The impact and moderating mechanisms of digital inclusive finance on public services in Chinese counties

Abstract

This article selects data from 1445 counties across three regions of China (Eastern, Central, and Western) between 2008 and 2022. It constructs evaluation index systems for the quality of livelihood and infrastructure public services and uses the entropy weight method to measure and analyze the spatiotemporal evolution characteristics of these two types of public services. The article employs a Spatial Durbin Model with two-way fixed effects to explore the spatial spillover effects and moderating mechanisms of the digital inclusive finance on the quality of livelihood and infrastructure public services in Chinese counties, while also analyzing the regional heterogeneity of the digital inclusive finance’s impact across the Eastern, Central, and Western regions. The results show that the digital inclusive finance significantly improves the quality of both livelihood and infrastructure public services in Chinese counties, with a more pronounced positive effect on livelihood services. In the counties of Western and Eastern China, the digital inclusive finance exhibits a significant positive spatial spillover effect on livelihood public services. In contrast, for infrastructure public services, a significant positive spatial spillover effect is observed in Western counties, while the spillover effect is significantly negative in other regions. From the perspective of moderating effects, the development of the primary industry in counties enhances the digital inclusive finance’s positive impact on infrastructure services and strengthens its spillover effect on livelihood services in neighboring counties. An increase in local fiscal revenue boosts the digital inclusive finance’s positive impact on both livelihood and infrastructure services within the county but suppresses its spillover effect on livelihood services in neighboring counties. This article provides empirical evidence from a Chinese perspective on the regional impact of the digital inclusive finance on public services.

Introduction

Public services are critical for ensuring social welfare and equity while supporting national economic and social development¹. As providers of public services, governments at all levels are tasked with delivering high-quality services to meet the needs of residents. In recent years, China has achieved significant progress in public service provision, including steady increases in education funding and improved healthcare conditions². However, challenges remain, such as uneven service quality, excessive focus on infrastructure over livelihood services, and disparities in service provision due to regional economic and fiscal constraints^3,4. The "14th Five-Year Plan" emphasizes the need to enhance public service quality, innovate service delivery, and strengthen policy guarantees, highlighting the urgency of improving public services to advance the modernization of national governance.

Digital inclusive finance, as a key component of the digital economy, is progressively integrating into various levels of public services. This is especially evident at the county level, where digital inclusive finance shows immense potential. First, digital inclusive finance can reduce the cost and entry barriers associated with providing financial services during public service investment or financing processes through digital means⁵. This is particularly beneficial in remote areas such as counties and rural regions beyond urban centers. It facilitates greater access to financial services such as loans, insurance, and payments for both businesses and residents. The inclusive nature of digital finance significantly enhances financial accessibility, further promoting equitable and inclusive development of the economy, society, and public services⁶. Second, the application scenarios for digital inclusive finance are continuously expanding, covering various sectors such as industrial upgrading and urbanization in counties⁷. Digital currency, particularly the digital yuan, is one of the key tools in China’s promotion of digital inclusive finance. The widespread use of digital currency ensures that policy subsidies, social assistance, and other funds can be distributed accurately, in real time, and through clear channels, thereby improving the efficiency and transparency of fund usage. Finally, digital inclusive finance also contributes to the development of credit information sharing platforms, advancing the establishment of local credit systems⁸. This provides more financing opportunities for county residents and small businesses, fostering the sustainable development of the local economy. In turn, this indirectly stimulates county governments to invest more in public services.

With the rapid development of the digital inclusive finance, its profound impact on public services has become increasingly evident. However, existing research primarily focuses on how the digital inclusive finance enhances public service quality at the provincial and municipal levels or within specific regions, often overlooking its potential spatial spillover effects at the county level and across regions. In reality, improvements in a county’s public service quality and digital inclusive finance not only enhance local residents’ well-being but also influence neighboring counties through economic linkages and policy diffusion, ultimately reshaping the regional service landscape⁹. This cross-regional interdependence and spillover effect present new opportunities for optimizing public services and promoting coordinated regional development. The varying economic levels, industrial foundations, and fiscal conditions of different counties play a significant moderating role in the impact of digital inclusive finance on public services. This moderating mechanism not only reveals how local economic conditions, industrial structures, and fiscal capacities influence the effectiveness of public service upgrades during the promotion of digital inclusive finance but also emphasizes the need to account for regional differences when formulating targeted policies. For instance, counties with weak economic foundations, outdated industrial structures, and limited fiscal revenues may face challenges such as insufficient coverage of digital financial services and delayed infrastructure development¹⁰. These issues can prevent the full realization of digital finance benefits and may even exacerbate inequalities in public service access. Therefore, when advancing the application of digital inclusive finance, it is crucial to consider the specific economic, industrial, and fiscal conditions of each county. Tailored strategies should be implemented to ensure that the benefits of digital finance reach a broader population, particularly residents in counties and remote rural areas.

Building on this background, this article focuses on the spatial effects, moderating mechanisms, and regional heterogeneity of the digital inclusive finance’s role in enhancing county-level public service quality, addressing three core research questions. First, how does the digital inclusive finance influence overall regional public service quality through cross-county mechanisms, namely its spatial spillover effect? Second, to what extent do regional economic factors, such as primary industry development and fiscal revenue, moderate the impact of the digital inclusive finance on public service quality, thereby revealing the specific pathways of this moderating mechanism? Third, is there significant heterogeneity in the impact of the digital inclusive finance on public services across different regions—namely, the eastern, central, and western areas—and what regional characteristics and developmental trends underlie these differences?

To address these issues, this article sets the following research objectives: First, to systematically analyze the spatial effects of the digital inclusive finance on county-level public services and explore how cross-regional influence mechanisms facilitate either positive or negative spillover effects on neighboring counties. Second, to clarify the role of economic and fiscal factors in shaping the impact of the digital inclusive finance on public service quality, with a particular focus on how these effects vary across different industrial and fiscal contexts. Third, to conduct an empirical analysis of regional heterogeneity by examining county-level samples from the eastern, central, and western regions, thereby revealing how disparities in digital inclusive finance development influence public services. Finally, based on empirical findings, to propose practical policy recommendations that support regional coordination and public service optimization, offering both theoretical and practical pathways for the deep integration of the digital inclusive finance and high-quality public service development.

Literature review

Digital inclusive finance, as a crucial component of the digital economy, has garnered significant attention in academic circles in recent years. Existing research primarily explores its definition, measurement methods, and its impact across various sectors. Regarding the definition of digital inclusive finance, Tay et al.⁶ assert that it is a financial model that integrates digital technologies with inclusive finance, aiming to reduce barriers to financial services, enhance their accessibility, and promote economic inclusive growth.

In terms of measurement methods, one of the most commonly used tools is the "Digital Inclusive Finance Index," developed through a collaboration between Peking University’s Digital Finance Research Center and Ant Group’s Research Institute. This index encompasses dimensions such as the breadth of digital financial coverage, depth of use, and the degree of digitalization¹¹. It provides a quantitative tool for academic research and has been widely applied to analyze the effects of digital inclusive finance on various groups¹², businesses¹³, and regional economic activities¹⁴.

