Abstract
Counties in western China face persistent challenges in delivering equitable public services, while the digital economy offers opportunities to enhance service provision. This study examines 752 counties across 12 provinces, using composite indicators for the digital economy and public services, spatial analysis, Spatial Durbin Models, and threshold effect models with robustness checks. Results show that the digital economy generally improves public services, but population density above 298 persons per square kilometer and fiscal surplus weaken gains and intensify negative spillovers, per capita GDP above 26,515 yuan and industrial upgrading index above 0.9175 strengthen contributions to education, culture, and social welfare, and in fiscally constrained counties with low revenue-to-expenditure ratios, a unit increase in informatization improves services by only 0.0063 units. To reduce disparities, densely populated counties should implement shared digital service hubs and joint e-governance platforms, while fiscal-constrained counties can deploy targeted transfer payments to fund IT infrastructure, digital literacy programs, and online public service applications.
Similar content being viewed by others
Introduction
In recent years, the Chinese government has regarded the promotion of high-quality and equalized development of public services as an important pathway to addressing livelihood concerns, advancing social equity, and achieving the 2035 goal of common prosperity. Against this backdrop, public service provision and its economic and social effects have gradually become focal issues in academic research. Existing studies have conducted relatively systematic analyses of public service levels, regional disparities, and their determinants at the national scale, most of which adopt the conventional eastern–central–western regional framework (Xiong et al., 2025). While this line of research provides valuable macro-level insights into the overall landscape of public services in China, it largely remains confined to relatively coarse spatial divisions such as the east–central–west classification, with limited attention paid to the structural constraints, operational characteristics, and heterogeneous manifestations within underdeveloped regions. Yet it is precisely these underdeveloped regions where contradictions in the public service provision system are most concentrated and shortcomings most pronounced. From both academic and policy perspectives, therefore, these regions merit further in-depth investigation in order to uncover their distinctive operational mechanisms and potential pathways for improvement.
The western region, particularly at the county level, vividly encapsulates the real-world difficulties faced by underdeveloped areas in China with respect to public service provision, and thus exhibits strong representativeness. Public services play a critical role in narrowing regional disparities, improving residents’ welfare, and enhancing labor productivity, a role that is especially pivotal in underdeveloped counties in western China. Constrained by long-standing weaknesses in the economic base, a highly concentrated industrial structure, and insufficient investment in public services, western counties commonly experience relatively low levels of public services and imbalanced supply structures, and face more pronounced constraints in the process of coordinated regional development (Huang et al., 2025). As the basic administrative units responsible for delivering fundamental public services within China’s urban–rural system, counties are simultaneously tasked with addressing livelihood needs and reducing regional disparities, while also confronting multiple challenges such as sustained population outflows, limited fiscal capacity, and relatively lagging infrastructure. As a result, internal tensions within the public service provision system at the county level are more concentrated and pronounced.
It is precisely against this backdrop that western counties not only constitute a typical representation of underdeveloped regions, but also exhibit pronounced complexity and heterogeneity. On the one hand, substantial differences exist across counties in terms of economic development levels, industrial foundations, and the initial conditions of digital infrastructure. On the other hand, under conditions of constrained economic development, tight fiscal capacity, and limited governance capability and flexibility in public service provision, the impact of the digital economy on public services is more likely to display nonlinear or conditional characteristics. Such mechanisms are often difficult to fully identify in national-scale analyses or in studies based on coarse regional classifications that rely on averaged effects.
Accordingly, this article does not simply replicate existing findings at the national level. Instead, it adopts the perspective of underdeveloped regions and focuses on western counties as grassroots units characterized by both typicality and diversity. It not only provides an in-depth spatial examination of disparities in public service provision and their constraining factors, but also systematically analyzes the mechanisms through which the digital economy influences public services, with particular attention to potential moderating and threshold effects. Through this approach, the article extends the existing literature along multiple dimensions—including the theoretical framework, methodological design, and empirical analysis—thereby offering new insights into digital economy governance and public service optimization in underdeveloped regions.
The digital economy has emerged as a transformative force in this context. Characterized by high efficiency, low cost, and expansive coverage (Sturgeon, 2021), it holds considerable promise for enhancing public service delivery. The Chinese government has actively promoted digital economic growth and its integration with the real economy, viewing it as essential for building globally competitive digital industry clusters (Zhang et al., 2021). In county regions, especially in the west, the digital economy facilitates the reconfiguration of industrial chains, boosts factor productivity, and supports new urbanization. These dynamics not only promote high-quality local development but also enable counties to share in the benefits of digital transformation. Through advanced data processing and resource integration, the digital economy enables more targeted, intelligent, and inclusive service provision (Dener et al., 2021).
However, the spatial impacts of the digital economy are highly complex, particularly at the county level. While digital platforms can reduce geographic barriers, enhance regional coordination, and produce positive spillovers (Wang et al., 2022), the uneven distribution of digital resources may also generate negative externalities. Core areas may attract disproportionate attention and investment, siphoning resources from surrounding counties and worsening existing inequalities (Gladkova et al., 2020). These dual effects call for comprehensive assessments of how the digital economy reshapes public service landscapes across different regions.
Scholarly attention to the relationship between the digital economy and the equity of public services has expanded rapidly, producing valuable insights into service quality, spatial inequality, and the mechanisms behind these dynamics. Existing research has developed diverse evaluation frameworks through approaches such as DEA, the entropy method, and the SERVQUAL model (Shi et al., 2023; AlOmari, 2021), and has examined public services in sectors including education, healthcare, and social security (Lee et al., 2021). However, most of these studies have been conducted at the provincial or municipal level (Li et al., 2020; Ma et al., 2020).
Although provincial- and city-level studies have confirmed that the digital economy exhibits spatial relevance and spillover effects—for instance, through the diffusion of digital infrastructure across jurisdictions and the cross-boundary delivery of online services—they generally treat such relationships as linear and homogeneous, leaving the potential for nonlinear interactions and spatially differentiated mechanisms largely unexplored. Against this backdrop, county-level research remains scarce, despite counties being the core units for the direct provision and fiscal implementation of basic public services. In western China, where counties are numerous, economically fragile, and institutionally diverse, understanding how digitalization influences local public services and how these effects extend spatially has become increasingly urgent. By situating the analysis within this underexplored context, this article contributes to the literature in several respects. It adopts a finer county-level scale to reveal the spatial dependence and spillover effects of the digital economy on public service provision, extending beyond the aggregate insights of provincial and municipal studies. Moreover, rather than limiting the analysis to linear identification, it incorporates nonlinear effect detection to reveal the heterogeneous pathways through which different dimensions of the digital economy—such as informatization, internet development, and digital transactions—affect basic public services. This approach enriches both the empirical understanding and theoretical interpretation of how the digital economy reshapes public service systems in less-developed western regions.
By integrating perspectives from public administration, welfare economics (Pigou, 1912; Arrow, 1948; Samuelson, 1938), and spatial science (Krugman, 1991), this article develops a cross-disciplinary framework to interpret the digital economy’s role in transforming public services. The study advances theory by linking digital infrastructure with welfare outcomes and distinguishing the differential effects of the digital economy on specific public service domains such as livelihood security, education, and social welfare. On the practical side, the research provides locally adapted evaluation tools and a replicable model for developing countries seeking to address regional imbalances and strategically enhance service delivery through digital means. These insights collectively underscore the need for tailored, region-specific strategies to balance digital development with public service equity.
This article systematically examines the impact of the digital economy on county-level public services. Section 2 develops the theoretical framework and research hypotheses, including potential moderating effects. Section 3 introduces data, methods, and model specifications, covering spatial econometric models and variable construction. Section 4 presents empirical results, including spatiotemporal patterns, spatial spillovers, and moderating mechanisms. Section 5 concludes with key findings, contributions, and governance implications.
Theoretical analysis and hypotheses
Theoretical analysis of the impact of the digital economy on basic public services
The impact of the digital economy on basic public services in western counties can be understood within the framework of the techno-economic paradigm theory. Its development involves not only the application of digital technologies within a given county but also potential spatial effects on surrounding counties through factor diffusion and knowledge spillovers (Griliches, 1991).
From the informatization dimension, information infrastructure and data platforms enhance the operation, management, and information-processing capacity of public services within the county, covering areas such as livelihood security, social welfare, and education, culture, and sports. These include the informatized management of employment information, medical data, housing administration, social assistance, and educational and cultural resources. Meanwhile, the informatization experience and platform structures accumulated in core counties can diffuse to neighboring areas through data transmission and information sharing, forming potential spatial spillover effects, the specific impacts of which will be examined empirically later.
From the internet dimension, the widespread adoption of internet platforms extends the provision of public services to online and mobile terminals, reshaping information circulation and service interaction patterns. Within the county, the internet enables online management of employment, healthcare, housing security, social assistance and education, culture, and sports services, thereby facilitating the digital governance of public services. At the same time, internet platforms and operational mechanisms developed in core counties may spread to surrounding counties through node connections, interface integration, and cross-regional data transmission, generating potential spatial spillovers. The techno-economic paradigm emphasizes that technological change not only improves production tools but also entails institutional and organizational restructuring (Perez, 1983; Freeman and Louçã, 2001), offering theoretical support for explaining the diffusion of internet technologies and the extension of service models across counties. The specific impacts will also be tested empirically.
From the digital transaction dimension, digital payments, e-procurement, and data management influence financial flows, process management, and transaction activities related to public services within the county, including wage and subsidy disbursement, medical payments, housing allowances, social welfare transfers, and educational and cultural product transactions. Meanwhile, the digital transaction systems in core counties may affect neighboring counties through cross-regional transactions and data circulation, generating potential spatial spillovers. Guided by the externality theory (Griliches, 1991), the processes, standards, and information flows associated with digital transactions not only influence resource allocation within a county but may also affect public service provision in adjacent counties through spatial diffusion. The specific impacts will be examined empirically later.
Based on the above analysis, this article proposes Hypothesis 1:
H1: The digital economy and its three dimensions—informatization, internet development, and digital transactions—exert both local and potential spatial spillover effects on basic public services in western counties.
Theoretical analysis of the moderating effects of external environments on the impact of the digital economy
The impact of the digital economy on basic public services in western counties is neither linear nor universally consistent, but rather embedded in the overall development environment of each county. Significant differences exist among counties in terms of fiscal capacity, population size, and economic foundations. Although these macro-environmental factors do not directly determine the outcomes of digital economy applications in the public service sector, they may influence the conditions and effectiveness under which such impacts occur. Under different environments, the intensity and form of the digital economy’s influence on public service quality may vary. Therefore, incorporating macro-environmental factors into the analysis helps avoid treating the digital economy as a single exogenous variable with uniform effects.
