Impact of digital economy on urban land green use efficiency: evidence from Chinese cities

Improving urban land green use efficiency (ULGUE) is essential for achieving the sustainable use of land resources and the high-quality economic development of cities. The digital economy has overcome spatial and locational constraints, blurred the boundaries between industries, and created a collaborative and open economic model, inevitably affecting urban land utilization. However, whether the digital economy will affect ULGUE and how such an effect would take place are still unclear. This study evaluates the digital economy and ULGUE of Chinese cities from 2011 to 2019 and systematically analyzes both the direct and indirect impact mechanisms through which the digital economy affects ULGUE. The results show that the digital economy can significantly enhance ULGUE, and these results prove to be reliable, as shown by various endogeneity treatments and robustness tests. The digital economy enhances ULGUE by optimizing the industrial structure (structural effect), increasing green technology innovation (technical effect), and agglomerating digital talent (scale effect). Moreover, land finance dependency plays an adverse moderating role in the relationship between the digital economy and ULGUE. Further heterogeneity analysis shows that the promoting effect of the digital economy on ULGUE takes full effect in the eastern cities, larger cities, and cities with high levels of digital economy development and land marketization. This paper presents recommendations for supporting the balanced and integrated development of the digital economy across regions and provides differentiated development strategies to enhance ULGUE in the context of digitization.


Introduction
Accelerating green transformation and improving urban land use efficiency under the constraints of land resources and the ecological environment are inevitable choices for realizing high-quality and environmentally sustainable urban development in China (Koroso et al 2020, Jiang et al 2022b).Since the 1990s, local governments in China have accelerated the process of local industrialization and urbanization through a massive increase in the supply of urban construction land, creating an economic growth miracle (Zhao and Zhang 2018).However, the effectiveness of this model has continued to decline, and the conflict between land supply and demand, ecological degradation, and the environmental pollution brought by the uncontrolled expansion of urban land and the inefficient use of such land have become increasingly severe (Yin et al 2023).According to the third national land resource survey, the total scale of land for construction reached 408,666 km 2 in 2022, an increase of 85,333 km 2 or 26.5% compared to the figures from the second national land resource survey ten years earlier.Current research shows that for every 1% increase in China's urbanization rate during the 2006-2030 period, the construction land area required as high as 3,460 km 2 under the current development model, which is much higher than that in other developed countries (Xiong et al 2019).The characteristics of limited and scarce 2. Literature review 2.1.Literature on ULGUE In contrast to evaluating the efficiency of urban land use only in terms of economic outputs, urban land green use is a new mode that considers economic development, environmental protection, and resource conservation.In the reality of increasingly scarce urban land resources, it has been widely recognized that the mindless and inefficient mode of urban land utilization has led to adverse consequences (Liu et al 2014).So, how to improve the efficiency of urban land under the consideration of the ecological environment has become a hot spot among scholars and administrators.ULGUE refers to integrated economic, social, and environmental benefits as outputs, considering land, labor, and capital as total inputs within environmental and land-carrying capacity constraints.The UN Habitat's World Cities Report (2016) states that the fundamental principle of sustainable urban development is promoting the efficiency of urban land; it is not only the key to overcoming the critical global challenge of the contradiction between population growth and limited land supply but also important for guiding sustainable land use (Wei and Ewing 2018).
The quantitative measurement of ULGUE is the foundation for further investigation.The methods are mainly divided into two categories.One category consists of the single ratio method, in which economic outcomes per unit of urban land area are taken as land use efficiency (Lin and Hülsbergen 2017).The other category involves first establishing a system of indicators based on the definition and connotation, and then quantifying the efficiency of a specific region based on non-parametric models (Zhu et al 2019), such as the data envelopment analysis (DEA) models, which can integrate total factor inputs, economic and socially desirable outputs, and environmentally undesirable outputs during urban land utilization.He et al found the economic output value per unit area evolution with an average annual growth rate of 7.55% in China from 2000 to 2015; additionally, the internal differences in efficiency between cities were increasing (He et al 2020).Lu et al calculated that the average ULGUE value was 0.5215 from 2003 to 2016; it first decreased and then increased, with a significant spatial difference (Lu et al 2020).
Exploring the factors influencing ULGUE is another hot topic.Current studies have confirmed that economic development (Yang et al 2021), technological innovation (Song et al 2018), industrial structure upgrading (Yu et al 2019), and government intervention (Liu et al 2021b) can significantly impact ULGUE, and the impact effect varies heterogeneously across regions and cities.ULGUE is also affected by policies.For example, excessive dependence on land finance and non-market-based land transactions can place ULGUE at a disadvantage (Fang and Guo 2021, Wang et al 2021).Furthermore, ULGUE is affected by the characteristics of cities.

Overview of the digital economy
The digital economy is progressively playing a crucial role in urban development by reconfiguring production factors, reshaping the economic structure, and changing the competitive landscape (Pan et al 2022).Current research related to the digital economy has mainly concentrated on the concept, quantitative assessment, and effects of the digital economy.
Regarding the definition of the digital economy, Tapscott first introduced the concept in 1996 as a new information and communication technology (ICT)-based paradigm for the functioning of the economy and society in the age of networked intelligence.Since then, academics have presented various manifestations of the digital economy.The G20 Digital Economy Development and Cooperation Initiative defined the digital economy, taking digital data and knowledge as production elements, using information networks as carriers, and taking ICT as the leading economic form to increase efficiency and optimize the economic structure.Although a systematic definition has not yet been formed, the two main components of the digital economy regarding which a consensus has been reached are digital industrialization (expansion of the digital industry) and industrial digitization (digital transformation of traditional sectors) (Sturgeon 2021).In terms of measuring the digital economy, there are primarily two approach categories.The direct approach calculates the digital economy level in a given region following national accounting principles and the range of industries specified within regulations (Watanabe et al 2018).The other approach is the comparative method, which constructs an evaluation indicator system by selecting relevant indicators based on the concept of the digital economy and then evaluating the current status of digital economy development in an area (Gao et al 2022).Compared to the direct approach, the comparative method is more comprehensive, open, and practical since it is compatible with different measurement ideas and applicable to cross-sectional and longitudinal comparisons between time series and regions (Deng et al 2022).Current digital economy evaluation indicator systems tend to concentrate on Internet development, digital finance, digital industrialization, and industrial digitization (Allam and Jones 2021, Deng et al 2022).Regarding the impact of the digital economy, researchers primarily focus on socioeconomic and environmental aspects.According to relevant studies on the social and economic effects, the development of the digital economy can replace traditional factors (Litvinenko 2020), correct resource distortions (Gao et al 2022), improve factor efficiency and economic outputs (Ding et al 2022), optimize the industrial structure, and promote technological innovation (Liu et al 2022).The results of studies on the environment indicate that the development of the digital economy can minimize pollution (Xu et al 2022), reduce carbon emissions (Li andWang 2022, Feng et al 2023b), and improve carbon efficiency (Zhang et al 2022), and green total factor productivity in forestry (Chen et al 2023).

