[go: up one dir, main page]

Next Article in Journal
Vit-Traj: A Spatial–Temporal Coupling Vehicle Trajectory Prediction Model Based on Vision Transformer
Previous Article in Journal
Patent Openness Decisions and Investment Propensities of Frontier Enterprises in Asymmetric Competition
Previous Article in Special Issue
Design Thinking in Innovation Processes: A Market Segmentation Tool in Social Networks Research
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Digital Finance, Digital Usage Divide, and Urban–Rural Income Gap: Evidence from China

1
School of Business, Hunan University of Science and Technology, Xiangtan 411201, China
2
School of Architecture and Art Design, Hunan University of Science and Technology, Xiangtan 411201, China
*
Author to whom correspondence should be addressed.
Systems 2025, 13(3), 145; https://doi.org/10.3390/systems13030145
Submission received: 2 January 2025 / Revised: 17 February 2025 / Accepted: 19 February 2025 / Published: 21 February 2025

Abstract

:
Digital finance can reduce the urban–rural income gap, but the digital divide may limit this effect. This study develops a theoretical framework to explore the interactions between digital finance, the digital usage gap, and income disparity. Using data from 274 Chinese cities, the research applies a two-way fixed-effects and threshold effect model. The results indicate that disparities in digital usage not only diminish but may also distort the convergence benefits of digital finance, producing a U-shaped relationship that exhibits variability across dimensions and regions. Additionally, traditional financial systems appear to moderate this U-shaped pattern by delaying the point at which digital finance begins to widen the urban–rural income gap. However, the extent of this alleviation is influenced by the digital usage is divisive. Once digital technology adoption exceeds a threshold, the negative effect becomes positive, narrowing the urban–rural income gap. Consequently, policy initiatives should prioritize improving financial conditions in rural areas, accelerating the digital transformation of conventional finance, bolstering digital education in rural regions, and addressing the disparities in digital usage.

1. Introduction

Eliminating poverty, reducing inequality, and ensuring basic living standards and peace worldwide are essential conditions for sustainable development [1]. Since the early 1990s, both developed and emerging economies have experienced a notable increase in income disparities, which in recent years has become a significant global challenge to sustainable development [2]. In response, the United Nations Agenda of 2030 on Sustainable Development Goals has emphasized the need to eradicate all forms of poverty and reduce inequalities both within and between countries [3,4]. Governments worldwide have prioritized poverty alleviation and coordinated development efforts. In developing nations, the urban–rural income gap, driven by rural poverty, challenges sustainable development [5]. Reducing this gap hinges on improving the income of rural populations, particularly disadvantaged groups and the so-called “long-tail” population. A major challenge for countries like China is finding sustainable ways to increase rural income, improve economic conditions, and reduce the urban–rural income divide.
Financial system development is a key factor in the income gap, with traditional finance contributing to uneven growth [6]. On the one hand, the inherent “preference for the affluent” in traditional finance has restricted long-tail populations from accessing financial services effectively, exacerbating income disparities [7]. On the other hand, the reliance on conventional financial systems has further entrenched China’s dual-tiered financial structure, ultimately leading to a skewed allocation of financial resources [8]. Urban centers tend to secure a larger portion of available resources, whereas rural and less developed regions often grapple with high transaction costs, restricted financial access, and exclusion from mainstream financial services [9]. This imbalance hinders rural development and widens the urban–rural income gap. However, digital finance, enabled by information technology, overcomes the spatial limitations of traditional finance [10]. Leveraging the internet and modern technology, digital finance enhances access to financial services, reduces information asymmetry, and mitigates financial exclusion [11]. This has enabled underserved populations to access financial resources, stimulated rural economies, and addressed the dual economic structure, offering opportunities to narrow the urban–rural income gap [12]. Promoting digital finance can drive economic growth, reduce poverty, and strengthen the structure, supporting sustainable development [13].
However, the relationship between digital and traditional finance can be defined by two opposing dynamics: “competitive substitution” and “competitive cooperation”. The competitive substitution view suggests that digital finance bridges the gaps left by traditional finance—characterized by high transaction costs and limited convenience—thereby challenging conventional banking systems and gradually eroding their market share [14]. Conversely, the competitive cooperation perspective suggests that traditional finance can provide a supportive environment for digital finance development [15]. Digital finance, in turn, acts as a catalyst for reform and spurs innovation within conventional financial institutions. This transformation enhances both the accessibility and affordability of financial services, especially for groups that were historically underserved [16]. Together, they complement each other, optimizing financial services and fostering economic growth [17].
Nonetheless, the digital divide, a critical issue associated with digital finance, cannot be ignored. As digital finance develops, the negative impact of the urban–rural digital divide grows [18]. In China, the vast wealth disparity, diverse resource endowments, and underdeveloped rural financial infrastructure, coupled with limited human resources and low financial literacy, place rural areas at a disadvantage. The technical demands of innovations like AI and big data may worsen digital exclusion in rural areas, deepening the urban–rural digital gap [19]. The digital divide is commonly broken down into three tiers: the first level (access and coverage), the second level (skills and usage), and the third level (outcomes). For example, according to the China Statistical Yearbook, as of 2023, China had 1.727 billion mobile phone users, with rural residents owning an average of 271.2 mobile phones per 100 households. Although mobile phones are now commonplace in rural regions, addressing the first-level divide, the gap in terms of digital skills and usage remains a major obstacle. Differences in digital technology usage lead to disparities in digital finance outcomes, excluding regions with lower educational and financial literacy levels from digital dividends. This creates a “Matthew effect”, where wealthy regions grow richer while poorer areas lag, distorting digital finance’s benefits and sometimes widening the urban–rural income gap.
Scholars remain divided on whether digital finance narrows the urban–rural income gap, with some citing its inclusivity and technological advantages as key factors [20,21]. Digital finance can reach long-tail populations [15], lower the barriers to financial services, alleviate information asymmetry, stimulate rural entrepreneurship [22], promote household consumption [23], and reduce poverty rates [24]. These factors contribute to significant effects on income growth and distribution [25]. Digital finance benefits rural residents more than urban populations, narrowing the urban–rural income gap and reducing wealth inequality [26]. Some studies suggest that digital finance may increase the urban–rural income gap [27,28,29]. For example, Prieger [30] argues that while digital finance boosts socioeconomic participation through improved network coverage, its benefits are skewed toward urban residents, creating more employment opportunities for them and failing to achieve equitable outcomes. Instead, it amplifies the income gap between urban and rural populations. Others argue that digital finance amplifies the income gap, with some scholars viewing its effects as dual—positive “digital dividends” and negative “digital divides”—resulting in an inverted-U or U-shaped relationship with the urban–rural income gap [4,31,32].
In summary, previous research has mainly focused on digital finance’s impacts, giving less attention to how the digital divide affects the urban–rural income gap. However, the digital divide is likely a crucial factor in explaining the non-linear and regionally diverse outcomes, as well as a key element in understanding the variations in regional mechanisms.
This study makes three key contributions to the literature. First, this study uses China as a case to explore how digital finance influences the urban–rural income gap, emphasizing the threshold effect of the digital usage divide. The findings show that as digital adoption increases and the divide narrows, digital finance becomes more effective in reducing income inequality, providing key insights for developing regions. Second, while much research has explored the overall effects of digital finance, few studies consider its interaction with traditional finance and its combined impact on the urban–rural income gap. This study fills that gap by developing a theoretical framework based on the Barro and Cobb–Douglas models [33], clarifying the mechanisms linking digital finance, traditional finance, and the digital usage divide. Finally, while previous studies have broadly discussed the digital divide, research on the threshold effect of the digital usage divide remains scarce. This study specifically examines how these threshold effects influence the relationship between digital and traditional finance, uncovering key barriers that prevent digital finance from effectively narrowing urban–rural income disparities. These findings provide fresh perspectives for advancing digital finance in China and other developing economies.

2. Theoretical Model and Research Assumptions

Digital finance is a modern financial system that uses technologies like the internet, big data, and AI to innovate financial services and processes. It enhances the efficiency and quality of financial services, expands their scope and accessibility, and better meets diverse financial needs. Digital finance is characterized by convenience, efficiency, and inclusivity [10]. It primarily functions through digital tools, which enable various financial services. Common digital financial tools include online banking and mobile banking apps, which facilitate digital payments and digital lending. Additionally, digital currency wallets serve as another type of digital financial tool, allowing for the storage and use of digital currencies.
As digital finance advances rapidly, internet-enabled technologies are increasingly seen as a fifth production factor alongside capital, labor, land, and technology. Building on Barro’s framework and the Cobb–Douglas function, this study integrates digital technology as capital into a streamlined two-sector model to examine the micro-level channels through which digital finance impacts the urban–rural income gap.