This perspective is consistent with the logic of the Technology–Organization–Environment (TOE) framework. Tornatzky and Fleischer (1990) pointed out that the adoption and effectiveness of new technologies are jointly influenced by technological conditions, organizational capabilities, and external environments. In this study, the digital economy is regarded as the technological condition, the county governments and public service provision systems correspond to organizational capabilities, while macro-environmental factors such as population density, per capita GDP, the level of industrial upgrading, and the fiscal revenue–expenditure ratio represent the external environment. According to the TOE framework, different environments alter the difficulty, pathways, and outcomes of technology adoption. Therefore, macro-environmental factors are likely to exert conditional influences on the role of the digital economy in the field of public services.
Building on the basic understanding that the impact of the digital economy on public services is conditional in nature, this article adopts an integrated perspective centered on the county-level development environment and structural constraints to systematically characterize the relevant moderating factors. The influence of the digital economy on public services is neither linear nor unconditional; its effects on service provision depend not only on the level of digital economy development or the scale and structure of public service demand, but are also deeply embedded in multiple structural conditions, including population distribution density, the economic development base, industrial structure, and fiscal support capacity at the county level. It is precisely these interrelated conditions that jointly shape the real-world context and boundaries within which the digital economy affects public services. Therefore, it is necessary to comprehensively depict differences in county-level development environments within a unified analytical framework in order to reveal the conditional characteristics of the digital economy’s impact on public services.
Population density, per capita GDP, the level of industrial upgrading, and the fiscal revenue–expenditure ratio characterize the key conditions underpinning the operation of county-level public services from the respective dimensions of demand, supply, structure, and support. Population density is used to capture the scale, intensity, and spatial concentration of public service demand, reflecting how demand-side pressures constrain the efficiency and coverage of public service provision. Per capita GDP depicts the level of county-level economic development and the capacity it provides to sustain public services, indicating the material foundation and hierarchy of needs upon which service provision is based. The level of industrial upgrading reveals, from a structural perspective, changes in modes of economic operation and resource allocation efficiency; by influencing the logic of public resource allocation, it indirectly shapes the structural features of public service provision. The fiscal revenue–expenditure ratio, in turn, reflects local fiscal self-sufficiency and the sustainability of public spending from the perspective of institutional and financial constraints, thereby delineating the basic boundaries within which the public service system operates.
From an overall logical perspective, the above variables correspond to demand conditions, supply foundations, structural environments, and support constraints in the formation and operation of public services, jointly constituting a conditional influence framework of “demand–supply–structure–support” that systematically encompasses the key links and major constraints within the public service provision system. This framework emphasizes that the effects of the digital economy on public services are not homogeneous outcomes, but are embedded in heterogeneous county-level development conditions and manifest through differentiated impact pathways shaped by multiple structural constraints. Accordingly, adopting a structured perspective helps to more accurately identify the underlying mechanisms through which the digital economy influences public services, rather than reducing these effects to a single or averaged outcome, thereby providing a clear and robust theoretical foundation for the subsequent empirical identification of moderating effects.
Population density reflects the degree of factor agglomeration and the potential size of service recipients within a county. Counties with high population density usually possess more developed communication networks and lower marginal infrastructure costs, which facilitate the promotion of digital platforms. At the same time, a larger user base can generate stronger network effects, enhancing the depth and coverage of public service platforms. In contrast, although low-density counties face higher infrastructure investment costs, their smaller market size may lead to more precise service provision, resulting in differentiated applications. Therefore, population density may exert a conditional influence on the effectiveness of the digital economy, and the specific effects will be examined empirically later.
County-level per capita GDP reflects the degree of economic development and the capacity of residents to meet their needs. Counties with higher economic levels tend to have better infrastructure and more mature market mechanisms, which are conducive to the application and diffusion of the digital economy in public service areas such as education and healthcare. Conversely, counties with weaker economic foundations face constraints in investment and payment capacity, and thus rely more on government subsidies or inter-county cooperation to advance digital projects. Therefore, per capita GDP may affect the strength and coverage of the digital economy’s impact, and the specific effects will be examined empirically later.
The level of industrial upgrading reflects the stage of county development and the efficiency of resource allocation. In regions with a high proportion of manufacturing, the digital economy is more likely to be applied in public management, monitoring, and safety services. In contrast, in counties where the service and cultural industries are more developed, applications of the digital economy tend to focus on e-commerce, online education, and public participation. Therefore, under different industrial structures, the direction and focus of the digital economy’s role may vary, and the specific effects will be examined empirically later.
The fiscal revenue–expenditure ratio reflects the degree of financial constraint within the public service provision system. Counties with sufficient fiscal resources are more capable of sustaining investment in digital infrastructure construction, platform operation, and talent cultivation. In contrast, counties facing fiscal pressure may be more inclined to rely on social capital or interdepartmental collaboration to advance digital projects. Therefore, fiscal conditions may influence the depth and efficiency of digital economy applications, and the specific effects will be examined empirically later.
In summary, population density, per capita GDP, industrial upgrading level, and the fiscal revenue–expenditure ratio, as key indicators of the macro environment, may exert conditional influences on the process through which the digital economy promotes basic public services. Therefore, this article proposes Hypothesis 2.
H2: Population density, per capita GDP, industrial upgrading level, and the fiscal revenue–expenditure ratio have potential moderating effects on the process through which the digital economy influences basic public services.
Materials and methods
Data sources
Under the premise of ensuring the official nature and comparability of the data, this study selected relevant indicators from the annual reports on government information disclosure and the Statistical Communiqués on National Economic and Social Development for each county from 2014 to 2022, supplemented by the China County Statistical Yearbook and the China County Urban Construction Statistical Yearbook. In terms of data processing, for the issue of missing values in some indicators, this article employs the multiple imputation (MI) method with repeated simulations. Missing data imputation is performed using the mice package in R. Additionally, to eliminate the impact of inflation on economic-related data, the GDP deflator index for each province is used to adjust the economic data, with 2018 as the base year. To ensure the consistency and comparability of the indicators, the article applies a logarithmic linear standardization method to transform the raw data into dimensionless values, ensuring the data is suitable for subsequent analysis.
Before conducting spatial econometric regression, necessary variable tests were performed in this article. First, to eliminate multicollinearity issues among the explanatory variables, the variance inflation factor (VIF) test was conducted using Stata software. The results showed that the maximum VIF value was 1.66, with an average of 1.25, both of which were significantly below the threshold of 5, indicating no significant multicollinearity problem. Secondly, to avoid the risk of spurious regression, the article used the HT test and IPS test to perform unit root stationarity tests on the panel data. The results indicated that all variables passed the stationarity test, ensuring the data’s suitability for subsequent analysis using spatial econometric models. These data processing and testing methods provide solid technical support for this study, ensuring the scientific rigor and reliability of the research findings.
Research methods
The weights of the sub-indicators for the dependent and explanatory variables are determined using the entropy weighting method, a widely adopted objective approach. This method standardizes each sub-indicator, calculates its entropy based on variability across counties, and derives its weight according to its relative information contribution, so that more informative indicators receive higher weights. Due to space constraints, the calculation formulas are not presented here, but the procedure follows this standard approach.
This article follows the standard procedures in the existing literature to further determine the appropriate type of spatial econometric model (Anselin, 2022). As shown in the likelihood ratio (LR) tests (Table 1), the core explanatory variable, the digital economy, and its sub-dimensions all exhibit significant results, indicating that the null hypothesis—i.e., the simplification of the Spatial Durbin Model (SDM)—should be rejected. Accordingly, this article adopts the SDM to fully account for the spatial spillover effects when examining the impacts of the digital economy and its dimensions on basic public services and their subcomponents.
Spatial Durbin model
Based on the results of the spatial econometric model test, a two-way fixed effects spatial Durbin model is ultimately constructed (Zhao et al., 2022).
In Eq. (1), i represents the county, t represents the year, \(BP{S}_{it}\) denotes the dependent variable, which is the composite quality index of basic public services (and its subdimensions), \(D{E}_{it}\) represents the core explanatory variable, which is the composite quality index of the digital economy (and its subdimensions), \(co{n}_{it}\) denotes the control variable group, α is the intercept term, \(\rho\) is the spatial autoregressive coefficient, W is the composite spatial weight matrix that integrates both economic distance and geographic distance, \({\mu }_{i}\) represents the individual effects, \({\lambda }_{t}\) represents the time effects, and \({\varepsilon }_{it}\) is the independently and identically distributed random disturbance term.
Geographically and Temporally Weighted Regression Model (GTWR)
The Geographically and Temporally Weighted Regression (GTWR) model extends the classic Geographically Weighted Regression (GWR) by incorporating a temporal factor, enabling the analysis of relationships between variables across both spatial and temporal dimensions. It is suitable for scenarios where data vary simultaneously across both time and space. The GTWR model assigns different weights to various geographic locations and time points, overcoming the limitations of traditional regression analysis that neglect spatial and temporal variations. It allows regression coefficients to vary across locations and times, capturing the local features and dynamic changes in both the spatial and temporal dimensions. The formula is as follows (Jiang et al., 2023):
In Eq. (2), \({y}_{i}\) represents the dependent variable at the i-th observation point, and \({x}_{ik}\) represents the k-th independent variable at the i-th observation point. \(({u}_{i},{v}_{i})\) denotes the spatial coordinates of the i-th observation point, with \({u}_{i}\) and \({v}_{i}\) representing longitude and latitude, respectively. \({t}_{i}\) corresponds to the time point at the i-th observation point. \({\beta }_{k}({u}_{i},{v}_{i},{t}_{i})\) is the regression coefficient for the k-th independent variable at the i-th observation point, which is a function of geographic location \({u}_{i}\), \({v}_{i}\) and time \({t}_{i}\), rather than a global constant. This coefficient varies with the spatial location and time of each observation point, meaning the local regression coefficient may differ across observation points. \({\varepsilon }_{i}\) is the error term.
Moderating effect model
To examine the moderating effects of external environmental factors (e.g., population density, per capita GDP, industrial upgrading, and fiscal revenue-to-expenditure ratio) on the impact of the digital economy across various dimensions, the following moderating effect model is constructed (Anselin, 2022; Zhao et al., 2022):
In Eq. (3), i represents the county, t denotes the year, \(se{r}_{it}\) is the dependent variable indicating the overall level of basic public services (and its sub-dimensions), \(di{g}_{it}\) is the core independent variable representing the overall level of the digital economy (and its sub-dimensions), \(co{n}_{it}\) refers to the set of control variables, \(m{o}_{it}\) represents the moderating variable within the external socio-economic environment, and \(di{g}_{it}\times m{o}_{it}\) denotes the interaction term between the digital economy and the moderating variable.
α represents the intercept term, \(\rho\) is the spatial autoregressive coefficient, W is the composite spatial weight matrix that integrates both economic distance and geographic distance, \({\mu }_{i}\) represents individual effects, \({\lambda }_{t}\) denotes temporal effects, and \({\varepsilon }_{it}\) is the independently and identically distributed random disturbance term.
This article focuses on the regression coefficient of the core independent variable \({\beta }_{1}\) and the coefficient of the interaction term β3. When β1 > 0 and β3 > 0 or β1 < 0 and β3 < 0, the moderating variable enhances the impact of the digital economy on public services. Conversely, when β1>0 and β3<0 or β1<0 and β3>0, the moderating variable weakens the impact. Variables are mean-centered to address potential endogeneity issues in the moderating effect test.