Studies on the nexus between the digital economy and ULGUE
Existing studies have discussed the correlation between the digital economy and urban land utilization.For example, 'the end of geography' (Graham 1998) and 'the death of distance' (Cairncross 2002) are important arguments about how the application of information technology (IT) can change the spatial structure of cities.The development of IT impacts urban land use patterns through changes in the types and locations of human activities (van Wee et al 2013).There are different views on the impact of IT on land use patterns.A study in the Chicago area suggested that IT applications will lead to decentralization, followed by regrouping in regions with better information infrastructure (Sohn et al 2002).Other research findings indicated that the widespread use of IT will lead urban land use patterns toward multicentricity (Lee et al 2014).It has also been argued that information infrastructure, as a new category of urban infrastructure added to cities (Barns et al 2017), is rapidly expanding and becoming a key element in the urban fabric and that its physical equipment, such as cables, towers, and base stations, has an impact on the urban spatial pattern (Wiig 2013).Some studies have also confirmed that the effects of the digital economy are highly compatible with the characteristics and requirements for improving ULGUE (Tan et al 2021).However, Current research examining the relationship between the digital economy and ULGUE is limited.It is essential to investigate whether and how the digital economy affects ULGUE and to understand the green value of the digital economy, which is crucial for promoting the sustainable use of urban land resources and fostering sustainable and high-quality economic growth.

Direct effect of the digital economy on ULGUE
The digital economy has a direct effect on ULGUE in the following ways: (1) It saves land production factors: Digital technology creates a virtual digital space beyond the original three-dimensional physical space, completely overcoming the constraints of time and space in terms of the act of being productive and living, weakening the limitations of factors flow due to geographical distance, and significantly reducing the reliance of urban development on urban land resources (Saker and Frith 2020).Moreover, data production factors have a substitutive effect on land and other traditional production factors while upgrading production and management modes through digital technology, making them dependent on new production factors such as data, technology, and talent, and shifting away from the original mode of 'maintaining economic growth by increasing the supply of construction land (Yang et al 2020).' (2) The digital economy optimizes resource allocation: With the widespread use of data elements and Internet platforms, the digital economy can ameliorate the fragmentation and asymmetry of information in land factor markets, thereby reducing factor search and transaction costs, preventing human-induced resource waste and encouraging the free flow of land elements among urban and rural areas, industries, and sectors (Zhu and Chen 2022).Furthermore, data elements can activate traditional production factors through innovative business models and the delivery of new products and services, thereby generating more value with fewer urban land-resource inputs.(3) The digital economy increases economic outputs: Data elements that are integrated into the manufacturing, distribution, circulation, and consumption processes can create enormous additional value due to their replicable sharing and infinite growth (Shahbaz et al 2022).Meanwhile, the new digital industries, business models, and patterns generated by the digital economy yield new supply and demand for products and services, transforming the previous economic industrial system and expanding the market scope.Integrating the digital economy with traditional industries effectively extends the industrial chain, promotes the iterative upgrading of products and services, creates economic and social values, and significantly increases the efficiency of urban land.(4) The digital economy enhances the ecological environment: Compared to the traditional industries, especially the manufacturing and processing sectors, digital industries and digitally enabled traditional industries have higher levels of effectiveness, efficiency, and productivity, as well as lower levels of energy consumption, pollution, and emissions.The environmental losses associated with urban land utilization can be reduced by optimizing the mode of environmental supervision and fostering environmental protection awareness among urban residents (Xu et al 2022).
3.2.Indirect effect of the digital economy on ULGUE 3.2.1.Mediation mechanism of industrial structure optimization Digital economy development can optimize the industrial structure.First, the digital industry drives the market to the middle and upper tiers by cutting costs, achieving economies of scale, and making more efficient use of resources (Teece 2018).Second, digital industrialization primarily refers to using digital technology to deliver digital goods and services.The new types of supply and demand transform people's production and lifestyles, significantly optimizing the industrial structure (Koch andWindsperger 2017, Heo andLee 2019).Third, industrial digitization alters traditional industries with digital information technology, as well as advanced Internet, and artificial intelligence technologies in an all-around, all-angle, and all-chain manner, speeding up the rational flow and the ideal combination of traditional resource elements, modifying the rate of return on investment, and accelerating technological advancement in traditional industries, thereby strengthening the coupled coordination between the economic structure and resource utilization, which results in industrial structure upgrading (Su et al 2021).
Urban land resources are a vital production factor and the primary carrier of the evolution of the industrial structure.Consequently, modifications to the industrial structure will inevitably be reflected in changes in urban land utilization, which will have additional effects on ULGUE (Liu et al 2022).Meanwhile, the combined strategy involves production elements and the management mode of industries linked to digital technology has the advantage of requiring fewer land factors, and improving the intensive and economical level of land resource utilization, further reducing pressures on natural resources and the environment during the use process.In addition, according to bidding-rent theory, each type of land for construction has a bidding-rent curve (Alonso 1964), and industrial structure optimization encourages the functional conversion of land components across sectors, introduces land elements into higher-returning production areas, optimizes the land allocation structure, and forms a new urban spatial layout, thereby increasing the economic and social outputs per unit area of urban land (Liu et al 2021a).(2) It contributes to improving the innovative capacity of green technology.The digital economy provides innovation talent with greater access to knowledge, reducing the cost of seeking innovative information and enhancing communication efficiency (Agrawal andGoldfarb 2008, Dana andOrlov 2014).The demand for high-tech and high-skilled human capital in digital industries forces employees to improve their innovation capability through learning (Chen et al 2022b).(3) It accelerates the diffusion and application of green innovation technologies.Knowledge and technology mobility are significantly improved in the context of the digital economy, allowing for the rapid proliferation of green innovation technologies through learning and competitive effects.Additionally, online platforms enable the rapid and accurate capture and matching of innovative technologies and market demand information, making technical cooperation and exchange more convenient, enhancing interregional and interindustry interaction and remote collaboration, reducing inefficiencies and unnecessary losses, and creating a virtuous cycle for green technology innovation (Cao et al 2021).

Mediation mechanism of green technology innovation
Green technology innovation incorporated into the production process can optimize the supply and demand structure of traditional production elements, including land, labor, and capital, thus improving the return per unit of production factors (Chen et al 2023).Utilizing green innovation technologies to enhance production and operation processes can significantly increase the marginal conversion rate of production factors and the return on investment (Ran et al 2023).Green technology innovation provides technological support for urban environmental protection supervision and the scientific management of pollution, reducing environmental losses in land use by reducing the consumption of traditional resources and reducing the pressure on ecological and environmental systems.