2.1. Assumptions

(1) Assume that the economy consists of two sectors: the more developed urban sector (u) and the relatively underdeveloped rural sector (r). Due to institutional barriers, the two sectors are segmented and operate independently.
(2) Assume the economy consists of competitive firms with constant-returns Cobb–Douglas production, yielding only normal profits.
Given the critical role of finance in providing the capital required for social production, this study defines the capital input in the production function as financial capital (F), specifically traditional financial capital. Additionally, recognizing the growing importance of accessing and utilizing digital technologies for business operations, digital technology capital (I) is incorporated into the production function. The production function for the representative firm in the urban sector is defined as Q u = A u F u α I u β L u 1 α β , and the production function for the representative firm in the rural sector is defined as Q r = A r F r γ I r λ L r 1 γ λ .
Here, Q, A, F, I, and L represent total output, production technology, traditional financial capital, digital technology capital, and labor capital, respectively. The subscripts u and r denote the urban and rural sectors. The parameters α and β are the output elasticities of traditional financial capital and digital technology capital for the representative firm in the urban sector, while γ and λ represent the same for the rural sector. These parameters satisfy the following conditions: 1 >   α ,   β ,   γ   , λ   > 0, and 1 >   α + β > 0, 1 >   γ + λ   > 0. The production technology (A) is treated as an exogenous variable.
(3) Assume that the total labor force (L) is fixed, such that L = L u + L r , where μ is defined as the proportion of urban labor to total labor, calculated as μ = L u / ( L u + L r ) .
The financial inequality rate (θ) measures the share of rural traditional financial capital in total traditional financial capital, reflecting the inequality in financial development between urban and rural areas as follows: θ = F r / ( F r + F u ) . In practice, traditional financial resources are often disproportionately allocated to urban sectors, making θ < 0.5. A smaller θ indicates that the rural sector receives less financial capital, signifying greater financial inequality.
The digital divide (η) is defined as the share of rural digital technology capital in total digital technology capital, representing the imbalance in digital technology development between urban and rural areas, a follows: η = I r / ( I r + I u ) . A smaller η indicates that the rural sector has less access to digital technology capital, reflecting a lower level of digital technology utilization and a wider urban–rural digital divide. Typically, η < 0.5.
The level of financial development is measured as the ratio of total traditional financial capital to total social output, calculated as φ = F / Q .

2.2. Model Derivation

2.2.1. The Impact of Digital Finance on the Urban–Rural Income Gap

By defining q = Q / L , f = F / L , i = I / L , the intensive forms of the production functions for the two sectors are expressed as
q u = A u f u α i u β
q r = A r f r γ i r λ
In a perfectly competitive market, firms make production decisions to maximize profit. The prices of financial and labor capital are equal to their marginal products. The price of labor capital is the wage w, which is defined as the marginal product of labor: w = Q L . Since Q = q L , the actual wage in equilibrium can be expressed as
w = Q L = ( q L ) L = q + L ( q L ) = q + L q , ( F / L ) L = q q , f
The actual wages for the urban and rural sectors can be further derived as
w u = q u q u , f u = q u α f u A u f u α 1 i u β = q u α q u = ( 1 α ) q u
w r = q r q r , f r = q r γ f r A r f r γ 1 i r λ = q r γ q r = ( 1 γ ) q r
The urban–rural income gap (π) is measured as the ratio of urban wages to rural wages, reflecting the earnings disparity between urban and rural residents [34].
π = w u / w r = ( 1 α ) q u / ( 1 γ ) q r
When financial development is still emerging and overall capital is scarce, the distinct urban–rural structure tends to channel financial resources predominantly toward urban areas, leaving rural regions with limited access to capital. Concurrently, rural residents lack sufficient capital accumulation to meet the high thresholds required by traditional financial services. The scarcity of financial services in rural regions further deepens financial disparities and widens the income gap. Financial development spurs growth, extends services to disadvantaged communities, and helps reduce urban–rural income disparity.
The financial development level is expressed as
φ = F Q = F r + F u Q r + Q u = F u ( 1 θ ) Q u 1 + Q r / Q u = F u 1 θ Q u 1 α 1 μ 1 γ π μ + 1
Terms yields are reorganized as follows:
F u / Q u = φ 1 θ Q u 1 + 1 α 1 μ μ π 1 γ
Define the implicit function as follows:
F 1 = φ 1 θ Q u 1 + 1 α 1 μ μ π 1 γ F u / Q u
To determine the effect of financial development on the urban–rural income gap, take the partial derivative of π with respect to φ , as follows:
π φ = F 1 / φ F 1 / π = 1 θ Q u 1 + 1 α 1 μ μ π 1 γ 1 θ Q u 1 + 1 α 1 μ μ π 2 1 γ
If π φ > 0, it implies that financial development exacerbates the urban–rural income gap under the condition of uneven financial development between urban and rural sectors.
Similarly, the partial derivative of the urban–rural income gap π with respect to financial inequality θ is calculated as
π θ = F 1 / θ F 1 / π = φ Q u 1 + 1 α 1 μ μ π 1 γ 1 θ Q u 1 + 1 α 1 μ μ π 2 1 γ
If π θ < 0, it indicates that improving financial development in the rural sector can help reduce the urban–rural income gap.
Digital finance overcomes the limitations of traditional systems and fuels financial growth while reshaping the unequal allocation of capital between urban and rural areas. Assume digital finance is represented as D = D u + D r , where the primary target of digital finance services is financially excluded populations. Thus, the financial services provided to rural areas exceed those to urban areas, meaning D u < D r .
Under these circumstances, financial development is expressed as φ 1 = F + D / Q , and financial inequality is expressed as θ 1 = F r + D r / ( D + F r + F u ). It follows that φ / D > 0 and θ / D > 0 .
Adding Equations (10) and (11), we obtain
π D = π φ φ D + π θ θ D = φ D 1 θ θ D φ 1 + 1 α 1 μ μ π 1 γ φ 1 θ 1 + 1 α 1 μ μ π 2 1 γ
Digital finance promotes growth and narrows the urban–rural financial gap, enhancing financial development ( φ increases) and reducing inequality ( θ increases).
When digital finance primarily promotes financial development at a certain stage, θ / D < φ / D , which may result in π D < 0. Under some circumstances, the advancement of digital finance might actually increase the income disparity between urban and rural regions [35].
Conversely, when the inclusive nature of digital finance dominates, primarily alleviating financial exclusion, θ / D > φ / D , which may result in π D < 0. Under specific conditions, digital finance may help bridge the income divide between urban and rural areas.
Based on this, the following hypothesis is proposed:
H1. 
Digital finance has a nonlinear impact on the urban–rural income gap.

2.2.2. The Impact of Interaction with Traditional Finance on the Urban–Rural Income Gap

Digital finance employs advanced digital tools to streamline service operations and reduce disparities in information availability [36]. On the one hand, it reduces transaction costs; on the other hand, it enhances accessibility. Additionally, digital finance targets populations severely affected by financial exclusion, particularly rural residents, making financial services more focused and directing funds toward rural areas, alleviating the imbalance in urban–rural financial services [37].
However, the development of digital technologies is often uneven, with urban areas typically advancing first while remote rural areas remain excluded from the digital world. Economically developed regions enjoy an advantage in digital technology adoption, which in turn influences digital finance development. Uneven digital access limits digital finance’s reach in rural areas, reducing its impact on the urban–rural income gap [16,17].
By assume that the capacity to access digital finance is represented as V = F / I = f L / i L = f / i and by combining Equations (1) and (2), the intensive forms of production functions become the following:
q u = A u V u α i u α + β ,   q r = A r V r γ i r γ + λ
Substituting into Equations (4) and (5), we obtain
w u = 1 α A u V u α i u α + β , w r = 1 γ V r γ i r γ + λ
Taking derivatives separately, we obtain
w u V u = α 1 α A u V u α 1 i u α + β , w r V r = γ 1 γ A u V r γ 1 i r γ + λ
If w u V u > 0 , w r V r > 0 , this indicates that improving the ability to obtain financial capital from digital technology resources increases income in both the urban and rural sectors.
The evolution of digital finance is closely linked to traditional finance. Conventional institutions excel in risk management and credit evaluation, while digital finance leverages technology and data to enhance inclusion. This partnership can counteract the bias of traditional finance toward wealthier clients, which often deepens urban–rural disparities. By collaborating, both sectors can pool their strengths and resources to advance the financial industry overall.
Meanwhile, advancements in traditional finance lay the groundwork for digital finance. As urbanization spreads, conventional banks have expanded nationwide—especially in rural areas—improving financial service reach. This favorable environment enables both urban and rural residents to easily access digital financial benefits. Compared to urban areas, rural regions hold significant financial potential, and traditional finance has a stronger impact on these areas.
When digital finance reaches a certain level and urban and rural digital technology capital ( I u and I r ) remain constant, the improvement in traditional financial development allows rural residents to obtain more financial services from digital technology. This increases wages in rural areas more significantly than in urban areas. Specifically, compared to w u V u , there will be a larger increase in w r V r , such that w r V r >   w u V u . Consequently, rural incomes rise, strengthening digital finance’s role in narrowing the urban–rural income gap. Thus, the following hypothesis is proposed:
H2. 
The level of traditional financial development influences the nonlinear relationship between digital finance and the urban–rural income gap.

2.2.3. The Threshold Effect of the Digital Usage Divide on the Interaction Between Digital Finance and Traditional Finance

Digital finance’s rapid expansion brings with it a significant digital divide that must not be ignored. The digital usage divide refers to disparities among different groups in terms of their ability, opportunities, and proficiency in using digital technologies. These disparities lead to uneven access to digital financial products, including online credit, digital insurance, and electronic payment systems. Consequently, individuals with strong digital skills can more easily and effectively participate in economic activities, while disadvantaged groups may become further marginalized in the economic landscape [19]. While access to digital technology is now relatively equal between urban and rural areas, a significant gap remains in digital literacy and usage. This second-level divide limits rural residents’ ability to fully benefit from digital financial services.
On the one hand, when the digital usage divide is significant, w r V r <   w u V u . Under these circumstances, urban areas absorb more financial capital, deepening financial inequality and weakening digital finance’s role in narrowing the urban–rural income gap.
The digital usage divide can undermine the positive effects of traditional finance, leading to digital exclusion, especially in rural areas with weaker financial infrastructure. Lower education levels, insufficient digital skills training, and limited technology adoption further widen this gap. Even in well-developed financial regions, marginalized groups like farmers often face restricted access to financial services and digital resources [38].
As a result, the long-tail and impoverished populations face low efficiency in utilizing digital finance. This causes digital resources to flow disproportionately toward advantaged groups targeted by traditional finance, undermining the original intention of digital finance to serve disadvantaged groups. Consequently, rural populations remain excluded from financial services, financial inequality intensifies, and w r V r <   w u V u .
Additionally, the benefits of internet and AI advancements vary greatly between urban and rural areas. In such cases, traditional finance may hinder digital finance’s effectiveness in reducing the urban–rural income gap [39]. Therefore, the following hypothesis is proposed:
H3. 
The current digital usage divide is a key factor influencing the directional mechanism of traditional finance. Specifically, after surpassing the second threshold of digital technology usage, the negative suppressive effect of traditional finance will reverse into a positive synergistic effect.
In conclusion, the conceptual framework of this study is shown in Figure 1.