Threshold model
To further quantify the specific moderating effect of population density on the influence of the digital economy and its dimensions, the threshold effect model is established as follows (Yang et al., 2024):
In Eq. (4), \(thre{s}_{it}\) represents the threshold variable “depth of digital financial usage,” and \(I(\cdot )\) is an indicator function that takes the value of 1 if the condition in parentheses is satisfied and 0 otherwise. \({v}_{1}\) to \({v}_{m-1}\) represent the threshold values to be estimated. These threshold values divide the sample into multiple intervals, within which the regression coefficients of explanatory variables, from \({\eta }_{1}\) to \({\eta }_{m}\), vary across different intervals.
The threshold effect test involves two main steps. First, the existence of thresholds and the number of threshold values are examined. Second, the statistical significance of the estimated threshold values identified in the previous step is tested. The bootstrap method was employed with 1000 iterations to determine the statistical significance of the threshold test and to estimate the critical threshold value of digital financial usage depth.
Models and variables
Dependent variable
The “Basic Public Service Standards of China (2023 Edition)” clearly define the foundational standards for the country’s basic public services, covering nine areas: early childhood education, compulsory education, labor income, medical care, elderly care, housing, support for vulnerable groups, military service guarantees, and cultural and sports services. These standards provide a clear policy framework and measurement benchmarks for the study, facilitating a comprehensive analysis of the content and distribution of basic public services. Based on these standards and referencing existing literature (Huang et al., 2025; Jiang et al., 2020), this article constructs an evaluation system for basic public services (Table 2).
To systematically analyze the breadth, depth, and functional attributes of basic public services, the authors categorize the nine areas into three broad categories: livelihood security services, social welfare services, and cultural, educational, and sports services. Each category is further subdivided into several subcategories to more clearly reflect different service objectives. This classification method highlights the functional characteristics of public services while providing a structured analytical framework for subsequent research.
Furthermore, to eliminate the influence of population size differences across counties, all indicators in this article are measured on a per capita basis. Using per capita indicators, rather than total indicators, aims to remove the impact of population scale differences, making inter-county comparisons more objective. Total indicators often exaggerate or obscure the actual service level due to variations in population size and fail to comprehensively reflect the fairness of resource allocation or efficiency of usage. By analyzing per capita indicators, the authors can more accurately reveal the average service level received by individuals, providing a more scientific assessment of the balance between digital economy and public services across counties, thus enhancing the reliability and practical value of the research conclusions.
Explanatory variables
This article constructs a digital economy evaluation index system (Table 3) suitable for western counties, drawing on existing literature (Liu et al., 2020) and focusing on three dimensions: information development, internet development, and digital transaction development. The system takes into account the economic and social characteristics of the western region, addressing both the foundational conditions for digital economy development and its practical application in economic activities and transactions. This multidimensional approach not only broadens the perspective for assessing digital economy development but also provides a basis for more accurately capturing regional disparities.
Moderating variables
Given the pronounced macro-conditional dependency of the digital economy’s impact on public services, this paper eschews the view of it as a uniform exogenous shock. Instead, it systematically examines its moderating effects through the prism of development conditions and structural constraints at the county level. Whether, and to what extent, the digital economy translates into enhanced public service delivery is not merely a function of digital advancement or demand structures; rather, it is contingent upon a county’s capacity to assimilate digital dividends. This process is deeply embedded within multi-dimensional structural factors, including spatial population distribution, economic development stages, industrial organization, and fiscal capacity.
From the perspective of public service operational logic, supply performance is the collective outcome of multiple interacting conditions. On one hand, the scale and agglomeration characteristics of public service demand determine the marginal efficiency of service provision. On the other hand, population density, the foundation of economic development, industrial structure, and fiscal guarantee capacity collectively constrain the capacity and sustainability of public service supply. Only when these aforementioned conditions are relatively coordinated can the technological progress and resource allocation optimization brought by the digital economy be effectively translated into substantive improvements in service quality and coverage. Therefore, the purpose of introducing moderating variables is not merely to empirically test isolated factors, but to identify the critical constraints and the “implementation space” within which the digital economy influences public services.
Based on the aforementioned analysis, this paper selects four indicators—population density, per capita GDP, the level of industrial structure upgrading, and the fiscal balance ratio—to characterize the structural foundations that enable counties to translate the potential advantages of the digital economy into public service performance. Specifically, population density is employed to reflect the scale and spatial agglomeration of public service demand, determining whether the practical conditions exist for digital public services to achieve economies of scale. Per capita GDP is used to measure the level of local economic development and its capacity to support public service investment and digital infrastructure construction. The level of industrial structure upgrading reflects shifts in economic structure and resource allocation, represents the degree of alignment between digital technologies and the public service supply system. Finally, the fiscal balance ratio characterizes the sustainability and flexibility of local finance, defining the institutional and financial boundaries for the long-term operation of public services.
In summary, the aforementioned indicators correspond to the demand foundation, economic support, structural conditions, and security capacity within public service operations, collectively outlining a conditional analytical framework composed of multiple constraints. This framework suggests that the impact of the digital economy on public services is neither unidirectional nor uniform; rather, it manifests through diverse causal paths and varying effect intensities, contingent upon the developmental conditions of each county. Through this lens, the internal mechanisms by which the digital economy influences public services can be more systematically identified, thereby providing a clear and consistent theoretical foundation for the subsequent empirical analysis of moderating effects.
Regarding specific measurement, population density is calculated as the ratio of the permanent resident population at year-end to the built-up area, measured in persons per square kilometer. The level of economic development is represented by per capita GDP, calculated by dividing the Gross Regional Product (GRP) by the total population, with the unit being yuan per person. The level of industrial structure upgrading is characterized by the ratio of the added value of the tertiary industry to that of the secondary industry. Finally, the fiscal balance ratio is measured by the ratio of general public budget revenue to general public budget expenditure.
Control variables
In the context of spatial econometric regression analysis, the factors influencing basic public service levels are multifaceted, spanning various macroeconomic and social dimensions. To enhance the explanatory power of the model and mitigate omitted variable bias, this study follows established literature (Liu et al., 2021; Li et al., 2022) by selecting key general development characteristics—encompassing consumption, urbanization, transportation, and ecology—as control variables (Table 4).
The distinction between control and moderating variables in this paper follows a rigorous theoretical logic: moderating variables are conceptualized as the “structural environmental conditions” that shape the impact intensity of the digital economy, whereas control variables represent “general development characteristics” that influence the level of public services themselves. Consumption, urbanization, transportation, and ecology are designated as control variables rather than moderators because, while these factors may potentially shift the baseline level of public services, they do not fundamentally alter the causal pathways or marginal efficiency through which the digital economy empowers public service delivery. For instance, while transportation and ecology provide the foundational conditions for the transmission and diffusion of public services, they do not directly intervene in the internal mechanisms by which the digital economy reconstructs public services. By adopting this configuration, the model effectively isolates external environmental interference, avoids model over-specification, and ensures a more accurate capture of the “net effect” of the digital economy on basic public services.
Results
Analysis of spatiotemporal evolution characteristics of basic public services
From a temporal perspective (Table 5), the overall quality of basic public services has continuously improved, with an increase of 78.30% over time. Among the subdimensions, science, education, culture, and sports records the fastest growth of 79.41%, followed by social welfare at 78.36%. In comparison, livelihood infrastructure shows the slowest growth of 76.49%. From the perspective of Maslow’s hierarchy of needs (Maslow, 1943), subsistence-oriented public services and social welfare primarily correspond to basic needs at the levels of physiological well-being, safety, and belonging, whereas services related to science, education, culture, and sports are more closely associated with higher-order needs such as esteem and self-actualization. The current configuration of basic public services in western counties suggests that, once foundational needs are largely addressed, policy priorities have gradually shifted toward higher-order demands. As a result, resource inputs and service provision in the fields of science, education, culture, and sports have experienced the fastest growth. This pattern not only indicates an ongoing evolution of the basic public service system in western counties toward an orientation emphasizing capability enhancement and sustainable development (Sen, 1999), but also reflects a stage-specific transition from ensuring basic subsistence to fostering more comprehensive human development.
Although all dimensions have progressed, livelihood infrastructure consistently remains the weakest in both comprehensive quality and improvement pace. This indicates persistent deficiencies in essential services such as income security, healthcare, and housing in western counties, which—given generally low income levels and limited medical and housing support—call for more targeted investment and resource allocation in this dimension.
This article uses MATLAB to plot the three-dimensional Kernel density time evolution of basic public service quality and its subdimensions in western counties from 2014 to 2022 (Fig. 1), aiming to reveal their evolutionary characteristics. Basic public services in western counties have steadily improved, with the Kernel density curve shifting rightward over time, indicating a general movement from lower to higher service levels. From 2014 to 2017, service levels remained relatively stable, after which they increased sharply, reflecting the impact of the national Western Development Strategy and increased county-level development attention. Although the COVID-19 pandemic caused a temporary decline in 2020, service levels quickly recovered and stabilized at higher levels by 2021. Over the observation period, the curve height decreased while its width expanded, signaling growing disparities among counties.
a Basic public services; b Livelihood infrastructure; c Social welfare; d Science, education, culture, and sports.
Among the subdimensions, livelihood infrastructure showed slower growth and persistent imbalance, with some counties lagging due to low income levels and insufficient healthcare and housing. Improvement accelerated after 2018 with rural revitalization policies, but the dimension still lags behind others. Social welfare progressed steadily, particularly after 2019, with broader distribution indicating increasing variation among counties. Science, education, culture, and sports (SECS) experienced significant improvement, especially between 2018 and 2019, yet the widening curve width highlights growing disparities similar to other dimensions.
From a spatial perspective (Table 6), Among the western provinces, Inner Mongolia exhibits the highest overall quality of basic public services, with consistently strong performance across all indicators. Its vast territorial area combined with a relatively small population allows public service resources to be allocated more generously on a per capita basis. From the perspective of welfare economics, such a distribution reflects the contribution of increased per capita allocable resources to aggregate social welfare through the social welfare function (Pigou, 1912; Arrow, 1948). When county-level populations are relatively small, investments in basic public services—such as education and healthcare—can more fully meet individual needs, thereby facilitating a more favorable balance between efficiency and equity.
Similar patterns are observed in northwestern provinces such as Qinghai, Ningxia, and Shaanxi, where comparable geographic and demographic characteristics are associated with service levels above the regional average. In contrast, southwestern provinces, including Chongqing, Sichuan, and Guizhou, characterized by higher population density, tend to maintain service levels around the regional median. Meanwhile, provinces such as Gansu, Yunnan, and Guangxi exhibit relatively lower overall service quality, which may be related to constraints in public investment and structural development bottlenecks.