Mediation mechanism of digital talent agglomeration
Talent agglomeration is an economy-of-scale phenomenon.Digital economic development encourages the agglomeration of digital talent since labor, which is an essential factor of production, is highly influenced by the evolution in productivity and since it positively correlates with industrial concentration (Han et al 2019).The status of the digital economy reflects the potential for urban development and the availability of digital employment opportunities, which are vital in forming digital talent agglomeration.The proliferation of the digital economy in government administrations has improved the effectiveness of social management and services, further enhancing the attraction of digital talent (Liang and Li 2023).
Production factor agglomeration is a prerequisite for optimal resource allocation and is essential for effective economic operation.Agglomerating digital talent can increase the efficiency of allocating labor and other resources are located by better matching labor with additional elements (Ge et al 2021).Digital talent agglomeration provides intellectual support to guide the cultivation of green development models in cities and provides new strategies and models for the exploitation and utilization of urban land (Fang and Wolski 2021).Furthermore, through the learning and sharing effects of agglomeration, knowledge diffusion and the exchange and interaction of various factors are strengthened, and then, the input-output efficiency of land factors is enhanced.

Moderating mechanism of land finance dependence
Land finance dependence is depicted as the degree of regional governments' budgetary reliance on land transfers and land-related taxes (Liu et al 2018).As a derivative of the land production factor in economics, its impact on ULGUE is unquestionable (Wang et al 2021).Land-related revenue has appeared to be a vital source of financing for promoting urban development in China, playing a crucial role in affecting urban land utilization.According to statistics, in 2021, more than 60% of Chinese cities' land finance revenues accounted for more than 50% of the general public expenditure; in some cities, the percentage was as high as 80%.Under this background, local governments tend to adopt the differentiated strategy of granting industrial land at lower prices while granting commercial and residential land at higher prices to obtain higher returns, which may impair ULGUE.Excessive commercial and residential land prices tend to push the real estate industry to rapidly expand, attracting an enormous inflow of land, capital, and labor elements, which will severely crowd out the production factors needed for digital industries and other traditional industries (Lu et al 2019), resulting in distorted factor allocation, causing a decline of other industrial sectors, lowering the output of society and hindering an improvement in ULGUE.Meanwhile, local governments providing industrial land at low prices make up for the fiscal gap by attracting investment and increasing tax revenue (Tong et al 2022), thus resulting in low-quality investment attraction and an increase in nonmarket-oriented land transactions, in addition to hindering an improvement in ULGUE (Fan et al 2020).
Based on the literature review and the theoretical analysis, we draw the framework diagram below (figure 1) and propose the following hypotheses: Hypothesis 1 : The digital economy is conducive to improving ULGUE;  Hypothesis 5 : The moderating effect of land finance dependence in the digital economy negatively affects ULGUE.

Econometric methods
Considering the characteristics of the independent and dependent variables, we further performed the linear correlation, multicollinearity, homoscedasticity, and autocorrelation tests on the variables of the models, and the results suggested that the multivariate linear model was appropriate for this case.

Benchmark model
The following econometric model is used to examine whether the digital economy impacts ULGUE: where α k , kä(1,7) are the estimate coefficients; i and t refer to the city and the year; Ulgue it represents urban land green use efficiency; Dige it represents the level of the digital economy development; Controls it represents a set of control variables, including economic development level (lnpgdp it ), Government effect (Gove it ), Population density (Popd it ), Financial pressure (Finp it ), Infrastructure (Infra it ), which will be demonstrated explicitly in the following subsection; μ i is the individual fixed effect, δ t is the time fixed effect respectively; ε it indicates the random error term.

Mediating effect model
The following mediating effect model is built to explore the mechanism through which the digital economy enhances ULGUE: where Medi it is the mediating variable, and represents the industrial structure (Indus it ), green technology innovation (Ginno it ), and digital talent agglomeration (Dtat it ); The remaining variables are the same as those in equation (1).If the coefficients β 2 in equation (2) and j 3 in equation (3) are both significant and j 2 is not significant, and a complete mediating effect of the mediator variable is demonstrated; If j 2 is significant and its absolute value is less than β 2 in equation (2), the variable Medi it plays a partial mediating role in the nexus between the digital economy and ULGUE.

Moderating effect model
We also construct the following moderating effect model to verify the moderating mechanism of land finance dependency in terms of the influence of the digital economy on ULGUE: where Mode it is the moderator variable, which refers to land finance dependency (Lfin it ), and Dige it * Mode it is the interaction term of the core explanatory variable (Dige) and the moderator variable (Lfin it ); The remaining variables are the same as those in equation (1).When η 2 and η 4 in equation (5) are significant, denoting that the variable Lfin it plays a moderating role in the relationship between the digital economy and ULGUE.

Dependent variable
Urban land green use efficiency (Ulgue) is the dependent variable.Following the relevant studies (Dong et al 2020, Xie et al 2021), We designed an evaluation indicator system to measure the ULGUE, adopting the area of urban land for construction, the fixed capital stock, and the number of employees in the secondary and tertiary industries to characterize the land, capital, and labor factor inputs of land utilization, respectively; the value added of the secondary and tertiary industries and the per capita income of urban residents reflect the desired economic and social outputs; The composite pollution index consisting of emissions of industrial sulfur dioxide, industrial wastewater, and industrial nitrogen oxides reflects the undesirable output, as shown in table 1.
Referring to some related literature (Yu et al 2019, Kuang et al 2020), we chose the super-efficiency SBM-UN model, which is an improved version of the traditional DEA, to calculate the ULGUE of cities since it has the following advantages compared with the other models: (1) It is a non-radial and non-angle model, which overcomes the problem of the traditional CCR and BCC models being based on radial and angle and thus unable to account for the slack variables, resulting in a higher efficiency value than the actual value.(2) The results of the super-efficiency SBM-UN model are not constrained to the interval [0, 1]; hence, it solves the problem of order difficulty when there are multiple efficient frontier units and when such units are subject to fewer restrictions in the process of selecting the econometric model.(3) It can introduce undesirable outputs account for the negative environmental impacts caused by land use, and more accurately measure ULGUE values.Detailed calculation steps are omitted, refer to the cited literature (Zhao et al 2020).