3. Study Design

3.1. Model Construction

Building on the theoretical analysis of digital finance’s impact on the urban–rural income gap, this study constructs a model from three key perspectives: baseline effect, interaction effect, and threshold effect.

3.1.1. Baseline Effect Model

To analyze the impact of digital finance on the urban–rural income gap and test H1, this study employs a two-way fixed-effects regression model, defined as follows [4]:
t h e i l i , t = β 0 + β 1 d i f i , t + β 2 d i f i , t 2 + ω X i , t + μ i + δ t + ε i , t
In model (16), the Theil index ( t h e i l i , t ) measures the urban–rural income gap, while the digital finance index ( d i f i , t ) and its squared term ( d i f i , t 2 ) capture the potential nonlinear effects of digital finance. Control variables ( X i , t ) include factors such as openness, education expenditure, government intervention, human capital, and consumption levels. The model also incorporates regional fixed effects ( μ i ), time fixed effects ( δ t ), and an error term ( ε ). Here, t represents time and i denotes provinces. In the empirical analysis, all continuous variables are winsorized at the 1% level to address outliers, and robust standard errors are applied for statistical testing.

3.1.2. Interaction Effect Model

To test H2, interaction terms for traditional finance are added to the fixed-effects model, yielding the following equation:
t h e i l i , t = β 0 + β 1 d i f i , t + β 2 d i f i , t 2 + β 3 f d i , t d i f i , t + β 4 f d i , t d i f i , t 2 + β 5 Z i , t + ω X i , t + μ i + δ t + ε i , t
In this model, f d i , t represents traditional finance, while other variables remain consistent with the previous model. To reduce the correlation among interaction terms and mitigate multicollinearity, both digital finance and traditional finance are mean-centered.
The evaluation focuses on three aspects. First, if β 4 is significant, it confirms the validity of f d i , t as a mechanism variable. Second, for a U-shaped relationship, a significantly negative β 4 indicates that the curve flattens as f d i , t increases, while a significantly positive β 4 suggests that the curve steepens. In contrast, the effects are reversed for an inverted U-shaped relationship. Third, to assess the impact on the curve’s inflection point, the inflection point is calculated as d i f i , t = β 1 + Z β 3 2 β 2 + 2 Z β 4 . Third, to assess the impact on the curve’s inflection point, the inflection point is calculated as l n d i f Z = β 1 β 4 β 2 β 3 2 ( β 2 + β 4 Z ) 2 . Third, to assess the impact on the curve’s inflection point, the inflection point is calculated as β 1 β 4 β 2 β 3 : if positive, the inflection point shifts left; otherwise, it shifts right.

3.1.3. Threshold Effect Model

Digital and traditional finance may either compete, substituting for each other, or collaborate to enhance financial inclusion. Additionally, the digital usage divide significantly influences the convergence of urban–rural income disparities. To examine its phased impact on the interaction between traditional and digital finance, this study tests H3 by incorporating it as a threshold variable. Following the Hansen panel threshold regression method [40], the study constructs the following model:
g t f y i , t = β 0 + β 1 Z i , t d i f i , t × d ( M i , t γ 1 ) + β 2 Z i , t d i f i , t × d ( γ 1 < M i , t γ 2 ) + β 3 Z i , t d i f i , t × d ( M i , t > γ 2 ) + ω X i , t + μ i + ε i , t
In this equation, M i , t represents the threshold variable, which is digital technology usage. d denotes the indicator function, γ refers to the estimated threshold value, and X i t includes the same control variables as in the previous models. To reduce multicollinearity, both digital finance and traditional finance have been mean-centered.

3.2. Selection of Variable Indications and Descriptive Statistics

3.2.1. Core Explanatory Variable and Measurement: Digital Finance Index ( d i f )

The digital finance index ( d i f ) is based on the Inclusive Digital Finance Index compiled by the Digital Finance Research Center of Peking University [41], covering 274 prefecture-level cities in China from 2011 to 2022. To eliminate dimensional effects, d i f is divided by 100.
The description of the level of digital finance in China for 2011 and 2022 is shown in Figure 2. The legend represents the average digital finance level of each province, calculated based on selected cities and mapped using ArcGIS software (version 10.8). As only Urumqi, the capital city, was selected for Xinjiang, and Xining for Qinghai, these two provinces lack horizontal comparability. Apart from these two provinces, data from 2011 show that digital finance was more developed in the eastern coastal regions than in inland areas. This trend is largely driven by the rapid advancement and widespread adoption of information technology in coastal regions. However, overall, the level of development was still low, with indices ranging from 0.614 to 0.812.
To compare digital and traditional finance across regions, we calculated average index values for 2022. The results show that digital finance is most developed in the eastern region (3.044), followed by the central (2.872) and western (2.775) regions, which have similar levels. In terms of coverage breadth, the eastern region leads (3.251), while the central (3.068) and western (2.992) regions follow closely. The average usage depth index is 2.610, 2.405, and 2.271, while the average digitalization level is 3.150, 3.077, and 2.973. These values exhibit a decreasing trend from east to west. Among these dimensions, coverage breadth contributes the most to digital finance development, whereas usage depth remains the lowest overall. Therefore, future efforts to enhance digital finance should focus on increasing usage depth. Similarly, traditional finance development follows a comparable pattern, with average levels at 3.529 in the east, 2.904 in the central region, and 2.968 in the west.
By 2022, apart from the eastern coastal regions, provinces such as Anhui, Shandong, Henan, Hubei, and Chongqing also exhibited relatively high levels of digital finance development, with indices ranging from 2.861 to 3.064. Other regions also showed significant improvements compared to 2011. Beijing, Shanghai, Jiangsu, and Zhejiang saw the largest increases, while western regions such as Yunnan and Gansu experienced relatively smaller improvements. This suggests that while digital finance has expanded rapidly across all regions in recent years, disparities in development levels persist.

3.2.2. Explained Variable and Measurement: Urban–Rural Income Gap

Building on the work of Yuan et al. [42], the urban–rural income gap, this study uses the Theil index to quantify income disparity between urban and rural areas. The calculation model is as follows:
t h e i l i , t = i 1 2 y i t y t × ln y i t y t x i t x t
where i =  1 represents urban areas and i =  2 represents rural areas, y denotes disposable income, and x represents the total population. The Theil index is always greater than or equal to 0, with smaller values indicating lower inequality. When the index equals 0, it signifies no disparity.
Figure 3 depicts the evolution of China’s urban–rural income gap from 2011 to 2022. In 2011, the Theil index ranged between 0.004 and 0.181, with the lowest value observed in Urumqi, Xinjiang, and the highest in Yunnan Province. The urban–rural income gap showed a gradual increase from east to west, with particularly large gaps observed in western regions such as Gansu, Guizhou, Yunnan, and the Guangxi Zhuang Autonomous Region. By 2022, the Theil index ranged from 0.016 to 0.109, with the lowest value in Heilongjiang Province and the highest in Gansu Province. Apart from Urumqi in Xinjiang and Beijing, where the urban–rural income gap expanded, all other provinces experienced a decline in disparity. A comparison of the 2011 and 2022 data reveals a trend of narrowing urban–rural income disparities across China. While lower Theil index values increased and higher values decreased, the overall urban–rural income gap showed signs of further contraction. This suggests that alongside China’s economic growth, rising urban economies have also contributed to improving rural income levels. The increase in lower values may indicate that urban economic growth outpaced that of rural areas, leading to a relatively balanced state of urban–rural income disparity.
When compared with digital finance development levels, the eastern coastal regions, which lead in digital finance development, continued to maintain relatively low urban–rural income gaps, though the degree of improvement was modest. In contrast, regions with faster digital finance development, such as Anhui, Henan, and Hubei, experienced relatively greater reductions in the urban–rural income gap. Additionally, while western regions like Yunnan and Guizhou still had relatively large urban–rural income gaps, they also achieved significant improvements. Overall, regions with more advanced digital finance development tended to have smaller urban–rural income gaps. While digital finance may help reduce the urban–rural income gap, its impact varies by region, suggesting that regional and developmental factors play a key role. This theoretical hypothesis requires further empirical testing.

3.2.3. Threshold Variables and Measurement: Digital Technology Usage ( d g )

The fundamental prerequisite for using digital technology is digital access. On this basis, whether individuals choose to connect to and use the internet reflects their willingness and ability to utilize digital technology, highlighting differences in usage capabilities. Additionally, regional disparities in resource endowments and communication infrastructure can lead to variations in digital access. To ensure a more accurate measurement of digital technology usage, it is essential to account for differences in digital infrastructure rather than relying solely on internet usage rates. Therefore, digital technology usage ( d g ) is quantified as the ratio of “households with internet access to the number of mobile phone users”. A higher value reflects a narrower digital usage divide, whereas a lower value indicates a wider gap. The spatial distribution of digital technology usage in China is shown in Figure 4.
In 2011, digital technology usage was more advanced in the eastern coastal regions compared to inland areas. Western regions like Gansu, Guizhou, Qinghai, and the Guangxi Zhuang Autonomous Region had lower levels of usage, highlighting a larger digital divide. By 2022, the digital technology usage level in Urumqi, Xinjiang, decreased from 0.714 in 2011 to 0.383, reflecting an overall widening of the digital usage divide. This may explain why the region’s digital finance level improved while its urban–rural income gap widened. Conversely, in provinces leading in digital finance development, such as Shanghai, Beijing, and Guangdong, the digital usage divide remained relatively large, with limited improvement. This could be a factor in the slower pace of urban–rural income gap reduction in these regions despite advancements in digital finance.
Regions with faster digital finance development, such as Anhui, Henan, and Hubei, saw significant improvements in digital technology usage, indicating a narrowing of the digital usage divide. Western regions like Yunnan and Guizhou still had relatively low levels of digital technology usage and larger divides. However, as the divide narrowed, these areas saw improved convergence in reducing the urban–rural income disparity.
Overall, regions with advanced digital finance often show wider digital usage gaps, leading to slower reductions—or even increases—in urban–rural income disparities. This suggests that the digital divide may hinder digital finance’s effectiveness in closing the income gap, though this hypothesis needs further empirical validation.