Dimension-specific analysis shows that social welfare and SECS services generally exceed basic livelihood services. Inner Mongolia leads across all dimensions, supported by sustained investment. Other northwestern provinces follow similar patterns, confirming higher per capita levels of public services. The Sichuan-Chongqing region stands out in social welfare, particularly for vulnerable groups, while Gansu and Guizhou prioritize basic livelihood needs alongside improvements in higher-level services. By contrast, Yunnan and Guangxi face pronounced challenges due to remoteness, limited fiscal capacity, and unique demographic conditions, highlighting the need for increased investment to ensure equitable and multidimensional public service development.
This article employs ArcGIS’s natural break classification method to map the spatial distribution of basic public services and their dimensions in western counties for 2014 and 2022 (Fig. 2). Warm colors indicate higher service levels, while cool colors represent lower levels. From 2014 to 2022, counties with high levels of basic public services were concentrated in southeastern Xinjiang, northern Gansu, northern Qinghai, and eastern Inner Mongolia, with additional improvements in southern Qinghai and southwestern regions including Sichuan, Chongqing, northeastern Guizhou, and Yunnan. These areas gradually moved from low to moderate or higher service levels, reflecting strengthened construction and investments, particularly under the Belt and Road Initiative. In contrast, Ningxia, and Guangxi remained at moderate-to-low levels, constrained by remoteness, limited fiscal capacity, and geographic challenges, highlighting the need for targeted central support.
a 2014, b 2022. This map is based on the standard map No. GS (2022) 1873 downloaded from the standard map service website of China National Bureau of Surveying, Mapping and Geoinformation (NSGMGI). The same below.
Dimension-specific maps show that livelihood infrastructure historically led in eastern Xinjiang, Gansu, northern Qinghai, Ningxia, northern Shaanxi, and eastern Inner Mongolia (Fig. 3). While improvements occurred in Yunnan and Guizhou by 2022. Social welfare levels were highest in Qinghai, Inner Mongolia, and the Sichuan-Chongqing region (Fig. 4), with notable improvements in eastern Xinjiang and the Gansu-Ningxia-Shaanxi region by 2022. Border regions such as Yunnan, Guizhou, and Guangxi continued to face low social welfare provision. SECS remain concentrated at low per capita levels in most western counties (Fig. 5), except in sparsely populated or resource-rich areas like Qinghai, Inner Mongolia, and northern Shaanxi. By 2022, economically stronger provinces such as Shaanxi, Sichuan, and Chongqing advanced SECS from low to moderate levels, with neighboring provinces also benefiting. This indicates that development of higher-tier public services is closely linked to local economic strength, with national policies such as the Western Development Strategy reinforcing these trends.
a 2014, b 2022.
a 2014, b 2022.
a 2014, b 2022.
Analysis of spatial autocorrelation and aggregation characteristics of basic public services
The spatial autocorrelation of basic public services in western counties is calculated using Moran’s I index (Table 7).
During the study period, the global Moran’s I index for basic public services and its dimensions was consistently positive and significant at the 1% level, indicating a clear spatial positive autocorrelation effect across all dimensions. Among these, the spatial autocorrelation value for the social welfare dimension was the highest, followed by the science, education, and cultural dimensions, while the spatial autocorrelation for the basic livelihood dimension was the lowest, though still showing significant positive autocorrelation. Notably, over time, the spatial autocorrelation for basic public services and each dimension gradually diminished, suggesting a weakening of spatial autocorrelation effects. This indicates that the mutual influence and spatial agglomeration between counties have slowed over time. This may be due to the improvement in public service levels within individual counties, where certain dimensions have achieved self-sufficiency, thereby reducing the spatial spillover effects between counties. However, counties with similar geographical proximity and economic conditions still exhibit a certain degree of positive spatial interaction in public services and their dimensions.
Specifically, the spatial spillover effect for the basic livelihood dimension decreased the most, followed by social welfare, while the spatial spillover effect for the science, education, and cultural dimension showed the smallest decline. Analyzing the absolute values of spatial autocorrelation and their rates of decrease across these dimensions reveals that the spatial autocorrelation for the basic livelihood dimension is the weakest, while the social welfare and science, education, and cultural dimensions exhibit relatively stronger spatial autocorrelation. This disparity may stem from the nature of the services provided within each dimension. Basic livelihood services, which primarily involve labor, health, and housing, are closely tied to residents’ daily lives. As such, residents generally prefer to access these services within their own county, as the cost of obtaining them in neighboring counties is relatively high. Consequently, the spatial autocorrelation for the basic livelihood dimension is smaller. However, despite this, a certain level of spatial autocorrelation remains, particularly in counties with poorer economic conditions, where residents may still opt to obtain better basic services from nearby counties.
The social welfare dimension primarily concerns vulnerable groups such as children, the elderly, and the disabled, including facilities like kindergartens, nursing homes, and social welfare institutions. While these services are less closely related to daily life, they are crucial to meeting the basic needs of specific populations. As a result, residents are more willing to seek better welfare services across county lines, which contributes to the relatively strong spatial autocorrelation of social welfare. Similarly, the science, education, and cultural dimension exhibits similar characteristics. Education, culture, and sports are higher-level needs, particularly education resources. Parents are often willing to send their children to schools with better facilities, even beyond their own county. With the increasing prevalence of boarding schools, seeking educational resources across county boundaries has become common. Parents typically only need to cover transportation costs, which lowers the expense of accessing educational opportunities outside their immediate area. As a result, the spatial autocorrelation effect for the science, education, and cultural dimensions is also relatively strong.
This article uses ArcGIS to generate LISA cluster maps for basic public services and its dimensions in western counties for 2014 and 2022 (Fig. 6), to examine the spatial clustering differences in public service quality.
a Basic Public Services, b Livelihood Infrastructure, c Social Welfare, d Science, Education, Culture, and Sports.
In the spatial clustering of basic public services and their subdimensions, significant regional disparities are evident. Although there are “low-low” clusters in regions such as Ningxia, Guangxi, southern Sichuan, and western Yunnan, these areas also exhibit a relatively even distribution of “high-low” clusters, with high-level counties in more accessible locations. These counties can leverage regional radiation effects to promote the development of other low-level counties, thus forming a more effective regional coordination effect. Therefore, in the future, Ningxia, Guangxi, southern Sichuan, and western Yunnan are likely to improve basic public service levels through “high-low” clustering. In contrast, southern Shaanxi, Chongqing, and Guizhou exhibit “high-high” clustering, with good regional coordination, enabling the promotion of basic public services through clustering effects. Similarly, regions such as Gansu, Inner Mongolia, northern Shaanxi, and northeastern Xinjiang also show a “high-high” clustering phenomenon, indicating a strong effect of regional collaborative development in these areas.
In terms of spatial clustering in the livelihood infrastructure dimension, northern Qinghai, Gansu, and western Inner Mongolia mainly show “high-high” clustering, where the livelihood infrastructure has been mutually promoted in these regions. The situation in Shaanxi Province is more complex, with both “high-high” and “low-high” clustering, suggesting that some counties in Shaanxi still have relatively weak livelihood infrastructure and require the radiation effect of higher-level counties to foster regional progress. Similarly, Guizhou Province faces a situation akin to Shaanxi, where the radiation effect of high-level counties within the province is essential to improving the livelihood infrastructure in lower-level counties. In regions such as the southern part of Ningxia, the northwest of Sichuan, the southeast of Qinghai—areas bordering Ningxia, Qinghai, Sichuan—the counties exhibit a “low-low” clustering pattern. These regions lack sufficient regional driving forces, and thus, a comprehensive improvement in the livelihood infrastructure is necessary for sustainable development.
In the social welfare dimension, a noticeable “high-high” cluster has emerged in the northeastern part of Tibet, near the Qinghai border, indicating an enhanced regional radiative effect. Other counties with “high-high” clustering in social welfare are mainly distributed in Gansu, eastern and southern Inner Mongolia, northern Shaanxi, and northern Ningxia. These areas demonstrate positive regional coordination development effects, promoting the shared growth of counties. In contrast, the clustering of counties in Chongqing, Guizhou, and eastern Sichuan is more complex, with a mixture of “high-high” and “low-high” clusters. This suggests significant developmental disparities in these regions, which need to be addressed through further regional radiative effects to compensate for the deficiencies of low-level counties. Additionally, areas such as Ningxia, southern Sichuan, Yunnan, and Guangxi also exhibit “low-low” clustering, but the distribution of “high-low” clustering in some regions is relatively even, allowing the higher-level counties to promote the improvement of social welfare in the lower-level counties through radiative effects.
In the dimension of education, science, culture, and sports, regions such as the southern Qinghai, southern Ningxia, the northern part of Sichuan, northwestern Yunnan, and southern Guangxi generally show a “low-low” clustering pattern, similar to the pattern observed in the basic livelihood dimension. However, in the “low-low” clusters of southern Qinghai, southern Ningxia, northern Sichuan, and northwestern Yunnan, there is a presence of a certain number of “high-low” clusters. This indicates that these regions have relatively high-level counties in education, science, culture, and sports, which, through radiative effects, can drive the development of lower-level counties. In contrast, regions such as Guangxi show “low-low” clustering without a significant concentration of high-level counties in education, science, culture, and sports, indicating a lack of radiative effects from such areas. At the same time, areas like Gansu, Inner Mongolia, and northern Shaanxi generally exhibit “high-high” clustering, reflecting a strong mutual promotion of education, science, culture, and sports in these regions. Nevertheless, some regions still display a “low-high” distribution, indicating that, although there are high-level counties promoting each other, there is still a need to further improve the education, science, culture, and sports levels in the lower-level counties.
In summary, the spatial autocorrelation of basic public services and their subdimensions demonstrates the agglomeration and diffusion paths of public services across western counties: subsistence-oriented services exhibit weak spatial linkages, while the dimensions of social welfare and science, education, culture, and sports benefit from stronger spillover effects. Combined with insights from new economic geography, when some counties take the lead in upgrading public services, neighboring counties often achieve associated improvements through policy coordination, fiscal diffusion, and the flow of technology and human capital (Krugman, 1991). Therefore, to promote the balanced development of basic public services in western counties, it is necessary to strengthen cross-regional governance and resource sharing. In particular, for counties located in the “high–low” or “low–low” quadrants, balanced development can be advanced by enhancing inter-county cooperation mechanisms and leveraging information technologies to optimize the allocation and sharing of public service resources, thereby improving the quality and equity of basic public services.
The impact and spatial spillover effects of digital economy on basic public services and its dimensions
This article evaluates the direct effects and spatial spillover effects of the digital economy and its dimensions, thereby empirically verifying Hypothesis 1, exploring their influence on basic public services and each of its dimensions. The findings indicate that the digital economy, overall, exerts a significantly positive direct effect on basic public services and their respective dimensions across counties, with this effect confirmed at the 1% significance level (Table 8). This result suggests that the digital economy can significantly enhance the level of basic public services, which is consistent with the findings of Chen et al. (2024) and further verifies this relationship at the county level. From the perspective of the techno-economic paradigm, the digital economy, as the currently dominant technology, exhibits characteristics consistent with the “paradigm formation” stage in its investment and diffusion processes—namely, declining costs and the gradual emergence of network effects—which in turn drive the development of basic public services (Freeman and Louçã, 2001).