Core explanatory variable
The digital economy (Dige) is the core explanatory variable.Based on the definition proposed by the G20, we measure the digital economy by accounting for the following four dimensions (table 2): (1) Digital infrastructure is mainly reflected by Internet applications and its carrying capacity, and it is measured by the number of Internet broadband users per 100 people and the number of mobile phone users per 100 people.(2) The digital environment is characterized by digital governance and digital innovation.Specifically, the amount of government spending on science, technology, and education, the number of digital economy patents granted per ten thousand people, the performance of government websites in terms of city governments, and the sum of undergraduate students per ten thousand people are selected.(3) Digital industrialization consists primarily of products and services enabled by digital technologies, as characterized by the percentage of employees in computer services and software, and the per capita total revenue of telecommunications and postal services.(4) The digital transformation of traditional industries is measured by digital finance and the integration of the digital economy and the entity economy, as exemplified by the digital inclusive finance index and the digital transformation index of listed companies.We chose the PCA to evaluate the level of the digital economy during the research period, which is widely used in the research of multi-indicator comprehensive measurement in various disciplines, including in the evaluation of the digital economy (Luo et al 2022).It can overcome the defects of the subjective assignment method, eliminate the correlation effect between the assessment indicators, and reduce the workload of indicator selection, also has many advantages, such as easy to calculate, objective, and fair evaluation conclusions.Prior to conducting the PCA, the data are standardized using the extremum methodology, and the Kaiser-Meyer-Olkin (KMO) results and result of Bartlett's test of sphericity are KMO = 0.79 and p < 0.000, suggesting that the PCA is a suitable methodology in this case.

Mediator variables
Industrial structure optimization (Indus).Under the broader context of IT, the servicification of the economic structure is an essential feature of the process of updating the industrial structure.Following current studies, it is represented as the ratio of the value of the tertiary industry to that of the secondary industry (Guo et al 2021).
Green technology innovation (Ginno).The number of new green patents granted can indicate a region's capacity for green innovation.There are three types of granted patents: innovations, utility models, and design patents (Ding et al 2022).We chose the number of granted green innovation patents per ten thousand people to represent the green technology innovation of a city due to the indicator's 'substantial development' and 'exceptional materiality' characteristics.
Digital talent agglomeration (Dtag).The Global Digital Talent Development Annual Report (2020) describes digital talent as individuals with sophisticated digital skills, excluding those with only rudimentary digital literacy.Based on current studies (Nikitaeva and Chunlei 2021), we first calculated the employment ratio in each city's information transmission, computer services, and software sectors as a percentage of total employment.Then, we divided this ratio by the national average to obtain the result.

Moderator variable
Land finance dependence (Lfin) is a variable used to measure how dependent the local economy is on land concession revenue (Fan et al 2020).This study selects the ratio of land premiums to general government budget revenues in cities to measure a city's reliance on land finance.Hence a higher ratio indicates a greater reliance on land finance by local governments.

Control variables
To reduce the potential of bias in the estimation results due to omitted variables, we referred to the current relevant literature and chose the following control variables: (1) Economic development (lnpgdp), is an essential, and desirable output of urban land utilization and the most intuitive reflection of ULGUE.It is expressed as the logarithm of GDP per capita (Chen et al 2022a).
(2) Government effect (Gove), is a function of the government's planning, allocation, and administration of urban land, and it significantly impacts ULGUE.It is measured by the ratio of general government spending to GDP (Pu et al 2021).
(3) Population density (Popd), manifests as a rise in the density of people, and it promotes the intensity of other production factors; the resulting scale economy could further alter the distribution and productivity of urban land resources, It is defined as the number of people divided by the size of the administrative area (Masini et al 2019, Koroso et al 2020).
(4) Financial pressure (Finp), land concession revenue primarily fills the local fiscal gap in the context of increased local fiscal expenditure.Such revenue is an essential funding source for urban development.This variable is expressed as the ratio of the gap between general public expenditure and revenue to total public revenue (Yu et al 2019, Zhu et al 2019).
(5) Infrastructure (Infra), improving urban infrastructure can attract high-quality investment, thereby enhancing land efficiency.However, the expansion of construction land and environmental loss issues may have a specific inhibiting effect on UGLUE.This variable is expressed as the road area per capita (Chen et al 2022a).

Data sources and descriptive statistics
China's 286 cities above the prefecture level were chosen as the research samples, and 2011 to 2019 was selected as the research period.The primary sources of the panel data were the China City Statistical Yearbook and the China Economic Network statistical database.Various missing statistics were added through consultation with a statistical communique of national economic and social development from provinces and cities with corresponding data.
In addition, urban land utilization data from the Landsat remote sensing image-based 30 m land cover dataset (CLCD) for China (1990China ( -2020)), which was provided by Wuhan University (http://irsip.whu.edu.cn/recent_achi/recent_show.php?16), were further obtained for each city by using the Spatial Overlay Analysis tool of ArcGIS 10.2 software.The data of 'the website performance of cities' governments' were obtained from the 'Report on the Evaluation of the Performance of Chinese Government Websites' published by the China Software Evaluation Centre of the Ministry of Industry and Information Technology, the Institute of International Governance of Tsinghua University and the School of Public Administration; 'The quantity of digital economy patents granted per ten thousand people' was derived from the China Research Data Service (CNRDS) platform.The digital finance inclusion index was obtained from the research results of Peking University (https://www.idf.pku.edu.cn/yjcg/zsbg/513800.htm).The digital transformation index for listed companies was obtained from the text analysis of the annual reports of listed companies, and transformed into city panel data in accordance with the location of the company's registered city.Table 3 reports the descriptive statistics of the variables.

Analysis of the digital economy and ULGUE
Based on the indicator systems and the evaluation methods described in the previous section, we calculated the level of digital economy development level and ULGUE of 286 cities in China from 2011 to 2019.The national and subregional dynamic evolution of the digital economy and ULGUE are shown in figure 2.
Overall, the average value of ULGUE was 0.422 during the research period, which was not considered efficient and still had much room for improvement.The average ULGUE values showed a general upward trend, with a slight fluctuation from 2013 to 2015, and significant spatial differences.Specifically, the average ULGUE value of eastern cities was 0.464, higher than the mean of cities nationwide and cities in the central and western regions by 12.19% and 22.0%, respectively.The growth rate of cities in the central and western regions reached 22.30%, while it reached 18.04% in eastern cities.This result was mainly due to the implementation of the Strategy for the Rise of Central China, the Development of Western China, and a series of environmental protection policies, which significantly stimulated the economic development of central and western cities while emphasizing the protection of natural resources and the environment, thereby increasing ULGUE.
The digital economy, with an average yearly growth rate of 10.44% from 2011 to 2019 showed an upward trend, suggesting that China's digital economy has experienced dramatic growth and has become the most important contributing factor to China's economic development.However, there was a significant digital divide among the regions.With extremely high digital economy levels, eastern cities consistently held the leading position among regions.Although the level in central and western cities continued to show substantial expansion momentum during the research period, it was still lower than that of eastern cities and the nationwide average value, and the digital divide has yet to demonstrate a convergence trend.Thus, considerable scope exists for balanced and coordinated interregional development in the digital economy.