3.2.4. Mechanism Variable

The level of traditional financial development ( f d ) is measured by the ratio of “year-end loan and deposit balances of financial institutions to GDP”.

3.2.5. Control Variables

The control variables include openness ( o p e n ), education level ( e d u ), government intervention ( g o v ), human capital ( h u m ), and consumption level ( c p i ). Openness is measured by the proportion of total trade (imports and exports) to regional GDP. Education level is assessed using the student–teacher ratio in high schools. Government intervention is indicated by local fiscal expenditure as a share of regional GDP. Human capital is reflected in the ratio of higher education enrollments to the permanent population. The consumption level is represented by total retail sales relative to regional GDP.
The specific measurement indicators for each variable are detailed in Table 1.

3.3. Sample Selection and Data Sources

This study analyzes data from 274 Chinese cities (2011–2022) using Peking University’s Digital Finance Index and additional data from national statistical sources.
All data processing and model estimations were conducted using StataSE 17(64-bit) software, which was also used to generate the relevant tables.

4. Empirical Analysis

4.1. Baseline Effect Analysis

Table 2 shows that digital finance has a significantly negative coefficient and a significantly positive squared term (both at 1%), regardless of controls.
These findings indicate a U-shaped relationship between digital finance and the urban–rural income gap. Initially, digital finance helps reduce income disparities by expanding financial access and overcoming traditional geographic and institutional barriers, lowering service costs for rural residents. However, as digital finance advances, it demands higher digital literacy and skills, exacerbating the digital usage divide. Rural populations often struggle with these requirements, and agricultural-based economies face higher costs and challenges in digital transformation, limiting their ability to benefit from digital finance. As a result, beyond a certain threshold, the urban–rural income gap starts to widen, supporting H1 on the nonlinear effects of digital finance.
Openness significantly narrows the urban–rural income gap, while education level, government intervention, and consumption level significantly widen it. Human capital shows a positive but insignificant effect, likely due to ongoing resource disparities and the digital divide.

4.2. Heterogeneity Analysis

To further explore how digital finance influences the urban–rural income gap, a regional and dimensional heterogeneity analysis was conducted, with results shown in Table 3.
Columns (1) and (2) show how digital finance affects the urban–rural income gap in regions grouped by economic development. Prefecture-level cities with per capita GDP above the mean are considered developed regions, while those below are categorized as less-developed. In both types of regions, digital finance follows a U-shaped relationship with the urban–rural income gap though the turning points differ. In developed regions, where the average digital finance index in 2022 was 3.108 —above the inflection point of 2.917—digital finance has reached a stage where it widens the income gap due to internal disparities despite high levels of coverage and usage. Conversely, in less-developed regions, the 2022 digital finance index stood at 2.786, comparable to the level of developed regions in 2020. Since this surpasses the inflection point of 2.214, most less-developed areas have also entered the stage where digital finance contributes to widening income inequality.
Columns (3)–(5) show a U-shaped relationship across regions, with the eastern region’s 2022 average index at 3.044. With strong fintech adoption, rapid growth, and deep industry integration, this remains below the inflection point of 3.150, indicating that digital finance is still helping to reduce income disparities. Meanwhile, in the central region, where the 2022 index reached 2.872—surpassing the inflection point of 2.500—digital finance has begun to widen the urban–rural income gap. At this stage, technological adoption is accelerating, mobile internet is widely accessible, and financial inclusion has significantly improved. However, regional disparities remain evident, with cities like Wuhan and Zhengzhou exhibiting higher levels of digital finance development, while significant differences persist within provinces. In the central region, many areas have already crossed to the right side of the U-shaped curve, meaning that digital finance is contributing to a growing urban–rural income gap in most places. Similarly, in the western region, the average digital finance index in 2022 was 2.775, exceeding the inflection point of 2.594. Despite government policies aimed at promoting digital finance development and continuous improvements in digital infrastructure, internal regional disparities remain significant. The level of digital finance development varies considerably, with some remote and economically underdeveloped areas still lacking adequate digital infrastructure, exacerbating the digital divide. As a result, in the majority of the western region, digital finance is leading to an increasing urban–rural income gap.
Additionally, the regional inflection points show that developed regions have a higher inflection point of 2.917 compared to 2.214 in less-developed areas. The eastern region’s inflection point of 3.150 is also higher than those of the central and western regions. This could be attributed to the stronger economic foundations, advanced urban and rural development, and well-established financial markets in the eastern and developed regions, which support better integration of digital finance and more effective financial inclusion. Furthermore, better digital infrastructure, broader network coverage, and higher overall digital literacy allow rural areas to benefit more from high-quality digital financial services. As a result, digital finance reduces the urban–rural income gap more effectively and for longer in developed regions, while central, western, and less-developed areas face challenges like weaker economies, underdeveloped infrastructure, and lower digital literacy, limiting its impact. The difficulties in improving digital skills limit the ability of digital finance to drive economic growth in these regions, causing them to enter the stage of widening urban–rural income disparities more quickly.
Columns (6)–(8) show that both coverage breadth and usage depth of digital finance follow a positive U-shaped pattern, initially narrowing the income gap before widening it past a certain threshold. On the other hand, digitalization presents a different trend: while the coefficient for digital finance is positive, its squared term is negative and insignificant. This suggests that digitalization initially increases the income gap before gradually narrowing it, though the effect is relatively weak. Specifically, the inflection point for coverage breadth is 2.955, while the mean value in 2022 was 3.112, indicating that coverage breadth currently contributes to widening the income gap. Similarly, the inflection point for usage depth is 0.938, with a 2022 mean of 2.440 also on the curve’s right-hand side, showing that usage depth increases the income gap. For digitalization, the inflection point is around 15, while the 2022 mean is 3.073, placing it on the left-hand side of the inverted U-shaped curve, implying that digitalization reduces the income gap but with weak effects due to insignificant coefficients.
This result may stem from the varying aspects of digital finance. Coverage breadth represents service availability, encouraging greater financial participation among rural residents. However, the low frequency of online financial product usage among rural residents and their lower effective demand for deeper usage compared to urban residents exacerbate the digital divide. This weakens digital finance’s ability to reduce the urban–rural income gap. Similarly, usage depth, which measures engagement with digital financial services, shows disparate impacts across regions, further widening the divide. Digitalization, which represents the convenience and cost-effectiveness of digital financial services, does offer potential benefits to long-tail populations, but its current level remains insufficient to significantly reduce the urban–rural income gap.

4.3. Robustness and Endogeneity Tests

4.3.1. Robustness Tests

The urban–rural income gap was reassessed using the urban-to-rural income ratio. Column (1) of Table 4 shows a significantly negative coefficient for digital finance’s linear term and a significantly positive coefficient for its squared term, confirming a robust U-shaped relationship.
Beijing, Shanghai, Tianjin, and Chongqing, as municipalities, have unique statuses and policy orientations. To account for this, data for these four municipalities were excluded, and the regression was re-estimated. Column (2) of Table 4 reports a digital finance coefficient of −0.073 and a squared term of 0.015, both of which are significant, confirming prior results.
In 2015, China introduced its “National Big Data Strategy”. To examine the effects under this new context, the sample period was adjusted to 2015–2022 for empirical analysis. Column (3) of Table 4 indicates that even with an adjusted time frame, digital finance’s linear term remains significantly negative, and its squared term is significantly positive, further affirming the U-shaped relationship.
Urbanization and industrial structure significantly influence the urban–rural income gap and are included as control variables. Urbanization (urb) is measured by the urbanization rate, and industrial structure upgrading (str) is represented by the ratio of the tertiary to secondary sector. Column (4) of Table 4 confirms robustness, showing a negative linear term and positive squared term for digital finance. The urbanization coefficient of −0.020, significant at the 1% level, indicates a narrowing income gap, while the industrial structure coefficient (0.002) is not significant. This may be due to disparities in China’s industrial base and the digital usage divide. Cities not only have a strong industrial foundation but also higher digital technology adoption, enabling them to rapidly transition and develop emerging industries during industrial upgrading. In contrast, rural areas have a weak industrial base and a high proportion of traditional agriculture and face high costs and difficulties in industrial upgrading, leading to slower income growth. This obstructs the narrowing of the urban–rural income gap.