Specifically, the most notable effect is observed in the dimension of science, education, culture, and sports, followed by the social welfare dimension, while the impact on the livelihood infrastructure dimension is comparatively smaller. The science, education, culture, and sports dimension involves scientific technology and cultural activities, which heavily rely on the dissemination and innovation of knowledge. This is closely linked to the digital economy, as it provides a rapidly developing platform for knowledge dissemination, significantly accelerating the speed at which knowledge spreads. In contrast, the livelihood infrastructure and social welfare dimensions primarily concern the real economy and industries, with less reliance on virtual platforms. Although the digital economy is gradually influencing these traditional sectors—such as the intelligentization of healthcare, grid-based management of labor protection, and the application of elderly care robots—the transition from the real economy to digitalization requires time. Consequently, the transformation in these dimensions is slower, and while affected by the digital economy, their pace of change is not as fast as that in sectors like science, education, culture, and sports, which are more dependent on virtual platforms.
Moreover, while the digital economy has played a positive role in the development of public services within the counties, its impact on neighboring counties has exhibited significant negative effects. These effects are notably evident across all dimensions, especially in the livelihood infrastructure and science, education, culture, and sports dimensions, where the impact reaches a 5% significance level. This phenomenon suggests that although the development of the digital economy within the county has facilitated the digital transformation of its public services, it has also created a clear siphoning effect on neighboring counties, resulting in a localized Matthew effect. In counties with more developed digital economies, the digital transformation of basic public services has been advanced, allowing them to surpass surrounding counties and attract residents from those areas seeking higher-quality services. As a result, the development of public services in neighboring counties is hindered, potentially exacerbating the lag in their service development.
This phenomenon indicates that the development of the digital economy in western counties is still in its early stages. While some relatively developed counties have made progress in this area, the overall level remains low, and their radiative effect on surrounding regions is not yet significant. Therefore, as the digital economy continues to develop, although some more advanced counties will further transform, those counties lagging behind in digital economic development may lose the public service development opportunities they should have benefited from due to the siphoning effect. This development pattern could result in only a few counties benefiting from the digital economy, thereby exacerbating the imbalance in public services between regions.
Therefore, governments at all levels in western counties should pay closer attention to the localized effects of the digital economy, particularly its negative impact on the development of public services in neighboring counties. Efforts should be made to strengthen cooperation and experience sharing between counties to prevent further imbalance in regional public services due to the uneven development of the digital economy. Particular attention should be given to dimensions such as science, education, culture, and sports, which are more strongly influenced by the digital economy, as their spatial spillover effects are quite evident. This should be closely monitored to prevent exacerbating the public service gap between regions.
Endogeneity test
To mitigate potential endogeneity concerns to the greatest extent possible, this study follows the methodology of Nunn and Qian (2014) by constructing an instrumental variable (IV) for the digital economy. Specifically, this article generates an interaction term between the “Comprehensive Quality Index of the Digital Economy” and the historical data of “the number of Weibo-certified enterprises per 10,000 people across counties in 2007” (Wang et al., 2022). Subsequently, a Two-Stage Least Squares (2SLS) regression analysis is conducted to validate the results (Table 9).
The selection of this instrumental variable is predicated on the dual logic of relevance and exogeneity. Regarding relevance, the adoption of social media platforms such as Weibo by enterprises is not merely an application of digital technology; more profoundly, it reflects the early “digital awareness” and information processing capabilities of a county. Drawing upon the conceptual framework that treats historical fixed-line telephone and post office data from 1984 as “digital genes” (Huang et al., 2019), this study argues that 2007 marked the nascent stage of mobile internet development in China. The number of Weibo-certified enterprises across counties during this period captures the initial endowment for transitioning from traditional infrastructure to advanced digital applications. Through path-dependent effects, this early-formed preference for digital applications laid the technological foundation and environmental support for the subsequent explosive growth of the current digital economy, thereby satisfying the relevance condition.
Regarding exogeneity, as 2007 significantly predates the rapid expansion period of the digital economy analyzed in this study (2014–2022), the lead time of this historical marker ensures that current digital economic developments cannot retroactively influence enterprise digital behavior in 2007. This temporal precedence effectively eliminates reverse causality interference. Simultaneously, the number of Weibo-certified enterprises in 2007 reflects the historical internet activity level at the county level and does not directly determine contemporary public service provision. Its impact on public services is exercised indirectly through the structural transformation of the current digital economy. By constructing an interaction term between cross-sectional data and a time-varying index, this study captures the dynamic evolutionary path of the digital economy across different spatial and temporal dimensions. This approach satisfies the exclusion restriction while enabling a more accurate identification of the “net effect” of the digital economy on public services.
In the DWH test, the endogeneity test statistic for the core explanatory variable, digital economy, passes at the 5% significance level, indicating a strong endogeneity effect of the digital economy on basic public services. Therefore, conducting an endogeneity test is necessary.
First, the validity and reliability of the selected instrumental variable need to be assessed. The testing process includes the identification test and the weak instrument test. In the identification test, regardless of whether socio-economic variables are controlled for, the Kleibergen-Paap rk LM statistic passes at the 1% significance level, suggesting that the selected instrument is correlated with the core explanatory variable, digital economy.
In the weak instrument test, the Cragg-Donald Wald F statistic far exceeds the critical value of 16.38 proposed by Stock and Yogo, regardless of whether other variables are controlled. This confirms that the selected instrument is not weak and is strongly correlated with the core explanatory variable, digital economy. The instrument captures most of the information related to the digital economy and does not cause significant interference or bias into the results.
Both the identification test and the weak instrument test demonstrate that the selected instrumental variable is reasonable and valid, and can be used as a proxy for the digital economy in the next stage of the analysis.
In the second-stage test, the instrumental variable for digital economy still exerts a positive effect on basic public services at the 1% significance level. Regardless of whether other socio-economic variables are controlled for, the effect size remains significantly positive, and the direction and significance of this effect align with the impact of digital economy itself on basic public services. This indicates that, after addressing the endogeneity issue, the effect of digital economy on basic public services remains robust, and the endogeneity problem does not affect the main conclusion.
Robustness check
Based on the above research results and endogeneity tests, the direct effect and spatial spillover effect of the digital economy on basic public services have been verified. To further ensure the robustness of the regression results, this study employs two methods for robustness checks: (1) replacing the spatial weight matrix and (2) removing outliers. These checks help verify the reliability of the results and eliminate potential biases.
Replacing the spatial weight matrix
To further validate the robustness of the regression results, this study adopts the method of replacing the spatial weight matrix. In previous sections, this article used a spatial weight matrix that combines economic and geographical distances, aiming to capture the mutual learning effects between counties with similar economic levels and the spillover effects between geographically proximate counties. However, in practice, the influence of geographical distance on inter-county interactions is often more direct and rapid than economic distance. Most counties are primarily affected by the geographic radiation of neighboring counties.
Furthermore, existing studies have shown that as the geographical distance between counties increases, the impact effect between them generally decreases exponentially. Therefore, using an inverse-distance squared geographical weight matrix better reflects the actual situation. Compared to the standard inverse-distance geographical weight matrix, the inverse-distance squared matrix more accurately captures the impact of geographic proximity on spatial spillover effects. As a result, this study ultimately adopts the inverse-distance squared geographical weight matrix to replace the previous economic + geographical distance matrix for the robustness check (Table 10).
After replacing the “economic–geographic” composite weight matrix with the inverse distance second-order geographic weight matrix, the direct effects of the digital economy on basic public services and their subdimensions remain significantly positive at the 1% level, with magnitudes highly consistent with those in the baseline model, confirming the robustness of the results. The spatial spillover effects of the digital economy on neighboring counties’ basic public services remain significantly negative, and the total effects across the entire region are also negative. This finding is consistent with the baseline regression results, indicating that the directions of the direct, spillover, and total effects do not change notably under the robustness test, thereby demonstrating the strong reliability of the baseline estimates.
Removal of extremely high-value outliers
The western region of China is vast, and the counties within it exhibit significant disparities in the stages and levels of digital economy and basic public services due to differences in geographic location, economic conditions, and natural endowments. According to the 2022 China County Digital Economy Top 100 list published by the DCIC 2023 China Data Center Industry Development Conference and the “China County Digital Economy Development Report,” some counties in the western region have already entered the ranks of the top 100 counties in digital economy development. This indicates that these counties have begun and advanced their digital economy development earlier and more rapidly, gradually surpassing most other counties in the western region. As a result, their digital economy levels are significantly higher than those of other western counties, diminishing their representativeness.
To avoid the disproportionate influence of these high-ranking counties on the results, this article decides to exclude the top 100 counties in digital economy development in order to remove the interference of extreme outliers. A robustness check is then conducted on the remaining counties to more accurately examine the spatial effects of the digital economy on basic public services in a context that is more representative of the majority of counties in the western region (Table 11).
After excluding county samples with exceptionally high digital economy levels—those listed among the Top 100 Counties—from the western region, the direct effects of the digital economy on county-level basic public services and their subdimensions remain significantly positive at the 1% level, while the spatial spillover effects remain significantly negative. This indicates that, after removing these extreme-value counties, the regression results for the remaining, more representative western counties are largely consistent with the baseline results in both magnitude and direction. Therefore, the positive impact of the digital economy on local public services and its negative spillover effect on neighboring counties are robust features in western China, unaffected by the exclusion of counties with exceptionally advanced digital economy development.
Spatial heterogeneity of the impact of digital economy and its dimensions on basic public services
First, the diff index for digital economy and its dimensions is calculated to identify which dimension is a local variable. The spatial mismatch of digital economy and public service facilities in western counties is a complex phenomenon, influenced by various factors throughout its evolution. This article first uses the diff index to determine whether the selected influencing factors are global or local variables (Table 12).
Then, the authors introduce the geographical and temporal weighted regression model (GTWR) to conduct local spatial regression analysis on the factors affecting the spatial fit of digital economy and public service facilities. The analysis explores the regression coefficients of the influencing factors (Table 13) and the spatial variation in their effects.
As shown in Table 11, the Diff of Criterion values for the main explanatory variables of digital economy and its sub-dimensions are all negative, indicating significant spatial local correlation. Therefore, these variables can be treated as local variables, and the geographic and temporal weighted regression method is used to explore the spatial heterogeneity of the impact of the digital economy and its sub-dimensions on public services at the spatial unit level.
The AICc values of the GTWR model are small, indicating that the model demonstrates a good overall performance with a reasonable goodness of fit. Further examination of the spatial heterogeneity of the impact of digital economy and its dimensions reveals the localized nature of their effects.