Benchmark regression analysis
We examine the direct effect of the digital economy on ULGUE using the pooled ordinary least squares (OLS) and fixed effect models.According to the regression results in table 4, the estimated coefficient of the core explanatory variable (Dige) is positive and significant at the 1% level based on various methods.The result of the Hausman test indicates that the fixed effect model is the better model to adopt in this case.Hence, the regression parameters shown in column (4) serve as the benchmark for subsequent analysis.The estimated parameter of the core explanatory variable (Dige) is 0.0739, indicating that each unit of growth unit in the digital economy will contribute an average increase of 0.0739 units of ULGUE, assuming that all other conditions remain constant.The values in parentheses are robust standard errors cluster at city level; * , ** and *** indicate significance at the confidence levels of 10%, 5% and 1%, respectively.(The same applies to other tables.) The digital economy develops and continuously penetrates the production and life of human society and urban governance, promotes the evolution of economic organization and spatial structure, and pushes the integration and renewal of urban spatial layout.Based on Chinese city data, this paper confirms the relationship between the two, and the results are similar to current research on the impact of the digital economy.Czernich et al empirically investigated that for every 1% increase in broadband penetration, the economy's per capita annual growth rate increased by 0.09 to 0.15% in Organization for Economic Co-operation and Development (OECD) member countries (Czernich et al 2011).Some scholars believe that the emergence of the digital economy form centered on data and supported by IT could improve the ecological environment through 'the substitution effect' and 'dematerialization' to overcome the conflict between economic development and the ecological environment (Asongu et al 2018).Fan confirmed that the digital economy's development significantly contributes to the urban land green use efficiency.The ideas above support previous results according to which digital economy growth can achieve the multiple goals of maximizing resource use efficiency, minimizing environmental pressure, and coordinating ecological and economic development in urban land use.
Regarding the control variables, economic development (Lnpgdp) positively affects ULGUE.Cities with higher levels of economic development attract capital, talent, and technology, which can increase economic and social outputs, thus improving ULGUE.The government effect (Gove) negatively affects ULGUE due to the distortion of land prices caused by local government interventions, such as expanding the amount of land supply or lowering the transfer price, thus leading to an excessive supply of land and low-quality investment, and consequently decreasing the efficiency of urban land (Gao et al 2021).Population density (Popd) positively impacts ULGUE because it can cause an inflow of factors, and the scale effect created by the resulting increase in demand can positively impact ULGUE.The parameter of financial pressure (Finp) shows a positive effect on ULGUE due to the government's efforts to alleviate financial stress by supplying more construction land and because market-oriented land transactions can optimize resource allocation and can then improve resource efficiency (Pu and Zhang 2022).Infrastructure (Infra) has a nonsignificant negative correlation with ULGUE; the possible reason is that infrastructure construction can accelerate the flow of factors and improve the efficiency of resource utilization.At the same time, however, it requires a large amount of land resource inputs, leading to ecological and environmental losses, and the two effects cancel each other out.

Endogeneity
This work selects the instrumental variable (IV) approach to mitigate the bias in estimation results caused by potential endogeneity problems.Historical fixed-line phone penetration data were chosen as the IV of the core explanatory variable (Dige) since these data satisfied the IV requirements of relevance and exclusivity (Lyu et al 2023, Wang andShao 2023).The growth of the digital economy depends on the development of the digital infrastructure, which is reflected by the level of Internet development (Litvinenko 2020).Moreover, the development of Internet technology began with the proliferation of fixed-line telephones.Consequently, cities with historically high telephone penetration rates are expected to have greater Internet penetration and application in the future, positively influencing the development of the digital economy, and thus meeting the relevance requirement of the IV.Additionally, historical postal data do not directly influence current land utilization and have no significant correlation with the random interference term of equation (1), satisfying the assumption of exogenous exclusion of the IV after controlling for the pertinent city characteristic variables.Because the 1984 fixed-line phone penetration rate is cross-sectional data, it is incompatible with the panel data of the benchmark model.It is multiplied by the number of internet users nationwide in the last year to create a panel IV, and it is denoted as Dige_IV.Due to changes in the layout of administrative divisions, missing samples are excluded from the IV regression analysis using the IV-two-stage least squares (2SLS) method.
The first-stage regression outcomes are presented in column (5) of table 4, showing a significant positive relationship between Dige_IV and Dige.In the second stage, the estimated coefficient of Dige presented in column (6) is significantly positive, indicating that the positive effect between the two remains valid after accounting for the endogeneity issue.Furthermore, the chosen IV passes various tests for nonidentifiable IV, weak IV, and robust identification tests, demonstrating that the IV is appropriate.

Replacement of the dependent variable
The previous regression results suggest that the digital economy development has the potential to improve ULGUE.This subsection attempts to perform a regression-based robustness test by the method of replacement of the dependent variable.Specifically, we utilized the Malmquist -Luenberger model to recompute the variation in ULGUE of each city throughout the research period based on the evaluation indicator system constructed in subsection 4.2.1 and denoted it as Ulgue1.We then replaced the original explanatory variable with Ulgue1 and re-estimated it using equation (1).It is essential to point out that the Malmquist -Luenberger index describes the changing efficiency status from period t to t+1.Hence the computation result is for period t-1, and the final analysis is carried out with the data of 8 years.As reported in column (1) of table 5, the estimated coefficient of Dige is 0.2777, which passes the significance level test of 1% and is consistent with the baseline regression results, indicating that the digital economy continues to have a positive impact on ULGUE after replacing the explanatory variables.The findings of the study will not be changed as the explanatory variables are measured by different methods.The robustness of the benchmark regression analysis is verified.

Excluding provincial capitals
As provincial administrative centers, provincial capitals possess advantages in terms of economic level, industrial structure, infrastructure, urban area, and resource endowments, making it difficult for other ordinary cities to compare.Such provincial capitals can also have a strong siphoning effect on various resources in neighboring ordinary cities, which makes them more likely to become cities with a high level of the digital economy, and the pilot cities of land policies, such as Nanjing, Hangzhou, Chengdu, Fuzhou, Guangzhou, and other provincial capitals are all listed in the pilot cities of low utility land redevelopment by the Ministry of Natural Resources of China.It is reasonable to presume that the city level is an important influencing factor of the digital economy development and then correlated with the ULGUE.Neglecting such an effect may make the mean regression curve of the sample cities biased towards extreme values, leading to the fitting of the retrospective equation deviating from reality.Therefore, excluding provincial capitals can reflect the average treatment effect of samples in a more objective manner.We re-estimate using samples that exclude provincial capitals.According to the results presented in column (2) of table 5, it can be seen that after controlling for city and year fixed effects, the estimated parameters of the core explanatory variables are still significantly positive, indicating that excluding provincial capital cities, the development of the digital economy also significantly improves ULGUE, which is consistent with the results of the benchmark regression.