In terms of the impact of digital inclusive finance, it is generally regarded as having a positive effect on high-quality economic development. Wang et al.¹⁵ found that digital inclusive finance indirectly influences economic growth and innovation by increasing disposable income, enhancing fiscal expenditure levels, and promoting the development of education, with significant regional heterogeneity in its effects. Ahmad et al.¹⁶ examined the impact of digital inclusive finance and human capital on provincial economic growth in China, concluding that the development of human capital and the upgrading of digital inclusive finance provide services to previously financially excluded populations, thus driving economic growth.

Additionally, digital inclusive finance is also considered to impact economic resilience and income inequality. Du et al.¹⁷ found that digital inclusive finance enhances regional economic vitality and resilience by fostering innovation and promoting consumption development, with a more pronounced effect in China’s central and eastern regions and large, economically developed cities. Das and Chatterjee¹⁸ argued that digital inclusive finance plays a positive role in reducing rural–urban poverty through information and communication technology channels, but its impact on alleviating income inequality exhibits regional differences, particularly being somewhat diminished in urban areas. In conclusion, as a new financial service model, digital inclusive finance is profoundly influencing various aspects of the economy and society. Future research can further explore its mechanisms of action in different sectors and among different groups.

Research on public services focuses on their evaluation, influencing factors, and economic impacts. Scholars construct evaluation frameworks either from the supply perspective, considering government functions like education and healthcare, or from the demand perspective, based on resident satisfaction surveys. Methods such as the entropy method, Theil index, and kernel density estimation are commonly applied to assess service quality and analyze spatial trends. Influencing factors range from urbanization and fiscal capacity to population dynamics, with studies highlighting both positive and negative effects depending on the context¹⁹.

Existing studies on the relationship between the digital inclusive finance and public services have made significant progress, yet several key gaps remain. First, most research focuses on the impact of the digital inclusive finance at the provincial or municipal levels, emphasizing large-scale trends while overlooking county-level dynamics. However, counties possess distinct administrative, economic, and infrastructural characteristics, necessitating a more granular investigation. Second, while some studies acknowledge the role of the digital inclusive finance in regional development, they rarely explore its spatial spillover effects on neighboring counties. Given the interconnected nature of public services, improvements in one county may extend beyond administrative boundaries, reshaping the regional service landscape. Third, existing research often assumes a uniform impact of the digital inclusive finance on public services, yet variations in local economic structure, fiscal capacity, and industrial composition likely mediate these effects. Few studies systematically incorporate such factors into their analytical frameworks, limiting the understanding of regional heterogeneity.

This article addresses these gaps by systematically analyzing the spatial effects and moderating mechanisms of the digital inclusive finance on county-level public services. By shifting the focus to county-level dynamics, it provides a more detailed understanding of how digital inclusive finance development influences public service quality in localized contexts. Furthermore, it explicitly examines spatial spillover effects, revealing cross-county interactions that shape regional public service distribution. In addition, by incorporating economic and fiscal factors—such as primary industry development and fiscal revenue—this article uncovers key moderating mechanisms that condition the digital inclusive finance’s effectiveness in public service provision. Lastly, the findings offer targeted policy insights for optimizing regional public services and fostering balanced development in the digital era.

Materials and methods

Model

The digital inclusive finance possesses distinct advantages in temporal and spatial compression, while public services also exhibit spatial interconnections. Therefore, before conducting spatial econometric analysis, it is necessary to first examine whether there is spatial autocorrelation between the digital inclusive finance and the quality of public services. Firstly, this article constructs a spatial weight matrix W, representing the spatial weight relationship of an n × n dimension. Three different types of spatial weight matrices are employed to measure the relationships between geographical units, including the inverse distance matrix (W1), the inverse distance squared matrix (W2), and the 0–1 adjacency matrix (W3). These three matrices can serve as the basis for robustness tests. Specifically, the inverse distance matrix indicates the reciprocal relationship of geographical distances between counties; the inverse distance squared matrix represents the reciprocal relationship of squared geographical distances between counties; and the 0–1 adjacency matrix indicates whether counties are adjacent, with a value of 1 indicating adjacency and 0 otherwise.

Following that, a global $Moran^{\prime}sI$ will be employed to conduct a global spatial autocorrelation test on both digital inclusive finance and the quality of public services. The global $Moran^{\prime}sI$ model is as follows:

$$I = \frac{{\sum\limits_{i = 1}^{n} {\sum\limits_{j = 1}^{n} {W_{ij} \left( {x{}_{i} - \overline{x} } \right)\left( {x{}_{j} - \overline{x} } \right)} } }}{{S^{2} \sum\limits_{i = 1}^{n} {\sum\limits_{j = 1}^{n} {W_{ij} } } }}$$

(1)

In Eq. (1), $n$ represents the number of spatial units (counties); $x_{i}$ and $x_{j}$ respectively denote the values of spatial factor $x$ in spatial units $i$ and $j$; $\overline{x}$ represents the mean of factor $x$; $W_{ij}$ stands for the spatial weight matrix, and $S^{2} = \frac{{\sum\limits_{i = 1}^{n} {\left( {x - \overline{x} } \right)^{2} } }}{n}$ denotes the sample variance. The $Moran^{\prime}sI$ values range from $\left[ { - 1,0} \right)$, $0$, to $\left( {0,1} \right]$, representing spatial negative correlation, spatial non-correlation, and spatial positive correlation, respectively.

To eliminate spatial scale differences, the weight matrix $W_{ij}$ needs to be standardized. The purpose of standardizing the weight matrix is to ensure that the sum of each row of the weight matrix equals 1. If the spatial weight matrix is standardized, then $\sum\limits_{i = 1}^{n} {\sum\limits_{j = 1}^{n} {W_{ij} } } = n$, and $S^{2} = \frac{{\sum\limits_{i = 1}^{n} {\left( {x - \overline{x} } \right)^{2} } }}{n}$ denotes the sample variance. In this case, $Moran^{\prime}sI$ can be expressed as:

$$I = \frac{{\sum\limits_{i = 1}^{n} {\sum\limits_{j = 1}^{n} {W_{ij} \left( {x{}_{i} - \overline{x} } \right)\left( {x{}_{j} - \overline{x} } \right)} } }}{{\sum\limits_{i = 1}^{n} {\left( {x{}_{i} - \overline{x} } \right)^{2} } }}$$

(2)

Assuming that the values $x_{i}$ of all regions $i$ deviate from the mean $\overline{x}$ by the same distance, the sum of squares of the denominator, $\sum\limits_{i = 1}^{n} {\left( {x{}_{i} - \overline{x} } \right)^{2} }$, represents the total variability within the region. By standardizing the weight matrix, this article ensures scale consistency between different regions, avoiding measurement bias caused by differences in the number of regions.

This article chose the Spatial Durbin Model (SDM) over other spatial models based on the following considerations: First, the SDM can comprehensively account for the bidirectional spatial effects, simultaneously considering the spatial lag effects of both the dependent and independent variables²⁰. In contrast, other common spatial models (such as the spatial lag model and spatial error model) typically capture only the spatial effects of either the dependent or independent variable, neglecting the spatial dependence of the other. Therefore, the SDM can more comprehensively reflect the bidirectional interaction of spatial effects, making it more suitable for addressing issues that involve multiple variables and require the simultaneous consideration of spatial autocorrelation between both independent and dependent variables.