In all the western counties studied, the digital economy positively influences both basic public services and their sub-dimensions, with stronger effects in certain regions (Fig. 7). Qinghai, Gansu, and Inner Mongolia show the most significant impact, with regression coefficients for the digital economy on basic public services generally exceeding 0.45. A 1% increase in the digital economy in these counties raises basic public service levels by at least 0.45%. The large marginal effects in these provinces are likely due to their inland, sparsely populated nature, which results in a more noticeable impact from the digital economy’s development. Despite having relatively higher per capita basic public services, these provinces still have weak overall public service strength due to vast areas and low population density, making the digital economy-driven transformation more apparent. Similar effects are seen in Ningxia, Shaanxi, and central and eastern Xinjiang, with coefficients ranging from 0.32 to 0.45. However, in the southwest, such as Sichuan, Chongqing, and Yunnan, the coefficients range from 0.26 to 0.32, while in more remote areas like Guizhou and Guangxi, they fall between 0.20 and 0.26. This weaker impact in the southwest suggests that, as the digital economy matures, its marginal effects may diminish, indicating a slowdown in benefits and a potential bottleneck.
a Basic Public Services, b Livelihood Infrastructure, c Social Welfare, d Science, Education, Culture, and Sports.
In conclusion, the effects of the digital economy on basic public services differ significantly across development stages. In particular, during the early stage of digital economy development, the marginal effects are more pronounced, whereas in the mature stage, the effects tend to level off. This stage-specific pattern is highly consistent with the findings of Hui et al. (2023), who show that although digital economy development can significantly improve the quality of public services, the magnitude of this improvement diminishes as the level of digital economy development increases, resulting in regional heterogeneity in outcomes. Building on this literature, this study further characterizes the differentiated effects of the digital economy on public services from a developmental stage perspective, providing new empirical evidence for related conclusions. Within the framework of the techno-economic paradigm, counties in the northwestern region are generally situated in the “early paradigm” stage, where the digital economy exhibits larger marginal effects and can rapidly advance the digitalization of public services, thereby elevating overall service levels. By contrast, counties in the southwestern region are gradually entering the “paradigm maturity” stage, in which the marginal effects of the digital economy tend to moderate (Perez, 1983). Looking ahead, efforts should focus on accelerating the transition of western counties from the “early paradigm” toward “paradigm shift,” leveraging innovation diffusion in the digital economy to drive the digital development of basic public services.
From a basic livelihood perspective, the digital economy has the strongest impact in Qinghai, Gansu, and Inner Mongolia, particularly in Gansu, where the impact coefficient ranges from 0.16 to 0.25, and in Qinghai and Inner Mongolia, where it ranges from 0.10 to 0.16. These regions, still transitioning from traditional stages of development, experience substantial positive effects from the digital economy, as it has not yet fully penetrated their livelihood foundations. In contrast, regions like Sichuan, Chongqing, Guizhou, and Guangxi show minimal effects (coefficients ranging from 0.04 to 0.06), where the digital economy’s impact is limited due to already established livelihood foundations or underdeveloped digital economies. In Shaanxi, Ningxia, Yunnan, and eastern Xinjiang, the digital economy’s impact is moderate but still positive, with coefficients ranging from 0.06 to 0.10.
In the social welfare dimension, counties in Gansu and eastern Inner Mongolia experience the strongest impact from the digital economy, with coefficients ranging from 0.26 to 0.40. Qinghai, Shaanxi, Ningxia, and eastern Xinjiang also show a significant impact, with coefficients between 0.14 and 0.26. In other regions, the effect of the digital economy on social welfare is weaker, particularly in Sichuan, Chongqing, and Yunnan, where coefficients range from 0.11 to 0.14. The weakest impacts are seen in Guizhou and Guangxi, where coefficients fall between 0.07 and 0.11. These southwestern and border regions, characterized by remote locations, complex terrain, and economic underdevelopment, face challenges in developing social welfare, making it harder for the digital economy to generate substantial positive effects. This highlights the need for further exploration of the digital economy’s impact on social welfare in these areas.
The digital economy’s influence on basic public services and sub-dimensions exhibits notable spatial differences between the northwest and southwest. This is particularly evident in the dimension of education, science, culture, and sports. In the northwest, counties in Qinghai, Gansu, and Inner Mongolia experience strong effects from the digital economy, with coefficients ranging from 0.23 to 0.47. Other northwest counties, such as eastern Xinjiang, Ningxia, and Shaanxi, show considerable influence, with coefficients between 0.10 and 0.23. In contrast, the digital economy’s impact in the southwest—regions like Sichuan, Chongqing, Yunnan, Guizhou and Guangxi—is positive but significantly weaker, with coefficients generally between 0.06 and 0.10. This suggests that southwestern counties should focus on better integrating the digital economy with public services to maximize their mutually supportive development. This presents a major challenge for the southwest region, where coordinated development is crucial for improving basic public services (Huang et al., 2025).
Moderating mechanisms
The following analysis examines the moderating effects of macro-environmental factors on the impact of the digital economy, thereby empirically verifying Hypothesis 2. Some estimated effects are statistically significant, while others are not. Non-significant coefficients do not imply the absence of a relationship; they may arise from factors such as lagged or gradually accumulating effects, the complexity of the underlying mechanisms, or condition-dependent influences. Even when a coefficient does not reach statistical significance, its direction and magnitude can still reveal underlying tendencies consistent with theoretical expectations, potentially reflecting patterns worthy of attention.
Population density as a moderator in the impact of the digital economy on public services
In the process through which the digital economy influences basic public services and their sub-dimensions in western counties, the interaction mechanism of population density exhibits complex variations (Table 14). An increase in population density may weaken the positive effects of the digital economy on the livelihood infrastructure and social welfare dimensions within the county, yet it strengthens the promotion effect on the science–education–culture–sports dimension. This differentiated outcome indicates that population agglomeration exerts a dual interaction mechanism in the process of digital empowerment of basic public services. High population density may intensify the supply pressure for livelihood and social welfare services and reduce resource allocation efficiency, thereby weakening the optimizing effect of the digital economy. Although digital technologies can enhance service management and scheduling efficiency, their effectiveness still depends on physical infrastructure capacity and resource input. When facilities are close to saturation or resources are overly concentrated in densely populated areas, the digital economy is less able to accommodate excessive demand, leading to diminishing marginal improvements in livelihood infrastructure and social welfare. In contrast, high population density facilitates the concentration of demand for knowledge and innovation, thereby creating stronger demand pull and application scenarios for digital development in the science–education–culture–sports dimension. Western counties should therefore optimize population distribution and spatial carrying capacities to prevent excessive concentration from weakening the positive effects of the digital economy on basic public services.
From the perspective of spatial spillover effects, population density also interacts with the diffusion path of the digital economy in a complex manner. A rise in population density intensifies the negative spillover effects of the digital economy on the livelihood infrastructure and science–education–culture–sports dimensions in neighboring counties, whereas in the social welfare dimension it provides a certain buffering effect. The results indicate that when population density becomes excessive, the negative spatial spillover effects of the digital economy on livelihood and science–education–culture–sports services may be amplified. This occurs because the development of the digital economy tends to concentrate in areas with dense populations and strong market demand, attracting capital, technology, and information elements toward these high-density counties. Surrounding counties then face a disadvantage in the competition for public resources related to livelihood and science–education–culture–sports services, causing relative contractions in service provision and exacerbating “crowding-out effects.” Western counties should enhance the cross-county mobility and allocation of digital economy resources, promote the spatial balance of basic public services, and guide the rational flow of resources and services among counties to mitigate the intensified negative spillover effects caused by population overcrowding.
As population density changes, the effects of the digital economy and its sub-dimensions on basic public services exhibit clear multi-stage threshold characteristics, confirming that population density plays a conditional role in the mechanism of influence (Table 15). When population density in western counties is below the threshold of 298 persons/km², the digital economy significantly promotes basic public services, although its effect gradually weakens as density increases. When population density is below 37 persons/km², the enabling effect of the digital economy is strongest: each one-unit increase in the digital economy level enhances basic public services by 0.322 units. When density ranges from 37 to 164 persons/km², the effect coefficient declines to 0.199; when density further increases to 164–298 persons/km², the coefficient drops to 0.105—approximately 33% of the effect observed in low-density counties. Once population density surpasses 298 persons/km², the positive effect of the digital economy on basic public services becomes statistically insignificant, suggesting that excessive population concentration diminishes the marginal contribution of the digital economy.
These findings reflect the structural role of population density differences within the mechanism. The threshold of 298 persons/km² delineates an “optimal density zone” versus an “overcrowded zone” for spatial population distribution in western counties. When population density is below this threshold, counties tend to have dispersed populations and greater spatial distances; investments in digital technologies and infrastructure reduce spatial barriers to information transmission and factor mobility, generating significant efficiency gains by lowering distance costs and improving the accessibility of online public services. In contrast, when population density exceeds the optimal range, excessive agglomeration can lead to service information overload, network congestion, and intensified resource competition, reducing the fluidity and diffusion of digital elements, and thereby weakening the overall promotion effect of the digital economy on basic public services.
The moderating mechanism of economic development in the impact of digital economy on basic public services
Per capita GDP plays a significant supportive role in the process through which the digital economy influences basic public services and their sub-dimensions in western counties. Increases in per capita GDP strengthen the positive effect of the digital economy on basic public services within the county at the 1% significance level (Table 16). At the sub-dimensional level, this supportive effect is most prominent in the science–education–culture–sports dimension, followed by the social welfare dimension, indicating that these dimensions are more sensitive to changes in economic development compared with the livelihood infrastructure dimension. Higher per capita GDP also alleviates the negative spatial spillover effects of the digital economy on basic public services and their sub-dimensions in surrounding counties. Notably, in the science–education–culture–sports dimension, an increase in per capita GDP significantly reduces negative spillovers at the 10% level, suggesting that economic development helps optimize the diffusion environment and external effects of the digital economy.
Mechanistically, higher per capita GDP provides solid support for the technological penetration and service transformation of the digital economy. Higher per capita GDP not only represents stronger local economic capacity, but also reflects greater residents’ willingness to pay and consumption capability for digital services. This enhances the enabling effect of the digital economy on basic public services within the county. In addition, improved per capita GDP may strengthen economic linkages and factor mobility across counties, thereby alleviating the “central siphon effect” that occurs when the digital economy disproportionately attracts resources to core counties. Counties with higher per capita GDP tend to be more open and cooperative, which promotes the sharing of digitally enabled public service benefits across counties while mitigating adverse spillover effects on neighboring counties.
Economic development also exhibits a significant threshold effect in the process through which the digital economy and its dimensions influence basic public services (Table 17). As per capita GDP increases, the positive effect of the digital economy on basic public services intensifies, indicating that economic development considerably amplifies the enabling role of the digital economy. At the aggregate level, the digital economy demonstrates a triple-threshold mechanism with respect to per capita GDP.
When per capita GDP is below 12,693 yuan per person, the effect of the digital economy on basic public services is positive but not statistically significant. This suggests that in counties with low per capita GDP, the digital economy fails to exert meaningful influence due to insufficient economic capacity, inadequate resource allocation, lagging infrastructure, and low digital penetration, which restrict the development and delivery of digital services. Once per capita GDP exceeds 12,693 yuan per person, the digital economy begins to significantly improve basic public services, reflecting improvements in infrastructure and wider application of digital technologies. As per capita GDP continues to rise, the influence of the digital economy strengthens further, indicating that counties with higher economic capacity are better able to unleash the potential of digital technologies in enhancing public service provision.
