.

Sources and evolution trends of overall Gini coefficient in coordinated development level of both economies in China

Within-regional differences in coordinated development level of both economies

Figure 6 also reflects the G_w in CDL during 2012–2021, which shows an overall decrease trend for the four regions, ranking “Western Eastern Northeastern Central”. It is noteworthy that the G_w in the Western region is significantly greater than in the other three regions, and only its difference level exceeds the national average to a large extent, indicating severe polarization and distinct non-uniform characteristics of CDL in the Western region. However, the G_w in this region has shown a noticeable decrease trend, particularly during 2012–2014, dropping from 0.183 to 0.144. This is because all severely disordered cities situated in the Western region were gradually moving towards a basically disordered state in this period. The G_w in the Northeastern has also shown a general decreasing trend but is relatively gradual with a low difference level. Additionally, the Central region has the lowest G_w, averaging 0.068 with a slight and stable decline. Finally, the Eastern region has a relatively high G_w level, being 0.090 in both 2012 and 2021, and fluctuating slightly during this period. This might be due to the presence of moderately coordinated cities like Shanghai, Guangzhou, and Shenzhen in the region, making it challenging for other cities with lower CDL to catch up with these cities’ pace of development, resulting in relatively small changes in G_w in the Eastern region.

Between-regional differences in coordinated development level of both economies

The evolution of Gb in the CDL from 2012 to 2021 (as shown in Fig. 8) generally shows a downward trend, but there is still a serious problem of inter-regional differentiation. Specifically, the ranking of these differences levels from high to low is: Western-Northeastern, Eastern-Western, Central-Western, Eastern-Central, Eastern-Northeastern, Central-Northeastern. Moreover, the Gb in the Western-Northeast, Eastern-Western, and Central-Western combinations witnessed a notable decline but remains far greater than that of other regional combinations on the whole. From 2014 to 2021, the pace of reduction in these differences slowed. It is worthy of attention that the Gb between the Central and Western regions progressively drew closer to that between the Eastern and Central regions, which indicates a favorable trend of CDL between the two regions. Additionally, the Gb of other combinations are relatively low, with small fluctuations but an overall decreasing trend from 2012 to 2021, indicating good synergistic interaction of both economies. In general, although the gap between the Western region and others has been narrowed to some extent, it remains notable, especially between the Western region and the Northeastern and Eastern regions. The large Western–Northeastern gap mainly stems from two factors. First, differences in policy responsiveness. Interprovincial coordination within the Western region is relatively weak and internal development is uneven, which makes it difficult to form scale effects and leads to lower policy implementation efficiency. By contrast, the Northeastern region has a higher urban density, and its accumulated research institutes and equipment manufacturing base make it easier to align with national green-transition projects, such as low-carbon retrofits in heavy industry. Second, inadequate digital infrastructure delays the application of low-carbon technologies in the Western region. Constraints related to geography, population density, and economic costs limit digital infrastructure coverage, thereby restricting the deeper application of digital technologies in low-carbon fields such as energy management and carbon-emissions monitoring. The Northeastern region, in contrast, has a stronger foundation for industrial digitalization, and earlier automation upgrades by large state-owned enterprises have provided data accumulation and technological-upgrading scenarios for low-carbon transition. Therefore, the Western region needs to take corresponding measures as soon as possible to narrow the gap with other regions to effectively enhance the CDL between regions.

Fig. 8.

Evolution trends of between-region Gini coefficient in coordinated development level of both economies in China

Driving mechanisms test of coordinated development level of both economies

Proposal of driving mechanisms

Given the significant regional difference in CDL of China, revealing the driving mechanisms for promoting this coordination is crucial for explaining these phenomena. The integration of both economies is intricately connected with a series of factors, including national policy support, resource market allocation, technological innovation capability, economic system openness, and industrial structure transformation. Therefore, the possible driving mechanisms are summarized as government guidance, market regulation, technological innovation, openness, and structural optimization mechanisms. Then, a spatial panel-data econometric model is constructed, which uses CDL as the dependent variable for empirically testing and provides insights into the differential characteristics and underlying patterns of CDL in China.

(1) Government guidance mechanism. The government can provide guidance for economic entities to make increased investment in both economies through policies, financial support, development platforms, and cooperation mechanisms, facilitating the application and dissemination of digital technologies in low-carbon sectors.

(2) Market regulation mechanism. Generally, higher marketization levels mean more refined trading systems, environmental policies, and technological infrastructure, guiding resources to sectors that are both highly efficient and carbon-efficient. The coordination of both economies can better achieve resource optimization through market mechanisms.