Second, the SDM, when considering spatial dependence, can not only reveal the spatial spillover effects between regions but also effectively identify and quantify the spatial heterogeneity within regions. This article focuses on the spatiotemporal effects of digital inclusive finance on public services in the western counties, and the SDM can better describe the interaction and transmission effects between different counties, thus more accurately capturing the relationship between digital inclusive finance and public services, as well as their spatial agglomeration characteristics.

Therefore, the SDM is highly applicable and advantageous in this article, providing a more detailed analysis of spatiotemporal effects, revealing the mechanisms through which digital inclusive finance impacts the balance of public services, and offering more reliable conclusions, especially when considering spatial spillover effects and interactions between counties.

Introduce spatial interaction terms for the dependent variable, independent variables, and control variables, constructing the spatial Durbin model as follows:

$$ser_{it} = \alpha + \rho Wser_{it} + \beta_{1} dig_{it} + \varphi_{1} Wdig_{it} + \beta_{2} con_{it} + \varphi_{2} Wcon_{it} + \mu_{i} + \lambda_{t} + \varepsilon_{it}$$

(3)

In Eq. (3), $i$ represents counties, $t$ represents years, $ser_{it}$ represents public service quality, $dig_{it}$ represents the level of digital inclusive finance, $con_{it}$ represents the control variable group, $\alpha$ represents the intercept term, $\rho$ represents the spatial autoregressive coefficient, $W$ represents the spatial weight matrix. This article introduces individual fixed effects and time fixed effects to control for unobserved heterogeneity between counties. The individual fixed effects primarily control for time-invariant factors at the county level, such as local governance and economic structure, which may vary across counties but remain relatively constant over time. This helps to eliminate the influence of unique county-specific factors on spatial autocorrelation. On the other hand, time fixed effects are used to control for common shocks or trends that may affect all counties over time, such as national policies, local strategies, or global economic changes. By incorporating both fixed effects, this article is able to more accurately estimate the relationship between digital inclusive finance and public services, ensuring that the results are not influenced by county characteristics or time trends, $\mu_{i}$ represents individual effects, $\lambda_{t}$ represents time effects, and $\varepsilon_{it}$ represents independently and identically distributed random disturbances.

Building upon this, to examine the heterogeneous spatial effects of the digital inclusive finance on public service quality under the moderating effect of external environmental variables, the following moderated panel model is constructed:

$$\begin{gathered} ser_{it} = \alpha + \rho Wser_{it} + \beta_{1} dig_{it} + \varphi_{1} Wdig_{it} + \beta_{2} mo_{it} + \varphi_{2} Wmo_{it} \hfill \\ + \beta_{3} dig_{it} \times mo_{it} + \varphi_{3} Wdig_{it} \times mo_{it} \hfill \\ + \beta_{4} con_{it} + \varphi_{4} Wcon_{it} + \mu_{i} + \lambda_{t} + \varepsilon_{it} \hfill \\ \end{gathered}$$

(4)

In Eq. (4), $mo_{it}$ represents the moderating variable in the external environment, $dig_{it} \times mo_{it}$ represents the interaction term between the digital inclusive finance and the moderating variable, and other variables are explained as in Eq. (3). Equation (4) focuses on observing the magnitudes of the coefficients of the core explanatory variable $\beta_{1}$ and the interaction term $\beta_{3}$. When $\beta_{1}$ > 0 and $\beta_{3}$ > 0 or $\beta_{1}$ < 0 and $\beta_{3}$ < 0, it indicates that the moderating variable strengthens the impact of the digital inclusive finance on public services. When $\beta_{1}$ > 0 and $\beta_{3}$ < 0 or $\beta_{1}$ < 0 and $\beta_{3}$ > 0, it indicates that the moderating variable weakens the impact of the digital inclusive finance on public services.

Variable selection

Dependent variables

The high-quality development of public services is a powerful support for achieving a high-quality life for residents, including continuously improving education quality, steadily improving medical standards, making the ecological environment more beautiful, and improving road traffic facilities. Drawing on the ideas of Li²¹, this article’s evaluation index system of public service quality involves education, healthcare, ecology, and transportation aspects. Among them: compulsory education, medical care, and ecological protection cover residents’ basic living, daily social interaction, and physical and mental health needs, closely related to people’s livelihoods, serving as indicators for measuring the quality of public services related to people’s livelihoods. Road traffic facilities cover the basic infrastructure needs of element circulation and serve as the driving force for economic and social development, serving as indicators for measuring the quality of infrastructure-related public services. Based on the above index system, this article calculates the Comprehensive Quality Index of Livelihood Public Services ($Livser$, People’s Livelihood Public Service Quality Index) and the Comprehensive Quality Index of Infrastructure Public Services ($Infser$, Infrastructure Public Service Quality Index), respectively, as the dependent variables included in the spatial Durbin model for parameter estimation. The indicator systems for the composite quality index of livelihood public services and infrastructure public services are presented in Tables 1 and 2, respectively.

Table 1 Evaluation index system of livelihood public service quality.

Table 2 Evaluation index system for infrastructure public service quality.

Explanatory variables

This article uses the China Digital Inclusive Finance Index published by the Peking University Digital Finance Research Center to represent the level of digital inclusive finance ($dig$) at the county level. This index consists of three sub-indices: the breadth of digital inclusive finance coverage, the depth of digital inclusive finance usage, and the degree of financial inclusiveness digitization²².

Digital inclusive finance specifically addresses one of the most pressing challenges: financial inclusion. As counties, particularly in less developed regions, often face significant barriers in accessing traditional financial services, digital inclusive finance plays a pivotal role in promoting economic participation and addressing key development gaps.

Research at the county level often emphasizes the role of digital inclusive finance in enhancing public service delivery, as it provides the necessary financial tools to improve access to essential services, including education, healthcare, and infrastructure. In this context, digital inclusive finance is not just an economic indicator but a means of addressing the most pressing issues of financial exclusion, which remain a major challenge in many counties.

Moderator variables

In this article, two moderator variables are incorporated to examine how local economic characteristics influence the impact of the digital inclusive finance on public service delivery. The first moderator variable is the output value of the primary industry (Pri, measured in yuan per person), operationalized as the per capita output value of agriculture, forestry, animal husbandry, and fisheries. This indicator captures the productivity and economic strength of the county’s traditional sectors, which may interact with digital inclusive finance advancements in unique ways. The second moderator variable is the local fiscal revenue, defined as the per capita general budgetary revenue of local finance (Rev, measured in yuan per person), this variable reflects the financial capacity of local governments to invest in and sustain public services. By including these moderators, the analysis aims to provide a more comprehensive understanding of how the economic base and fiscal resources condition the relationship between the digital inclusive finance and public service outcomes.

Control variables

Considering that there may still be other external variables that could affect public service quality, in order to minimize potential endogeneity issues caused by omitted variables, this article refers to the research of Zhang et al.²³, as well as Su and Li²⁴, and selects control variables as shown in Table 3.

Table 3 Control variables.