Specifically, when per capita GDP ranges between 12,693 and 26,515 yuan per person, the effect coefficient is 0.0324, meaning that a one-unit increase in the digital economy index results in a 0.032-unit improvement in basic public services. When per capita GDP exceeds 26,515 yuan but remains below 53,865 yuan per person, the effect coefficient rises to 0.066, implying that economically stronger counties are more capable of integrating the digital economy into public service development, thereby generating more effective digital empowerment. In counties where per capita GDP surpasses 53,865 yuan per person, the enabling effect further intensifies, with the coefficient increasing to 0.12, showing that a stronger economic foundation creates more favorable conditions for the full realization of digital economy benefits.
These threshold effects reflect the multi-stage, nonlinear nature of the digital economy’s influence on basic public services. In counties with low per capita GDP, the lack of infrastructure, insufficient digital resources, and limited technological adoption prevent the digital economy from exerting significant influence, even though the potential exists. As economic development improves, the digital economy begins to play a more visible enabling role. Higher per capita GDP indicates strengthened fiscal capacity, improved public investment ability, higher resident consumption levels, and greater market dynamism and industrial upgrading. This provides financial support for the construction of digital infrastructure and technological upgrading, facilitating the effective utilization of digital resources in the public service domain and enabling the digital economy to perform more effectively in an improved economic environment.
The regulatory mechanism of the advanced industrial structure factor in the impact of the digital economy on basic public services
The transition of industrial structure toward the tertiary sector in western counties significantly enhances the positive effect of the digital economy on the science–education–culture–sports dimension at the 1% significance level (Table 18). The expansion of the tertiary sector not only provides a more solid foundation of emerging industries for digital applications, but also strengthens the internal driving force of digital empowerment in basic public services through digital integration in education, culture, and science and technology services. Although the interaction effect of industrial upgrading on the livelihood infrastructure and social welfare dimensions is not statistically significant, a clear trend is observed. Industrial upgrading appears to enhance the effect of the digital economy on livelihood infrastructure but may weaken its positive impact on the social welfare dimension. This may be because, as western counties shift toward high value-added service industries, industrial upgrading encourages digital economy applications to first embed in livelihood infrastructure services—such as online medical services, employment platforms, and public housing information systems—thereby improving the efficiency of service provision. In contrast, social welfare services, such as programs for vulnerable groups, are constrained by scattered demand, long investment cycles, and higher management complexity, making digital transformation slower. Thus, while industrial upgrading strengthens digital empowerment in livelihood infrastructure, it may temporarily reduce the diffusion of digital benefits in the social welfare dimension.
From the perspective of spatial spillover effects, industrial upgrading consistently intensifies the negative spillover effects of the digital economy on basic public services and their sub-dimensions in surrounding counties. Industrial upgrading is typically accompanied by further concentration of factor resources and emerging industries. Counties with a more developed tertiary sector tend to attract more capital, talent, information, and technology, generating a strong siphon effect on neighboring counties. Although such agglomeration enhances development within the county, it may reduce neighboring counties’ access to resources and exacerbate the negative spatial spillover of the digital economy. Mechanistically, counties with a more advanced tertiary sector gain early advantages in digital infrastructure and data resources, and this advantage persists over time. As a result, these counties maintain a leading position in digital economy development and public service provision, further widening inter-county disparities—reflecting a “Matthew effect.” To mitigate the risks of spatial divergence caused by industrial agglomeration, western counties should promote advanced industrial transformation with a focus on digital and knowledge-intensive service industries, accelerate the digital transformation of traditional industries, and build inter-county industrial collaboration networks. Establishing shared digital service platforms may help strengthen the supportive role of industrial upgrading in the digital economy and enhance its enabling effect on basic public services.
Industrial upgrading also exhibits a significant threshold effect in the influence of the digital economy and its sub-dimensions on basic public services (Table 19). In the impact of the digital economy on basic public services as a whole, industrial upgrading follows a double-threshold mechanism. When the industrial transformation index exceeds the threshold value of 0.9175—meaning that the output value of the tertiary sector approaches that of the secondary sector—the role of the tertiary sector in the county economy becomes increasingly prominent. At this stage, the effect of the digital economy on basic public services shifts from insignificant to significantly positive at the 1% level, with a coefficient of 0.0404. The reason is that when the ratio of tertiary to secondary sector output is below 0.9175, the county economy remains dominated by resource-based industries and processing manufacturing, where digital technology adoption in production and management is limited, and mature digital application scenarios in basic public services are lacking. As the tertiary-sector share increases and the service industry expands, social demand related to education, healthcare, culture, and sports grows, creating a more conducive market environment for the coordinated development of the digital economy and basic public services. This enables the positive effect of the digital economy to be fully released.
Fiscal revenue and expenditure ratio as a moderating mechanism in the impact of the digital economy on basic public services
As a key institutional constraint reflecting fiscal balance, the fiscal revenue–expenditure ratio weakens the positive effect of the digital economy on basic public services and their sub-dimensions within the county (Table 20). The fiscal revenue–expenditure ratio reflects the relationship between fiscal revenue and expenditure, and its increase may result from either rising revenue or reduced expenditure. In western counties, however, due to natural endowment limitations and slower economic development compared with the eastern and central regions, fiscal revenue is generally limited. Therefore, increases in the fiscal revenue–expenditure ratio mainly stem from reductions in fiscal expenditure rather than increases in fiscal revenue. This conclusion echoes the findings of Zhan et al. (2022), who point out that local fiscal pressure exerts a nonlinear effect on public service quality. Once fiscal pressure exceeds a certain threshold, local governments, in an effort to maintain fiscal balance, tend to adjust the structure of fiscal expenditures, thereby significantly reducing the quality of public services within their jurisdictions. At the county level, this study empirically corroborates the mechanism linking fiscal pressure, expenditure structure adjustment, and the decline in public service quality, providing empirical support that aligns with the conclusions of the existing literature.
Reductions in fiscal expenditure restrict the local government’s ability to secure key resources needed for digital economy development. Such constraints weaken financial support for digital infrastructure and service platform construction, thereby limiting the digital economy’s capacity to expand the coverage and depth of basic public services. Consequently, when the fiscal revenue–expenditure ratio rises—implying fiscal expenditure contraction—the positive effect of the digital economy on education, healthcare, and social security may diminish, reducing its ability to improve basic public services within the county. When fiscal expenditure does not match the investment required for digitalization, even if digital technologies are available, the digital economy cannot effectively empower public service delivery. This mismatch between fiscal capacity and digital development needs may weaken the effectiveness of digital initiatives in improving public services. Reduced fiscal expenditure may hinder the construction and maintenance of online education platforms, telemedicine systems, and social security digitalization, thereby suppressing the effectiveness of the digital economy in the public service sector.
A rising fiscal revenue–expenditure ratio further amplifies the negative spatial spillover effect of the digital economy on basic public services in surrounding counties, particularly in livelihood infrastructure and social welfare dimensions. When a county maintains fiscal balance by reducing expenditure, investment in digital infrastructure, online service platforms, and digitalized social security systems becomes constrained. Insufficient fiscal resources not only weaken the digital economy’s enabling effects within the county, but also exacerbate negative spillovers due to restricted information flow and reduced digital service efficiency. The results confirm that fiscal expenditure contraction not only dampens the positive effect of the digital economy on basic public services within the county, but also aggravates negative spillovers due to insufficient sustainability of digital projects and limited service provision. Thus, western counties should allocate fiscal expenditure more strategically to support digital development while ensuring fiscal sustainability, seeking coordination between improvements in basic public services and digital economy development.
In the relationship between informatization development and basic public services, the fiscal revenue–expenditure ratio demonstrates a single threshold value of 0.2915 (Table 21). When the ratio is below 0.2915, informatization significantly promotes basic public services at the 10% level, with a coefficient of 0.0063, meaning that a one-unit increase in informatization improves basic public services by 0.0063 units. At this stage, although fiscal capacity remains limited, expenditure is relatively flexible, allowing investment in foundational digital projects such as informatization construction, and fiscal transfers from higher-level governments play a key supporting role. Once the fiscal revenue–expenditure ratio exceeds the threshold, fiscal expenditure becomes constrained by budget balance requirements. Driven by economic performance considerations, local governments prioritize spending on sectors that generate quick and visible economic returns—such as industrial development and industrial park construction—while reducing investment in digitalization projects, which require long-term funding and yield slower returns. Consequently, the positive effect of informatization weakens and may become insignificant, indicating that informatization in western counties has not yet formed a self-sustaining development mechanism and still heavily depends on fiscal investment.
A similar threshold pattern appears in the effect of internet development on basic public services. The fiscal revenue–expenditure ratio shows a single threshold of 0.3398. When the ratio is below 0.3398, internet development significantly promotes basic public services at the 1% significance level, with a coefficient of 0.0333, implying that a one-unit increase in internet development increases basic public services by 0.0333 units. As the fiscal revenue–expenditure ratio rises, fiscal expenditure tightens. Because internet infrastructure requires long construction cycles and yields delayed returns, and because the importance of internet development is often underestimated in western counties, fiscal resources tend to be allocated to areas with short-term economic gains. Under these conditions—where fiscal constraints and performance pressures coexist—investment in internet development becomes marginalized. Insufficient construction and maintenance weaken the role of internet development in supporting basic public services, causing its positive impact to become insignificant.
Conclusion
This study comprehensively examines how the digital economy influences basic public services in western China’s counties, revealing both progress and persistent disparities. While social welfare, education, and science-cultural services have improved, fundamental livelihood services such as income security, healthcare, and housing remain underdeveloped, particularly in remote and ethnic regions like Guangxi, where lower-level counties exhibit persistent “low-low” clustering. Spatial patterns indicate pronounced regional clustering: high-level counties function as growth hubs, whereas lower-level counties remain under-resourced, highlighting the need for optimized inter-regional resource allocation and coordinated development. The risk of a “siphon effect,” where digital benefits concentrate in core counties, further underscores the importance of strengthening digital infrastructure and promoting inclusive digitalization to diffuse benefits more evenly.
Empirical results reveal that macro-contextual factors critically shape the digital economy’s impact. Population density exhibits a dual mechanism: excessive local density weakens the positive effects of informatization and internet development on basic livelihood and social welfare, while spatial over-concentration intensifies negative spillovers on neighboring counties in basic livelihood and education-culture-sports dimensions, yet mitigates spillovers in social welfare. When population density exceeds 298 persons/km², the positive impact of informatization and internet development diminishes, whereas digital transactions gain in influence. Higher per capita GDP (above 26,515 yuan) amplifies positive effects on education-culture-sports and social welfare, and an industrial upgrading index above 0.9175 significantly enhances the contribution of digitalization, particularly internet development, although industrial upgrading may reduce direct effects on social welfare under certain conditions. Fiscal constraints also matter: with a revenue–expenditure ratio below 0.2915, a one-unit increase in informatization improves basic public services by 0.0063 units; tighter fiscal conditions diminish these effects and can intensify negative spillovers. Collectively, moderate population concentration, higher economic development, mature industrial upgrading, and stable fiscal support form a multidimensional threshold system enabling the digital economy to fully enhance county-level public service provision.