4.3.2. Endogeneity Test

Although various factors influencing the urban–rural income gap have been controlled for, the regression results may still be biased due to omitted variables and measurement errors. To address this issue, the instrumental variable (IV) method is employed. Given the potential lagged effect of digital finance, its one-period lag is used as an instrumental variable for endogeneity testing. The results, shown in Column (5) of Table 4, indicate an F-statistic of 565.28, which exceeds the critical threshold of 10, confirming its validity.
Additionally, geographic factors are introduced as instruments. Using Geographic Information System (GIS) data [35], the spherical distance between each city and Hangzhou, Zhejiang Province, is calculated. Since geographic distance is a constant and unsuitable for panel data analysis, it is interacted with digital finance to construct a time-varying panel instrumental variable. The results, presented in Column (6) of Table 4, yield an F-statistic of 37.86, surpassing 10 and verifying its validity. The linear and quadratic terms of digital finance align with the baseline regression, confirming the U-shaped relationship’s robustness.
Furthermore, a historical instrumental variable is constructed. The number of fixed telephone lines per 100 people in 1984 is used as a proxy for historical regional internet infrastructure [43], interacting with the previous year’s internet user count. As shown in Column (7) of Table 4, the F-statistic is 74.51, confirming its reliability. The significantly negative linear term and positive quadratic term further confirm the U-shaped relationship between digital finance and the urban–rural income gap, even after addressing endogeneity concerns.
To address dynamic effects, a system GMM estimation was conducted, incorporating the first- and second-period lagged urban–rural income gap as explanatory variables. In Column (8) of Table 4, the linear digital finance term is significantly negative, and the quadratic term is significantly positive at the 1% level. Diagnostic tests confirm model validity: the AR(1) P-value is below 0.1, the AR(2) P-value is above 0.1, and the Hansen test P-value exceeds 0.1, verifying the instrumental variables [44].

4.4. Interaction Effect Analysis

The nonlinear interaction effect regression results are presented in Table 5.
Column (1) shows that traditional finance’s interaction with digital finance is positive (0.002), while its interaction with digital finance’s squared term is significantly negative (−0.001 **). Additionally, β 1 β 4 β 2 β 3 > 0 , indicating that traditional finance significantly mitigates the impact of digital finance on the urban–rural income gap. This flattens the U-shaped curve, reducing the marginal effect and altering digital finance’s influence on the urban–rural income gap. It lessens digital finance’s initial narrowing impact while delaying the widening effect, shifting the inflection point to the right. These findings support H2.
Columns (2) and (3) highlight regional differences. In developed regions, the interactions are not statistically significant, suggesting that traditional finance has no substantial impact on digital finance’s relationship with the urban–rural income gap. In underdeveloped regions, however, the interaction with digital finance is significantly positive (0.006 **) and its squared term is significantly negative (−0.002 **), showing a clear moderating effect. Additionally, β 1 β 4 β 2 β 3 > 0 indicates that in underdeveloped regions, traditional finance moderates the influence of digital finance. It dampens digital finance’s initial role in narrowing the urban–rural income gap and lessens its later effect in widening it. This adjustment shifts the U-shaped curve’s inflection point further right, postponing the stage when digital finance starts to expand the gap. As a result, more cities remain in the phase where digital finance continues to reduce the urban–rural income disparity.
Columns (4)–(6) examine the dimensional heterogeneity in how traditional finance influences the impact of digital finance. The findings reveal that the interaction term with coverage breadth is positive (0.001), while its interaction with the squared term is significantly negative (−0.000 *). Furthermore, the condition β 1 β 4 β 2 β 3 > 0 suggests that traditional finance significantly moderates the effect of coverage breadth. This makes the positive U-shaped curve flatter, reduces marginal effects, weakens the initial positive impact of coverage breadth in reducing the urban–rural income gap, and obstructs its later effect in widening the gap. It also shifts the inflection point to the right. For usage depth and its squared term, as well as digitalization and its interaction with traditional finance, the coefficients are not significant. This suggests that traditional finance primarily exerts its influence through the coverage breadth of digital finance.

4.5. Threshold Test Based on the Digital Usage Divide

4.5.1. The Impact of the Digital Usage Divide on Baseline Effects

While digital finance reduces the urban–rural income gap by providing digital benefits; it also widens the digital usage divide, limiting income gains. This study explores how digital usage thresholds affect digital finance’s baseline impacts and its sub-dimensions.
Using the Bootstrap method with 300 resampling iterations, F-statistic values and corresponding digital technology usage thresholds were calculated for digital finance and its components, as shown in Table 6.
The results show a single threshold of 0.113 for digital finance, with no significance for double or triple thresholds. For sub-dimensions, both single and double thresholds are significant, with no triple threshold. Specifically, the threshold values for coverage breadth are 0.109 and 0.343, for usage depth are 0.113 and 0.293, and for digitalization level are 0.109 and 0.166. The use of digital finance requires internet connectivity, and rural residents’ willingness and ability to use digital technology determine whether they can effectively leverage digital finance. This confirms the threshold effect of the digital usage divide on digital finance’s ability to narrow the urban–rural income gap.
Table 7 indicates that below the 0.113 threshold, digital finance’s coefficient is significantly negative (−0.012 ***), suggesting it helps reduce the urban–rural income gap. Beyond this threshold, the coefficient becomes even more negative (−0.018 ***), suggesting that higher digital adoption further enhances digital finance’s gap-reducing effect.
The underlying reason is that as residents improve their ability and willingness to use digital technology, the digital divide gradually narrows. This reduces the self-exclusion of rural residents and other underserved groups from internet finance, allowing the demand for digital technology to emerge. Consequently, the cost-reduction and efficiency-enhancing effects of digital finance become more evident, stimulating economic vitality in less developed rural areas. This strengthens digital finance’s role in reducing the urban–rural income gap.
Specifically, as digital technology usage surpasses the thresholds of 0.109 and 0.343, the convergence effect of coverage breadth first intensifies and then diminishes. As residents’ digital literacy and adoption improve, the rising demand for digital services aligns with the expanded coverage of digital finance, enhancing its effectiveness in narrowing the urban–rural income gap. Similarly, when digital technology usage exceeds thresholds of 0.113 and 0.293, the convergence effect of usage depth strengthens. Likewise, surpassing thresholds of 0.109 and 0.166 enhances the impact of digitization on reducing the income gap.
This highlights the digital divide, where rural residents have lower digital proficiency than urban counterparts. Despite improvements in digital finance, regional disparities persist, limiting its impact on narrowing the urban–rural income gap. However, as digital adoption surpasses key thresholds and the divide narrows, rural economies gain momentum, unlocking growth potential in less developed areas. This, in turn, strengthens the role of digital finance—particularly in usage depth and digitization—in bridging the income gap.

4.5.2. The Impact of the Digital Usage Divide on the Interaction Between Digital Finance and Traditional Finance

The digital usage divide not only directly affects the relationship between digital finance and the urban–rural income gap but also influences it indirectly through the role of traditional finance.
As shown in Table 8, digital technology usage exhibits a threshold effect on digital finance’s ability to narrow the urban–rural income gap [45]. The presence of a significant digital usage threshold is confirmed. Specifically, when digital finance, coverage breadth, usage depth, and digitization level contribute to income convergence, a double threshold effect is evident, while no triple threshold is observed, reinforcing the existence of a significant dual digital threshold across all cases.
As shown in Table 9, when digital technology usage remains below the 0.227 threshold, the interaction terms between digital finance, coverage breadth, usage depth, and traditional finance all have positive coefficients. However, once the threshold of 0.227 is crossed, the signs of these interaction term coefficients reverse and become significantly negative (−0.008 ***, −0.006 ***, −0.008 ***, −0.008 ***). At this stage, traditional finance strengthens digital finance’s convergence effect, altering its impact direction.
When digital technology usage is below the second threshold, a wider digital divide causes traditional finance to hinder digital finance’s role in narrowing the urban–rural income gap. However, as digital adoption improves, this negative impact weakens, eventually turning positive and reinforcing digital finance’s convergence effect. These results validate H3.
The key reason is that a significant digital usage divide—marked by a wide gap in digital adoption between urban and rural areas—leads to a substitution effect between digital and traditional finance. This effect is more pronounced in regions with well-developed traditional financial systems. At this point, digital finance inherits the “preference for the affluent” characteristic of traditional finance, mainly serving the advantaged groups that traditional finance has already benefited. Combined with the high efficiency of digital technology, the marginal benefits to advantaged groups are further amplified. Conversely, digital finance fails to provide substantial financial assistance to disadvantaged groups. Even if digital finance reaches rural and less developed areas, the generally low digital technology usage capabilities of long-tail groups result in severe digital exclusion among disadvantaged groups, with no improvement in their financing environment. Thus, traditional finance weakens digital finance’s ability to narrow the urban–rural income gap.
However, as digital adoption exceeds the threshold, the technology usage gap between urban and rural areas, as well as across groups, diminishes. At this stage, traditional finance and digital finance primarily exhibit a synergistic and mutually reinforcing effect. Traditional finance then supports digital finance in realizing its inclusive potential, stimulating rural residents’ financial enthusiasm, unlocking the economic development potential of rural and less developed areas, and improving the urban–rural income gap.

5. Discussion

5.1. Key Findings

Digital finance’s impact on China’s urban–rural income gap is U-shaped. Initially, its expansion reduces physical barriers, facilitates resource flow, and boosts rural economic growth, narrowing the income gap through the digital dividend. However, in later stages, the growing digital usage divide exacerbates regional disparities, reversing this effect. Currently, most regions in China fall on the right side of the U-curve, where digital finance contributes to widening the income gap, consistent with previous studies [4]. Among its components, coverage breadth and usage depth tend to increase the gap, while digitization has a minimal but slightly mitigating effect.
Mechanism analysis shows that traditional finance significantly influences digital finance’s convergence effect, pushing the U-curve’s turning point further right and partially weakening digital finance’s role in reducing income disparities. This effect is particularly evident in underdeveloped areas, with the strongest impact observed in coverage breadth. Strengthening traditional finance, especially in rural regions, could help counteract digital finance’s negative effects on income inequality.
The threshold effect further underscores the role of the digital usage divide in shaping both the traditional finance mechanism and digital finance’s impact on income convergence. When digital adoption surpasses a critical threshold, the previously restrictive influence of traditional finance transitions into a supportive role, allowing both financial systems to work together in narrowing the income gap. These findings highlight how digital barriers in rural areas hinder urban–rural integration, suppress digital finance’s convergence effect, and distort traditional finance’s role, ultimately exacerbating income disparities in later stages. This reinforces the study’s conclusions.