Data source and processing

Considering data availability, samples with severe data deficiencies during the study period were excluded. Ultimately, this article utilizes panel data from 1,445 counties in China spanning 2008 to 2022 for empirical analysis. Furthermore, China is divided into three regions—eastern, central, and western—based on geographic location and economic development level. The sample includes 364 counties from the eastern region, 392 from the central region, and 689 from the western region.

The data for the dependent variable, public service quality, as well as all control variables, primarily come from the Annual Statistical Bulletin on National Economic and Social Development of counties in China, the China County Statistical Yearbook, and the China County Urban Construction Statistical Yearbook. The core explanatory variable, digital inclusive finance, is mainly sourced from the Digital Inclusive Finance Index (PKU-DFIIC) published by Peking University. To address the issue of missing data in the Digital Inclusive Finance Index for certain years, this article employs multiple methods to complete the dataset to ensure the integrity of the study. First, the PKU-DFIIC index has been published since 2011, initially covering data at the provincial and prefecture-level city scales, while county-level data became available starting in 2014. Therefore, there is no direct county-level data for the period from 2011 to 2013. To address this gap, this article applies a spatial proxy matching method, assigning the index values of the superior prefecture-level cities to the counties under their jurisdiction as approximate county-level values, which can to some extent reflect the development level of digital inclusive finance at the county level. Second, for the period from 2008 to 2010, due to the absence of direct digital inclusive finance index data, this article uses interpolation based on trends to extrapolate the missing data.

To eliminate the impact of inflation on price factors, this article adopts China’s GDP Deflator Index based on the year 2008 as the base year to adjust economic indicators. The indicators adjusted for inflation include "Carbon Emissions per 10,000 Yuan of GDP" from the livelihood-oriented public services indicator system, as well as "the output value of the primary industry", "the per capita general budgetary revenue of local finance" from the moderator variables, and "Per Capita GDP", "Per Capita Retail Sales of Consumer Goods", and "Per Capita Financial Loan Balance" from the control variables. These indicators are all influenced by inflation, and adjusting them ensures that changes in the variables are not simply due to inflationary effects but reflect the actual underlying trends in the respective dimensions. After adjusting for inflation, the indicators are standardized, and the weights of the indicators are calculated using the entropy weighting method to derive the comprehensive quality index of public services in counties in both livelihood and infrastructure dimensions.

Adjusting these economic indicators for inflation is necessary because inflation can distort the real value of economic variables over time, leading to misleading conclusions if not accounted for. For instance, "Carbon Emissions per 10,000 Yuan of GDP" is linked to economic output and prices; without adjusting for inflation, it could inaccurately reflect trends in emissions relative to economic growth. Similarly, adjusting "the output value of the primary industry" and "the per capita general budgetary revenue of local finance" ensures that these variables reflect real changes in the economy, not just inflationary effects. Variables like "Per Capita GDP", "Per Capita Retail Sales of Consumer Goods", and "Per Capita Financial Loan Balance" are also sensitive to inflation, and adjusting them allows for a more accurate assessment of economic development and the quality of public services. These adjustments are crucial for maintaining the reliability and validity of the analysis, ensuring that the results reflect the true dynamics of public service quality and economic development, rather than the mere impact of inflation. The descriptive statistics of the variables are presented in Table 4.

Table 4 Descriptive statistics of standardized variables.

Results

Spatiotemporal evolution characteristics of public services

This article employs the natural breaks classification method in ArcGIS to spatially classify the livelihood-oriented public services and infrastructure public service quality for 2008 and 2022, generating spatial classification maps (Figs. 1, 2) to illustrate geographic disparities in public service quality. Additionally, Local Moran’s I cluster maps (Figs. 3, 4) are presented to reveal the spatial clustering of hot and cold spots in public services.

Fig. 1

Geographical Distribution of Livelihood Public Service. This map is based on the standard map No. GS(2020)4619 downloaded from the standard map service website of China National Bureau of Surveying, Mapping and Geoinformation (NSGMGI). The same below.

Fig. 2

Geographical distribution of infrastructure public service.

Fig. 3

Spatial clustering characteristics of livelihood public service.

Fig. 4

Spatial clustering characteristics of infrastructure public service.

In terms of time span handling, the data for all years were standardized uniformly to ensure comparability. To maintain consistency in weights across years, the entropy method did not calculate indices separately for each year. Instead, the data for all years were processed in a balanced manner to calculate a unified objective composite weight. This approach effectively eliminates fluctuations in weights between years, ensuring that the spatial–temporal evolution of public service quality is accurately reflected.

To further reveal the trend of public service quality changes, this article specifically selects the weights for the starting year of the study period (2008) and the ending year (2022). This choice facilitates the visual representation of the evolution of public service quality over the study period and reflects the impact of factors such as the digital inclusive finance on public services. By comparing the weights for these two years, the spatial–temporal changes in public service quality can be clearly observed, providing a basis for subsequent spatial clustering analysis and the investigation of mismatch phenomena.

Figure 1 reveals that, by 2022, the average quality of livelihood-oriented public services in China’s county-level regions has steadily improved compared to 2008. This improvement is particularly evident in western regions, notably in Xinjiang, Shaanxi, and Gansu provinces, largely due to sustained financial subsidies and policy support from the Chinese government in recent years. In contrast, the central regions have maintained a moderate and relatively stable development in county-level livelihood-oriented public services over the 15-year period.

However, some counties in the eastern region, particularly along the southeastern coast, exhibit lower public service quality. This may be attributed to the region’s economic prosperity, which attracts a large influx of population, coupled with limited land area compared to the western regions. As a result, the per capita value of public services in eastern counties remains relatively low, highlighting a persistent issue even in developed areas: despite a high total volume of public services, the per capita distribution is inadequate, which warrants further attention.

In contrast to the high-quality development of livelihood-oriented public services, the progress of infrastructure public services in China’s county-level regions has been less satisfactory. Figure 2 shows that, with the exception of a few provinces along the eastern coast and the Inner Mongolia Autonomous Region, most regions exhibit low-quality infrastructure public services, with little improvement from 2008 to 2022. This underscores a critical constraint on the development of county-level regions in China—insufficient infrastructure public services.

The inadequacy of basic and transportation infrastructure in counties, compared to urban areas, hampers the efficient flow of production factors. To address this issue, greater emphasis must be placed on the development of infrastructure public services. Governments at all levels in China should consider allocating more resources to enhance county-level infrastructure, thereby promoting regional development.

From the perspective of spatial clustering characteristics, both livelihood-oriented (Fig. 3) and infrastructure public services (Fig. 4) in China’s county-level regions exhibit significant clustering patterns. For livelihood-oriented public services, counties in the eastern region predominantly show "low-low" clustering, which aligns with the earlier discussed issue of low per capita public service values due to limited land and high population density in the east.

In 2008, most counties in the central and western regions displayed "high-high" clustering, with a few showing "low–high" clustering. This suggests that these regions, benefiting from national policy support and the advantages of low population density, had relatively higher per capita values of livelihood-oriented public services than eastern counties. However, by 2022, counties in the northwest, particularly in Shaanxi Province, began shifting toward "low-low" and "high-low" clustering. This indicates a decline in per capita values of livelihood-oriented public services, potentially driven by increasing population density.