(3) Technological innovation mechanism. For the development of both economies, advancing environmental technology research and innovation can apply digital technologies to low-carbon fields, reducing carbon emissions and resource consumption while enhancing production efficiency and reducing costs.

(4) Openness mechanism. Implementing opening-up to the external can bring cutting-edge ideas, technologies, and capital to the economic development of China but also pose risks such as technological dependency, resource loss, and development imbalances. Moreover, openness might cause energy-consuming and highly polluting industries to relocate into China, imposing a challenge upon the goal of the low-carbon economy.

(5) Structure optimization mechanism. Introducing advanced digital technologies like big data and AI may boost the upgrade of conventional industries to digital formats, enhancing overall economic efficiency and competitiveness while facilitating green, low-carbon progress.

Test result analysis of driving mechanisms

By constructing a spatial adjacency matrix as the weight matrix, a spatial econometric model is established to explore the driving effects of related variables on CDL. Before making spatial econometric regression, it is necessary first of all to utilize the Moran’s I index for examining whether the CDL shows spatial autocorrelation across cities. As shown in Table 4, the Moran’s I index from 2012 to 2021 is positive and passes the significance level test of 1%, suggesting a strong positive spatial correlation in CDL among the 282 prefecture-level cities of China. Furthermore, the LM test is used to assess the spatial lag and spatial error effects of studied variables on the CDL. Both SLM and SEM significantly reject the original OLS mixed regression model at 1% level, which indicates the existence of spatial lag and spatial error effects. Therefore, the SDM is selected to research the driving effects of various variables on CDL. Subsequently, robustness tests using LR and Wald models show that SDM cannot degenerate into SLM or SEM, confirming the validity of using SDM. Finally, according to the Hausman test results, the random effects null hypothesis is rejected at 1% significance level, indicating that it is necessary to select a fixed effect model. Thus, the fixed-effect SDM is ultimately used to examine the driving mechanisms behind the CDL.

Table 4.

The global moran’s I index

I	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021
	0.336	0.326	0.325	0.324	0.312	0.317	0.316	0.315	0.318	0.308
z	9.086	8.846	8.79	8.775	8.455	8.59	8.572	8.544	8.61	8.362
P-value	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000

The spatial regression results of the driving mechanisms for CDL are as given in Table 5. As the spatial autoregressive coefficient, ρ is 0.446 and passes the test at 1%, indicating a significant positive spatial correlation in the CDL across regions once again, consistent with the Moran’s I index analysis. GOV, MAR, INN, and STR all significantly positively impact the CDL, while OPE exerts significantly negative impact. Specifically, INN contributes the most to the CDL with a regression coefficient of 0.202 and passes the test at 1%. Thus, technological innovation is dominant in CDL and a key driver of the coordination. By continuously advancing the exploration and innovation of digital and green technologies, green technologies can improve resource allocation efficiency and supply chain transparency with digital technology support, effectively cut down energy use and carbon discharge, and promote high-level coordination between both economies. The regression coefficients for MAR and STR are both 0.006 at the 1% significance level, with MAR playing a significantly positive spatial spillover effect, suggesting that market regulation mechanism and structural optimization mechanism also significantly promote the CDL, and market regulation mechanism can drive the CDL in neighboring regions. The coefficient of GOV is 0.004, passing the test at 5% significance level. This suggests that while the government guidance mechanism can significantly promote CDL, its effect is relatively small. Such limited effect may be due to the lack of coordination among different departments, leading to inconsistent policies related to both economies and affecting the coordination effect. Additionally, lacking breakthrough advancements in key technologies within the government’s guidance process hinders the depth and breadth of coordinated development.

Table 5.

Regression results and effect decomposition of Spatial Durbin model (SDM)

Variables	(1)	(2)	(3)	(4)	(5)
	Main	Wx	LR_Direct	LR_Indirect	LR_Total
	0.004**	0.003	0.004**	0.008**	0.013***
	(0.05)	(0.18)	(0.02)	(0.01)	(0.00)
	0.006***	0.006***	0.007***	0.014***	0.021***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
	0.202***	−0.091	0.206***	−0.004	0.203
	(0.00)	(0.21)	(0.00)	(0.97)	(0.10)
	−0.091***	−0.118**	−0.110***	−0.255***	−0.365***
	(0.00)	(0.02)	(0.00)	(0.00)	(0.00)
	0.006***	−0.003	0.006***	−0.001	0.005
	(0.00)	(0.20)	(0.00)	(0.87)	(0.19)
	0.446***	0.446***
	(0.00)	(0.00)
sigma2_e	0.000***	0.000***
	(0.00)	(0.00)
Observations	2,820	2,820	2,820	2,820	2,820
R-squared	0.096	0.096	0.096	0.096	0.096
Number of id	282	282	282	282	282