5.2. Implications

This study provides key insights for developing countries seeking to narrow the urban–rural income gap through digital finance.
To mitigate the challenges of the digital usage divide and its effect on income disparities, efforts should prioritize enhancing the rural financial environment and accelerating digital transformation. While China has made significant strides in promoting digital finance to support urban–rural integration, practical evidence suggests that while the basic digital access gap has largely been bridged, the digital usage divide remains a growing concern. The degree of digitization, particularly in terms of convenience and affordability, has not significantly reduced the urban–rural income gap, which continues to widen. Therefore, the current focus should remain on rural areas. Developing countries can draw lessons from China by enhancing rural digital infrastructure and fostering a more supportive financial environment. These efforts would leverage the inclusive potential of digital finance, encouraging rural residents to embrace online financial products and converting potential disadvantages into economic benefits.
Additionally, strengthening the integration of traditional and digital finance while optimizing the financial system is crucial. Digital finance builds upon the foundation of traditional finance, and under appropriate conditions, these two systems can exhibit cooperative effects. Policymakers should strengthen the role of traditional financial services by upgrading physical branches with intelligent and accessible designs. Concurrently, the financial system should be improved, with regulatory frameworks standardized to ensure orderly and healthy competition between traditional and digital finance. These efforts will better integrate traditional and digital finance, amplifying their impact on the urban–rural income gap.
Bridging the digital usage divide requires a strong focus on rural digital education. This divide can distort digital finance’s role in income convergence, making it essential for governments in developing countries to address its risks. The gap primarily stems from disparities in financial literacy and the ability to adopt and utilize digital technologies between urban and rural populations. Prioritizing policies to close this divide is crucial. Government initiatives should intensify education campaigns to enhance digital knowledge and consumer awareness among rural populations. Improving financial literacy, reducing urban–rural cognitive disparities, and empowering disadvantaged groups such as farmers with the skills to navigate digital platforms would not only raise their income levels but also foster a more balanced and inclusive financial ecosystem.

6. Limitations

(1) This study examines digital finance’s effect on China’s urban–rural income gap. However, it does not extend its analysis to developed nations or other developing economies, which may limit the generalizability of its conclusions. Nonetheless, as the world’s largest developing country and the second-largest digital economy, China serves as a representative case, offering valuable lessons for understanding the role of digital finance in addressing income disparities. Future research could expand the sample and conduct cross-country comparative studies;
(2) This study examines the U-shaped relationship between digital finance and the urban–rural income gap but does not include a structural decomposition analysis. Such an approach could offer deeper insights into the underlying factors and characteristics of the income gap, providing a stronger foundation for understanding the influence of digital finance. However, conducting such an analysis requires detailed data on different types of household income sources and their contributions, which poses challenges in data availability. Therefore, this study does not delve into this aspect. Future research could further explore structural decomposition analysis and examine changes in income structure before and after the digital finance inflection point.
Additionally, analyzing digital finance’s impact at the micro level, including household income and business development, is essential. A more granular analysis could provide deeper insights into how digital finance influences individual economic well-being and entrepreneurial growth. Future research could use household and firm-level data to explore digital finance’s impact on income and corporate financing constraints.

Author Contributions

Y.X. (Yanfei Xiao) and H.W. conceptualized the study. M.Y. analyzed the data. Y.X. (Yunbo Xiang) drawed the map. All authors were involved in drafting and revising the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This research is supported by the National Social Science Foundation General Project (23BJL095).

Institutional Review Board Statement

This study utilized secondary data that are publicly available and anonymized. As such, it does not involve direct interaction with human subjects, and ethical approval was not required. The research was conducted in compliance with all relevant ethical guidelines and regulations.

Informed Consent Statement

Not applicable. The study utilized secondary data that are publicly available and anonymized, and no direct interaction with human participants was involved.

Data Availability Statement

All supporting data can be found on the National Bureau of Statistics of China website (http://www.stats.gov.cn/), the Digital Finance Research Center of Peking University (https://idf.pku.edu.cn/).