For infrastructure public services, counties in the southwestern region exhibit a pronounced "low-low" clustering pattern, with many counties falling into this category. This highlights the particularly weak quality of infrastructure public services in southwestern counties. In contrast, counties in the eastern region, especially along the eastern coast, benefit from robust economic development, resulting in higher-quality infrastructure public services. Most counties in the east display a "high-high" clustering pattern, with fewer showing "low–high" clustering characteristics.

Spatial effects of digital inclusive finance on public service quality

Spatial autocorrelation

The previous section provided a preliminary exposition on the theoretical spatial autocorrelation of public services among counties. In this article, the spatial clustering of the dependent variables was assessed by calculating the global $Moran^{\prime}sI$ (Table 5). The results reveal that from 2008 to 2022, both livelihood public services and infrastructure public services exhibited significantly positive spatial autocorrelation effects under three different spatial weight matrices. This indicates that there is a notable spatial positive correlation effect among counties for both types of public services, and they demonstrate distinct dynamic changes over time. Consequently, employing ordinary panel regression models may lead to estimation errors, thus necessitating further analysis using spatial econometric models.

Table 5 Global $Moran^{\prime}sI$.

Spatial effects

Based on the model selection criteria proposed by Anselin²⁵, the results presented in Table 6 indicate that the LM test is significant at the 1% level for livelihood public services and at the 10% level for infrastructure services. This suggests that both the Spatial Error Model (SEM) and Spatial Lag Model (SAR) are appropriate, with the Spatial Durbin Model (SDM) being preferred for its incorporation of both effects. The Hausman test shows significance at the 1% level for livelihood public services and at the 10% level for infrastructure services, leading to the selection of fixed effects. Additionally, the LR test indicates significance at the 1% level for livelihood services and at the 5% level for infrastructure services, supporting the use of the SDM without degradation. Fixed effects testing further reveals significance for livelihood public services at the 1% level and for infrastructure services at the 10% level, resulting in the choice of a two-way fixed effects model incorporating both time and individual effects. Therefore, this article adopts the Spatial Durbin Model with two-way fixed effects to explore the spatial impact of digital inclusive finance on livelihood and infrastructure services.

Table 6 Selection results of spatial econometric model types.

The results from the spatial Durbin model estimation, shown in Table 7, provide insights into the effects of the digital inclusive finance on public services. The analysis reveals that, both before and after controlling for other variables, the digital inclusive finance positively impacts both livelihood and infrastructure public services, with a more pronounced effect on livelihood services. This suggests that the digital inclusive finance enhances the quality of these services at the county level. However, an interesting spatial spillover pattern emerges. While the digital inclusive finance positively impacts local infrastructure services, it has a significantly negative spillover effect on the infrastructure services of neighboring counties. In contrast, no significant spatial spillover is observed for livelihood services in neighboring areas. This implies that, while the digital inclusive finance improves public services within counties, its benefits do not extend to adjacent counties, especially in terms of infrastructure development.

Table 7 Results of spatial Durbin model regression.

To further understand the underlying mechanisms, the article decomposes the spatial Durbin model to examine the direct, indirect, and total effects of the digital inclusive finance on infrastructure services. The direct effect, representing the impact within the county, is positive and significant, suggesting that the digital inclusive finance fosters improvements in local infrastructure. However, the indirect effect, which captures the spillover on neighboring counties, is significantly negative. This indicates that the benefits of the digital inclusive finance do not extend beyond county borders, resulting in a negative total effect on neighboring counties’ infrastructure services. This phenomenon can be attributed to the concentration of financial resources such as capital, talent, and technology in more developed counties, which may lead to a “siphon effect” that draws financial resources away from surrounding areas. This, in turn, inhibits the development of digital inclusive finance infrastructure and the enhancement of service capabilities in these regions. In contrast, the application of digital inclusive finance in public welfare sectors does not exhibit a significant negative spatial spillover effect. Services such as telemedicine payments, online education loans, and mobile payments, which leverage digital finance platforms, can extend beyond administrative boundaries to broader regions. These services do not rely on traditional physical outlets or heavy asset investments, enabling residents in neighboring areas to conveniently access financial services and social security support. As a result, this promotes the sharing and balanced distribution of financial resources, contributing to a more widespread and equitable allocation of digital inclusive finance benefits.

The findings underscore the complex, regionally differentiated effects of the digital inclusive finance on public services. While it significantly enhances both livelihood and infrastructure services within counties, it also generates spatial imbalances, particularly in infrastructure development. The concentration of resources in counties with more advanced digital inclusive finance leads to negative spillovers in neighboring areas, further highlighting the need for coordinated regional policies. Conversely, the inclusive nature of livelihood services reduces regional disparities by allowing neighboring counties to share in the benefits without competing for local resources. Given these uneven spatial effects, it is crucial to adopt a balanced approach to regional development. Promoting livelihood services through digital inclusive finance offers a model for more equitable resource distribution, but the negative spillover effects on infrastructure highlight the necessity for policies that address regional inequalities. Future strategies for the digital inclusive finance should prioritize the equitable distribution of benefits, focusing on coordinated efforts to meet the infrastructure needs of neighboring counties. By doing so, it is possible to achieve a more balanced, inclusive, and sustainable improvement in public service quality across regions.

Moderation effects

The relationship between the digital inclusive finance and public services is far from static. It is shaped and influenced by various external factors, particularly the output value of the primary industry and local fiscal revenue. These factors do not only have direct consequences within the county but also extend their influence through spatial spillover effects, impacting surrounding regions. The primary industry, which remains a cornerstone of economic activity in many counties, plays a significant role in the interplay between the digital inclusive finance and public services. At the same time, local fiscal revenue plays a crucial role in determining the total amount of financial resources available to support the development of digital inclusive finance. The way these external factors interact with the digital inclusive finance has profound implications for the realization of its potential benefits on public services. The regression results for the moderation effects of digital inclusive finance on public services are presented in Table 8.

Table 8 Spatial Durbin model regression results of moderation effects.

In many counties, the primary industry, which includes sectors such as agriculture, forestry, animal husbandry, and fishery, still serves as the backbone of the local economy. The output value of this industry remains a key determinant of the county’s overall economic development. A strong primary industry boosts the economic capacity of the county, providing a solid foundation for the growth of both the digital inclusive finance and public services. However, the effect of this industry on the digital inclusive finance’s impact on public services is complex and varies across different service sectors. On the one hand, an increase in the output value of the primary industry can dampen the positive impact of the digital inclusive finance on the quality of livelihood-oriented public services, such as education, healthcare, and social welfare. This may occur because local governments, in an effort to support the growth of the primary industry, often allocate a disproportionate share of fiscal resources to enhance the competitiveness of these industries. As a result, investments in social welfare services may be deprioritized. On the other hand, the expansion of the primary industry can have a positive effect on infrastructure-oriented public services, such as transportation, utilities, and public buildings. The rationale behind this is that as the primary industry grows, local governments are more likely to invest in infrastructure to support this growth, thus improving the overall quality of public services in these areas.