Significance levels are indicated as follows: *, **, and *** represent significance at 10%, 5%, and 1% respectively. Values within brackets are standard errors

The regression coefficient of −0.091 for openness (OPE) is significant at the 1% level, indicating that the openness mechanism exerts a relatively strong inhibiting effect on the coordinated development of the digital economy and the low-carbon economy, and that this inhibition features a significant negative spatial spillover effect. Several factors may account for this result. First, at the level of resource allocation, the dominance of traditional industries may crowd out integrated innovation activities. When traditional foreign-invested and export-oriented industries remain dominant, local governments’ policy resources and factor supply often tilt toward them to maintain short-term stability, thereby squeezing out digital–green integrated innovation. This is consistent with the “path dependence” argument in technological change [51], which suggests that systems tend to allocate resources along existing technological trajectories, thereby inhibiting the emergence of breakthrough integrated technologies. Second, in terms of technological effects, the expansion of digital infrastructure may entail implicit carbon costs. Although foreign-investment-driven digital infrastructure construction can improve operational efficiency, its high energy intensity—when combined with a carbon-intensive energy mix—may instead increase regional emissions. Meanwhile, to maintain investment attractiveness, some regions may strategically relax the enforcement of environmental regulations. Related studies indicate that global data center energy consumption continues to rise markedly [52], and that environmental regulation intensity may weaken under growth pressure [53]. Together, these factors can lead to an underestimation of the carbon costs of digital expansion.Third, regarding division-of-labor position and global trade, within global production networks, foreign investment may embed local economies in manufacturing or low-end service segments, providing limited incentives for full-chain green transition and digital technology transfer. Developed economies may also leverage their economic and technological advantages to impose “technology barriers”, meaning that core low-carbon and digital technologies are often retained by foreign parent companies, constraining host countries’ capacity to enhance the synergistic development of the digital and low-carbon economies. From a global value chain perspective, lead firms can lock developing economies into low value-added segments through control of core technologies and standards [54], thereby inhibiting digital–low-carbon synergy from the division-of-labor channel. In global trade, prior research shows that early-stage international investment and trade may be accompanied by the relocation of pollution-intensive industries, generating a “pollution haven” effect [55]. In addition, the EU’s Carbon Border Adjustment Mechanism may exacerbate carbon emissions in developing countries, mainly for two reasons. One is industrial relocation and carbon leakage: if the policy restricts developing economies’ exports of carbon-intensive products to the EU, related industries and capacities may shift to countries and regions with laxer environmental standards and without comparable carbon constraints. This implies “carbon transfer” rather than genuine “carbon reduction”. The other is the crowding-out of resources for green transition: for firms in developing economies seeking to remain in the EU market, the compliance costs induced by carbon border policies can be substantial, potentially consuming financial resources that would otherwise be invested in long-term abatement projects such as renewable energy deployment and energy-efficiency improvements, thereby slowing overall decarbonization. Fourth, at the level of capital attributes, a short-term return preference is inherently misaligned with long-cycle green investment. Foreign capital—especially financial capital that prioritizes near-term financial performance—often cannot match the long horizon and high-risk nature of deep digital–green integrated innovation. As a result, capital is more likely to flow to rapidly monetizable digital applications rather than long-term green low-carbon technologies. Related studies have also shown that although carbon risk is gradually being priced, short-term financial performance remains the dominant factor in investment decisions [56], which reveals the insufficient driving force for the integration and development of the digital economy and low-carbon economy from the financing side.

In the SDM, spatial effects can be decomposed into direct, indirect, and total effects. The direct effect is the immediate impact of local explanatory variables on the dependent variable in that area plus received feedback effects. The indirect effect is the impact of local explanatory variables on the dependent variables in adjacent regions. Total influence is the sum of direct and indirect influences. Table 5(3)-(5) show the decomposition of various driving factors’ spatial effect on CDL. The result indicates that all variables except have significant direct effects on local CDL with positive coefficients. Additionally, and have positive indirect effects on CDL in neighboring areas, while has a significantly negative spatial spillover effect. Notably, and have insignificant indirect effects (−0.004 and − 0.001 respectively), suggesting they boost local CDL but mildly inhibit that in neighboring areas. This might be due to the attraction of capital and innovative talent to the local area, thereby hindering the coordination in adjacent regions. Meanwhile, as structural optimization promotes the green and efficient integration of industrial digitization and low-carbonization in this region, it may force high-energy-consuming and low-efficiency industries into surrounding areas, negatively impacting the CDL. What’s more, the indirect effects of , , and on the CDL of neighboring regions are about twice as large as their direct effects on the local area. Thus, government guidance and market regulation not only advance local coordination but also effectively promote deeper coordination in neighboring regions, while openness mechanisms inhibit local coordination and have a greater suppressive effect on coordination in neighboring areas.