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Klarin, T. The Concept of Sustainable Development: From its Beginning to the Contemporary Issues. Zagreb Int. Rev. Econ. Bus. 2018, 21, 67–94. [Google Scholar] [CrossRef]
  2. Wade, R. Global trends in income inequality. Challenge 2011, 54, 54–75. [Google Scholar] [CrossRef]
  3. Szymańska, A. Reducing Socioeconomic Inequalities in the European Union in the Context of the 2030 Agenda for Sustainable Development. Sustainability 2021, 13, 7409. [Google Scholar] [CrossRef]
  4. Jiang, Q.; Li, Y.; Si, H. Digital Economy Development and the Urban–Rural Income Gap: Intensifying or Reducing. Land 2022, 11, 1980. [Google Scholar] [CrossRef]
  5. Tu, Z.; Kong, J.; Sun, L.; Liu, B. Can the Digital Economy Reduce the Rural-Urban Income Gap? Sustainability 2024, 16, 938. [Google Scholar] [CrossRef]
  6. Shah, A. Fiscal Policies for Coordinated Urban-Rural Development and Their Relevance for China. Public Financ. Manag. 2016, 16, 51–74. [Google Scholar] [CrossRef]
  7. Conroy, J. APEC and financial exclusion: Missed opportunities for collective action? Dev. J. 2006, 12, 53–79. [Google Scholar] [CrossRef]
  8. Feng, S.; Zhang, R.; Li, G. Environmental decentralization, digital finance and green technology innovation. Struct. Change Econ. Dyn. 2022, 61, 70–83. [Google Scholar] [CrossRef]
  9. Berger, A.N. The Economic Effects of Technological Progress: Evidence from the Banking Industry. J. Money Credit Bank. 2003, 35, 141–176. [Google Scholar] [CrossRef]
  10. Luo, D.; Luo, M.; Lv, J. Can Digital Finance Contribute to the Promotion of Financial Sustainability? A Financial Efficiency Perspective. Sustainability 2022, 14, 3979. [Google Scholar] [CrossRef]
  11. Ji, X.; Wang, K.; Xu, H.; Li, M. Has Digital Financial Inclusion Narrowed the Urban-Rural Income Gap: The Role of Entrepreneurship in China. Sustainability 2021, 13, 8292. [Google Scholar] [CrossRef]
  12. Gao, J.; Wu, Y.; Li, H. Digital Inclusive Finance, Rural Loan Availability, and Urban–Rural Income Gap: Evidence from China. Sustainability 2024, 16, 9763. [Google Scholar] [CrossRef]
  13. Manioudis, M.; Meramveliotakis, G. Broad strokes towards a grand theory in the analysis of sustainable development: A return to the classical political economy. New Political Econ. 2022, 27, 866–878. [Google Scholar] [CrossRef]
  14. Beck, H. Banking is essential, banks are not. The future of financial intermediation in the age of the Internet. Netnomics 2001, 3, 7–22. [Google Scholar] [CrossRef]
  15. Haddad, C.; Hornuf, L. The emergence of the global fintech market: Economic and technological determinants. Small Bus. Econ. 2019, 53, 81–105. [Google Scholar] [CrossRef]
  16. Le, S.; Congmou, Z. Impact of Digital Inclusive Finance on Rural High-Quality Development: Evidence from China. Discret. Dyn. Nat. Soc. 2022, 2022, 7939103. [Google Scholar] [CrossRef]
  17. Zhang, X.; Tan, Y.; Hu, Z.; Wang, C.; Wan, G. The Trickle-Down Effect of Fintech Development: From the Perspective of Urbanization. China World Econ. 2020, 28, 23–40. [Google Scholar] [CrossRef]
  18. Jie, Y.; Hu, S.; Zhu, S.; Weng, L. How Digitalization and Its Context Affect the Urban–Rural Income Gap: A Configurational Analysis Based on 274 Prefecture-Level Administrative Regions in China. Land 2024, 13, 2118. [Google Scholar] [CrossRef]
  19. Scheerder, A.; van Deursen, A.; van Dijk, J. Determinants of Internet skills, uses and outcomes. A systematic review of the second- and third-level digital divide. Telemat. Inform. 2017, 34, 1607–1624. [Google Scholar] [CrossRef]
  20. Faizah, C.; Yamada, K.; Pratomo, D. Information and communication technology, inequality change and regional development in Indonesia. J. Socioecon. Dev. 2021, 4, 224. [Google Scholar] [CrossRef]
  21. Pengju, L.; Yitong, Z.; Shengqi, Z. Has Digital Financial Inclusion Narrowed the Urban–Rural Income Gap? A Study of the Spatial Influence Mechanism Based on Data from China. Sustainability 2023, 15, 3548. [Google Scholar] [CrossRef]
  22. Soldatos, J.; Kefalakis, N.; Despotopoulou, A.-M.; Bodin, U.; Musumeci, A.; Scandura, A.; Aliprandi, C.; Arabsolgar, D.; Colledani, M. A digital platform for cross-sector collaborative value networks in the circular economy. Procedia Manuf. 2021, 54, 64–69. [Google Scholar] [CrossRef]
  23. Li, J.; Wu, Y.; Xiao, J.J. The impact of digital finance on household consumption: Evidence from China. Econ. Model. 2020, 86, 317–326. [Google Scholar] [CrossRef]
  24. Heping, G.; Lianzhen, T.; Xiaojun, Z.; Decai, T.; Valentina, B. Research on the Effect of Rural Inclusive Financial Ecological Environment on Rural Household Income in China. Int. J. Environ. Res. Public Health 2022, 19, 2486. [Google Scholar] [CrossRef] [PubMed]
  25. Xu, L.; Yueying, M.; Wenyu, Z. Digital inclusive financial services and rural income: Evidence from China’s major grain-producing regions. Financ. Res. Lett. 2023, 53, 103622. [Google Scholar]
  26. Parker, E.B. Closing the digital divide in rural America. Telecommun. Policy 2000, 24, 281–290. [Google Scholar] [CrossRef]
  27. Fall, F.S.; Orozco, L.; Akim, A.M. Adoption and use of mobile banking by low-income individuals in Senegal. Rev. Dev. Econ. 2020, 24, 569–588. [Google Scholar] [CrossRef]
  28. Janine, A. Mobile Money and the Economy: A Review of the Evidence. World Bank Res. Obs. 2018, 33, 135–188. [Google Scholar]
  29. Correa, T.; Pavez, I.; Contreras, J. Digital inclusion through mobile phones?: A comparison between mobile-only and computer users in internet access, skills and use. Inf. Commun. Soc. 2018, 23, 1074–1091. [Google Scholar] [CrossRef]
  30. Prieger, J. The Broadband Digital Divide and the Economic Benefits of Mobile Broadband for Rural Areas. Telecommun. Policy 2012, 37, 483–502. [Google Scholar] [CrossRef]
  31. Salazar-Cantú, J.; Jaramillo-Garza, J.; Rosa, B.Á.-D.l. Financial Inclusion and Income Inequality in Mexican Municipalities. Open J. Soc. Sci. 2015, 3, 29–43. [Google Scholar] [CrossRef]
  32. Peng, Z.; Dan, T. Digital dividend or digital divide? Digital economy and urban-rural income inequality in China. Telecommun. Policy 2023, 47, 102616. [Google Scholar] [CrossRef]
  33. Barro, R.J. Government Spending in a Simple Model of Endogeneous Growth. J. Political Econ. 1990, 98, S103–S125. [Google Scholar] [CrossRef]
  34. Dong, Y.; Luo, W.; Zhang, X. Information and communication technology diffusion and the urban–rural income gap in China. Pac. Econ. Rev. 2023, 29, 159–186. [Google Scholar] [CrossRef]
  35. Lianying, Y.; Xiaoxiao, M. Has digital finance widened the income gap? PLoS ONE 2022, 17, e0263915. [Google Scholar]
  36. Hau, H.; Huang, Y.; Lin, C.; Shan, H.; Sheng, Z.; Wei, L. FinTech Credit and Entrepreneurial Growth. J. Financ. 2024, 79, 3309–3359. [Google Scholar] [CrossRef]
  37. Xu, Q.; Zhong, M.; Dong, Y. Digital finance and rural revitalization: Empirical test and mechanism discussion. Technol. Forecast. Soc. Chang. 2024, 201, 123248. [Google Scholar] [CrossRef]
  38. Shen, Y.; Hu, W.; Hueng, C.J. The Effects of Financial Literacy, Digital Financial Product Usage and Internet Usage on Financial Inclusion in China. MATEC Web Conf. 2018, 228, 05012. [Google Scholar] [CrossRef]
  39. Hongbo, Z.; Xiao, Z.; Lin, Y. Does Digital Inclusive Finance Narrow the Urban-Rural Income Gap through Primary Distribution and Redistribution? Sustainability 2022, 14, 2120. [Google Scholar] [CrossRef]
  40. Hansen, B.E. Threshold effects in non-dynamic panels: Estimation, testing, and inference. J. Econom. 1999, 93, 345–368. [Google Scholar] [CrossRef]
  41. Li, W.; Cai, J.; Zhu, Y.; Li, J.; Li, Z. Can digital finance development drive green transformation in manufacturing? Evidence from China. Environ. Sci. Pollut. Res. Int. 2024, 31, 23876–23895. [Google Scholar] [CrossRef]
  42. Yuan, Y.; Wang, M.; Zhu, Y.; Huang, X.; Xiong, X. Urbanization’s effects on the urban-rural income gap in China: A meta-regression analysis. Land Use Policy 2020, 99, 104995. [Google Scholar] [CrossRef]
  43. Bin, L.; Jing, Z.; Aoxiang, Z. Empowering rural human Settlement:Digital Economy’s path to progress. J. Clean. Prod. 2023, 427, 139243. [Google Scholar]
  44. Renyi, Y.; Changbiao, Z.; Zisheng, Y.; Qiuju, W. Analysis on the Effect of the Targeted Poverty Alleviation Policy on Narrowing the Urban-Rural Income Gap: An Empirical Test Based on 124 Counties in Yunnan Province. Sustainability 2022, 14, 12560. [Google Scholar] [CrossRef]
  45. Liu, S.; Zhang, S.; Xian, R.; Yang, Y. Research on the Impact of Digital Inclusive Finance and Rural Household Human Capital Investment on Urban-rural Income Gap. Front. Humanit. Soc. Sci. 2023. [Google Scholar] [CrossRef]
Figure 1. Conceptual framework.
Figure 1. Conceptual framework.
Systems 13 00145 g001
Figure 2. Comprehensive index of digital finance development levels in China’s provinces for 2011 and 2022.
Figure 2. Comprehensive index of digital finance development levels in China’s provinces for 2011 and 2022.
Systems 13 00145 g002
Figure 3. Comprehensive index of urban–rural income gap development in China’s provinces for 2011 and 2022.
Figure 3. Comprehensive index of urban–rural income gap development in China’s provinces for 2011 and 2022.
Systems 13 00145 g003
Figure 4. Comprehensive index of digital technology usage development in China’s provinces for 2011 and 2022.
Figure 4. Comprehensive index of digital technology usage development in China’s provinces for 2011 and 2022.
Systems 13 00145 g004
Table 1. Descriptive statistics of variables.
Table 1. Descriptive statistics of variables.
Variable NameVariable SymbolMeanStandard DeviationMinimumMaximum
Urban–rural Income Gap t h e i l 0.0720.0400.0070.203
Digital Finance d i f 1.9390.7560.3643.303
Coverage Breadth c o v 1.8830.8000.2443.516
Usage Depth d e p 1.8610.7150.3503.120
Degree of Digitalization d i g 2.2640.8310.0364.379
Traditional Finance f d 2.5651.1611.0376.845