Based on these findings, several targeted policy measures are recommended. First, counties with high population density should expand internet infrastructure, including bandwidth upgrades, support for digital transaction platforms, and edge caching deployment, while remote or sparsely populated counties require targeted digital grants and mobile service delivery models to ensure access to essential services such as telemedicine and online education. Second, inter-county collaboration should be promoted through co-investment in regional digital platforms, enabling wealthier counties to share technological dividends with neighbors. Third, in regions undergoing industrial restructuring, digital public service initiatives should align with local industrial strengths while mitigating potential negative spillovers in social welfare. Finally, fiscal policy should prioritize precision targeting: unallocated surpluses should be converted into conditional grants to expand digital public services in underdeveloped counties, fostering balanced county-level development.
Theoretically, this study contributes by focusing on the county level, a relatively underexplored but policy-relevant unit. Unlike previous research at provincial or urban scales (Li et al., 2024; Lee et al., 2021; Li et al., 2020; Ma et al., 2020), it integrates spatial econometric modeling and addresses endogeneity to provide a nuanced understanding of the mechanisms through which digitalization affects public services. By extending New Public Management Theory (Hood, 1991) and Public Choice Theory (Buchanan et al., 1965) to county-level digital transformation, and by adapting regional economic concepts such as geographic proximity and spillover effects (Krugman, 1991; Florida, 1995), the study enriches the discourse on public service efficiency, resource allocation, and regional coordination under digital conditions, offering a foundation for future interdisciplinary research on digital transformation, service equity, and county-level governance.
Despite these contributions, the study has several limitations. This study provides a general view of the digital economy’s impact without distinguishing its effects at different development stages. Future work should examine the initial, transitional, and mature phases of digital development to better understand stage-specific mechanisms. In addition, although the robustness tests conducted in this article enhance the reliability of the results to some extent, they could be made more comprehensive. Future studies may adopt more diverse approaches, such as testing for omitted variables and re-estimating the models by replacing core variables to further validate the robustness of the findings. Finally, micro-level studies—drawing on household surveys, qualitative interviews, and case studies—are needed to uncover how digitalization affects individuals, businesses, and communities. These methodological expansions will improve the explanatory power of the analytical framework, inform more effective policy design, and advance the synergy between the digital economy and public services.
Data availability
The data from this article cannot be shared as it will be used for the team’s upcoming research.
References
AlOmari F (2021) Measuring gaps in healthcare quality using SERVQUAL model: challenges and opportunities in developing countries. Meas Bus Excell 25(4):407–420. https://doi.org/10.1108/MBE-11-2019-0104
Anselin L (2022) Spatial econometrics. In Handbook of spatial analysis in the social sciences (pp 101–122). https://doi.org/10.1016/0165-1765(78)90097-6
Arrow KJ (1948) The possibility of a universal social welfare function
Buchanan JM, Tullock G (1965) The calculus of consent: logical foundations of constitutional democracy (Vol. 100). University of Michigan Press. https://doi.org/10.2307/1913055
Chen MG, Liu YS, Zhang Q (2024) The impact of the digital economy on basic public services. Res Financ Econ Issues (4), 81–93. https://doi.org/10.19654/j.cnki.cjwtyj.2024.04.007
Dener C, Nii-Aponsah H, Ghunney LE, Johns KD (2021) GovTech Maturity Index: the state of public sector digital transformation. World Bank Publications
Florida R (1995) Toward the learning region. Futures 27(5):527–536. https://doi.org/10.1016/0016-3287(95)00021-N
Freeman C, Louçã F (2001) The strange attraction of tides and waves. As time goes by: from the industrial revolutions to the information revolution (pp 288–322). Oxford University Press
Gladkova A, Ragnedda M (2020) Exploring digital inequalities in Russia: an interregional comparative analysis. Online Inf Rev 44(4):767–786. https://doi.org/10.1108/OIR-04-2019-0121
Griliches Z (1991) The search for R&D spillovers (NBER Working Paper No. 3768). National Bureau of Economic Research
Hood C (1991) A public management for all seasons? Public Adm 69(1):3–19. https://doi.org/10.1111/j.1467-9299.1991.tb00779.x
Huang Q, Yu Y, Zhang S (2019) Internet development and manufacturing productivity improvement: Internal mechanisms and evidence from China. China Ind Econ (8), 5–23. https://doi.org/10.19581/j.cnki.ciejournal.2019.08.001
Huang X, Lan F (2025) The influencing factors and regional heterogeneity of county-level public services quality in China. Int J Urban Sci, 1–33. https://doi.org/10.1080/12265934.2025.2452502
Huang X, Lan F (2025) Spatial–temporal mismatch between the digital economy and public services: case study of counties in Shaanxi, China. J Urban Plan Dev. https://doi.org/10.1061/JUPDDM/UPENG-6038
Huang X, Lan F (2025) Navigating digital transformation: synergistic pathways and challenges in county-level public service development. Sage Open 15(4):21582440251397672. https://doi.org/10.1177/21582440251397672
Hui N, Ning N (2023) Effects and mechanisms of digital economy-driven improvements in public service quality. J Beijing Univ Technol (Soc Sci Ed) 23(1):109–124
Jiang F, Chen B, Li P et al. (2023) Spatio‑temporal evolution and influencing factors of synergizing the reduction of pollution and carbon emissions: utilizing multi‑source remote sensing data and GTWR model. Environ Res 229:115775. https://doi.org/10.1016/j.envres.2023.115775
Jiang XP, Kang J (2020) Degree of achievement: a new perspective and indicator construction for evaluating the equalization of basic public services. Chin Public Adm 10:73–79. https://doi.org/10.19735/j.issn.1006-0863.2020.10.10
Krugman P (1991) Increasing returns and economic geography. J Political Econ 99(3):483–499. https://doi.org/10.1086/261763
Lee SM, Lee D (2021) Opportunities and challenges for contactless healthcare services in the post-COVID-19 era. Technol Forecast Soc Change 167:120712. https://doi.org/10.1016/j.techfore.2021.120712
Li H, Dong YL (2020) Measurement and trend evolution of China’s basic public service equalization: a study based on the dimension of high-quality development. China Soft Sci 10:74–84. https://doi.org/10.3969/j.issn.1002-9753.2020.10.008
Li JX, Liu LT, Huo JJ (2022) Spatiotemporal pattern and influencing factors of rural public service supply quality in China. Econ Geogr 42(6):132–143
Liu J, Yang Y, Zhang SF (2020) Research on the measurement and driving factors of China’s digital economy. Shanghai Econ Res 81 (6), 81–96
Liu XC, Li C, Xiong HJ (2021) Analysis of regional basic public service equalization and its influencing factors in China. J Jiangxi Soc Sci (6), 77–88
Li X, Ma Y, Zhang W (2024) Enhancing urban public service efficiency through the national civilized city policy: an empirical analysis of 282 cities in China from 2005 to 2019. J Asian Public Policy, 1–30. https://doi.org/10.1080/17516234.2024.2335434
Maslow AH (1943) A theory of human motivation. Psychol Rev 50(4):370. https://doi.org/10.1037/h0054346
Ma WJ, Li L, Gu J, Zhu YG (2020) Spatial layout and accessibility of basic public service facilities in the 15-minute living circle of Shanghai. Planner 20:11–19
Nunn N, Qian N (2014) US food aid and civil conflict. Am Econ Rev 104(6):1630–1666. https://doi.org/10.1257/aer.104.6.1630
Perez C (1983) Technological revolutions and financial capital: the dynamics of bubbles and golden ages. Edward Elgar Publishing
Pigou A (1912) Wealth and welfare. Macmillan
Samuelson PA (1938) Welfare economics and international trade. Am Econ Rev 28(2):145–161
Sen A (1999) Development as freedom. In The globalization and development reader: perspectives on development and global change (p. 525)
Shi J, Dai X, Duan K, Li J (2023) Exploring the performances and determinants of public service provision in 35 major cities in China from the perspectives of efficiency and effectiveness. Socio-Econ Plan Sci 85:101441. https://doi.org/10.1016/j.seps.2022.101441
Sturgeon TJ (2021) Upgrading strategies for the digital economy. Glob Strategy J 11(1):34–57. https://doi.org/10.1002/gsj.1364
Tornatzky LG, Fleischer M (1990) The process of technological innovation. Lexington Books
Wang J, Wang B, Dong K, Dong X (2022) How does the digital economy improve high-quality energy development? The case of China. Technol Forecast Soc Change 184:121960. https://doi.org/10.1016/j.techfore.2022.121960. Article
Wang Y, Liu Y, Hu Y (2022) An empirical analysis based on a panel threshold model of the effect of Internet development on the efficiency of Chinese government public service supply. PLoS One 17(7):e0271390. https://doi.org/10.1371/journal.pone.0271390
Xiong X, He W, Li L (2025) Digital economy, basic public services, and regional income gap. J Soc Econ Dev, 1–29. https://doi.org/10.1007/s40847-025-00471-8
Yang L, Chen IP, Lee C, Ren M (2024) Panel threshold model with covariate-dependent thresholds and unobserved individual-specific threshold effects. Econ Rev 43(7):452–489. https://doi.org/10.1080/07474938.2024.2339147
Zhang W, Zhao S, Wan X, Yao Y (2021) Study on the effect of digital economy on high-quality economic development in China. PLoS One 16(9):e0257365. https://doi.org/10.1371/journal.pone.0257365
Zhan XY, Wang RR (2022) Fiscal pressure, expenditure structure, and public service quality: an empirical analysis based on panel data from 229 prefecture-level cities in China. Reform (2) 111–126
Zhao C, Wang B (2022) How does new‑type urbanization affect air pollution? Empirical evidence based on spatial spillover effect and spatial Durbin model. Environ Int 165:107304. https://doi.org/10.1016/j.envint.2022.107304. Article
Author information
Authors and Affiliations
Contributions
XH conceived the research idea, conducted the analysis, prepared all figures and tables, and wrote the manuscript. The author reviewed and approved the final version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The author declares no competing interests.
Ethical approval
This study did not involve human participants or identifiable personal data. Therefore, formal ethical approval was not required. All analyses were conducted using publicly available, de-identified data, in accordance with relevant institutional and national research guidelines.
Informed consent
This study did not involve human participants or identifiable personal data. As no data were collected directly from participants, obtaining informed consent was not applicable. All data used were publicly available and anonymized, and their use complied with relevant institutional and national regulations.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Huang, X. Spatiotemporal impacts and nonlinear moderating effects of the digital economy on county-level public services in western China. Humanit Soc Sci Commun 13, 480 (2026). https://doi.org/10.1057/s41599-026-06625-1
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1057/s41599-026-06625-1