Given that spatial econometric estimates may be affected by the specification of the spatial weight matrix, this study further conducts spatial-matrix robustness checks based on the baseline geographic adjacency matrix to examine the sensitivity of the conclusions to alternative representations of spatial linkages. Specifically, while keeping the two-way fixed-effects spatial Durbin model specification and the set of control variables unchanged, we estimate the model not only with the 0–1 geographic adjacency matrix based on a shared border, but also with the economic distance matrix and the gravity-model nested matrix , and conduct comparative estimations. Table 6 reports the regression results under the three matrices. It can be seen that the spatial autoregressive coefficient ρ is significantly positive at the 1% level under all three matrix specifications, with values of 0.608, 0.559, and 0.492, respectively. This indicates a significant and robust positive spatial dependence in the coordinated development level between the digital economy and the low-carbon economy, meaning that coordination in neighboring regions or regions with stronger economic linkages can significantly promote local coordination. Regarding the explanatory variables, the estimated results remain broadly consistent across the three matrices. The spatial lag of marketization is significantly positive under all three specifications, suggesting that more developed market mechanisms in surrounding or connected regions can enhance local coordination through channels such as factor mobility and institutional diffusion. The spatial lag of government size is also significantly positive, indicating that government investment and public service provision may generate cross-regional spillovers. Regional innovation is significantly positive in its local term under all three matrices, whereas its spatial lag is generally insignificant, implying that innovation mainly promotes local coordination directly and its cross-regional diffusion is relatively limited. Openness shows a significantly negative local effect, and its spatial lag is significantly negative under the economic distance matrix and the gravity-model nested matrix, suggesting that when economic linkages are stronger, openness may trigger factor competition and siphoning effects, thereby inhibiting coordination in both local and connected regions. The spatial lag of industrial structure upgrading is significantly negative under the economic distance matrix and the gravity-model nested matrix, implying that under broader linkage structures, structural adjustment may be accompanied by resource competition or pressures from industrial relocation, thus generating negative spillovers to connected regions. Overall, the core conclusions remain stable after replacing the spatial weight matrix, indicating strong robustness of our identification of spatial effects.

Table 6.

Regression results under alternative Spatial weight matrices

	(1)	(2)	(3)	(4)	(5)	(6)
	Geographic adjacency matrix		Economic distance matrix		Gravity-model nested matrix
Variables	Main	Wx	Main	Wx	Main	Wx
GOV	−0.00250	0.00859**	−0.00171	0.00640*	−0.00162	0.00773**
	(−1.049)	(2.341)	(−0.683)	(1.689)	(−0.652)	(2.244)
MAR	−0.000463	0.0118***	−0.000140	0.0135***	0.000677	0.0159***
	(−0.473)	(7.528)	(−0.147)	(8.978)	(0.713)	(11.52)
INN	0.458***	0.0375	0.425***	0.00521	0.452***	−0.0726
	(9.150)	(0.227)	(8.854)	(0.0302)	(9.326)	(−0.433)
OPE	−0.127***	−0.102	−0.0593	−0.578***	−0.0836**	−0.405***
	(−3.316)	(−0.804)	(−1.632)	(−4.831)	(−2.281)	(−4.066)
STR	−0.00114	−0.00389	0.000997	−0.0135***	0.000913	−0.0155***
	(−0.585)	(−0.845)	(0.547)	(−2.915)	(0.498)	(−3.745)
rho	0.608***		0.559***		0.492***
	(18.53)		(17.95)		(17.40)
sigma2_e	0.000421***		0.000412***		0.000423***
	(37.32)		(37.27)		(37.18)
Observations	2,820	2,820	2,820	2,820	2,820	2,820
R-squared	0.211	0.211	0.217	0.217	0.203	0.203

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are z-statistics

On this basis, to further clarify the economic implications of the spatial effects, we decompose the estimated results under alternative spatial weight matrices into direct effects, indirect effects, and total effects, and the results are reported in Table 7. It can be seen that government size shows significantly positive indirect and total effects under all three matrices, whereas its direct effect is insignificant, indicating that the role of government is mainly reflected in spillover-driven promotion to neighboring or connected regions. The indirect and total effects of marketization are significantly positive across the three matrices, suggesting that market mechanisms not only promote local coordinated development but also significantly drive neighboring regions through spatial spillovers. Regional innovation exhibits a significantly positive direct effect, and its indirect effect is also significantly positive under the geographic adjacency matrix, implying the existence of geographically proximate knowledge spillovers. However, after incorporating economic linkages, this spillover effect becomes insignificant, suggesting that innovation diffusion may be constrained by absorptive capacity or economic-threshold effects. Openness has a significantly negative direct effect, and shows significantly negative indirect and total effects under the economic distance matrix and the gravity-model nested matrix, further supporting its competitive spillover characteristics. Industrial structure upgrading displays significantly negative indirect effects under the two matrices incorporating economic linkages, indicating that structural upgrading may inhibit connected regions through cross-regional resource reallocation and pressures associated with industrial relocation. Overall, the effect decomposition results corroborate the regression findings in Table 6 and further strengthen the interpretation of the driving mechanisms and spatial spillover effects in this study.

Table 7.

Decomposition of Spatial spillover effects under alternative Spatial weight matrices

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Geographic adjacency matrix			Economic distance matrix			Gravity-model nested matrix
Variables	Direct effect	Indirect effect	Total effect	Direct effect	Indirect effect	Total effect	Direct effect	Indirect effect	Total effect
GOV	−0.00209	0.0180**	0.0159**	−0.00133	0.0123*	0.0110*	−0.00106	0.0134***	0.0123**
	(−0.863)	(2.496)	(2.216)	(−0.527)	(1.901)	(1.764)	(−0.426)	(2.606)	(2.435)
MAR	1.67e-05	0.0289***	0.0289***	0.000522	0.0296***	0.0301***	0.00171*	0.0308***	0.0325***
	(0.0180)	(12.88)	(13.55)	(0.575)	(14.63)	(15.57)	(1.927)	(18.20)	(20.51)
INN	0.478***	0.783**	1.261***	0.442***	0.531	0.974***	0.467***	0.279	0.746**
	(9.954)	(2.059)	(3.255)	(9.283)	(1.506)	(2.633)	(9.600)	(0.946)	(2.378)
OPE	−0.135***	−0.423	−0.558*	−0.0914**	−1.331***	−1.422***	−0.113***	−0.831***	−0.944***
	(−3.638)	(−1.401)	(−1.824)	(−2.547)	(−5.268)	(−5.424)	(−3.094)	(−4.531)	(−4.815)
STR	−0.00132	−0.0111	−0.0125	0.000352	−0.0283***	−0.0280***	−4.88e-05	−0.0284***	−0.0284***
	(−0.712)	(−1.029)	(−1.152)	(0.202)	(−2.817)	(−2.721)	(−0.0277)	(−3.640)	(−3.502)
Observations	2,820	2,820	2,820	2,820	2,820	2,820	2,820	2,820	2,820
R-squared	0.211	0.211	0.211	0.217	0.217	0.217	0.203	0.203	0.203
Number of id	282	282	282	282	282	282	282	282	282

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are z-statistics

Robustness checks for the driving-mechanism analysis

To further examine the robustness of the baseline regression results, this study conducts robustness checks from three aspects. First, a time-robustness test is implemented by excluding the years affected by the COVID-19 pandemic. Second, robustness is examined by introducing a dynamic spatial econometric model. Third, robustness is assessed by replacing the form of the spatial weight matrix. The results of these three approaches are discussed in turn below.

1. Time-robustness test by excluding pandemic years.

Considering that the COVID-19 pandemic may affect the relationship between the explanatory variables and the dependent variable and thus interfere with the baseline estimates, we first conduct a time-dimension robustness check. Specifically, we exclude the sample years from 2020 onward, when the pandemic impact was more pronounced, and re-estimate the baseline model using the remaining observations. As reported in Table 8, the signs and significance levels of the core variables remain broadly consistent with those in the main regression. This indicates that after excluding the pandemic years, the overall conclusions remain unchanged, suggesting good time robustness of the baseline regression results.

Table 8.

Time-robustness test by excluding the pandemic years

Variables	(1)	(2)	(3)	(4)	(5)
	Main	Wx	Direct effect	Indirect effect	Total effect
GOV	−0.00317	0.00777**	−0.00282	0.0146**	0.0117*
	(−1.326)	(2.106)	(−1.163)	(2.052)	(1.665)
MAR	1.40e-05	0.0139***	0.000575	0.0337***	0.0343***
	(0.0126)	(7.228)	(0.544)	(12.14)	(12.77)
INN	0.395***	0.0905	0.414***	0.781*	1.195***
	(7.327)	(0.502)	(8.013)	(1.925)	(2.888)
OPE	−0.123***	0.0406	−0.124***	−0.0488	−0.173
	(−2.710)	(0.279)	(−2.849)	(−0.147)	(−0.515)
STR	−0.00172	−0.0114**	−0.00222	−0.0296**	−0.0319**
	(−0.795)	(−2.092)	(−1.081)	(−2.381)	(−2.556)
rho	0.595***
	(16.88)
sigma2_e	0.000415***
	(35.41)
Observations	2,538	2,538	2,538	2,538	2,538
R-squared	0.202	0.202	0.202	0.202	0.202
Number of id	282	282	282	282	282

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are z-statistics

(2) Robustness check using a dynamic spatial model.

Based on the baseline spatial Durbin model, we further conduct a robustness check using a dynamic spatial model. Specifically, we include the one-period lag of the dependent variable, CDL, in the regression equation and estimate a dynamic spatial Durbin model to capture the dynamic adjustment process and path dependence of the coordinated development level between the digital economy and the low-carbon economy.

As reported in Table 9, the coefficient on the one-period lag of the coordinated development level is significantly positive, indicating persistence and accumulation in the coordinated development between the digital economy and the low-carbon economy. More importantly, under the dynamic spatial model, the signs, significance levels, and directions of the coefficients for the core explanatory variables and their spatial lags remain broadly consistent with those in the baseline regression, without notable deviations. This suggests that after incorporating dynamic effects, the main conclusions of this study remain robust.

Table 9.

Robustness check using the dynamic Spatial Durbin model

Variables	(1)	(2)	(3)	(4)	(5)
	Main	Wx	LR_Direct	LR_Indirect	LR_Total
LCDL	0.326***	−0.0316	0.331***	0.218***	0.549***
	(19.03)	(−0.718)	(18.85)	(3.079)	(7.552)
GOV	−0.000753	0.00301	−0.000756	0.00478	0.00403
	(−0.366)	(0.940)	(−0.387)	(0.998)	(0.904)
MAR	0.000518	0.00479***	0.000772	0.00930***	0.0101***
	(0.546)	(2.825)	(0.864)	(3.473)	(3.887)
INN	0.276***	0.192	0.285***	0.591**	0.876***
	(6.109)	(1.263)	(6.468)	(2.143)	(3.136)
OPE	−0.0983***	0.0955	−0.0964***	0.0813	−0.0152
	(−2.864)	(0.794)	(−2.891)	(0.384)	(−0.0705)
STR	−0.00104	−0.000266	−0.000981	−0.00147	−0.00245
	(−0.589)	(−0.0638)	(−0.572)	(−0.203)	(−0.336)
rho	0.467***
	(12.15)
sigma2_e	0.000308***
	(35.49)
Observations	2,538	2,538	2,538	2,538	2,538
R-squared	0.831	0.831	0.831	0.831	0.831
Number of id	282	282	282	282	282

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are z-statistics

(3) Robustness check by replacing with the 0–1 adjacency matrix.

To examine whether the specification of the spatial weight matrix and model-variable selection affects the conclusions, we further replace the baseline spatial weight matrix with a 0–1 adjacency matrix based solely on geographic contiguity. Under this setting, we exclude all control variables and re-estimate the model using only the core explanatory variables. As reported in Table 10, even when spatial linkages are defined only by whether two regions share a common geographic border and no control variables are included, the estimated direct and total effects of the core explanatory variables remain broadly consistent with those in the main regression. The coefficient signs do not change, and the significance levels stay within a reasonable range. This suggests that our conclusions are not driven by a particular set of control variables or by a single spatial matrix specification.

Table 10.

Robustness check using the 0–1 adjacency matrix without control variables

Variables	(1)	(2)	(3)	(4)	(5)
	Main	Wx	LR_Direct	LR_Indirect	LR_Total
GOV	0.000142	0.00725**	0.000722	0.00957***	0.0103***
	(0.0576)	(2.370)	(0.292)	(2.809)	(2.937)
MAR	0.00725***	0.0134***	0.00824***	0.0200***	0.0283***
	(8.319)	(11.90)	(10.32)	(19.84)	(34.91)
INN	0.499***	−0.0190	0.513***	0.154	0.667***
	(9.409)	(−0.202)	(10.09)	(1.390)	(5.354)
OPE	−0.129***	−0.125*	−0.140***	−0.202**	−0.341***
	(−3.199)	(−1.865)	(−3.577)	(−2.441)	(−3.697)
STR	0.00428**	−0.00917***	0.00380**	−0.0102***	−0.00642
	(2.122)	(−2.975)	(2.005)	(−2.655)	(−1.594)
rho	0.282***
	(12.24)
sigma2_e	0.000480***
	(37.30)
Observations	2,820	2,820	2,820	2,820	2,820
R-squared	0.149	0.149	0.149	0.149	0.149
Number of id	282	282	282	282	282

*, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. Values in parentheses are z-statistics

In summary, across the three robustness checks, excluding pandemic years, introducing a dynamic spatial model, and replacing the spatial weight matrix with the 0–1 adjacency matrix, the relationships between the explanatory variables and the dependent variable remain consistent with the baseline regression results, indicating that the empirical conclusions of this study are robust.

Conclusion and suggestions

In this study, the measurement characteristics and regional differences in CDL of China are analyzed, and an empirical study on the driving mechanisms and spatial effects affecting such coordination are carried out. The findings are summarized as below: (1) Regarding the measurement results, the CDL has consistently improved over the years. As of 2021, it generally reflects a basic stage of coordination, with only a few cities reaching moderate or even high coordination, and considerable development potential for cities in the basic disorder and basic coordination stages. The CDL of the four major regions are ranked as follows: Eastern > Northeastern > Central > Western. The Eastern and Northeastern regions significantly surpass the Central and Western regions. Many cities in the Central and Western regions are still in basically disordered coordination, while those in the Eastern and Northeastern regions hold great promise to achieve large-scale high coordination. (2) Regarding the regional differences and their sources, there is a significant overall regional difference in CDL, although this difference is decreasing. Among them, the difference within regions constitutes the main source for regional difference in CDL, and the difference level ranks the Western, Eastern, Northeastern and Central regions in descending order; and the between-regional difference levels from the highest to the lowest are: Western-Northeastern, Eastern-Western, Central-Western, Eastern-Central, Eastern-Northeastern, and Central-Northeastern. (3) Regarding the driving mechanisms and their spatial effects, government guidance, market regulation, technological innovation, and structural optimization mechanisms all significantly affect CDL in a positive manner. However, openness mechanism exerts a notable inhibiting effect. Additionally, direct effects reveal that government guidance, market regulation, technological innovation, and structural optimization mechanisms all strongly promote local coordination, while the openness mechanism has a considerable inhibitory effect. Indirect effects show that government guidance and market adjustment mechanisms have significantly positive spillover effects on coordination of neighboring areas, whereas openness mechanism has significantly negative spillover effect. In the spatial econometric setting, robustness checks are conducted by comparing the geographic adjacency matrix, the economic distance matrix, and the gravity-model nested matrix, and spatial effects are further decomposed. Moreover, a dynamic spatial Durbin model is estimated by introducing the one-period lag of CDL, and the core conclusions remain broadly stable, indicating strong robustness of the identified driving mechanisms.

Based on the above conclusions, this article proposes the following suggestions:

First, develop differentiated regional development strategies.

For the Eastern region, digital technologies should be further promoted and the carbon emissions trading market should be improved. Key measures include establishing a traceable carbon-emissions monitoring platform with blockchain technologies to enhance transparency and credibility, encouraging collaboration between digital platform firms and carbon exchanges to develop carbon-asset management tools for small and medium-sized enterprises, and leveraging digital finance to innovate carbon-finance products to channel social capital into low-carbon projects.

For the Northeastern region, the potential of industrial digitalization should be activated to accelerate low-carbon transition. This can be achieved by expanding low-carbon application scenarios based on existing industrial digitalization foundations, promoting digital twin technologies in equipment manufacturing to enable virtual simulation and energy-use optimization, building regional industrial internet platforms to integrate energy data along industrial chains and facilitate cross-firm utilization of waste resources, and providing technological-upgrading subsidies and tax incentives to support the transition of traditional industrial clusters toward a low-carbon circular model.

For the Central region, technological research and innovation should be strengthened to drive the green and digital transformation of industries. Measures include using locational advantages to enhance research, development, and industrialization of green digital technologies, relying on universities and research institutes to advance key areas such as low-power chips and smart energy management, and focusing on advantaged industries such as electronic information and new energy vehicles to promote digital carbon-footprint management along supply chains.

For the Western region, green data centers should be developed to reduce digitalization costs. This requires coordinated planning of data-center construction and clean-energy utilization, including the rational layout of data-center clusters with supporting power supply and energy storage systems to reduce energy consumption and operating costs, and leveraging the “East-to-West Computing Resource Transfer” initiative to shift computing demand from the East to the West, narrowing the regional digital divide and advancing the synergy between green energy use and low-carbon computing.

Second, improve institutional arrangements that support high-quality economic development.

To strengthen the government guidance mechanism, fiscal subsidies and green credit tilt policies should be implemented to guide resources toward low-carbon digitalization, and cross-departmental coordination platforms should be established to enhance data interoperability, regulatory coordination, information sharing, and resource integration, thereby allocating public resources toward integrated “digital plus green” domains.

To improve the market regulation mechanism, market-oriented reforms in energy and digital sectors should be deepened, price formation mechanisms should be refined, and data platforms should be used to balance energy supply and demand dynamically to improve resource allocation efficiency. In parallel, the carbon emissions trading market should be accelerated by linking carbon-emissions data with digital economy platforms and using digital technologies for precise monitoring and trading.

To activate the technological innovation mechanism, support should be increased for original innovation and key technological breakthroughs, especially for applications of digital technologies in scenarios such as energy management and carbon tracking. Cross-disciplinary innovation platforms should be built to integrate research resources for digital and low-carbon technologies, promoting deeper integration between algorithm optimization and energy-efficiency improvement while reducing duplicated research investment. Cross-regional innovation collaboration mechanisms should also be strengthened to facilitate industry–university–research cooperation, knowledge sharing, and technology transfer.

To deepen the structural optimization mechanism, industrial internet and big data tools should be used to optimize processes and supply chain management so that high-energy-consuming industries can reduce emissions through intelligent upgrading. Meanwhile, new business models integrating digitalization and low-carbon development should be fostered to accelerate the construction of a green, low-carbon circular economic system.

To optimize the openness mechanism, high-standard opening-up should be balanced with autonomous low-carbon development. This includes promoting regional coordination and international cooperation in digital infrastructure and low-carbon investment, actively participating in international rule-making to prevent cross-border relocation risks of high-carbon industries, and expanding exports of green low-carbon products through digital trade.

This study reveals the spatiotemporal characteristics and driving mechanisms of the coordinated development of digital economy and low-carbon economy, but there are still limitations. Firstly, although panel data based on prefecture level cities can reflect macro trends, it fails to include county-level unit data, making it difficult to capture regional differences, and may underestimate the imbalance of coordinated development between urban and rural areas and counties. In addition, the construction of the indicator system is limited by the availability of data, and some indicators that represent the penetration of digital infrastructure or low-carbon technologies have not been fully reflected, which may affect the comprehensiveness of index measurement. Secondly, although the coupling coordination degree model can evaluate the interaction status between systems, it fails to deeply analyze the evolutionary path of collaborative development. Moreover, regarding the exploration of driving mechanisms, the research focuses on the analysis of macro driving factors, and there is insufficient exploration of micro level mechanisms such as the behavior mechanisms of micro entities such as enterprises and the transmission paths of policy interventions, which restricts the policy refinement reference value of the conclusions. Therefore, future research can be deepened in the following directions: firstly, data expansion and refinement. Collect county-level and even enterprise level data, try to integrate multi-source data, such as satellite remote sensing and Internet big data, supplement dynamic indicators such as digital application scenarios and real-time carbon emission monitoring, and improve the timeliness and explanatory power of the indicator system. Secondly, further enrich and expand the relevant research content. Firstly, by combining social network analysis, further reveal the evolutionary path of the synergy between digital economy and low-carbon economy. Secondly, deepen the research on driving mechanisms, conduct case comparisons or quasi natural experiments, and evaluate the actual effects of regional pilot policies on collaborative development. Thirdly, we should strengthen interdisciplinary perspectives and incorporate theories such as institutional economics and behavioral science to explore the multidimensional synergy mechanism between the digital economy and low-carbon economy, in order to provide a more solid empirical basis for formulating differentiated and precise regional synergy policies.

Author contributions

Deliang Zhou contributed to Conceptualization, Funding acquisition, Project administration, Supervision, Writing–original draft, and Writing – review & editing. Yuhan Song contributed to Data curation, Formal analysis, Methodology, Software, Visualization, and Writing – original draft.

Funding

This work was supported by the Western Project of National Social Science Fund of China (Grant Number: 24XGL001): “Research on the Mechanism, Pathways, and Policies for the Integration of Digital and Real Economies to Empower the Development of New Quality Productive Forces in the Yellow River Basin”.

Data availability

Data will be made available on request.

Declarations

Ethics and consent to participate declarations

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.