Further investigation into the spatial spillover effects reveals that the primary industry’s growth has significant implications not only within the county but also for neighboring regions. The development of the primary industry in a county can strengthen the positive impact of the digital inclusive finance on the quality of livelihood-oriented public services in surrounding counties. This finding highlights the potential for regional collaboration and the transfer of benefits between counties. As the primary industry in a given county experiences rapid growth and sustained economic activity, it facilitates the expansion and deeper penetration of digital inclusive financial services. This growth momentum may spill over into surrounding areas, thereby enhancing the breadth and depth of inclusive financial coverage and providing stronger support for the development of public welfare-related sectors. This demonstrates the interdependent nature of regional economies and public service provision, where the success of one county can have positive externalities for others in the same economic network.

Local fiscal revenue is another critical factor in the relationship between the digital inclusive finance and public services. The amount of fiscal revenue a local government generates directly affects its ability to invest in public services and promote the growth of the digital inclusive finance. With the increase in local fiscal revenue, governments are able to allocate more resources to the development of digital inclusive finance. This, in turn, improves the accessibility and convenience of financial services, enabling a broader range of low-income groups and small and medium-sized enterprises to access safe and efficient financial services. Ultimately, this contributes to promoting inclusive economic growth. This financial flexibility fosters a positive interaction between the two, especially in infrastructure-oriented public services. The increase in fiscal revenue provides county governments with the necessary resources to enhance digital infrastructure, laying a solid foundation for the development of digital inclusive finance. By improving digital payment systems and financial service platforms, governments can effectively reduce the cost and entry barriers of financial services, enabling more residents and businesses in counties and rural areas to access convenient and efficient financial services.

However, the spatial spillover analysis also reveals a more nuanced effect of local fiscal revenue. While an increase in local fiscal revenue enhances the digital inclusive finance’s positive impact on public services within the county, it may have a limiting effect on surrounding regions. Specifically, as fiscal revenue rises, local governments may shift their focus toward consolidating their own interests, potentially at the expense of neighboring counties. The increased fiscal capacity may prompt the local government to prioritize the welfare of its own residents, reducing the support provided to neighboring regions. This phenomenon weakens the positive impact of the digital inclusive finance on the quality of livelihood-oriented public services in surrounding counties. The results suggest that local governments, as their fiscal capacity grows, may prioritize their own development over regional collaboration, leading to a reduction in the spillover effects of the digital inclusive finance.

The output value of the primary industry and local fiscal revenue play integral roles in enhancing the quality of public services through the digital inclusive finance. These external factors significantly influence how the digital inclusive finance affects public service provision. The fiscal authority and policy orientation of local governments determine the depth and breadth of this impact, with the potential to either enhance or limit the benefits of digital inclusive finance growth on public services. The support of the primary industry may reduce investment in livelihood-oriented services, but it often promotes infrastructure improvements, which are crucial for economic development. Similarly, while increased fiscal revenue facilitates the development of both public services and the digital inclusive finance, it may also limit the positive impact of the digital inclusive finance on neighboring counties as local governments prioritize their own interests. Therefore, the decisions made by local governments regarding fiscal policies, industrial support, and investments in the digital inclusive finance have profound implications for the success of the digital inclusive finance’s effects on public services, both within the county and across surrounding regions.

Regional heterogeneity

The impact of the digital inclusive finance on public services in China is far from uniform and exhibits significant regional heterogeneity. This heterogeneity arises from a variety of factors, including geographic location, economic development levels, resource endowments, government policies, and the specific challenges faced by different regions. Given China’s vast territory and significant regional disparities, administrative divisions have long been the primary basis for policy formulation and economic planning. Therefore, this article adopts the official eastern, central, and western regional classification, which aligns with China’s national development strategies and reflects long-standing economic and policy differences. Compared to grouping by per capita GDP or other economic indicators, this classification better captures the institutional and structural characteristics shaping regional public service provision, making it more representative of China’s national conditions. The analysis, based on a spatial Durbin model with bidirectional fixed effects, further elucidates how the digital inclusive finance influences public services differently across these regions (Table 9), providing insights aligned with China’s existing regional development framework.

Table 9 Regional Heterogeneity in the Impact of Digital Inclusive Finance on Public Service Quality.

In the western region, the digital inclusive finance has an especially positive impact on county-level livelihood public services. This area is characterized by a predominantly rural population and relatively underdeveloped digital infrastructure, making it more dependent on the penetration of digital inclusive finance to enhance access to public services. In regions with limited infrastructure, digital inclusive finance helps compensate for the shortcomings of the traditional financial system by offering services such as digital payments, online loans, and insurance. These services significantly lower the barriers to financial access, enabling county-level residents and small businesses to obtain more convenient financing support. By providing accessible and low-cost financial services, digital inclusive finance plays a crucial role in narrowing the urban–rural economic divide and promoting balanced regional economic development. This impact is particularly significant in areas where traditional public services are lacking, as it directly contributes to improving the quality of county-level public services in western regions.

Compared to the western region, the central region—characterized by economic transformation and uneven urban–rural development—has experienced a significantly negative direct impact of the digital inclusive finance on county-level livelihood public services. In this area, the benefits of the digital inclusive finance are predominantly concentrated in urban centers, thereby exacerbating the existing gap between urban and rural communities. The unequal distribution of resources, combined with the economic challenges inherent in the transitional process, has led to a situation where the digital inclusive finance tends to intensify inequality rather than alleviate it. Although urban areas benefit from digital inclusive finance advancements, rural areas remain marginalized, resulting in an overall diminished impact on public services.

At the spatial spillover level, the digital inclusive finance has generated significant positive spillover effects on county-level livelihood public services in both the western and eastern regions. This indicates that government policies and resource allocation are key factors driving regional disparities in the impact of the digital inclusive finance. In both the eastern and western regions, governments have prioritized the development of digital inclusive finance as a key strategic objective. By advancing digital infrastructure and promoting the application of financial technologies, they have amplified its positive impact on public service delivery. A range of supportive policies has been introduced across regions, including increased investment in digital financial platforms and encouragement for financial institutions to innovate in inclusive finance, all aimed at expanding the reach and efficiency of financial services. In the western region, efforts to improve network infrastructure and promote services such as mobile payments and online lending have effectively addressed the shortage of financial services in counties, rural areas, and other remote locations. In contrast, the eastern region has leveraged its higher level of economic development to boost investment in digital financial technologies, thereby further enhancing the quality and accessibility of financial services. As a result, digital inclusive finance has played a critical role in narrowing regional development gaps and strengthening public service capabilities.

In contrast, the spatial spillover effect of the digital inclusive finance on neighboring county-level livelihood public services in the central region is not significant. This is primarily due to the limited resources and policy support allocated to the digital inclusive finance in this area, which constrains its ability to fully leverage the potential of digital inclusive finance to expand public service coverage and improve service efficiency. The central government’s priorities, coupled with the region’s economic challenges, have resulted in a less favorable environment for digital inclusive finance growth. With reduced investments in digital inclusive finance and limited policy incentives, the central region struggles to achieve the positive outcomes seen in the western and eastern regions, further emphasizing the crucial role of government support in shaping the impact of the digital inclusive finance.

Regarding infrastructure public services, the digital inclusive finance exerts a significantly positive effect on county-level infrastructure services in the central and western regions, whereas its impact in the eastern region is negative. This discrepancy can be attributed to the fact that the eastern region already boasts well-developed infrastructure, resulting in limited marginal gains from further digital inclusive finance activities; most infrastructure needs have already been met, and the potential for digital innovations to drive further improvements is comparatively constrained. Moreover, intense market competition in the eastern region may diminish the incentive for additional investments in infrastructure. In contrast, the central and western regions, where infrastructure development remains relatively underdeveloped, stand to gain more from the digital inclusive finance. These areas have a greater demand for enhanced infrastructure, and the digital inclusive finance provides novel avenues for investment and growth in public services. The expansion of digital inclusive finance platforms helps identify gaps in financial service coverage, attracts financial resources to underserved counties, and accelerates the rollout of related financial products and services.

The spatial spillover effect of the digital inclusive finance on county-level infrastructure public services exhibits notable regional disparities across the three major regions. In the eastern and central regions, the digital inclusive finance generates a significant negative spillover effect on neighboring counties’ infrastructure services, echoing the negative spatial effect observed in the baseline regression when regions are not differentiated. In contrast, the western region presents an exception: here, the digital inclusive finance exerts a positive spillover effect on surrounding county-level infrastructure public services. This divergence can be attributed not only to the more robust policy support provided by the central government under the Belt and Road Initiative but also to the unique economic and geographical characteristics of the western region. With relatively lagging digital inclusive finance, increased investments in the west yield more pronounced marginal benefits. Moreover, the region has actively promoted the development of digital inclusive finance platforms and regional financial cooperation networks, fostering coordinated progress among counties in terms of financial service coverage and digital financial infrastructure. This has further strengthened the positive spatial spillover effects of digital inclusive finance, driving neighboring counties to collectively enhance financial accessibility and the quality of public services.

The impact of the digital inclusive finance on public services in China is shaped by a complex interplay of regional factors. Economic conditions, infrastructure demands, government policies, and competitive dynamics all influence how digital inclusive finance advancements affect the quality of public services. The results of this article underscore the importance of tailoring policies to regional characteristics to effectively harness the digital inclusive finance in addressing the unique challenges each area faces. By recognizing and responding to the distinct needs of the eastern, central, and western regions, policymakers can help maximize the potential of the digital inclusive finance to improve public services and foster more balanced national development.

To further validate the heterogeneity results, this article adopts an alternative regional classification method. Using the national average per capita GDP of 99,651 yuan in 2022 as a benchmark, and grouping based on ± 1 standard deviation (31,845 yuan), the approach distinguishes economic development differences across regions with varying income levels. High-income regions are defined as those with per capita GDP exceeding 131,495 yuan, medium-income regions as those with per capita GDP between 67,806 and 131,495 yuan, and low-income regions as those with per capita GDP below 67,806 yuan. This classification method, grounded in statistical principles, aids in examining whether the impact of the digital inclusive finance on public service quality varies with economic development levels.

This article adopts both the “eastern-central-western regional division” and the “per capita GDP grouping” for regional heterogeneity analysis, primarily based on the strong correlation between the two under China’s specific development context. From the perspective of administrative divisions, the classification of eastern, central, and western regions has long served as a fundamental framework for China’s regional development policies and thus carries significant policy implications. However, this administrative division also largely reflects underlying disparities in economic development levels across regions. According to successive editions of the Statistical Bulletin on National Economic and Social Development published by the National Bureau of Statistics of China, the eastern region consistently exhibits the highest levels of per capita GDP, followed by the central region, with the western region showing comparatively lower levels. This stratified distribution of economic development aligns closely with the geographical delineation of the eastern-central-western framework, thereby providing a practical and empirical rationale for introducing per capita GDP as an alternative grouping method in this article.

In contrast, although grouping based solely on administrative divisions carries policy relevance, it has limitations in capturing the actual differences in regional economic development. Therefore, this article introduces per capita GDP as a complementary classification criterion alongside the traditional East-Central-West division. This dual approach allows for a more comprehensive identification of spatial heterogeneity in the effects of digital inclusive finance, particularly in uncovering potential variations in its mechanisms across regions with different levels of economic development. As a quantifiable and stratifiable economic indicator, per capita GDP can transcend the constraints imposed by administrative boundaries, offering a more accurate depiction of regional economic structures and development stages. Accordingly, the adoption of this dual-grouping strategy is both theoretically grounded and practically meaningful, enhancing the explanatory power of the heterogeneity analysis presented in this article.

Table 10 reveals that the regression results based on per capita GDP grouping align closely with those based on the traditional eastern, central, and western regional divisions. Specifically, the eastern region, characterized by higher economic and income levels, mirrors the outcomes observed in high-income areas. Conversely, the western region, with its comparatively lower economic and income levels, exhibits results akin to those in low-income regions. The central region, with income levels between those of the eastern and western regions, shows regression outcomes similar to medium-income areas. This correspondence reflects China’s current economic landscape. While there are minor differences in the significance levels of certain regression outcomes, the overall trends and fundamental conclusions remain consistent. This consistency suggests that grouping based on per capita GDP closely aligns with the economic development disparities inherent in the eastern, central, and western regional classifications, thereby enhancing the credibility of the heterogeneity test results derived from these regional groupings.

Table 10 Heterogeneity test results by per capita GDP level.

Robustness test

This article employs three different spatial weight matrices: the inverse distance matrix (W1), the inverse distance squared matrix (W2), and the 0–1 adjacency matrix (W3), successively entering them into the Spatial Durbin Model to measure the spatial effects of the digital inclusive finance on public services under different definitions of geographic relationships. As shown in Table 11, the direction and significance of the direct effects and spatial spillover effects of the digital inclusive finance on livelihood public services remain unchanged across different spatial weight matrix measurement methods. Similarly, there are no changes in the direction and significance of the direct effects of the digital inclusive finance on infrastructure public services, with only slight variations in the spatial spillover effects, but the differences are not significant, indicating robust results.

Table 11 Robustness test results.

Endogeneity test

The authors first conducted a Hausman test and a heteroscedasticity-robust Durbin-Wu-Hausman (DWH) test on the two-stage least squares regression (2SLS). The results indicated that digital inclusive finance ($dig$) could be considered an endogenous variable. To address potential reverse causality, this article follows the approach of Liang and Liu²⁶ by using county-level internet penetration as an instrumental variable for digital inclusive finance ($iv$) . Internet penetration refers to the scope of infrastructure deployment and resident usage within a county, which the authors measure by the internet access rate—the proportion of the county’s permanent population with broadband connectivity.

This article adopts county-level internet penetration as an instrument for digital inclusive finance for two reasons. First, out of consideration for instrument relevance, the digital inclusive finance index is strongly correlated with internet development—this is because the primary distinction between the digital inclusive finance index and traditional finance lies in the former’s focus on inclusive finance development from an internet-finance perspective; the growth of digital inclusive finance is inseparable from the internet as its practical carrier, making internet penetration critically important.

Second, out of consideration for instrument exogeneity, county-level internet penetration as an instrument for digital inclusive finance is exogenous in that it is unrelated to the disturbance term of public service quality. County-level internet penetration is driven mainly by external factors such as policy initiatives and internet infrastructure investment, which do not directly affect changes in public service quality; even if public service quality fluctuates, internet penetration will not change significantly as a result. Therefore, the effect of county-level internet penetration on public service quality operates mainly through the channel of enhancing public services via digital inclusive finance, avoiding any direct or indirect association with the dependent variable and satisfying the exogeneity requirement.

Table 12 reports the weak-instrument test results. The Cragg-Donald Wald F-statistic far exceeds the Stock-Yogo 10% critical value of 16.38, indicating that weak-instrument bias is not a concern. Moreover, in the identification test, the Kleibergen-Paap rk LM statistic rejects the null hypothesis at the 10% level, confirming the instruments’ identifiability and supporting their validity and relevance.

Table 12 The estimation results using the Heckman two-step method.

In this article, the authors incorporate the instrument into the regression model and estimate it via two-stage least squares (2SLS). In both the analysis of digital inclusive finance’s impact on social-welfare public services and on infrastructure public services, the first-stage coefficients on the instrument are significant at the 10% level, indicating a strong positive correlation between the instrument and digital inclusive finance. In the second stage, digital inclusive finance significantly improves social-welfare public service quality at the 5% level and infrastructure public service quality at the 10% level, with effect sizes and directions closely matching the baseline regressions. These results demonstrate that, after addressing endogeneity, the baseline findings remain robust, confirming that endogeneity does not undermine the main conclusions of this article.

Discussion

This article develops an evaluation index system to assess the quality of livelihood and infrastructure public services, using panel data from counties across China’s three regions (2008–2022). The entropy weight method is employed to measure the overall quality of these services. Moreover, the article empirically examines the spatial effects of the digital inclusive finance on the quality of both types of public services and their moderating mechanisms, applying a Spatial Durbin Model with two-way fixed effects. The research reveals several key findings. For instance, the digital inclusive finance significantly improves the quality of both livelihood and infrastructure public services within the local counties, a result that remains robust through sensitivity tests.

However, the empirical analysis in this article reveals a significant negative spatial spillover effect of digital inclusive finance on the quality of infrastructure-related public services in neighboring counties, while the spillover effect on livelihood-related public services is mostly insignificant. This finding deviates from the initial theoretical assumption of a "positive spillover effect." The authors argue that the negative or insignificant spillover effects are primarily due to the underutilization of digital inclusive finance in certain regions. Although, in theory, digital inclusive finance should enhance the accessibility and efficiency of financial services, thereby improving public service quality, its development remains in the early stages in economically weaker areas. These areas face multiple constraints, including inadequate infrastructure, limited penetration of digital financial products and services, and relatively low financial literacy among residents. As a result, the positive spillover effect on public service improvement has not been fully realized. Additionally, economic disparities across regions play a critical role in the observed heterogeneity of spillover effects. Regions with more developed digital inclusive finance infrastructures exhibit more mature applications and more noticeable spillover effects, while economically weaker regions experience a "siphoning effect," where financial services are concentrated in central areas, leaving surrounding regions underserved. This imbalance has inhibited the full realization of the spatial spillover effect of digital inclusive finance. Therefore, while the empirical results deviate from the initial theoretical assumptions, this phenomenon sheds light on the complexity of digital inclusive finance across different stages of economic development and varying regional conditions, offering new perspectives for further exploration of its mechanisms.

The moderation effect analysis indicates that the primary industry strengthens the impact of the digital inclusive finance on the quality of livelihood public services in surrounding counties, whereas fiscal revenue has an opposing effect. Furthermore, regional heterogeneity analysis demonstrates that the digital inclusive finance generates positive spatial spillover effects on livelihood public services across all three regions. It also shows a significant positive spillover effect on infrastructure public services in the western region, while the effects in other regions are notably negative.

Based on these findings, the article provides several policy recommendations. First, this article contends that a coordinated county-level governance system should be established to mitigate the negative spillover effects of the digital inclusive finance on public services—especially in the eastern and central regions where these adverse effects are more pronounced. It is recommended that developed counties in the eastern region take the lead in piloting a collaborative digital inclusive finance system. This can be achieved by promoting inter-county agreements on the sharing of financial technology infrastructure, and by strengthening the joint construction and use of key facilities such as digital lending platforms, credit information systems, and data processing centers. Blockchain technology can be employed to ensure transparency and dynamic monitoring of financial resource utilization. This model aligns with the strategic goals outlined in China’s 14th Five-Year Plan for digital economy development—particularly in areas such as inclusive finance and the construction of “Gigabit Cities”—and integrates with the regional coordination framework of the “Eastern Data, Western Computing” initiative. Through a lead-county-driven approach, counties with more advanced digital inclusive finance systems and higher fintech penetration can help neighboring underdeveloped areas improve their financial service capabilities, thereby reducing the marginal costs of infrastructure and service system development. In addition, supporting monitoring platforms—such as an early warning system for digital financial service disparities—should be established to track siphoning effects in real time. This would help prevent the excessive concentration of fintech talent and data resources in developed counties, ensuring balanced and sustainable financial service development across regions.

Second, this article advocates for deepening the digital transformation of the primary industry at the county level by establishing a data chain tailored to the region’s unique agricultural profile. For example, counties could introduce agriculture loans under digital inclusive finance programs, offering low-interest loans based on the area of digitized farmland—resulting in lower rates compared to traditional loans. In addition, collaborating with successful digital enterprises such as Alibaba to develop a blockchain-based traceability system for agricultural products is recommended. This approach aligns with the Ministry of Finance’s "Agricultural Industry Integration Project" and the Rural Revitalization Bureau’s "Consumption Assistance Policy." It is particularly suited to counties with high annual grain production, while also requiring measures to prevent data monopolization and to further establish a robust system for securing county-level data asset rights.

Third, this article advocates for establishing a robust cross-county public service sharing mechanism to break down regional heterogeneity barriers. For instance, by creating a "remote healthcare collaboration network," core counties with relatively strong public service resources can extend their support to neighboring counties, thereby facilitating broader coverage of AI-assisted diagnostic systems and 5G consultation equipment within the region. Additionally, in more advanced county clusters such as those in the Yangtze River Delta and the Pearl River Delta, pilot projects like a "county-level government data sharing catalog" can be introduced. This approach aligns with the overall framework outlined in the "Digital China Construction Overall Layout Plan" and aims to effectively decentralize public service resources to the county level.

In conclusion, this research emphasizes the importance of addressing the spatial dynamics of public service quality in the digital era, offering actionable policy recommendations to improve public service provision, promote equitable development, and enhance regional cooperation. However, although this article examines the impact of the digital inclusive finance on public service quality from the perspectives of spatial effects and moderating mechanisms, it has certain limitations. First, the construction of the indicator system is constrained by data availability, which may limit its ability to fully capture the multidimensional impact of the digital inclusive finance on public services. Second, this article has not further explored nonlinear relationships or dynamic effects, which may overlook the varying impacts of the digital inclusive finance at different stages of development. Furthermore, while the spatial spillover effects of the digital inclusive finance on public services exhibit regional heterogeneity, the analysis of the underlying mechanisms remains superficial. Future research can enhance the indicators, expand the methods, and deepen the understanding of these mechanisms to provide more comprehensive and robust conclusions.