Digital Government Construction and Common Prosperity in China: Effect and Transmission Channel

Abstract

How to achieve common prosperity has become the key to enhancing residents’ well-being in digital government construction (DG), which is particularly important for developing countries with relatively large income gaps. Using Chinese provincial panel data from 2018 to 2023, this study employs the entropy weight method and two-way fixed effects models to empirically examine the nonlinear impact of digital government development on common prosperity. This study has found that DG has a significant U-shaped impact on common prosperity, which first inhibits and then promotes. This effect operates primarily through improving digital inclusive finance and increasing education expenditure. Regional heterogeneity analysis indicates that the U-shaped relationship is more significant in the eastern region. From a single dimension of DG, service supply capacity and service intelligence capacity have a significant U-shaped impact on common prosperity. This study enriches the theory of the relationship between DG and common prosperity, providing policy references for promoting common prosperity and sustainable development.

1. Introduction

Promoting economic growth and common prosperity are the main focus of the country’s economic policy [1]. The International Monetary Fund (IMF) Annual Report 2024 notes a marked slowdown in worldwide economic expansion, anticipating a prolonged phase of moderate growth. Multiple developing countries and emerging economies grapple with the combined pressures of slowing growth and persistent distributive inequities [2]. China similarly confronts issues of suboptimal economic development quality and income inequality [3]. As evidenced by official data from the National Bureau of Statistics of China, the urban–rural income ratio stood at 2.34 in 2024, while the Gini coefficient has exceeded 0.46 for four consecutive years, highlighting the urgency of transitioning from “getting rich first” to “common prosperity”. Research demonstrates that unequal income distribution and significant urban–rural disparities negatively impact citizens’ overall well-being, whereas high-quality public services prompt industry development and ensure equal opportunities [4,5]. In response, China has rolled out a set of policy initiatives to tackle inefficient economic growth and uneven wealth distribution, including the Plan to Promote Common Prosperity through Digital Economy and the Resolution of the Central Committee of the Communist Party of China on Further Deepening Reform Comprehensively to Advance Chinese Modernization.

Achieving common prosperity necessitates the provision of high-quality public services, the generation of sufficient employment opportunities, and the establishment of equitable mechanisms for wealth distribution [6]. However, conventional governance patterns often suffer from postponement of information and execution inefficiencies, limiting their capacity to meet public expectations. Therefore, leveraging digital government construction (DG) to foster common prosperity has become an international consensus.

Amid the global wave of DG, the governments of developed countries are accelerating the improvement of government service quality by widely leveraging digital technologies in management [7]. The e-Government Benchmark Report 2024 indicates that 37 European nations are providing and promoting digital government services. By employing strategies such as “Upstream Social Marketing”, governments can more effectively deploy mobile governance services and strengthen public involvement and policy responsiveness [8]. Moreover, comparative analysis of local governance models in Slovakia and Lithuania demonstrates that civic participation substantially shapes the implementation effect of digital democracy [9]. As one of the earliest developing countries to incorporate DG as a national strategy, China has also made remarkable advances in policy design. In 2022, the Guidelines for Strengthening the Construction of Digital Government designated that DG is a central instrument for enhancing public service capacity and the efficiency of public administration. Driven by this policy, China’s digital infrastructure and the state of DG digital infrastructure and digital governance in China have consistently improved. China attains the highest standing among developing nations in the United Nations E-Government Survey 2024, with its E-Government Development Index standing at 0.8718. The ongoing process of DG accelerates public sector modernization while creating novel pathways to advance common prosperity.

What effects does DG have on common prosperity? Existing academic studies converge on the economic benefits of DG for common prosperity. Literature on digital governance theory demonstrates that adopting diverse digital technologies effectively curbs corruption and improves governance performance [10]. In particular, the pervasive adoption of AI, big data, and similar digital solutions fosters a more accuracy implementation of common prosperity policies by advancing the digitization and intellectualization of the policy process [11]. The platformization of government services functions to overcome geographical barriers in resource allocation, which significantly extends the reach and improves the accessibility of rural public services [12]. This allows traditionally underserved populations to benefit more equitably from public goods and ensures that the economic fruits are distributed more broadly across society. However, the economic effects of DG remain uncertain. For instance, the digital divide has resulted in unequal distribution of benefits for different groups during the early phases of DG. The dividends of digital development have less reached farmers in regions with pronounced income inequality, while significantly narrowing the wealth gap among citizens in highly economically developed areas [13]. This divergence is especially evident across regions with different levels of human capital. Lower-skilled individuals must upgrade their competencies to fully benefit from DG [14]. In the domain of employment and income, while digital technologies drive growth in the Internet economy, they may also disrupt the real economy in the short term, producing dual effects of destruction and creation to employment [15]. Against this backdrop, a thorough examination of the impact mechanisms linking DG to common prosperity is critical.

While a body of prior research has yielded certain progress in DG and common prosperity, several limitations remain. First, the theoretical framework for how DG specifically influences common prosperity is not yet fully articulated. Most current studies treat these two domains separately rather than exploring their interconnections. Given that common prosperity represents a vital pathway toward enhancing citizen welfare, further investigation is urgently needed to clarify how DG serves as a key governance means in advancing this goal. Second, developing countries frequently face a shortage of sufficient motivation and resources to implement measures to prompt DG, which leads to existing research mostly focusing on developed countries while paying less attention to developing countries and emerging economies. These regions confront distinct challenges in pursuing DG and common prosperity, and must reconcile the need between the imperative of rapid economic growth and equitable wealth sharing.

The governance innovations driven by DG are transforming the mode and effectiveness of public service supply, thereby strengthening equitable resource allocation and enhancing citizen well-being. This study makes three primary contributions. First, this study focuses on developing countries with relatively large wealth disparities. Due to their relatively lower economic development and uneven resource distribution, these countries exhibit evident income inequality. Within this context, this study examines whether DG facilitates the attainment of common prosperity, thereby extending the relevant theories and existing literature on the subject. Second, this study establishes a multidimensional assessment index system applicable to common prosperity. Employing a panel data model, this study delineates the specific transmission channels governing the U-shaped relationship between DG and common prosperity. Third, this study further reveals the heterogeneous influence of core DG dimensions on common prosperity, while also examining spatial effects and regional heterogeneity. In addition, this study equips policymakers with targeted recommendations, delivering theoretical foundations and actionable guidance to steer DG and foster common prosperity.

The following outline summarizes the organization of the rest of this article. Section 2 reviews the pertinent literature on DG and common prosperity, and derives hypotheses. Section 3 describes the research methodology. Section 4 reports empirical analysis results. Section 5 provides the discussion. Finally, Section 6 summarizes the main findings, policy implications, and study limitations.

2. Theoretical Analysis and Hypothesis Development

2.1. Common Prosperity

Common prosperity is “a group of social development features that contribute to the well-being of citizens, taking into account the prosperity of life and the fair and just distribution of wealth”. According to justice theory, the realization of common prosperity depends on all citizens enjoying bottom-line guarantees and the satisfaction of their development needs [16]. To define whether the development state of common prosperity is reached, different scholars have proposed concepts such as welfare-efficiency balance, inclusive growth, sustainable development, shared prosperity, and other conceptual connotations. Welfare-efficiency balance is to achieve a dynamic balance between economic growth and reasonable distribution of social welfare [17]. Inclusive growth advocates equal access to development opportunities for all citizens and sharing the fruits of economic growth [18]. Sustainable development emphasizes that economic growth should take environmental protection into account, so that all citizens and their descendants can benefit from inclusive growth [19]. Share prosperity emphasizes the need to focus on the growth of household income of the poor and ensure that public services cover low-income groups [20]. The concept of common prosperity can be divided into static and dynamic interpretations. The static interpretation refers to the outcome of common prosperity, that is, the elimination of the wealth gap and the enjoyment of a decent life for all citizens [21]. In contrast, Yu et al. regard common prosperity as a dynamic process, encouraging all citizens to participate in creating wealth together [1].

With the international community’s increasing emphasis on wealth inequality, the concept of common prosperity has aroused extensive discussions among scholars from different countries and regions, and related evaluation research constitutes a central concern in current academic discourse. In material dimensions, common prosperity entails fairness in both economic growth and wealth distribution, and requires a rationalized industrial structure and layout [22]. This will help to promote the integration of the labor market, realize high-quality employment for the workforce, and thus promote the balanced advancement of the regional economic pattern [23,24]. Zhao divides common prosperity into growth and development, and the shareability [25]. Kakwani conceptualized common prosperity as encompassing both shared prosperity and shared growth, and developed the corresponding index along with a measure of opportunity equality [26]. Overall, due to the limitation of data availability, many studies have different focuses on the evaluation of common prosperity.

2.2. DG

National governments leverage information and communication technologies to elevate service efficiency while harnessing digitalization benefits. The government uses digital technology to promote the automation of public administration and provide high-quality public services for citizens [27]. DG refers to the promotion of management model innovation and service function reform within government departments through the application of digital technologies [28]. Its core features include data-driven decision-making models, efficient public service supply, and cross-departmental collaborative governance. DG can significantly enhance the efficiency of policy formulation and implementation, reduce administrative costs, and strengthen the government’s ability to deal with complex issues [29]. Notably, big data and AI technologies empower governments to track economic dynamics in real time, refine resource allocation, and deliver high-quality public services. Although certain farmers consider digital government services as an administrative burden, scholarly finding indicates that digital technology adoption by local governments improves the efficiency of targeted poverty alleviation [30].

2.3. DG on Common Prosperity

Recent scholarly work has increasingly focused on DG’s impact for realizing common prosperity. DG contributes to economic growth through improved output efficiency and the creation of higher-value jobs [31]. On the other hand, the informational effect of digital government optimizes the relationship between government governance and labor market resource allocation, which helps mitigate job market polarization and expands overall employment [32]. Furthermore, improving public service supply and enabling cross-departmental collaboration are important ways for DG to boost economic growth and maintain social stability.

However, early-stage DG frequently exhibits a distinct digital divide effect. According to institutional theory, DG implies profound changes in organizational structure, operational processes, and institutional environment [33]. Initially, due to institutional inertia, reliance on the original governance path, and an incomplete supporting policy system, digital technology often failed to be fully integrated into governance practices. The poor connection between the old and new systems may bring about additional transaction costs, which can dampen economic and social efficiency. Meanwhile, the innovation adoption curves indicate that the promotion of digital government often begins in regions and groups with a better economic foundation and higher resource endowment [34]. As a result, regions with abundant financial and talent reserves tend to pioneer digital infrastructure deployment, whereas economically disadvantaged regions face barriers from the outset. For example, inadequate digital public infrastructure in rural areas can lead to inefficient resource allocation during early construction phases, potentially exacerbating urban–rural disparities [13]. In other words, the insufficient adaptability of the system and the penetration rate of digital technology have restricted the coverage and achievements of DG [35,36]. Such limitations can preclude equitable access to public services and resource distribution for certain regions and vulnerable groups. This initial “starting line” gap may evolve into an imbalance in the regional distribution of DG, allowing the digital divide effect to dominate during early phases and ultimately hindering common prosperity.

As DG reaches a relatively advanced stage, its primary impact transitions from creating a digital divide toward generating a digital dividend. Digital governance maturity models indicate that upon entering the integration and optimization phase, DG significantly enhances institutional reinvention, data integration capabilities, and public service efficiency [37]. Specifically, a mature digital government optimizes the deployment of key productive resources through platform-based and intelligent governance, thereby strengthening economic resilience [38]. Following advances in the urban business environment and declining information search costs, DG has effectively stimulated corporate digital transformation and promoted urban economic growth [39]. These developments effectively promote qualitative and dynamic improvements in economic development, fostering a sound material underpinning for attaining common prosperity. On the other hand, continuous upgrades in digital infrastructure and rising digital literacy markedly increase the accessibility and utilization efficiency of digital public services. This enhances inter-regional resource mobility and helps narrow development disparities between urban and rural areas as well as among different social groups [40]. Simultaneously, by restructuring governance frameworks through digital technology, governments can better regulate administrative behavior, curb corruption, and strengthen public trust and social cohesion [41]. The evolution of DG thus creates more favorable conditions for broader societal sharing in economic outcomes and advances common prosperity.

In conclusion, DG acts as a double-edged sword in the pursuit of common prosperity. While the digital divide may initially impede progress toward common prosperity, the digital dividend ultimately contributes to enhancing it. This pattern illustrates the complex interplay between DG and common prosperity, manifesting as a nonlinear U-shaped effect. During early construction, institutional adaptive barriers and innovation adoption disparities suppress common prosperity. As digital infrastructure and technologies mature and achieve broader dissemination, however, a transition occurs from institutional reconstruction to dividend diffusion. This evolution fosters higher levels of common prosperity. Thus, this study proposes Hypothesis 1.

Hypothesis 1.

DG has a non-linear U-shaped impact on common prosperity.

2.4. DG, Digital Inclusive Finance and Common Prosperity

DG initially lacked effective integration with digital financial systems during its early development phase, resulting in an information isolated island [42]. This fragmentation increases the cost of financial services for businesses and citizens, thereby impeding the advancement of common prosperity. When digital government reaches a certain level, the strong penetration and wide coverage of data elements will gradually appear, and new financial tools such as digital finance will develop rapidly. Digital inclusive finance contributes to urban common prosperity through bridging the digital divide and fostering innovation and entrepreneurial activity, with particularly pronounced effects observed in eastern Chinese cities [43]. Furthermore, digital inclusive finance enhances access to and usability of financial products and services in rural regions. For example, residents in impoverished areas can access funds more easily to support household production activities, thereby raising agricultural incomes [44]. This mechanism effectively reduces urban–rural household income disparities and mitigates wealth inequality. Consequently, digital inclusive finance not only boosts productivity and stimulates high-quality economic growth but also restructures income distribution patterns toward greater equity. Thus, this study proposes Hypothesis 2.

Hypothesis 2.

Digital inclusive finance plays a mediating role between DG and common prosperity.

2.5. DG, Education Expenditure and Common Prosperity

Existing research indicates that educational expenditure significantly influences urban–rural income disparities [45,46]. During initial phases of DG, substantial financial resources are often directed toward building digital infrastructure and upgrading digital technological systems [47]. This investment demand may have diminished resources for public funding in public service fields such as science and technology and employment, potentially leading to a relative reduction in government education expenditure. Educational investment has a lag effect [48]. When education spending fails to prioritize digital skill development or address urban–rural allocation imbalances, governments might defer educational investments until digitalization yields measurable returns. As DG advances, however, the government enables more efficient allocation of educational expenditures and helps cultivate higher-quality talent. For example, government-led network skills training enhances citizens’ digital literacy, improves workers’ productivity and innovative capacity, and reduces inequality among higher-income groups [49]. Furthermore, increased education expenditure in digital skills training within rural and less developed regions can reduce regional disparities in technological adoption capacities, strengthen human capital. These developments allow a broader population to share in developmental gains, thereby advancing common prosperity. Thus, this study proposes Hypothesis 3.

Hypothesis 3.

Education expenditure plays a mediating role between DG and common prosperity.

3. Methodology

3.1. Measurement of Variables

3.1.1. Dependent Variables: Common Prosperity

The dependent variable is common prosperity. The criteria for measuring common prosperity should include all the features that may affect a prosperous life and the average wealth as far as possible [50,51]. For instance, Gong et al. focused on the income gap between regions to construct a measurement framework for common prosperity [52]. However, a single or narrow indicator is not sufficient. An assessment of common prosperity requires a multidimensional framework that transcends reliance on isolated metrics. For instance, Emmers et al. hold that the goal of common prosperity includes social equality and economic equity [53]. Zou et al. constructed an index system from development, sharing, and sustainability dimensions [43]. Furthermore, data availability limitations result in considerable homogeneity in the selection of core common prosperity indicators across existing studies [50]. Therefore, drawing on several widely recognized standards in the literature, this study conceptualizes common prosperity as the unity of sharing and prosperity, and constructs a systematic indicator system for common prosperity, which includes two dimensions. Each dimension is measured by three indicators, respectively, which can effectively reflect the overall effect, especially economic and shared value.

Table 1. Common prosperity index system.

Primary Indicators	Secondary Indicators	Measurements	Effect
Sharing	Regional differences	Disposable income of residents/The average per capita disposable income	Positive
	Urban–rural differences	Disposable income of rural residents/Disposable income of urban residents	Positive
	Income disparity	Gini coefficient based on night lights	Negative
Prosperity	Engel coefficient	Food consumption accounts for the proportion of total consumer spending	Negative
	Per capita disposable income	Disposable income of residents/Permanent resident population	Positive
	Per capita consumption level	Total personal consumption of residents/permanent resident population	Positive

presents the index system, detailing the indicators and their corresponding measurement procedures.

At the sharing level, common prosperity manifests through equitable material distribution and equality of opportunity. The spatial distribution of regional economic activities can be effectively inferred through the differences in the distribution of light brightness among regions [54]. Referring to Chen et al., this study employs nighttime light data to calculate the Gini coefficient as a measure of income disparity [55]. Regional and urban–rural differences directly reveal imbalances in development opportunities and outcomes across geographic and social groups, highlighting deviations from the principle of shared prosperity. The dimension of prosperity reflects the characteristics of high-quality economic growth and material abundance in common prosperity. The Engel coefficient indicates consumption structure upgrading and improved living standards by tracking the declining proportion of household food expenditure. Per capita disposable income reflects the actual economic resources available to individuals or households, demonstrating their purchasing power and material well-being. Meanwhile, per capita consumption level illustrates the abundance of material life from the perspective of actual spending, signifying a developmental stage that surpasses basic subsistence needs.

The quantification of the common prosperity index in this study is conducted using the standard entropy weight method (EWM) [56]. For index weighting, EWM determines the objective weights according to actual data distributions, so as to capture the information utility of each index more effectively and enhance the objectivity of measurement. Accordingly, this study first standardizes the secondary indicators of common prosperity to remove metric incongruities, then utilizes EWM for weights determination, and ultimately aggregates the weighted indicators to form a composite index of common prosperity.

The common prosperity index is calculated following Equations (1)–(8), beginning with the standardization of raw data using range standardization.

Indicators for which higher values reflect more desirable outcomes are normalized using the procedure defined in Equation (1).

$$Y_{i,j}^{\prime }={\displaystyle \frac{Y_{i,j}-Y_{min}}{Y_{max}-Y_{min}}}$$

(1)

Indicators for which lower values are preferable are standardized according to the method specified in Equation (2).

$$Y_{i,j}^{\prime }={\displaystyle \frac{Y_{max}-Y_{i,j}}{Y_{max}-Y_{min}}}$$

(2)

Let Y_i,j denote the j-th indicator of the i-th common prosperity unit, where i = 1, 2, …, n and j = 1, 2, …, m. Y’_i,j is the normalized Y_i,j value, which is obtained via min–max scaling using the indicator’s maximum (Y_max) and minimum (Y_min). All transformed values are within the closed interval [0, 1].

Then, the contribution weight of each evaluation index is calculated. Based on Equation (3), this study determines each index’s contribution by calculating the proportion of the common affluency index j value.

$$C_{i,j}=Y_{i,j}^{\prime }/\sum _{i=1}^n{Y}_{i,j}^{\prime }$$

(3)

This study calculates the information entropy (e_j) for each indicator through Equations (4) and (5).

$$e_j=z\sum _{i=1}^n{C}_{i,j}ln{C}_{i,j}$$

(4)

$$z=1/Lnn$$

(5)

The utility value of information is

$$\begin{array}{c}{P}_j=1-e_j\end{array}$$

(6)

This study computes the evaluation indicator weights (w_j) by applying Equation (7).

$$w_j=P_j/\sum _{j=1}^m{P}_j$$

(7)

This study computes the common prosperity index (CP_i) for the period 2018–2023 using Equation (8).

$${CP}_j=\sum _{j=1}^m{w}_j{Y}_{i,j}^{\prime }$$

(8)

3.1.2. Measurement of DG

The independent variable is DG. Drawing on Haug et al., this study conceptualizes DG as a process where governments respond to digital era demands by leveraging digital technologies for elevating public service quality [28]. Relevant literature employs a composite assessment across multiple dimensions to measure DG [57]. Hao et al. explored keywords related to digital government using text analysis techniques and constructed an index system [58]. Meanwhile, the international community has gradually established an authoritative framework for evaluating the global e-government development level. Paroški employed the EGDI, published by the United Nations, to assess e-government development in EU countries [59]. This system assesses the comprehensive level of the government’s application of digital technology in governance from three dimensions: service supply capacity (SS), service response capacity (SR), and service intelligence capacity (SI). As a comprehensive and authoritative indicator for measuring DG, its construction follows a rigorous research design and has gained broad recognition in both academic and policy practice.

Figure 1 depicts the relationship between DG and common prosperity through a scatter plot. The red solid line represents the U-shaped fitting curve of the degree of DG and common prosperity. The intersection of the red solid line and the dotted line is the extreme point. The results represent that the scattered points are distributed around the U-shaped curve, indicating that DG initially inhibits common prosperity and then promotes it. This graphical evidence provides preliminary support for Hypothesis 1. Subsequent analyses apply a two-way fixed effects model (FE) to further examine.

3.2. Model Construction and Data Sources

Considering the possible nonlinear relationship, this study constructs a benchmark regression model as shown in Equation (9).

$$Y_{it}={\beta }_0+{\beta }_1{DG}_{it}+{\beta }_2{DG}_{it}^2+{\beta }_3{control}_{it}+{\mu }_i+{\partial }_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(9)

In the model, i denotes the province and t indicates the year, with the data spanning a six-year period. The terms µ_i and ∂_t represent province and year fixed effects, respectively, while ε_it denotes the random disturbance term. Using variables Y_it and DG_it represents the degree of common prosperity and DG. The variables Y_it and DG_it refer to the level of common prosperity and the DG level in province i during year t. To prevent endogeneity concerns arising from omitted variables, this study introduces a set of control variables including labor productivity (LP), R&D expenditure (R&D), urbanization rate (Urb), transportation facilities (TrF), and total population employment rate (TPER).

Based on data availability and continuity, this study removes regions with missing data and employs a panel dataset comprising 25 Chinese provinces over the period 2018–2023 as the empirical basis for the analysis. The data of the dependent variable and the control variable are derived from the China Statistical Yearbook and the China Social Statistical Yearbook. In the benchmark model, all variables were transformed using natural logarithms.

Table 2. Descriptive statistical results.

Variables	Mean	Standard Deviation	Min	Max	Measurements
lnY	−1.6834	0.6191	−3.5352	−0.3509	EWM
lnDG	4.2085	0.1396	3.7079	4.4271
lnLP	2.5183	0.2947	1.8019	3.2769	GDP/Employed population
lnR&D	0.4561	0.4541	−0.5978	1.2641	R&D on expenditure/GDP
lnUrb	4.1254	0.1085	3.8571	4.3231	The number of permanent urban residents/The total number of permanent residents
lnTrF	3.6870	0.4757	2.8583	5.0140	Total mileage of highways/The total number of permanent residents
lnTPER	3.9561	0.0745	3.7618	4.1231	Employed population/total labor force

summarizes the descriptive statistics for the main variables. The mean lnDG is 4.0285 (i.e., DG = 67.27), and the mean and standard deviation of lnY are −1.6834 (i.e., Y = 0.1858) and 3.1843 (i.e., Y = 0.6749), respectively. The distribution reveals a relatively low level of common prosperity across provinces, and exhibits obvious regional disparities.

Before testing the hypothesis, this study conducts the correlation test among variables.

Table 3. Pearson correlation coefficient.

	Y	DGI	LP	R&D	Urb	TrF	TPER
lnY	1.0000
lnDG	0.3922 ***	1.0000
lnLP	0.7358 ***	0.5904 ***	1.0000
lnR&D	0.6638 ***	0.5156 ***	0.5788 ***	1.0000
lnUrb	0.7763 ***	0.3517 ***	0.7891 ***	0.5399 ***	1.0000
lnTrF	−0.7228 ***	−0.36512 ***	−0.3736 ***	−0.6428 ***	−0.3990 ***	1.0000
lnTPER	0.3361 ***	0.1593	0.1717 **	0.4060 ***	0.0602	−0.4053 ***	1.0000

Note: *** p < 0.01, ** p < 0.05.

demonstrates that the absolute values of all variable coefficients remain below 0.8, while multicollinearity tests reveal a mean variance inflation factor under 10, confirming the absence of severe multicollinearity concerns.

To ensure time series stationarity and prevent spurious regression, this study conducts unit root tests using the LLC, ADF, and Hadri methods. Results in

Table 4. The test results of variable stationarity.

Variables	LLC	Fisher-ADF	Hadri
lnY	−6.9403 ***	−5.7276 ***	2.3244 **
lnDG	−12.6619 ***	−7.1857 ***	7.7912 ***
(lnDG)2	−12.2366 ***	−7.1805 ***	7.7725 ***
lnDFI	−26.9402 ***	−5.1202 ***	11.5544 ***
lnEdu	−5.3156 ***	−5.4670 ***	5.5142 ***
lnLP	−5.3616 ***	−4.2404 ***	11.4014 ***
lnR&D	−6.2540 ***	−6.3643 ***	10.4279 ***
lnUrb	−12.2780 ***	−8.3544 ***	10.6129 ***
lnTrF	−41.6256 ***	−8.5864 ***	9.3247 ***
lnTPER	−6.8038 ***	−6.9214 ***	8.5382 ***

Note: *** p < 0.01, ** p < 0.05.

demonstrate that all variables are significant at a confidence level of 1% or 5%, passing the stationarity test.

4. Results Analysis

4.1. Baseline Result

Table 5. Baseline regression results of DG and common prosperity.

Variables	FE
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
lnDG	−0.3117	−8.2814 ***	−8.1510 ***	−8.1498 ***	−8.2192 ***	−8.4005 ***	−8.6079 ***
	(0.1880)	(1.9950)	(1.9325)	(1.9783)	(1.9142)	(1.7364)	(1.6500)
(lnDG)2		0.9879 ***	0.9715 ***	0.9713 ***	0.9788 ***	0.9984 ***	1.0258 ***
		(0.2479)	(0.2405)	(0.2468)	(0.2404)	(0.2209)	(0.2121)
lnLP	0.0592		−0.0984	−0.0981	−0.0968	−0.0938	0.1320
	(0.3136)		(0.2405)	(0.2345)	(0.2395)	(0.2310)	(0.2863)
lnR&D	0.0689			0.0010	0.0047	0.0038	0.0565
	(0.0612)			(0.0894)	(0.0914)	(0.0887)	(0.0591)
lnUrb	−0.2205				0.4857	0.0644	−0.2008
	(0.8923)				(0.7461)	(0.7378)	(0.7377)
lnTrF	0.2931 *					0.3074 *	0.3130 **
	(0.1606)					(0.1518)	(0.1495)
lnTPER	0.5813						0.6504
	(0.5777)						(0.4518)
Constant	−3.1130	15.5823 ***	15.5516 ***	15.5483 ***	13.7212 **	14.7309 ***	13.0507 ***
	(3.7018)	(4.0311)	(3.9922)	(4.0978)	(5.6162)	(4.9424)	(4.6618)
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Extreme point		4.1916 **	4.1951 **	4.1952 **	4.1987 **	4.2071 **	4.1956 **
Observations	150	150	150	150	150	150	150
R-squared	0.3912	0.4113	0.4129	0.4129	0.4155	0.4321	0.4459

Note: Standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1.

reports the baseline regression results of lnDG and lnY based on Equation (9). Column (1) displays estimates without (lnDG)², indicating no significant linear relationship. Columns (2) to (7) progressively incorporate control variables and include (lnDG)². Across all specifications, the lnDG coefficients exhibit a statistically significant negative value (p < 0.01), while the squared term (lnDG)² registers a significantly positive coefficient at the same significance threshold, confirming a U-shaped relationship between DG and common prosperity. Given that relying solely on the signs and significance of these coefficients may lead to erroneous conclusions regarding U-shaped relationships, this study employs the U-test procedure for further verification [60]. The results indicate that the significance of the U-test results holds at the 5% level, irrespective of the inclusion of control variables, thereby offering robust empirical support for a U-shaped relationship. According to Column (7), the U-test identifies an extreme point at 4.1956 (i.e., Y = 66.43), which is significant at the 5% level. The results indicate that DG initially inhibits common prosperity until a threshold, after which it begins to foster this outcome. Thus, Hypothesis 1 is preliminarily confirmed.

The U-shaped impact of digital government on common prosperity arises primarily because substantial upfront investment in digital infrastructure during its initial phase diverts fiscal resources away from near-term tax reductions, subsidies, and public expenditures [61]. Simultaneously, integrating traditional administrative systems with digital technologies requires time, and policymakers’ inexperience in digital governance may lead to suboptimal decisions that inhibit economic growth. For instance, early digital government systems often suffered from data fragmentation and poor interdepartmental coordination, which reduced policy implementation efficiency [62]. As DG advances, its scale effects on common prosperity become significantly positive. Improved data governance capabilities enable more precise use of digital tools, facilitating targeted resource allocation and reducing marginal costs of public service delivery. These enhancements create greater public value and contribute substantially to common prosperity.

4.2. Robustness Test

To ensure the reliability of the above findings, this study performs two robustness tests, building upon the baseline regression results. First, a random effects (RE) model is introduced to replace the FE model. Second, two control variables are incorporated stepwise. An advanced industrial structure (InA) influences common prosperity by increasing the proportion of technology-intensive industries and the share of labor compensation, thereby improving income distribution. Internet penetration rate (IPR) contributes to common prosperity by reducing information barriers, expanding access to opportunities and promoting efficient resource allocation. This study measures InA using the ratio of the output value of the tertiary industry to that of the secondary industry. IPR is measured as the number of broadband internet users divided by the year-end resident population. All related data are obtained from the China Statistical Yearbook.

Table 6. Robustness test results.

Variables	RE	Adding
	(1)	(2)	(3)
lnDG	−7.7740 ***	−7.4315 ***	−7.5948 ***
	(2.4067)	(2.2550)	(2.5513)
(lnDG)2	0.9331 ***	0.8865 ***	0.9054 ***
	(0.2932)	(0.2840)	(0.3151)
lnLP	0.0346	0.4857	0.4814
	(0.1530)	(0.4560)	(0.4676)
lnR&D	0.0610	−0.0023	−0.0089
	(0.0838)	(0.0937)	(0.0921)
lnUrb	1.7928 ***	−0.3816	−0.4880
	(0.4230)	(0.7516)	(0.9107)
lnTrF	−0.5486 ***	0.2979 *	0.2929 *
	(0.1307)	(0.1511)	(0.1467)
lnTPER	0.4142	0.8780	0.8977 *
	(0.2838)	(0.5361)	(0.4937)
lnInA		0.2156	0.2140
		(0.2330)	(0.2383)
lnIPR			0.0501
			(0.2260)
Constant	7.3620	9.5989	10.4006
	(4.7321)	(6.3398)	(8.8690)
Year fixed effects	No	Yes	Yes
Province fixed effects	No	Yes	Yes
Extreme point	4.1656 ***	4.1915 *	4.1942 *
Observations	150	150	150
R-squared	0.0513	0.4542	0.4547

Note: Standard errors in parentheses, *** p < 0.01, * p < 0.1.

presents the estimation results of the aforementioned robustness tests. Both the alternative regression model and the augmented model with additional control variables yield positive coefficients for (lnDG)², with a significance level (p < 0.01) for each. The U-test results further confirm the U-shaped relationship hypothesis. These findings provide robust evidence supporting the U-shaped effect.

4.3. Endogeneity Test

To address potential endogeneity, this study employs three identification strategies. The test results are presented in

Table 7. One period lag, Winsorized regression and system-GMM.

Variables	One Period Lag	Reduced Tail Regression	System-GMM
	(1)	(2)	(3)
L.lnY			0.2816 *
			(0.1452)
L.lnDG	−8.7213 ***
	(2.9836)
L.(lnDG)2	1.0117 **
	(0.3756)
lnDG		−9.5396 ***	−23.5951 ***
		(1.9050)	(8.8150)
(lnDG)2		1.1453 ***	2.9176 **
		(0.2370)	(1.1375)
lnLP	0.1045	0.1580	0.2057
	(0.2926)	(0.2619)	(0.2726)
lnR&D	0.0255	0.0211	−0.0218
	(0.0725)	(0.0637)	(0.1157)
lnUrb	−2.0158	−0.2996	1.7956 **
	(1.4067)	(0.5915)	(0.8913)
lnTrF	0.4739 *	0.3361 **	−0.3112 **
	(0.2383)	(0.1490)	(0.1361)
lnTPER	0.6905	0.7266	0.3710
	(0.4772)	(0.4552)	(0.4157)
Constant	20.5434 ***	14.8326 ***	38.1356 **
	(6.4057)	(4.4939)	(15.1579)
Year fixed effects	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes
Extreme point	4.3100	4.1648 ***	4.0437 **
AR (1)			0.007
AR (2)			0.599
Hansen			0.127
Observations	150	150	125
R-squared	0.4860	0.4498

Note: Standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1. L.lnY represents one-period lag of common prosperity.

.

First, we introduced the first-period lagged term of the linear DG variable L.lnDG and its squared term L.(lnDG)² as core explanatory variables and estimated the model using FE. This approach mitigates endogeneity due to reverse causality. Table 7, Column (1) reports a persistently positive and statistically significant coefficient (5% level) for L.(lnDG)², in alignment with the baseline results.

Second, to ensure the reliability of the findings and minimize the influence of potential outliers, this study conducted Winsorized regression. The relevant variables were Winsorized at the 1st and 99th percentiles. The results after Winsorization, presented in Column (2) of Table 7, indicate that the (lnDG)² coefficient estimate is positive and significant (p < 0.01). These results strengthen the robustness of Hypothesis 1

Third, Column (3) of Table 7 presents the results of the two-step system GMM estimation. Endogenous regressors were instrumented using their second-to fourth-period lags in levels. The AR (1) statistic is significant (p = 0.007), suggesting first-order serial correlation. The AR (2) test statistic is insignificant (p = 0.599). Moreover, the Hansen test result (p = 0.127) suggests that the overidentifying restrictions are valid and the instrument set is exogenous. After including L.lnY to control for potential omitted variable bias, DG continues to exert a positive effect at the 10% significance level. These findings collectively affirm Hypothesis 1.

4.4. Mechanism Analysis

As previously established, DG has a significant U-shaped relationship with common prosperity. This study employs a quadratic function-based mediation approach to examine the mechanisms. The corresponding mediating effect models are specified in Equations (10)–(12).

$${Med}_{it}={\alpha }_0+{\alpha }_1{DG}_{it}+{\alpha }_2{DG}_{it}^2+{\alpha }_3{control}_{it}+{\mu }_i+{\partial }_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(10)

$$Y_{it}={\delta }_0+{\delta }_1{DG}_{it}+{\delta }_2{DG}_{it}^2+{\delta }_3{Med}_{it}+{\delta }_4{control}_{it}+{\mu }_i+{\partial }_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(11)

$$\begin{array}{c}{IND}_{it}={\alpha }_2\ast {\delta }_3\end{array}$$

(12)

Med denotes the mediating variable. Equation (10) examines the potential nonlinear association between DG and the mediating variable. Equation (11) extends Equation (9) by incorporating the mediating variable to evaluate its influence on common prosperity. Equation (12) captures the magnitude of the nonlinear indirect role of DG in shaping common prosperity through the mediator. The significance and strength of the indirect effect (IND) serve to evaluate the presence and relevance of the mediating mechanism.

Based on the hypotheses in the previous text, the mediating variables set in this study are as follows:

(1)

Digital inclusive finance (DIF). The data is sourced from PKU_DFIIC.

(2)

Education expenditure (Edu), calculated by the ratio of education expenditure to general public budget expenditure. The data is sourced from the China Statistical Yearbook.

Column (1) of

Table 8. Analysis of influence mechanism.

Variables	lnDIF	lnY	lnEdu	lnY
	(1)	(2)	(3)	(4)
lnDIF		0.6564 *
		(0.3493)
lnEdu				0.3697 *
				(0.1899)
lnDG	−2.5623 ***	−6.9259 **	−4.1761 ***	−7.0639 ***
	(0.8732)	(2.5327)	(0.8764)	(1.8826)
(lnDG)2	0.3014 ***	0.8280 **	0.5070 ***	0.8384 **
	(0.1053)	(0.3164)	(0.1056)	(0.2414)
lnLP	−0.1315	0.2183	−0.0376	0.1459
	(0.1038)	(0.2797)	(0.1216)	(0.2818)
lnR&D	−0.0005	0.0569	0.0777	0.0278
	(0.0293)	(0.0604)	(0.0494)	(0.0655)
lnUrb	−0.6813 ***	0.2464	0.1285	−0.2483
	(0.5113)	(0.7031)	(0.4074)	(0.6912)
lnTrF	−0.0541	0.3485 **	0.0542	0.2930 *
	(0.0616)	(0.1556)	(0.1587)	(0.1620)
lnTPER	−0.2010 *	0.7823 *	0.2597	0.5544
	(0.1946)	(0.4162)	(0.1815)	(0.4513)
Constant	15.1407 ***	3.1124	9.6562 ***	9.4807
	(2.1896)	(7.4665)	(2.9933)	(4.5934)
Year fixed effects	Yes	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes	Yes
Extreme point	4.2503 *	4.1825 *	4.1181 ***	4.2129 *
Observations	150	150	150	150
R-squared	0.9755	0.4672	0.2768	0.4626

Note: Standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1.

indicates that the lnDG coefficient exhibits a statistically significant negative value (p < 0.01), while that of the (lnDG)² coefficient is significantly positive (p < 0.01), confirming a U-shaped relationship between DG and digital financial inclusion. Column (2) indicates that the coefficients of both (lnDG)² and lnDIF are significant (p < 0.05 and p < 0.1), respectively, suggesting a mediating effect of DIF. This implies that once the DG index exceeds a critical threshold, it exerts a strong positive impact on common prosperity by enhancing DIF. Thus, digital inclusive finance serves as a key mechanism through which DG influences common prosperity in a U-shaped manner, supporting Hypothesis 2.

Column (3) of Table 8 reveals that the coefficients of both lnDG and (lnDG)² are significant (p < 0.01), indicating that their effect on Edu first decreases and then increases. Column (4) indicates that the lnEdu coefficient is 0.3697 and significant (p < 0.1), suggesting that increased Edu promotes common prosperity. Hence, education expenditure is another key mechanism of the U-shaped impact, confirming Hypothesis 3.

This study employs the Bootstrap method to test the robustness of the above mediating effects. Using 2000 simulation repetitions and a 95% confidence interval (CI),

Table 9. Mediating effect Bootstrap test.

Variables	Effect	Observed Coefficient	Boot SE	95% CI
lnDIF	ind_eff	0.1979	0.1457	(0.0108, 0.6078)
lnEdu	ind_eff	0.2639	0.1145	(0.0807, 0.5998)

presents the results for the mediating roles of DIF and Edu. The Bias-corrected 95% CI for both mediators excludes zero, confirming the significance of the mediating pathways. Therefore, Hypotheses 2 and 3 are supported.

4.5. Further Analysis

To examine the spatial effects of DG on common prosperity, this study develops a spatial panel metrological model extending the baseline model in Equation (9) as follows:

$$\begin{array}{c}{lnY}_{it}={\gamma }_0+{\rho WlnY}_{it}+{\gamma }_1{WlnDG}_{it}+{\gamma }_2{WlnDG}_{it}^2+{\gamma }_{3}ln{DG}_{it}+{\gamma }_4{lnDG}_{it}^2+{\gamma }_{5}Wln{control}_{it}+{\gamma }_6{lncontrol}_{it}+{\mu }_i+{\partial }_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}\end{array}$$

(13)

Under the economic geographic matrix, Moran’s I is used to test the spatial correlation of common prosperity in the sample provinces, and the results are shown in

Table 10. Common prosperity Moran’s I 2018–2023.

Years	Moran’s I	Sd.	z
2018	−0.2362	0.0982	−1.9812
2019	−0.2314	0.0985	−1.9257
2020	−0.2186	0.0986	−1.7944
2021	−0.2337	0.0982	−1.9561
2022	−0.2104	0.0977	−1.7276
2023	−0.2459	0.0972	−2.1004

. Moran’s I for common prosperity is all less than 0, and all have passed the significance test. This means that the level of common prosperity is not randomly distributed in space, but rather shows a distinct “high-low” interlaced distribution pattern. That is, regions with a higher level of common prosperity tend to be surrounded by those with a lower level, and vice versa. This spatial model may reflect the “siphon effect” or “competitive relationship” in regional development. Developed regions have absorbed resources from surrounding areas rather than diffused them, indicating that the current drive toward common prosperity is hindered by significant regional inequalities, and the coordinated development mechanism remains incomplete.

Preliminary tests confirm significant spatial dependence in common prosperity. This study therefore employs Lagrange Multiplier (LM) tests to evaluate the suitability of alternative spatial panel metrological. As presented in

Table 11. Lagrange Multiplier (LM) tests results.

Tests	LM Value	p Value
LM_lag	3.3920	0.000
Robust_LM_Lag	25.3270	0.000
LM_error	28.5600	0.000
Robust_LM_error	50.4940	0.000

, the p-values for tests are all statistically significant at conventional levels. These results suggest the simultaneous presence of both spatial lag and spatial error autocorrelation. Consequently, the Spatial Durbin Model (SDM) emerges as the most appropriate specification.

Based on the preceding spatial correlation test results, this study further employs SDM to examine whether DG exerts spatial spillover effects on common prosperity.

Column (1) of

Table 12. Regression Results of SDM.

Variables	Direct effect	Spatial Spillover Effects
	(1)	(2)
lnDG	−8.0104 ***	1.4794
	(2.1153)	(8.1028)
(lnDG)2	0.9508 ***	−0.1109
	(0.2611)	(0.9820)
lnLP	0.1351	−1.2858 **
	(0.2242)	(0.5786)
lnR&D	0.1034	−0.2422
	(0.1110)	(0.4008)
lnUrb	−0.2165	2.0236
	(0.6391)	(2.0096)
lnTrF	0.4396 **	−0.5790
	(0.1748)	(0.5706)
lnTPER	0.6418 *	−2.3216 *
	(0.3651)	(1.3779)
rho	−0.2843 *
	(0.1530)
Observations	150	150
R-squared	0.0461	0.0461

Note: standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1.

reveals a significant U-shaped influence of DG on common prosperity. The lnDG estimate is −8.0104 (p < 0.01), and the (lnDG)² is 0.9508 (p < 0.01), supporting a U-shaped relationship. This suggests that initial digital government development may suppress common prosperity, yet promote it after a certain threshold. The spatial autocorrelation coefficient is −0.2843 (p < 0.1), indicating negative spatial dependence and a high-low clustering pattern across regions.

According to Column (2), neither lnDG nor (lnDG)² is statistically significant (p > 0.1), implying that no significant spatial spillover effects are detected. Thus, digital government in one region does not appear to influence common prosperity in neighboring areas. Notably, R&D in adjacent regions exerts a significant negative spatial spillover effect (−1.2858, p < 0.05), indicating that it inhibits common prosperity in surrounding regions.

4.6. Heterogeneity Analysis

4.6.1. Regional Heterogeneity

Although the level of DG is steadily improving, statistical data show that there is still a clear gap between eastern and western provinces, as evidenced by the 13 eastern provinces generally leading the 12 western provinces in the sample. Considering the differences among various provinces in terms of the level of digital infrastructure construction, economic development and market scale, this study further analyzed regional heterogeneity.

Columns (1) and (2) of

Table 13. Results of regional heterogeneity test.

Variables	Eastern Provinces	Western Provinces
	(1)	(2)
lnDG	−9.3202 ***	−5.1222 *
	(2.8433)	(2.7960)
(lnDG)2	1.1234 ***	0.5894
	(0.3436)	(0.3508)
lnLP	0.1525	0.0485
	(0.3319)	(0.6467)
lnR&D	0.0863	−0.0481
	(0.1064)	(0.2634)
lnUrb	−0.9856	−1.4581
	(0.8977)	(2.0746)
lnTrF	0.1061	0.4500
	(0.1955)	(0.2997)
lnTPER	0.0526	2.0803
	(0.4666)	(1.7485)
Constant	21.0156 ***	4.7040
	(6.6220)	(7.8096)
Year fixed effects	Yes	Yes
Province fixed effects	Yes	Yes
Extreme point	4.1482 ***	4.3453
Observations	78	72
R-squared	0.4576	0.5620

Note: Standard errors in parentheses, *** p < 0.01, * p < 0.1.

reveal a significant U-shaped relationship between DG and common prosperity in eastern provinces, whereas such an effect is not statistically significant in western regions. This divergence may be attributed to the eastern region’s superior initial conditions, which enable it to absorb initial digital government investment costs more efficiently and achieve long-term gains. In contrast, the western region’s relatively weaker economic foundation and insufficient scale and capacity for digital investment may hinder the effective realization of value from digital government services [63].

4.6.2. Multi-Dimensional Heterogeneity

To explore the heterogeneous effects of different sub-dimensions in DG, this study analyzes the impact of service supply capacity (SS), service response capacity (SR) and service intelligence capacity (SI) on common prosperity.

Columns (1) and (3) of

Table 14. Results of multi-dimensional heterogeneity test.

Variables	FE
	(1)	(2)	(3)
lnSS	−10.6649 ***
	(1.6440)
(lnSS)2	1.5891 ***
	(0.2474)
lnSR		−0.1558
		(0.4966)
(lnSR)2		0.0098
		(0.1129)
lnSI			−3.6884 ***
			(1.1045)
(lnSI)2			0.5755 ***
			(0.1802)
lnLP	0.1139	0.0682	0.1553
	(0.3219)	(0.2953)	(0.2602)
lnR&D	0.0350	0.0646	0.0537
	(0.0605)	(0.0583)	(0.0696)
lnUrb	0.0488	−0.7409	−0.5662
	(0.7632)	(0.9651)	(0.8159)
lnTrF	0.2616	0.2902 *	0.2946 **
	(0.1540)	(0.1443)	(0.1405)
lnTPER	0.6482	0.6325	0.7237
	(0.5194)	(0.5997)	(0.5147)
Constant	12.1394 **	−2.1695	2.1200
	(4.6227)	(4.0186)	(3.7385)
Year fixed effects	Yes	Yes	Yes
Province fixed effects	Yes	Yes	Yes
Extreme point	3.3557 ***	7.9727	3.2044 **
Observations	150	150	150
R-squared	0.4405	0.3892	0.4345

Note: Standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1.

report negative coefficients for lnSS and lnSI, respectively. The coefficients of (lnSS)² and (lnSI)² are positive and statistically significant (p < 0.01). The above results indicate U-shaped relationships between both service supply capacity and service intelligence capacity of DG and common prosperity, characterized by initial suppression followed by subsequent promotion. Furthermore, comparing the inflection points across models reveals that the threshold for service intelligence capacity (3.2044) is lower than that for service supply capacity (3.3557), suggesting that the promoting effect of service intelligence capacity manifests earlier and exerts a stronger influence. This may be because service intelligence leverages smart technologies to improve income distribution with greater cost-effectiveness and efficiency. It enhances policy precision in areas such as social security and taxation, thereby contributing more immediately to common prosperity. By contrast, enhancing service supply capacity requires substantial investment and longer-term development to effectively address regional imbalances.

Column (2) of Table 14 indicates that the service response capacity dimension does not exhibit a significant effect on common prosperity. Although improved service responsiveness can increase the efficiency of online consultation services, it may not directly mitigate urban–rural income disparities. One plausible explanation is that digitally disadvantaged groups, such as elderly individuals, still face barriers in using internet-based services [64]. As a result, the benefits of such improvements are neither widely accessible nor evenly distributed, limiting their overall impact on common prosperity.

5. Discussion

Against the backdrop of digital transformation, enhancing common prosperity has become a critical societal issue attracting widespread attention from scholars and governments globally. However, limited understanding exists regarding the impact and mechanisms of DG on common prosperity. This study addresses this under-explored area by examining how DG fosters both economic growth and distributional equity. It further elucidates pathways through which less developed regions and vulnerable groups can overcome developmental disadvantages and achieve more inclusive benefits from digitalization. Current measurement frameworks for common prosperity continue to evolve. This research innovatively develops a comprehensive evaluation system integrating sharing value and prosperity value, enabling multidimensional and precise assessment. EWM-based indices reveal considerable disparities in common prosperity across Chinese provinces from 2018 to 2023. The overall low average index indicates an urgent need for more balanced regional development. These findings align with recent studies [65]. Regions endowed with abundant resources and sound economic foundations typically provide stronger institutional support and sufficient resources for advancing common prosperity. These advantaged areas not only better meet basic public service demands but also supply higher-quality development opportunities. In contrast, less developed regions face multiple constraints—including weak infrastructure, limited fiscal capacity, scarce human capital, and ineffective policy implementation-which further exacerbate interregional disparities in common prosperity.

While studies on common prosperity in developed countries offer valuable insights, developing countries commonly face acute resource allocation imbalances and heightened vulnerability to digital divide escalation. These contexts underscore the particular importance of DG in unlocking developmental benefits. Empirical findings reveal a U-shaped relationship where DG initially suppresses but subsequently accelerates common prosperity. This result replenishes the conventional assumption of a purely linear association prevalent in existing literature [66].

Common prosperity is also constrained by underdeveloped digital finance and insufficient human capital. Promoting digital inclusive finance and strengthening education expenditure represent two key channels through which DG advances common prosperity. By leveraging the internet and digital technologies, DG facilitates information exchange between government agencies and financial institutions. This enhances the precision of credit allocation and contributes to equalizing economic opportunities [67]. Meanwhile, increased investment in education helps improve digital literacy and technical skills among the population, strengthening productive capabilities and technological adoption. These improvements raise labor productivity and stimulate innovation, thereby mitigating income inequality [68]. Although DG may require citizens to invest more time in acquiring digital skills during the process of public service optimization and resource allocation, the resulting accumulation of digital literacy ultimately generates substantial economic and social returns. These gains support the broader goal of common prosperity.

Furthermore, the multidimensional nature of DG offers a robust foundation for examining its differential effects on common prosperity. In terms of individual dimensions, both service supply capacity and service intelligence capacity show significant U-shaped relationships with common prosperity, whereas service responsiveness shows only a weak influence. These findings suggest that DG generates heterogeneous effects across distinct functional domains, thereby extending and enriching existing literature [38].

6. Conclusions, Implications and Limitations

6.1. Main Findings

DG represents a vital indicator of modernization in public governance, reflecting the expansion and integration of information technology across all stages and domains of government operations. Based on provincial-level panel data from China (2018–2023), this study systematically examines the impact of DG on common prosperity. Adopting multiple empirical approaches, it identifies and verifies the transmission channels through which DG influences common prosperity. These findings help to a deeper understanding of the conceptual and practical dimensions of digital government and enrich theoretical discourse on its effects within public governance. The study yields three main findings:

First, this study reveals that the relationship between DG and common prosperity is not simply linear, but rather follows a nonlinear U-shaped pattern characterized by initial suppression followed by promotion. This finding underscores the long-term and conditional nature of DG effectiveness. Further analysis confirms the robustness of this relationship, though the effect does not yet exhibit cross-regional spillovers. The above findings provide strong evidence that DG contributes to common prosperity through multiple pathways, thereby enriching theories and existing literature on the economic effects of DG. Solomon & van Klyton found that the coefficient of the impact of enterprises and governments’ use of ICT on economic growth in African countries is sometimes not significant, and our study supplements this finding [69].

Second, mechanism analyses indicate that DG exerts a U-shaped impact on common prosperity through two core channels: enhancing digital inclusive finance and increasing education expenditure. Digital inclusive finance expands service availability while lowering transaction expenses, effectively supporting income growth among vulnerable groups. Increased education expenditure strengthens human capital accumulation, laying a foundation for social mobility and long-term equitable development.

Third, further analysis indicates that two sub-dimensions of DG exert a significant U-shaped influence on common prosperity, whereas the service responsiveness dimension shows no statistically significant effect. This conclusion is consistently supported by U-test results. Moreover, the impact of DG exhibits notable regional heterogeneity. A significant U-shaped relationship is observed in the more economically developed eastern regions, while this association is not statistically significant in western regions. These findings underscore the dependence of DG on initial regional digital resource conditions and provide an empirical basis for formulating differentiated regional policies.

6.2. Theoretical and Practical Implications

This study not only identifies the complex effects of DG on common prosperity but also clarifies its underlying mechanisms and contextual boundaries. Based on empirical findings, this study proposes targeted policy implications. That offer theoretical support and empirical insights to assist governments in formulating precise and efficient DG strategies aimed at promoting common prosperity and sustainable social development.

First, implement differentiated regional strategies for DG and improve digital infrastructure. During the initial phase, efforts should prioritize addressing infrastructure gaps in underdeveloped regions to reduce access and usage barriers, thereby shortening the duration of the inhibitory effect. This can be achieved through targeted central fiscal transfers to support the construction of 5G networks, government cloud platforms, and other digital infrastructure in western and rural areas. In the middle and later stages, the focus should shift toward sustained institutional optimization. Meanwhile, equal access to digital public services should be incorporated into local government performance evaluation systems to establish binding mechanisms that ensure equitable advancement of DG across regions.

Second, promote the development of digital inclusive finance and increase education expenditure. Facilitate the growth of digital financial inclusion by integrating government databases with financial institution interfaces and enabling secure data sharing across departments. Such measures enable digital financial services to address credit and financing challenges more accurately and efficiently. Concurrently, education expenditure should be directed preferentially toward less developed regions and specifically allocated to digital skills training, thereby boosting human capital and advancing common prosperity.

Third, maintaining policy continuity and coordinating diverse governance tools to maximize the overall effectiveness of DG in advancing common prosperity. In addition, deepen a strategic integration of AI and blockchain to accelerate the maturation of DG and foster common prosperity. In more developed eastern regions, policy emphasis should transition from infrastructure “construction” to “deepened application.” Examples include adopting AI-powered approval systems and blockchain-based governance to transcend the inflection point and pioneer pathways to common prosperity underpinned by high-quality DG. In less developed western regions, policymakers should develop simplified digital service platforms adapted to local socioeconomic conditions. These efforts should strengthen public capacity to utilize digital tools in accessing services and opportunities, creating essential conditions for DG to improve income distribution.

6.3. Limitations and Future Directions

Although this study provides new empirical evidence on the U-shaped relationship between DG and common prosperity and its transmission mechanisms, several limitations remain that warrant further investigation in future research. First, this study conducts analysis at the provincial level and fails to capture the differential effects of DG on common prosperity across cities and counties. Future research can employ city-level data for more granular empirical investigations. Second, due to data availability constraints at the provincial level, this study employs a Gini coefficient based on night lights data as a proxy for common prosperity. Although widely adopted in the literature, this measure may not fully capture income distribution details directly relevant to common prosperity. Future studies could incorporate more accurate micro-level income survey data to enhance the empirical robustness of the findings. Third, the DG index used in this study is sourced from a third-party database, which has inherent limitations in statistical coverage and indicator selection. Future research could develop a more comprehensive evaluation index system, for instance, the fiscal expenditure for DG. Finally, while this study identifies significant regional heterogeneity, the model remains unable to fully account for all region-specific factors, such as institutional quality and social capital. Future research can introduce more abundant microscopic data to explore heterogeneous effects.