Alternative province fixed effects
Changes in land planning and industrial policies in different provinces may have differentiated impacts on digital economy development and land use efficiency.China's National Development and Reform Commission and the Central Office of Internet Information Technology constructed six national experimental zones for the innovative development of the digital economy in Hebei, Zhejiang, Fujian, Guangdong, Sichuan, and Chongqing in 2019.Similar implementation of digital industry policies may generate non-market-oriented policy arbitrage space, inducing some industry investment behaviors to have differentiated impacts on urban land use and efficiency.We attempt to control the province-fixed effect and year-fixed effect to eliminate the systematic changes in macro factors and other unobservables from interfering with the estimated results of this research.The sign and significance level of the estimated coefficient shown in column (3) of table 5 do not change, compared to the estimation results of the benchmark regression, confirming the validity of the benchmark regression.

Shortening the sample period
The research period chose the period from 2011 to 2019 because China is undergoing a significant transformation in resources allocated, and economic development.The digital economy has developed rapidly, and the land use pattern has changed dramatically.In 2015, the 'national cyber development strategy' was first incorporated into the 'Five-Year Plan for China's Economic Development', signaling that promoting Internet development ranked as a national policy.Since the Internet is the essential digital foundation for developing the digital economy, 2015 is regarded as a critical milestone that has a far-reaching impact on the development of the digital economy in China.Therefore, to ensure the research sample in a homogeneous socio-economic environment and to reduce the estimation bias of significant policy change, this study shortens the study period to 2015-2019 by tuning to examine the robustness of the impact of the digital economy on ULGUE during a specific period.The results in column (4) of table 5 show that the sign and significance level of the regression coefficients are unchanged, with the strength of the effect being slightly more significant than that of the baseline regression results, indicating that the influence of the digital economy development on ULGUE increases, and the level of significance is elevated after implementing the 'national cyber development strategy'.The robustness of the benchmark regression results is further verified.
5.5.Indirect effect of the digital economy on ULGUE 5.5.1.The mediating effect of industrial structure optimization Regarding the structural effect, we included the mediator variable industry structure optimization in equations ( 2) and (3).The Sobel test result is 0.0049, which is statistically significant at the 1% level, demonstrating that industrial structure optimization acts as a mediator.Specifically, the coefficient of the digital economy in column (2) of table 6 is positive, indicating that the development of the digital economy can optimize the industry structure.The high-technology service industry, denoted by the Internet and the information industry, is the main component of the digital economy, and the development of the digital economy can expedite the service-oriented transformation of traditional sectors.Furthermore, the coefficient of the mediator variable in column (3) is significantly positive, indicating that the industrial structure optimization can increase ULGUE.Current research findings show that the services industry has the advantage of demanding less land and yielding higher output than sectors such as manufacturing and mining (Su et al 2023).Therefore, under the current technological level and economic scale, optimizing the industrial structure by developing the digital economy is an ideal approach to enhancing ULGUE compared to the expansion of other industries.

The mediating effect of green technology innovation
Regarding the technical effect, we introduced green technology innovation (Ginno) as the mediator.The Sobel test result is 0.0270 and p < 0.01, demonstrating the valid mediating role of green technology innovation.The estimated coefficient of the core explanatory variable shown in column (5) of table 6 is significantly positive, suggesting that the digital economy has the potential to enhance green technology innovation.Digital industrial development can accelerate the efficiency of information sharing and resource exchange, enabling the integration of technologies across different fields.It also enhances the probability of obtaining information on key green technology and facilitates the leapfrogging and upgrading of green technologies, promoting green technological innovation.The estimated parameters of Ginno and Dige are all significantly positive in column (6), demonstrating that green technology innovation is a crucial transmission path between the two.Green technology innovation refers to new technologies, processes, or products that can alleviate ecological and environmental pressures while improving resource use efficiency.It can provide technical support for urban green economic development by reducing pollution and carbon emissions, promoting energy saving and consumption reduction, and effectively enhancing urban green economic efficiency.This conclusion supports the strategy of 'promoting green technology innovation to achieve green, low-carbon and sustainable development' proposed in the 'Implementation Plan on Further Improving the Market-Oriented Green Technology Innovation System (2023-2025)' issued by China's National Development and Reform Commission and the Ministry of Science and Technology.

The mediating effect of digital talent agglomeration
Regarding the scale effect, we introduced digital talent aggregation (Dtag) as a mediator variable.The Sobel test result is 0.0576 and p < 0.01, indicating that a mediating effect exists.Specifically, in the regression estimation results presented in column (4), the estimated coefficient of the core explanatory variable (Dige) is positive and significant at the 1% level, implying that digitization can significantly promote digital talent agglomeration.The rapid development of the digital economy has led to a large-scale demand for employment.Elsby & Shapiro's findings verified that communications technology development has created 1,585,000 new jobs in the United States.In particular, the digital economy has generated a series of new industries, sectors, and occupations that assemble a relatively high proportion of high-skilled digital talent (Elsby and Shapiro 2012).As shown in column (8), the estimated coefficient of the digital economy is nonsignificant, while the coefficient of the mediator variable (Datg) is significantly positive, indicating that digital talent agglomeration plays a fully mediating role, which means that digital talent agglomeration positively affects ULGUE.First, the assemblage of digital talents helps to reduce the cost of information collection and dissemination, making it easier to obtain advanced ideas, experience, and technology, all of which can improve labor productivity and enterprise performance through collective learning.The result of the agglomeration will expand the investment scale of cities, raise the value of land, and increase the output of urban land utilization (Wu and Yang 2022).

The moderating effect of land finance dependency
We incorporate land finance dependency (Lfin) as a moderator variable to examine Hypothesis 5.After separately centralizing the explanatory and moderator variables, the interaction item (Dige * Lfin) is constructed and included in equations (4) and (5).According to the estimated results in column (3) of table 7, the coefficient of the digital economy remains positive, while the estimated parameter of the interaction term is negative, implying that the moderating variable (Lfin) weakens the effect of the digital economy on ULGUE.A high level of dependence on land finance can trigger diseconomies of scale.To entice more enterprises to settle and invest, local governments participated in the competition for investment attraction with the advantage of a low land price cost.Additionally, focusing on the scale of land attraction rather than the high-quality development of the business model, lowers the entry threshold, resulting in an inefficient use of land resources.Furthermore, excessive reliance on land finance can push up land prices, inducing a rise in the production cost of enterprises, inhibiting an improvement in technology levels, and leading to a decline in land use efficiency.
5.6.Heterogeneous analysis 5.6.1.Heterogeneous effect of city location The differences in the regions of cities can be a source of heterogeneity.Following the categories of the National Bureau of Statistics of China, we classify the research samples as 114 eastern cities and 172 central and western cities.As shown in columns (1) and (2) of table 8, the coefficient of the digital economy is significantly positive in eastern cities but insignificant in central and western cities.These results indicate that eastern cities show the effects of the digital economy on the ULGUE enhancement, but such effects have yet to be noticeable in central and western cities.One possible reason for this difference is that the digital economy of eastern cities has been developing for a longer period, resulting in better-developed infrastructure and a broader adoption of cuttingedge digital technologies.These advantages of forwardness can boost social and economic development and further improve ULGUE.In addition, the digital economy has incremental marginal benefits, exacerbating the regional differences in terms of the effect on ULGUE among subregions, resulting in the positive impact of the digital endowment that has yet to be evident in the central and western regions.Hence, it is necessary to carefully consider and make strenuous efforts to overcome the coordination and equilibrium development issues within the various regions.

Heterogeneous effect of different city sizes
Following the 'Notice on Adjusting the Criteria for Urban Size Division' issued by the State Council of China in 2014, we classify all samples into 180 large and 106 small cities based on whether the population size of a city is more than 3 million or than this number, respectively.We perform subsample regressions based on equation (1).As shown in table 8, the coefficient of the digital economy is significant in column (3) and nonsignificant in column (4), indicating that the digital economy can optimize the scale and structure of urban land, further impacting the ULGUE in large cities.But it has yet to serve as a positive driving mechanism for ULGUE in those small cities.The reason is that market size is a critical threshold for digital economy development.In addition, investment in related infrastructure needs to be supported by sufficient market revenue to be effectively maintained.Its continuing development needs to be built on a comprehensive digital economy ecosystem, which can further strengthen the scale effect through the decreasing marginal cost of digital elements.For example, in e-commerce, mobile payment terminals, cloud computing, and other operating platforms, the marginal cost of consumer use continuously decreases; when it reaches a particular scale, the marginal cost even approaches zero (Goldfarb and Tucker 2019).Local governments should plan and grow based on a city's characteristics rather than blindly replicating other areas' approaches to developing the digital economy, thereby providing ULGUE with an opportunity to benefit from developing the digital economy.

Heterogeneous effect of the levels of digital economy development in cities
All cities are categorized into high-level and low-level subgroups, based on whether the digital economy index is above or below the median, respectively.Then, we perform subsample regressions based on equation (1).The estimated coefficients in columns (5) and (6) of table 8 show that the digital economy positively affects ULGUE in high-level cities but not in low-level cities.The reason is that low-level cities are undergoing the initial stages of digital economy development with poor Internet foundations.The construction of Internet infrastructure in these cities needs more inputs of factors and resources, resulting in a rapid expansion of urban land and severe environmental loss, partially mitigating the promoting effect of ULGUE.Furthermore, there is a time lag between the widespread use of data elements and the digitization of traditional industries.As a result, cities with a lower level of digital economy are not yet fully able to boost their ULGUE.Referring to Metcalfe's Law, the value of the internet develops rapidly in a multiplicative exponential manner as the quantity of participating users increases.Hence, in cities with a higher digital economy index, the Internet is integrated at a deeper level into all facets of economic production and life.Thus, they continue to release digital dividends with their technological and scale effects, which can significantly enhance ULGUE.
5.6.4.Heterogeneous effect of different levels of land marketization in cities Since China's reform and opening up, the method for transferring land use rights has changed from being solely nonmarket-based to being a combination of nonmarket and market based.The primary non-market-based method is land granted by allocation (hua bo), while the market-based method for conveying land use rights is allocation (hua bo), and market-based methods include public cautions (pai mai), tenders (zhao biao), and listings (gua pai).Because the allocation (hua bo) method mostly obtains the land for free, land granted in this way does not reflect investment preferences (Liu et al 2016).In contrast, market-oriented land supply methods can automatically regulate the demand for land through competitive and price mechanisms, raise the cost of land acquisition for enterprises, promote the survival of the fittest, and guide enterprises to increase their investment per unit of land, all of which can provide the possibility and opportunity for the digital economy to enhance ULGUE.We select the proportion of allocated land area to the total land area granted as a measure of land marketization.A higher proportion of allocated land suggests a lower degree of land marketization, while a lower proportion indicates a higher degree.Equation (1) is estimated based on two subsamples, and the coefficient of the digital economy in column (7) is significantly positive but nonsignificant in column (8).This result indicates that the digital economy has a more significant influence on ULGUE in cities with a higher degree of marketability in land transactions.

Conclusions and discussions
As the largest and fastest rate of land urbanization in human history occurs in China, alleviating the contradiction between land supply and demand and the problem of environmental degradation by improving the ULGUE is a necessary path.The rapid development of the digital economy provides an opportunity to increase ULGUE.This paper employs PCA and super-efficiency SMB-UN model to evaluate the level of digital economy development and ULGUE in China from 2011 to 2019 and adopts some econometric models to examine the direct relationship, transmission mechanism, and heterogeneous effects between the digital economy and ULGUE.The main conclusions are as follows: (1) Overall, the average value of ULGUE of the 286 cities studied was relatively low and still showed much room for efficiency The digital economy and ULGUE showed a general upward trend and similar spatial distribution characteristics of decreasing from the eastern coastal region to the western region.The average ULGUE measured in this study is slightly lower than that of research Jiang's (2021), mainly due to data structure and index system differences.Specifically, this study uses a multisource dataset, and it incorporates the undesirable output to account for the environmental loss in the land use process, which can more accurately describe ULGUE.However, we all find that ULGUE has yet to realize DEA effectiveness in general, which is related to the continuous expansion of urban land and the simultaneous existence of a large amount of inefficient and idle land in China in recent years.
(2) The regression results show that the digital economy significantly improved the ULGUE nationally, and the conclusion remains robust after applying IV-2SLS estimation, replacing the explanatory variable, excluding provincial capitals, using alternative province fixed effects, and shortening the sample period.This conclusion aligns with Fan et al (2023), who argued that the development of the digital economy created the information space to replace the traditional urban space and that the geographical distribution of labor, information formation, production diffusion, and flexibility of siting in the information environment determines the spatial distribution of the new industry, which in turn affects the urban land-use efficiency.This study verifies the enhanced effect of the digital economy development on ULGUE, which can provide a more prosperous solution for enhancing ULGUE under the background of digitalization and greening transformation.
(3) The mechanism analysis shows that the digital economy indirectly promotes ULGUE through structural, scale, and technical effects, specifically by optimizing the industrial structure, expanding green technology innovation, and promoting digital talent agglomeration.Meanwhile, land finance dependence has a negative moderating effect on the relationship between them.Many scholars have noted the positive impacts of industrial structure, technological innovation, and resource agglomeration on ULGUE (Tan et al 2021; Wang et al 2023).However, it differs from that of Liu et al's study in the Yellow River Basin of China, which concluded that the digital economy has improved technological innovation and that the resulting augmentation in consumer demand and production scale has increased energy consumption and pollution emissions.This 'rebound effect' reduces land eco-efficiency; at the same time, the optimization of industrial structure driven by the digital economy does not have a statistically significant effect on land eco-efficiency, and we believe that the source of this difference is mainly due to the study area, on the one hand, because the Yellow River Basin is a region of China where highly-polluting, high-energy-consuming.High-emission traditional manufacturing industries are concentrated, and the positive effects of the digital economy through technological innovation and structural optimization are offset.On the other hand, it also suggests that China's digital economy is less integrated with traditional industries.
(4) The promoting effect of the digital economy on ULGUE shows a significant characteristic of heterogeneity, as this effect is more effective in cities with better digital economies, cities with a higher degree of land marketization, larger cities, and eastern cities.Current researches have focused on the diverse qualities of the digital economy's impact effect as influenced by geographic differences (Qiu et al 2023).This study further classifies the cities by cities' characteristics to investigate different impacts among cities, which draws more detailed findings and can further provide the differentiation strategy based on the specific conditions of the cities.

Policy implications
This work has the following policy implications: First, governments should promote the advancement of the digital economy through multiple approaches, such as increasing the investment and support for digital infrastructure and technological innovation to accelerate the process of digital industrialization and its integration with traditional industries, establishing cross-regional data and information platforms to alleviate distortions in the market for land resources, achieving the optimal allocation of resources, and fostering a positive environment for enhancing the ULGUE.Comprehensively promoting the coordinated growth of the digital economy and sustainably releasing its effect on land utilization transformation.
Second, the governments should fully capitalize on the indirect effect of the digital economy, as improving ULGUE necessitates a multifaceted approaches.These efforts include optimizing the industrial structure while accounting for industrial characteristics and economic development stages, scientifically planning the spatial layout, encouraging green technological innovation, and assembling digital talent to directly guide the development of industries toward digitization, scale, intensification, and greening.They also include reducing land finance dependency, guiding the implementation of land pricing strategies under market-oriented management, and further increasing ULGUE by changing and optimizing the scale and structure of urban land utilization.
Third, differentiated urban digital economy and land utilization strategies should be implemented across regions and cities.Specifically, eastern cities should utilize research innovation to speed up industrial upgrading, promote economically sustainable growth, and play a leading role in driving nationwide growth and they should further guide the enhancement of the structure and pattern of urban land utilization.The government should strengthen the cooperation mechanism among subregions, and promote the harmonious development of the economy and the environment by facilitating the orderly transfer of digital resources to the central and western regions while considering resource-carrying capacity and the ecological environment.Cities with less developed digital economies and smaller populations should follow the general rules of urban construction and economic development, take into account their own characteristics and development stages, insist on both scale expansion and quality improvement, avoid unthinkingly following the trend, and try to explore the differentiated paths of digital economy development and land utilization strategies.
International, given the positive impact of the digital economy on ULGUE.First, governments should strengthen international exchanges and cooperation.The application of data elements and platforms can break down country-specific barriers and promote the efficient flow of digital resources, technologies, and talents among countries.Regional cooperation can realize the sharing of technological knowledge and skills in developed countries as well as the spillover effect of advanced digital technologies, and it is also an essential way for less-developed countries that lack efficient digital infrastructure, high-quality digital talents, and digital technology innovation capacity to realize the digital economy development at a lower cost.It can shrink the digital divide between developed and developing countries and ensure that countries at different stages of development can all benefit from the dividends of cleaner production and green, sustainable development brought about by the digital economy.Secondly, there are wide variations in digitalization and land resource utilization models between different countries, which calls for differentiated solutions for digital economy development and optimization of land utilization.Developing countries in the process of urbanization typically face the dual challenge of an intense conflict between urban land supply and demand, as well as inefficient land use.Governments should provide funding for digital economy development, focus on strengthening digital infrastructure construction, and actively introduce and apply digital technologies.However, it should refrain from mindlessly copying the development experience of other countries and develop infrastructure and digital industries according to its characteristics of resource endowment and industrial advantages to fully utilize the benefits of digitization in terms of resource utilization and environmental protection.For countries with a high level of development of the digital economy, on the one hand, governments should continue to encourage digital technological innovation to ensure the sustainable impetus for digital economy development.On the other hand, it should adopt goal-oriented policy measures to actively guide the integration of digitization into the process of land resource utilization, promote the integration of digital elements into the combination of production factors, realize the decoupling of economic growth and urban land expansion, reduce the high degree of dependence of the economic structure on land resources, and also pay attention to avoid the rebound effect that may be brought about by digitization.

Limitations and future perspectives
This paper is still limited in several ways.First, the panel data were last updated in 2019.Regarding the reality of the booming development of the digital economy in China, more valuable insights may be found by considering more recent data.Second, the digital economy can overcome the spatial and temporal limitations of production activities and accelerate the integration and flow of resource elements.Therefore, whether its development will have a spillover effect on the land utilization of neighboring districts is an important question to address.Thus, future research should try to explore the influence of the digital economy on ULGUE from a spatial perspective.
Digital economy development improves green technology innovation.(1) It improves the environment of green technology innovation.Digital industries are mainly knowledge intensive and innovation intensive, and their innovation environment is superior to that of other industries (Anttiroiko et al 2020, Feng et al 2023a).When green technology in the form of intermediate products is introduced into production, it can trigger another round of technological advancement.This constant stimulation of the self-innovation production model is highly effective in promoting the evolution and modernization of green technology (Ding et al 2022).

Hypothesis 2 :
The digital economy enhances ULGUE by optimizing industrial structure; Hypothesis 3 : The digital economy improves ULGUE by innovating green technologies; Hypothesis 4 : The digital economy enhances ULGUE by agglomerating digital talent;

Figure 2 .
Figure 2. Trend in the level of digital economy and ULGUE.

Table 1 .
Evaluation indicator system for ULGUE.

Table 2 .
Evaluation indicator system for the digital economy.

Table 3 .
Descriptive statistics of the main variables.

Table 4 .
Benchmark results and tests of endogeneity.

Table 6 .
Results of the mediating effect model.

Table 7 .
Results of the moderating effect model.

Table 8 .
Results of the heterogeneity analysis.