Openness o p e n 0.1740.2550.0021.472
Education Level e d u 0.1760.0380.0940.270
Government Intervention g o v 0.1980.0910.0810.567
Human Capital h u m 1.8591.9710.1039.802
Consumption Level c p i 0.3850.1070.0261.013
Table 2. Baseline regression effects.
Table 2. Baseline regression effects.
(1)(2)(3)(4)(5)(6)
d i f −0.079 ***
(0.007)
−0.078 ***
(0.007)
−0.076 ***
(0.007)
−0.072 ***
(0.007)
−0.074 ***
(0.007)
−0.074 ***
(0.007)
d i f 2 0.016 ***
(0.001)
0.016 ***
(0.001)
0.015 ***
(0.001)
0.015 ***
(0.001)
0.016 ***
(0.001)
0.016 ***
(0.001)
o p e n −0.009 ***
(0.003)
−0.009 ***
(0.003)
−0.009 ***
(0.003)
−0.010 ***
(0.003)
−0.010 ***
(0.003)
e d u 0.060 ***
(0.014)
0.079 ***
(0.014)
0.077 ***
(0.014)
0.077 ***
(0.014)
g o v 0.030 ***
(0.007)
0.029 ***
(0.007)
0.029 ***
(0.007)
c p i 0.009 ***
(0.003)
0.009 ***
(0.003)
h u m 0.000
(0.000)
Time/Region EffectsControlledControlledControlledControlledControlledControlled
Constant0.156 ***
(0.010)
0.157 ***
(0.010)
0.144 ***
(0.010)
0.127 ***
(0.011)
0.127 ***
(0.011)
0.127 ***
(0.011)
N 328832883288328832883288
R 2 0.9430.9430.9440.9440.9440.944
d i f −0.079 ***
(0.007)
−0.078 ***
(0.007)
−0.076 ***
(0.007)
−0.072 ***
(0.007)
−0.074 ***
(0.007)
−0.074 ***
(0.007)
Notes: *** indicates significance at the 1% level. Standard errors are reported in parentheses.
Table 3. Heterogeneity Analysis.
Table 3. Heterogeneity Analysis.
Regional HeterogeneityDimensional Heterogeneity
(1)(2)(3)(4)(5)(6)(7)(8)
Developed RegionsUnderdeveloped RegionsEastern RegionCentral RegionWestern RegionCoverage BreadthUsage DepthDigitalization
d i f −0.070 ***
(0.009)
−0.062 ***
(0.013)
−0.063 ***
(0.012)
−0.070 ***
(0.014)
−0.083 ***
(0.012)
d i f 2 0.012 ***
(0.001)
0.014 ***
(0.002)
0.010 ***
(0.002)
0.014 ***
(0.002)
0.016 ***
(0.002)
c o v −0.065 ***
(0.005)
c o v 2 0.011 ***
(0.001)
d e p −0.015 ***
(0.005)
d e p 2 0.008 ***
(0.001)
d i g 0.006
(0.005)
d i g 2 −0.000
(0.001)
Control VariablesControlledControlledControlledControlledControlledControlledControlledControlled
Time/Region EffectsControlledControlledControlledControlledControlledControlledControlledControlled
Constant 0.136 ***
(0.016)
0.113 ***
(0.017)
0.126 ***
(0.020)
0.137 ***
(0.022)
0.155 ***
(0.019)
0.128 ***
(0.008)
0.046 ***
(0.008)
0.040 ***
(0.007)
N 1224206411881152948328832883288
R 2 0.9100.9410.8960.9200.9560.9460.9410.937
Inflection   Point 2.9172.2143.1502.5002.5942.9550.93815.000
Notes: *** indicates significance at the 1% level. Standard errors are reported in parentheses.
Table 4. Robustness and endogeneity tests.
Table 4. Robustness and endogeneity tests.
(1)(2)(3)(4)(5)(6)(7)(8)
Replacing Measurement MethodExcluding MunicipalitiesChanging Sample PeriodInstrumental VariablesInstrumental Variable 1Instrumental Variable 2Instrumental Variable 3System
GMM
L 1 . t e z s 0.456 ***
(0.000)
L 2 . t e z s 0.234 ***
(0.000)
d i f −0.747 ***
(0.088)
−0.073 ***
(0.007)
−0.441 ***
(0.098)
−0.071 ***
(0.007)
−0.045 ***
(0.002)
−0.031 ***
(0.004)
−0.050 ***
(0.015)
−0.024 ***
(0.005)
d i f 2 0.166 ***
(0.012)
0.015 ***
(0.001)
0.124 ***
(0.015)
0.015 ***
(0.001)
0.007 ***
(0.000)
0.003 ***
(0.001)
0.008 **
(0.004)
0.004 ***
(0.001)
u r b −0.020 ***
(0.007)
s t r 0.002
(0.001)
Control VariablesControlledControlledControlledControlledControlledControlledControlledControlled
Time/Region EffectsControlledControlledControlledControlledControlledControlledControlledControlled
Constant2.834 ***
(0.147)
0.126 ***
(0.011)
2.510 ***
(0.173)
0.135 ***
(0.012)
0.095 ***
(0.005)
0.082 ***
(0.005)
0.094 ***
(0.008)
0.022 **
(0.039)
N 32883244219232883014328825322740
R 2 0.9160.9440.9530.9450.9380.9290.919
F 565.2837.8674.51
A R ( 1 ) 0.000
A R ( 2 ) 0.507
H a n s e n 0.239
Notes: *** indicates significance at the 1% level. ** indicates significance at the 5% level.Standard errors are reported in parentheses.
Table 5. Interaction effects.
Table 5. Interaction effects.
BaselineRegional HeterogeneityDimensional Heterogeneity
(1)(2)(3)(4)(5)(6)
Full SampleDeveloped RegionsUnderdeveloped RegionsCoverage BreadthUsage DepthDigitalization
f d 0.005 ***
(0.001)
0.007 ***
(0.001)
0.004 ***
(0.001)
0.005 ***
(0.001)
0.005 ***
(0.001)
0.004 ***
(0.001)
d i f −0.079 ***
( 0.008 )
−0.064 ***
(0.011)
−0.066 ***
(0.013)
d i f 2 0.017 ***
( 0.001 )
0.011 ***
(0.002)
0.015 ***
(0.002)
d i f _ f d 0.002
( 0.001 )
0.001
(0.001)
0.006 **
(0.003)
d i f 2 _ f d −0.001 **
( 0.000 )
−0.000
(0.000)
−0.002 ***
(0.001)
c o v −0.071***
(0.005)
c o v 2 0.013 ***
(0.001)
c o v _ f d 0.001
(0.001)
c o v 2 _ f d −0.0005 *
(0.000)
d e p −0.010 **
(0.005)
d e p 2 0.008 ***
(0.001)
d e p _ f d 0.003 *
(0.001)
d e p 2 _ f d −0.000
(0.000)
d i g 0.007
(0.004)
d i g 2 −0.000
(0.001)
d i g _ f d 0.002
(0.001)
d i g 2 _ f d −0.000
(0.000)
Control   Variables Controlled ControlledControlledControlledControlledControlled
Time / Region   Effects Controlled ControlledControlledControlledControlledControlled
Constant 0.130 ***
( 0.012 )
0.124 ***
(0.018)
0.119 ***
(0.017)
0.133 ***
(0.009)
0.038 ***
(0.008)
0.039 ***
(0.007)
N 3288 12242064328832883288
R 2 0.946 0.9150.9430.9470.9430.939
Notes: *** indicates significance at the 1% level. ** indicates significance at the 5% level. * indicates significance at the 10% level. Standard errors are reported in parentheses.
Table 6. Baseline effect threshold test results.
Table 6. Baseline effect threshold test results.
VariableThreshold TypeF-Statisticp-ValueThreshold Value95% Confidence Interval
d i f Single Threshold50.6400.0000.113(0.105, 0.116)
Double Threshold15.8100.1270.366(0.345, 0.372)
Triple Threshold9.4200.3270.455
c o v Single Threshold59.1700.0000.109(0.103, 0.113)
Double Threshold29.6700.0070.343(0.336, 0.348)
Triple Threshold14.2000.5930.236(0.218, 0.240)
d e p Single Threshold82.2100.0030.113(0.105, 0.116)
Double Threshold33.7100.0730.293(0.283, 0.296)
Triple Threshold20.0600.3030.127(0.124, 0.129)
d i g Single Threshold176.140.0000.109(0.103, 0.113)
Double Threshold68.6500.0030.166(0.162, 0.169)
Triple Threshold56.0800.5900.290(0.277, 0.293)
Table 7. Baseline effect threshold effect regression model.
Table 7. Baseline effect threshold effect regression model.
(1)(2)(3)(4)
Digital FinanceCoverage BreadthUsage DepthDigitization
d i f ( d g 0.113 ) −0.012 ***
(0.000)
d i f ( d g > 0.113 ) −0.018 ***
(0.000)
c o v ( d g 0.109 ) −0.014 ***
(0.000)
c o v ( 0.109 < d g 0.343 ) −0.018 ***
(0.000)
c o v ( d g > 0.343 ) −0.016 ***
(0.000)
d e p ( d g 0.113 ) −0.007 ***
(0.000)
d e p ( 0.113 < d g 0.293 ) −0.016 ***
(0.000)
d e p ( d g > 0.293 ) −0.017 ***
(0.000)
d i g ( d g 0.109 ) −0.006 ***
(0.000)
d i g   ( 0.109 < d g 0.166 ) −0.011 ***
(0.000)
d i g ( d g > 0.166 ) −0.015 ***
(0.000)
Control   Variables ControlledControlledControlledControlled
Time / Region   Effects ControlledControlledControlledControlled
Constant 0.087 ***
(0.000)
0.090 ***
(0.000)
0.081 ***
(0.000)
0.082 ***
(0.000)
Observations 3288328832883288
R 2 0.8910.8950.8740.884
Notes: *** indicates significance at the 1% level. Standard errors are reported in parentheses.
Table 8. Mechanism effect threshold test results.
Table 8. Mechanism effect threshold test results.
VariableThreshold TypeF-Statisticp-ValueThreshold Value95% Confidence Interval
d i f Single Threshold628.3500.0000.178(0.175, 0.181)
Double Threshold81.9800.0000.227(0.223, 0.230)
Triple Threshold39.5100.9030.119(0.103, 0.113)
c o v Single Threshold758.060.0000.178(0.175, 0.183)
Double Threshold70.5000.0070.227(0.223, 0.230)
Triple Threshold36.6800.8830.119(0.113, 0.122)
d e p Single Threshold400.4100.0000.166(0.163, 0.169)
Double Threshold78.0600.0170.227(0.223, 0.230)
Triple Threshold39.2700.9600.105(0.101, 0.109)
d i g Single Threshold366.2400.0000.151(0.146, 0.155)
Double Threshold89.5700.0300.227(0.223, 0.230)
Triple Threshold30.6000.9230.178(0.175, 0.181)
Table 9. Mechanism effect threshold effect regression model.
Table 9. Mechanism effect threshold effect regression model.
(1)(2)(3)(4)
Digital FinanceCoverage BreadthUsage DepthDigitization
d i f _ f d ( d g 0.178 ) 0.012 ***
(0.000)
d i f _ f d ( 0.178 < d g 0.227 ) 0.001
(0.354)
d i f _ f d ( d g > 0.227 ) −0.008 ***
(0.000)
c o v _ f d ( d g 0.178 ) 0.013 ***
(0.000)
c o v _ f d ( 0.178 < d g 0.227 ) 0.001
(0.224)
c o v _ f d ( d g > 0.227 ) −0.006 ***
(0.000)
d e p _ f d ( d g 0.166 ) 0.012 ***
(0.000)
d e p _ f d ( 0.166 < d g 0.227 ) 0.002 **
(0.024)
d e p _ f d ( d g > 0.227 ) −0.008 ***
(0.000)
d e p _ f d ( d g 0.151 ) 0.010 ***
(0.000)
d e p _ f d ( 0.151 < d g 0.227 ) 0.003 ***
(0.001)
d e p _ f d ( d g > 0.227 ) −0.008 ***
(0.000)
Control   Variables ControlledControlledControlledControlled
Time / Region   Effects ControlledControlledControlledControlled
Constant 0.052 ***
(0.000)
0.053 ***
(0.000)
0.050 ***
(0.000)
0.053 ***
(0.000)
Observations 3288328832883288
R 2 0.8650.8680.8580.858
Notes: *** indicates significance at the 1% level. ** indicates significance at the 5% level. Standard errors are reported in parentheses.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xiao, Y.; Yin, M.; Wang, H.; Xiang, Y. Digital Finance, Digital Usage Divide, and Urban–Rural Income Gap: Evidence from China. Systems 2025, 13, 145. https://doi.org/10.3390/systems13030145

AMA Style

Xiao Y, Yin M, Wang H, Xiang Y. Digital Finance, Digital Usage Divide, and Urban–Rural Income Gap: Evidence from China. Systems. 2025; 13(3):145. https://doi.org/10.3390/systems13030145

Chicago/Turabian Style

Xiao, Yanfei, Mengli Yin, Huilin Wang, and Yunbo Xiang. 2025. "Digital Finance, Digital Usage Divide, and Urban–Rural Income Gap: Evidence from China" Systems 13, no. 3: 145. https://doi.org/10.3390/systems13030145

APA Style

Xiao, Y., Yin, M., Wang, H., & Xiang, Y. (2025). Digital Finance, Digital Usage Divide, and Urban–Rural Income Gap: Evidence from China. Systems, 13(3), 145. https://doi.org/10.3390/systems13030145

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop