The Impacts of Digital Finance on Economic Resilience

Abstract

Enhancing economic resilience is crucial to sustainable development. However, issues such as resource misallocation and financing difficulties have severely impacted supply chain stability and security. The rise of digital finance presents potential solutions to these problems. Based on panel data from 30 provinces in China from 2011 to 2020, this study explores the mechanisms and pathways through which digital finance enhances economic resilience. The results reveal four key findings. First, the development of digital finance significantly enhances economic resilience by improving innovation capabilities and consumption vitality. Second, a high degree of marketization strengthens the promoting effect of digital finance on economic resilience. Third, in areas with higher levels of industrial advancement or lower levels of traditional financial development, the enhancement effect of digital finance on economic resilience is more significant. Fourth, digital finance significantly improved the economic resilience of geographically adjacent areas through spatial spillover effects. Overall, this study provides a new perspective on the impact of digital finance on economic resilience in developing countries such as China. In addition to its academic contributions, this study offers detailed practical implications.

1. Introduction

Currently, the world is undergoing significant changes unseen in the last century, such as anti-globalization, geopolitical conflicts, and epidemic disasters, all of which have brought serious challenges to the development of developing countries. For this reason, it has become an important driving force for stable development to enhance economic resilience. The word resilience is derived from the Latin word Resilire, which translates as “return to original state” and refers to the resilience of a system after an external disturbance [1]. Economic resilience not only refers to the ability to maintain its original structure and return to its original development level [2] after an economy suffers shocks but also reflects the ability to return to its original growth track and maintain its stable growth [3] and sustainable development [4] when an economy faces external shocks and internal fluctuations, and the resilient economy can represent greater resistance and adaptability to external shocks in this era of uncertainty and challenges. In other words, economic resilience refers to the continuous optimization of the economy through technological advances, thereby enhancing its overall competitiveness, influence, and control and ensuring long-term stable development. Nevertheless, the long-term stable development of the modern economy must depend on finance, which is a powerful driving force. As an advanced financial form formed by the integration of financial innovation and technological innovation, digital finance refers to a new financial model that uses modern information technology, especially the Internet, big data, cloud computing, artificial intelligence, and other technologies, to provide financial services and financial products. This is a vane for modern financial development. This has improved the efficiency of financial services, reduced transaction costs, enhanced user experience, and expanded the coverage of financial services. By supporting the digital transformation of enterprises, digital finance solves the problem of energy poverty and promotes green growth [5]. This plays an important role in promoting sustainable economic development. The rapid development and widespread application of digital finance have provided strong support for enhancing the adaptability, innovation, and risk-resisting ability of the economy.

How can economic resilience be improved? What is the mechanism and path? It is of important theoretical and practical significance to clarify these questions for enhancing economic resilience and promoting development.

2. Literature Review

We find that previous studies mainly analyze issues on how to improve economic resilience at the macro and micro levels by reviewing the existing literature. At the macro level, they mainly select indices to analyze economic resilience [6], emphasizing the importance of quantitative and qualitative assessments [7]. Moreover, from the perspective of long-term growth trajectory and recovery force to future shocks, some examined the resilience of economies by specifying the shock to a natural disaster [8]. Hu (2022) found that the regional economic resilience of developing countries is influenced not only by economic structural factors but also by national institutions in terms of economic restraint and recovery measures [9]. For developed countries, taking Eastern and Central Europe as an example, employment in most areas could not reach the original level within two years after the outbreak of COVID-19 [10]. Political or economic disasters are often harder to recover from. Wang (2021) found that fostering entrepreneurship could improve China’s regional economic resilience [11] by systematically analyzing the response of China’s regional economy to the economic crisis in 2008 by means of elasticity measuring resilience. For developed regions, population migration is the most positive factor affecting regional resilience, whereas population aging has a strong negative impact on regional resilience [12]. From the perspective of long-term economic development, the positive influence of the synergistic effect between digital finance and green innovation on economic resilience will gradually fade after reaching a peak [13]. Some scholars have also focused on the influence of urban agglomeration on economic resilience; for instance, Ye et al. (2021) [14] studied the regional economic resilience of the three major urban agglomerations in eastern China and found that it formed a more advanced and stable type of social capital as the spatial scale expanded, thereby enhancing the economic resilience of the entire urban agglomeration. Furthermore, for regions with high digitalization level, through economic network linkage [15] and E-commerce transformation [16], it can better enhance the economic resilience of urban agglomerations [17].

At the micro level, Li et al. (2023) [18] verified that enhancing the development level of the digital economy can significantly increase the resilience of the value chain of producer services through empirical study and proposed that we should strengthen the industrial foundation, improve the marketization level of factors, accelerate independent innovation [19] and attract foreign investment [20] by analyzing supply chain resilience in the manufacturing industry. In addition, in the agricultural sector, strengthening disaster control and promoting sustainable development of agriculture can promote the stable and healthy development of the regional ecological economy [21]. Enterprises should actively absorb scientific and technological innovation talents, increase investment in scientific and technological innovation, and improve the industrial scientific and technological innovation ability, thereby laying a good ecological foundation for the development of regional scientific and technological innovation [22]. In addition, enterprises should consider resilience while pursuing sustainable development; otherwise, it may harm the ability of participants to adapt and transform [23]. Supply diversity can strengthen an enterprise’s ability to respond to risk by providing more options in the face of disruption [24].

However, there are significant shortcomings in these studies on economic resilience, which mostly remain in the theoretical stage and lack data support and empirical tests. The object of these studies is a single industry, lacking a grasp of the whole. In addition, the influence mechanism and path proposed are relatively simple and unconvincing, even though some studies include empirical analysis and tests. In addition, there are relatively few studies on the influence of digital finance on economic resilience, and there are no in-depth analyses of its influence path.

Compared with existing studies, the possible marginal contributions of this paper are as follows: First, it expands the theoretical perspective of research on economic resilience. On the basis of the meaning of economic resilience, an economic resilience evaluation index system is constructed, and the entropy method is used for measurement and analysis, thereby quantitatively determining economic resilience. Second, taking China as an example, it empirically examines how digital finance enhances economic resilience by improving innovation ability and consumption vitality and how a high degree of marketization can enhance the promoting effect of digital finance on economic resilience. This can provide actionable implications for developing countries when formulating policies.

3. Theoretical Analysis and Research Hypothesis

3.1. The Direct Effect of Digital Finance on Economic Resilience

Digital finance is still essentially finance, although it is disguised as information technology [25]. Consequently, the theories of financial development and economic resilience can provide a reference for this study. Based on a review of the literature, most researchers believe that financial development contributes to economic resilience. On the one hand, finance can promote economic development by reducing transaction costs and optimizing capital allocation, thereby enhancing economic resilience. On the other hand, finance can reduce the possibility of risk occurrence to stabilize economic development [26] by optimizing the financing environment and reducing information asymmetry.

Compared with traditional finance, digital finance can overcome the limitations of time and space to play the functions of traditional finance further. First, digital finance can help increase financial inclusion and accessibility by providing more convenient and efficient financial services, particularly for those who previously had no access to financial markets. This provides the conditions for reducing inequality of opportunity, promotes higher household incomes, especially for low-income rural groups, helps narrow regional and rural–urban disparities, and promotes inclusive growth. Second, digital finance uses digital technology to lower the threshold of financial services and relies on big data, cloud computing, and other technologies to optimize credit business processes and reduce information search costs, labor costs, operating costs, and risk costs, which is conducive to improving economic output and promoting economic growth. Third, digital finance has improved the transparency and efficiency of the financial system, reduced information asymmetry, improved risk assessment and pricing capabilities, and contributed to financial stability. At the same time, the development of fintech has promoted the digital transformation of financial markets and provided more tools for financial regulation. By integrating the information flow, cash flow, and other credit characteristics of enterprises, quantitative risk analysis and risk pricing are carried out to alleviate the market failure of the credit market so that the economy can show stronger resistance and resilience, adaptation and adjustment, innovation and transformation in the face of shocks. Fourth, digital finance can improve economic resilience by guiding more factor resources to key areas and weak links. Digital finance can set up a financial service platform serving upstream and downstream enterprises in the supply chains with the help of cutting-edge technologies such as blockchain, the Internet of Things, and artificial intelligence, and further solve problems such as financing difficulties of enterprises and effectively enhancing economic resilience. Based on this, we propose the following hypothesis:

Hypothesis 1: 

Digital finance can significantly enhance economic resilience.

3.2. The Indirect Effect of Digital Finance on Economic Resilience

Digital finance not only has a direct influence on economic resilience but also has an indirect influence on economic resilience by optimizing urban innovation capacity and consumption vitality [27]. On the one hand, digital finance lowers the financial market access threshold and expands the coverage of financial services with the help of digital technologies, which helps alleviate problems that market subjects may face, such as credit discrimination, financing constraints, and so on, thereby optimizing the market competition environment and promoting the transformation of enterprises to sustainable development [28]. On the other hand, the emergence of digital finance also contributes to the rapid development of E-commerce, unleashing more business opportunities and greater market vitality, thereby significantly enhancing innovation capability and consumption vitality [29]. Meanwhile, positive urban innovation capacity and consumption vitality are also of vital importance for improving economic resilience. It can effectively reduce the difficulties and risks faced in the process of supply chain growth and improve the dynamic mechanism of supply chain growth by optimizing innovation capability and consumption vitality, thereby effectively enhancing economic resilience [30]. First, the innovation capability significantly enhances urban economic resilience through multiple functions, such as promoting technological innovation, stimulating new industries, attracting and cultivating talents, promoting financial development, enhancing the urban image, etc. Second, urban consumption vitality has a positive influence on economic resilience through various means, such as providing economic growth drivers, promoting employment and income stability, stimulating innovation and entrepreneurship, strengthening economic diversity, improving people’s quality of life, and promoting social capital accumulation. Based on the above analysis, we propose the following hypothesis:

Hypothesis 2: 

Digital finance promotes economic resilience by improving the innovation capability and consumption vitality.

3.3. The Moderating Effect of Marketization

Marketization level refers to the function and efficiency of the market mechanism in resource allocation in an economic system. It involves the competitiveness of the market, the degree of government intervention, the effectiveness of property rights protection, and the freedom of business activities. New start-ups are weak in terms of resource acquisition, and financing is often constrained; therefore, increasing the local market scale and marketization level is conducive to giving play to the functions of digital finance [31]. The improvement of marketization can help make digital financial services more efficiently reach the areas and enterprises with the most urgent needs. Digital finance can provide more appropriate financial products and services to small and medium-sized enterprises and individuals through accurate data analysis and risk assessment while also helping to break up monopolies and reduce unwarranted administrative interference to make more fintech companies and innovative financial institutions enter the market and provide diversified financial services. Meanwhile, for regions with high marketization levels, digital finance improves the availability of financial services for the “tail” group to promote the development of private enterprises, thus increasing the share of labor income [32]. Expanding the coverage of financial services, especially for remote areas and low-income groups, enhances their sense of economic security and the ability to respond to crises through digital finance. As a result, the improvement in the marketization level provides a favorable external environment for the development of digital finance, thereby promoting digital finance to improve the adaptability and resilience of the overall economy.

Hypothesis 3: 

Marketization level has a positive moderating effect on the relationship between digital finance and economic resilience.

Based on the above assumptions, we have drawn up a path map of the impact of digital finance on the economic resiliences we are developing in our research, as shown in Figure 1.

4. Research Design

4.1. Variables Selection

4.1.1. Dependent Variable: Economic Resilience

In this study, we combine the comprehensiveness and availability of data to construct an economic resilience index system by drawing on previous research results. In terms of resilience, the economic resilience proposed in existing studies means that the industrial chains and supply chains can maintain their own stable state after suffering risks and shocks, and economic resilience can be divided into three dimensions: resistance ability, recovery ability, and adjustment ability from the perspective of the ability to recover and adjust from shocks. In the meantime, the economic resilience proposed by Li et al. (2022) [33] includes autonomy and control of core technologies, self-sufficiency of basic products, sustainable use of resources, and so on. Based on this, economic resilience is complementally measured from the three dimensions of technological ability, self-sufficiency ability, and sustainability. Therefore, the entropy method is adopted for calculation. Specific indexes are listed in

Table 1. Economic Resilience Index System.

Index	First-Grade Index	Second-Grade Index	Index Interpretation	Index Attribute
Economic resilience	Resistance ability	GDP growth rate	Direct data	Positive
		Urbanization level	Urban population at the end of the year within the region/total population at the end of the year within the region	Positive
		Industrial diversification level	Measured by using the reciprocal of Herfindahl Index	Positive
		Industrial rationalization level	Measured by using the Theil Index	Negative
	Recovery ability	Industrial agglomeration level	Employed population/region area	Positive
		Social security level	Local financial expenditure on social security and employment/general financial expenditure	Positive
	Adjustment ability	Local financial self-sufficiency level	Local financial revenue/local financial expenditure	Positive
		Urban–rural income gap	Measured by using the Theil Index	Negative
		Social consumption capacity	Total retail sales of consumer goods/GDP	Positive
		Social investment level	Total fixed assets investment/GDP	Positive
	Technological ability	The proportion of researchers	Employed personnel in urban units engaged in scientific research and technological services/employed personnel in urban units	Positive
		The proportion of scientific research expenditure	Local science and technology financial expenditure/local financial expenditure	Positive
		Scientific research achievement	Number of authorized domestic invention patents/employed personnel in urban units of scientific research and technology services	Positive
	Self-sufficiency ability	External dependence on basic products	Import volume of goods/GDP	Negative
	Sustainability	Energy consumption elasticity coefficient	Energy consumption growth rate/GDP growth rate	Negative
		Wastewater generated per unit of output	Wastewater emissions/GDP	Negative
		Waste gas generated per unit of output	Sulfur dioxide emissions/GDP	Negative

.

(1)

Resistance Ability.

Resistance ability means that supply chains can maintain core business operations and will not be completely paralyzed because of unexpected circumstances when suffering from external shocks. The GDP growth rate is one of the key indexes of a country or region’s economic health. Higher GDP growth rates usually mean relatively stronger economic activities, healthier industrial development, and greater resistance to external shocks. A higher urbanization level indicates that it can provide better infrastructure, inseparable supply chain networks, and greater adaptability so as to respond and disperse risks more effectively in case of emergency. In general, an economy with a higher industrial diversification level can disperse risk and reduce dependence on a single industry, which means that it has a stronger ability to resist risks; a rational industrial structure means that various industries complement each other, with a stronger synergistic effect, and as a result, it has more advantages in resisting external shocks.

It should be noted that we use the reciprocal of the Herfindahl Index to characterize the industrial diversification level, and the specific formula is as follows:

$$indiv=1/HHI={\displaystyle \sum _{i=1}^n{S}_i^2}$$

(1)

As shown in Equation (1) $S_i$ represents the proportion of the gross product of the $i$ industry to GDP. It can effectively characterize the industrial diversification level by adopting this method and help analyze the resilience of the industry. Meanwhile, in view of data availability, we calculate the proportion of the gross products of the primary, secondary, and tertiary industries in GDP respectively, namely $i=1,2,3; n=3$.

In addition, we draw on the practice of Gan et al. (2011) [34] and use the Thiel Index to characterize the industry rationalization level using the following formula:

$$TL={\displaystyle \sum _{i=1}^n\frac{Y_i}{Y}}\ln ({\displaystyle \frac{Y_i}{L_i}}/{\displaystyle \frac{Y}{L}})$$

(2)

As shown in Equation (2) $Y$ represents the output value, $L$ represents employment, $Y/L$ represents productivity, $i$ represents industry, $n$ represents the number of industrial sectors, and $i=1,2,3$ represents the primary, secondary, and tertiary industries, respectively. Based on classical economic assumptions, the productivity levels of various industrial sectors are the same when the economy is ultimately in equilibrium, namely, $Y_i/L_i=Y/L$, thereby, $TL=0$, which shows that the industrial structure is in an equilibrium state and reasonable. In contrast, it shows that the industrial structure deviates from the equilibrium state, and the industrial structure is unreasonable when $TL$ is not zero.

(2)

Recovery Ability

Recovery ability refers to the ability to resume normal operations as quickly as possible after supply chains suffer losses or interruptions. Among them, the industrial agglomeration level refers to the degree of agglomeration of industries in the economy, which can also organize resources and partners more quickly to recover quickly even though it suffers external shocks when the industrial agglomeration is high. The social security level covers various protection mechanisms provided by the government and society. A high level of social security means that individuals and enterprises have stronger economic and social support in the face of external shocks and can recover from losses more quickly while reducing systemic risks caused by poverty or social instability.

(3)

Adjustment Ability

Adjustment ability refers to the ability to quickly make adaptive changes and adjustments when supply chains face environmental changes, market demand changes, or external challenges. Among them, the level of local fiscal self-sufficiency refers to the level of autonomous fiscal surplus without relying on central fiscal transfer payments, and the high self-sufficiency level means that local governments have more financial resources to support local development and respond to change. A small rural–urban income gap usually means that resources and opportunities are more evenly distributed, which helps the whole economy to cope better with change. A high social consumption capacity shows that people can support more economic activities, thus having more flexibility in adjustment ability; the economy with a high investment level can quickly adjust its industrial structure through investment in the face of external shocks so as to adapt to economic and social changes.

It should be noted that we use the Theil Index to characterize the urban–rural income gap in this study, and the calculation formula is as follows:

$$DIS={\displaystyle \sum _{i=1}^n\frac{I_i}{I}}\ln ({\displaystyle \frac{I_i}{P_i}}/{\displaystyle \frac{I}{P}})$$

(3)

As shown in Equation (3) $I$ represents total income, $P$ represents population, $I/P$ represents average income, and $i=1,2$ represents town and country, respectively. $I_i/P_i=I/P$ when urban and rural per capita income levels are the same, thereby, $DIS=0$, indicating that the income gap between urban and rural areas is zero, which is relatively reasonable; otherwise, it is unreasonable.

(4)

Technological Ability

Technological ability refers to the supply chain’s ability to independently conduct technology research and development, production, and application in key areas without relying on the supply and licensing of external technologies. Technological ability can be measured by the proportion of researchers, the proportion of scientific research expenditure, and per capita scientific and technological achievements from different perspectives.

(5)

Self-sufficiency Ability

A high self-sufficiency ability means that a country or region can rely on its own or local production capacity to meet internal demand for the production of key basic goods without relying on external supply. It includes some necessary raw materials, energy, and infrastructure; therefore, we use the proportion of imported goods to characterize it.

(6)

Sustainability

Sustainability emphasizes that the acquisition and utilization of resources in supply chains should be environmentally friendly and sustainable without causing serious damage to the environment. It involves rational development, recycling, and conservation of resources to ensure their long-term availability while reducing environmental pressures and ecological risks. Therefore, we use the energy consumption elasticity coefficient, wastewater, and waste gas generated per unit of output to characterize it.

4.1.2. Independent Variable: Digital Finance

Drawing on the common practice in the existing literature, we use the Digital Financial Inclusion Index compiled by Peking University to measure the level of digital finance [35]. The Digital Financial Inclusion Index includes a total of 33 indexes in three dimensions involving the coverage breadth of digital finance, the use depth of digital finance, and the digitization level of inclusive finance, which can represent the development level of digital finance in a comprehensive way. In this study, both the index and its sub-dimensions are divided by 100 to eliminate dimensional differences.

4.1.3. Mediating Variable: Innovation Capability and Consumption Vitality

(1)

Innovation Capability

Invention patents in a region are a relatively intuitive index for measuring innovation capability. We use the number of authorized patent applications/permanent residents to represent innovation capability by referring to existing research findings, which can eliminate the effects of regional population differences.

(2)

Consumption Vitality

We obtain the urban consumption vitality index by dividing the total retail sales of consumer goods by the resident population.

4.1.4. Moderating Variable

In this study, the marketization level is selected as the moderating variable, and in today’s economic environment, the development of digital finance not only helps to reduce the financing difficulties of small and medium-sized enterprises and private enterprises but also helps to reduce their operating costs, thereby improving the regional economic resilience. For areas with low marketization levels, the promoting effect of the development of digital finance on the improvement of economic resilience will be relatively insignificant, as there are many state-owned enterprises in regions with low marketization levels and the financing difficulty is relatively low, and the degree of marketization is measured by:

(The number of private employees + the number of self-employed employees)/(the number of private employees + the number of self-employed employees + the number of employees employed in urban units).

4.1.5. Control Variables

In this study, we identify the following four control variables by reviewing and analyzing previous research findings: human capital, foreign investment intensity, transportation infrastructure level, and population density.

(1)

Human Capital

Human capital is characterized by the ratio of the number of students in colleges and universities to the total local population. Students in colleges and universities receive a high level of education and training, which can play an important role in the workplace by leveraging knowledge and skills acquired at school, thereby enhancing supply chain resilience. In general, students in colleges and universities are more open and more likely to accept new ideas and technologies, thereby contributing to the generation of inventions and innovations.

(2)

Foreign Investment

Foreign investment is characterized by the ratio of foreign direct investment (FDI) to GDP. It can improve the productivity and quality level of local enterprises, as foreign direct investment usually introduces advanced technology, management experience, and production methods. In the meantime, foreign direct investment can promote the diversity of local markets and attract investors from different countries and regions, thereby helping to reduce dependence on a single market and effectively improving the resilience of supply chains.

(3)

Transportation Facilities

The level of transportation facilities is characterized by the logarithm of the highway mileage. Smooth highway networks can enhance regional communication and facilitate cross-regional coordination and cooperation, thereby improving the overall resilience of supply chains. Furthermore, favorable transportation infrastructure can expand market access to make it easier for products to enter different regions and markets. It can reduce market concentration and dependence on specific markets, thereby improving the security level of supply chains.

(4)

Population Density

Population density is characterized by the number of people per square kilometer. High population density areas may make it easier for supply chains to fulfill the principle of proximity, which means moving closer to consumers, thereby reducing shipping time and costs and enhancing supply chain resilience.

4.1.6. Categorical Variables

We take the median of the advanced industrial level and the traditional financial development level as the basis for grouping for heterogeneity analysis.

(1)

Advanced Industrial Level

The advanced industrial level is characterized by the ratio of the added value of the tertiary industry to that of the secondary industry. There are many small and medium-sized enterprises in the tertiary industry, and the development of digital finance can provide financing guarantees for small and medium-sized enterprises and effectively alleviate the financing difficulties faced by medium-sized enterprises, thereby enhancing the resilience of supply chains. For this reason, the advanced industrial level is taken as a grouping variable to analyze the influence of digital finance on economic resilience at different levels.

(2)

Level of Development of Traditional Finance

The traditional financial development level is characterized by the ratio of deposits and loans to GDP. Compared with traditional finance, digital finance can overcome the limitations of time and space and allocate more factor resources to key areas and weak links of the supply chains, thus enhancing economic resilience. Therefore, it is necessary to analyze the influence of digital finance on economic resilience at different levels of traditional financial development, thereby proving whether digital finance can effectively compensate for the shortcomings of traditional finance.

4.2. Data Collection and Descriptive Statistics

In this study, due to limited data availability, we used data from 30 provinces (autonomous regions and municipalities) in China from 2011 to 2020, sourced from the China Statistical Yearbook, China Environmental Statistical Yearbook, China Population and Employment Statistical Yearbook, the official website of the National Bureau of Statistics, and the statistical yearbooks and statistical bulletins of all provinces (autonomous regions and municipalities), and some missing values are completed using the linear interpolation method. The descriptive statistics of the relevant variables are presented in

Table 2. Descriptive Statistics of Variables.

Variable Attribute	Variable	Mean	SD	Min	Median	Max	N
Dependent Variable	Resilience	0.295	0.109	0.133	0.264	0.666	300
Independent Variable	Digital Finance	2.173	0.970	0.183	2.241	4.319	300
Mediating Variable	Innovation capability	11.504	13.244	0.860	5.991	74.383	300
	Consumption Vitality	2.156	1.168	0.538	1.868	6.878	300
Moderating Variable	Marketization	0.608	0.091	0.335	0.612	0.802	300
Control Variable	Human Capital	0.020	0.005	0.008	0.019	0.041	300
	Foreign Investment	0.019	0.015	0.000	0.017	0.080	300
	Transportation Facilities	11.681	0.848	9.400	11.965	12.885	300
	Population Density	0.047	0.070	0.001	0.029	0.392	300
Categorical Variable	Advanced Industrial Level	1.219	0.696	0.518	1.058	5.297	300
	Development Level of Traditional Finance	3.231	1.161	1.518	3.014	8.131	300

.

It can be seen from the table that the mean value, maximum value, minimum value, and standard deviation of the economic resilience of each province are 0.295, 0.666, 0.133, and 0.109, respectively, indicating that the resilience of supply chains in different regions varies greatly, and there is polarization in the aspects of digital finance, urban innovation capacity, and consumption vitality. In terms of control variables, there are obvious differences in human capital, foreign investment, transportation facilities, and population density in different regions.

4.3. Research Model

To empirically analyze the influence of digital finance on economic resilience, that is to say, to verify Hypothesis 1, the following benchmark regression model is constructed in this study:

$$Y_{it}=a_0+{\alpha }_1{X}_{it}+\alpha Control{s}_{it}+{\mu }_i+{\nu }_t+{\epsilon }_{it}$$

(4)

where $Y_{it}$ is the economic resilience of $i$ Province in $t$, $X_{it}$ is the digital financial level of $i$ Province in $t$, $Controls$ is other factors affecting economic resilience, ${\mu }_i$ is the time fixed effect, ${\nu }_t$ is the individual fixed effect and ${\epsilon }_{it}$ is a stochastic disturbance term.

To further verify the mechanism of action of digital finance in enhancing economic resilience, in this study, we build a mediating effect test model by referring to existing methods in the literature to test the path effect of innovation capability and consumption vitality, that is, to verify Hypothesis 2:

$$Z_{it}={\beta }_0+{\beta }_1{X}_{it}+\beta Control{s}_{it}+{\mu }_i+{\nu }_t+{\epsilon }_{it}$$

(5)

where $Z$ is the innovation capability and consumption vitality of mediating variable. We construct a “Two-step Method” model for mediating effect test in combination with Equations (4) and (5) by learning from Jiang Ting’s practice (2022) [36], Equation (5) represents the influence of digital finance on innovation capability and consumption vitality.

$$Y_{it}={\gamma }_0+{\gamma }_1{X}_{it}+{\gamma }_2{M}_{it}+{\gamma }_3{X}_{it}\cdot M_{it}+\gamma Control{s}_{it}+{\epsilon }_{it}$$

(6)

where M is the marketization level of the regulating variable, γ3 is the regulating effect, and Equation (6) is used to prove Hypothesis 3 by constructing a mediating effect model.

5. Results and Discussion

5.1. Baseline Test

Therefore, in view of the large number of variables selected in this study, it is necessary to test whether there is multicollinearity between variables before the regression analysis. It is found that they are all less than five after calculating the variance inflation factor (VIF) for each variable, which shows that there is no serious multicollinearity between the variables. On this basis, it is found that the P value is less than 0.01 through the Hausman test on these data; therefore, the fixed effects model is used for regression analysis in this study, and the results are shown in columns 1, 2, and 3 of

Table 3. The impact of digital finance on resilience.

	(1)	(2)	(3)	(4)	(5)
	Resilience	Resilience	Resilience	Innovation Capability	Consumption Vitality
Digital Finance	0.030 ***	0.095 ***	0.096 ***	23.121 ***	2.269 ***
	(0.002)	(0.032)	(0.027)	(3.655)	(0.285)
Human Capital			1.006
			(1.665)
Foreign Investment			0.418 **
			(0.178)
Transportation Facilities			0.063
			(0.061)
Population Density			0.833
			(1.437)
Time Fixed Effects	NO	YES	YES	YES	YES
Individual Fixed Effects	NO	YES	YES	YES	YES
Constant	0.229 ***	0.208 ***	−0.582	245.81	−9.789
	(0.018)	(0.015)	(0.683)	(100.221)	(9.474)
N	300	300	300	300	300

Note: (1). Standard errors are in parentheses; (2). ** p < 0.05, *** p < 0.01.

.

To ensure the robustness of the results, only the regression analysis of digital finance on economic resilience is conducted. Second, individual fixed effects and time-fixed effects are introduced for regression analysis. Finally, control variables are introduced for regression analysis. It can be found that digital finance always has a significant positive effect on economic resilience; in other words, digital finance can significantly enhance economic resilience. Accordingly, it verifies Hypothesis 1.

In addition to the explanatory variables, the intensity of foreign investment also has a significant positive effect on economic resilience, which is in line with the previous analysis. Human capital, transport infrastructure level, and population density all have positive effects on economic resilience. However, they are not significant, indicating that they can enhance economic resilience to some extent. The reasons for its insignificance may lie in the following: First, the quality of human capital is also important, although the ratio of the number of students in colleges and universities to the total local population can reflect the amount of human capital input to some extent. It may not adequately improve the quality of labor if the quality of training and education is not high, even if there are a large number of students in colleges and universities, which thereby affects its substantive influence on economic resilience. Second, it does not necessarily increase economic resilience significantly simply by increasing highway mileage. This is because the highways need supporting facilities, management, and maintenance to ensure high circulation efficiency. The increasing highway mileage may not work as well as it should if there is no effective management or maintenance. Third, the areas with high population density usually have more abundant human resources, which may provide potential advantages such as abundant labor and the consumer market to some extent. Nevertheless, it is equally important to determine whether resources are allocated and utilized efficiently and whether there is an appropriate infrastructure and management level to support the development of supply chains. A high population density may cause excessive dispersion of resources, resulting in an imbalance between supply and demand and even bringing about enormous pressure on the environment and infrastructure, thereby affecting economic resilience.

5.2. Mediating Effect Analysis

Based on the mediating effect model constructed earlier, we test whether digital finance can optimize urban innovation capacity and consumption vitality, thereby enhancing economic resilience. The results are shown in columns 4 and 5 of Table 3, and control variables are controlled. It can be found that digital finance has a significant promoting effect on innovation capability and consumption vitality. Ren (2024) [37] incorporated the innovation dimension into an economic resilience assessment system based on the theoretical interpretation of the connotation of economic development resilience, and economies with high innovation capacity can adapt more quickly to change and return to growth in the face of external shocks and challenges, avoiding or mitigating the negative effects to a certain extent. City panel data prove that the development of innovation and entrepreneurship can enhance urban economic resilience. The increase in consumption directly led to the expansion of the economy, as consumer expenditure accounts for a large share of the gross domestic product. A dynamic consumer market can provide stable support for economic growth, and digital finance can reduce income inequality by increasing online shopping [38]. Generally, the improvement in consumption vitality is accompanied by an increase in employment opportunities and household income. It can not only increase families’ economic security sense but also enhance their ability and willingness to spend, thereby forming a positive economic cycle and enhancing economic stability and resilience. In conclusion, digital finance can have a positive promoting effect on economic resilience by improving innovation capability and consumption vitality; accordingly, Hypothesis 2 holds.

5.3. Moderating Effect Analysis

In this study, we conduct a regression analysis on whether marketization level promotes or inhibits the positive effect of digital finance on economic resilience based on the moderating effect model constructed earlier. The results are presented in

Table 4. Moderating effect results.

	(1)	(2)
	Resilience	Resilience
Digital Finance	0.096 ***	0.080 ***
	(0.027)	(0.026)
Marketization	0.005	0.030
	(0.035)	(0.032)
Digital Finance*Marketization		0.072 ***
		(0.025)
Control Variables	YES	YES
Time Fixed effects	YES	YES
Individual Fixed Effects	YES	YES
Constant	−0.581	−0.257
	(0.684)	(0.635)
N	300	300

Note: (1). Standard errors are in parentheses; (2). *** p < 0.01.

.

It can be seen from columns (1) and (2) of Table 4 the influence of digital finance on economic resilience is still positive and significant when the adjustment variable regression is introduced, and it is still significant when the interaction term coefficient is positively significant after the interaction term regression is introduced, showing that it has a positive moderating effect, and the higher the marketization level, the greater the promoting effect of digital finance on economic resilience. Accordingly, it proves Hypothesis 3.

5.4. Endogenous Processing

In this study, we conduct endogenous processing using the Two-stage Least Square Method to solve possible endogenous problems. First, we use the digital inclusive financial index with a one-stage lag as an instrumental variable for testing by referring to the instrumental variable method. There is a strong correlation between the development level of digital inclusive finance with a one-stage lag and that of digital inclusive finance in the current period, as the influence of digital inclusive finance may lag, and its correlation with economic resilience in the current period is not high, meeting the requirements of the necessary correlation and exogeneity of instrumental variables. On this basis, we use the two-stage least square method for endogeneity analysis in this study, and the results are shown in columns 1 and 2 of

Table 5. Endogeneity results.

	(1)	(2)	(3)	(4)
	Digital Finance	Resilience	Digital Finance	Resilience
Digital Finance		0.081 ***		0.214 ***
		(0.025)		(0.020)
One Phase Lagged Digital Finance	0.730 ***
	(0.054)
Distance From Hangzhou			−0.155 ***
			(0.020)
Control Variables	YES	YES	YES	YES
Time Fixed Effects	YES	YES	YES	YES
Individual Fixed Effects	YES	YES	YES	YES
Constant	1.019	−0.512 *	0.789	−0.006
	(0.669)	(0.281)	(0.185)	(0.042)
First-stage F	184.750 ***		58.281 ***
N	300	300	300	300

Note: (1). Standard errors are in parentheses; (2). * p < 0.1, *** p < 0.01.

.

It can be found that the F value is greater than 10 and is significant at the 1% level in the first stage, indicating that there is no problem with weak instrumental variables. Digital finance still has a significantly positive effect on economic resilience after controlling for endogenous problems, which is in line with the benchmark regression results.

Second, considering that digital finance originated in Hangzhou, we use the distance (taking logarithm) between the provincial capital of each province and Hangzhou as instrumental variables. This is because there is a high correlation between the distance from the provincial capital of each province to Hangzhou and the local digital financial development level; in other words, the farther away from Hangzhou, the less vulnerable it is to the influence of Hangzhou’s digital financial development. Moreover, the distance has little relationship with local economic resilience, which is a relatively exogenous variable, meeting the requirements of instrumental variable selection. Nevertheless, it cannot be included in the panel data set used in this study for the regression analysis because the distance between the provincial capital of each province and Hangzhou will not change. Here, we start from the dimension of the cross-section and only control for the time-fixed effect to estimate, and the results are shown in columns 3 and 4 of Table 5.

The results show that the distance between the provincial capital of each province and Hangzhou has a significant positive effect on digital finance. The influence of digital finance on economic resilience remains significant and positive after the instrumental variables are adopted, which is in line with the benchmark regression results. It can be found from further analysis that the F value is greater than 10 and is significant at the 1% level in the first stage, showing that there is no problem with the weak instrumental variables, and the results are valid.

5.5. Robustness Checks

We conduct the following tests to ensure the robustness of the regression results.

5.5.1. Replace the Samples

Compared with other provinces, each municipality has a strong advantage in terms of policies and economy. We delete the samples from Beijing, Shanghai, Tianjin, and Chongqing and then perform regression analysis again, and the results are shown in columns 1, 2, and 3 of

Table 6. Robust test (1).

	(1)	(2)	(3)
	Resilience	Innovation Capability	Consumption Vitality
Digital Finance	0.099 ***	14.554 ***	2.229 ***
	(0.301)	(3.983)	(0.379)
Control Variables	YES	YES	YES
Time Fixed Effects	YES	YES	YES
Individual Fixed Effects	YES	YES	YES
Constant	−0.365	70.016	−4.903
	(0.822)	(76.190)	(7.464)
N	260	260	260

Note: (1). Standard errors are in parentheses; (2). *** p < 0.01.

. From the regression results, digital finance still has a significant positive effect on economic resilience after the samples of the four municipalities are deleted, while urban innovation capacity and consumption vitality have a significant mediating effect in terms of how digital finance affects economic resilience, which shows that the research conclusions of this study still hold.

5.5.2. Replace Independent Variable

We use the digitization level, a digital inclusive finance sub-indexes, as a proxy variable of the independent variable and incorporate it into the model for regression analysis, the results of which are shown in columns 4, 5, and 6 of

Table 7. Robust test (2).

	(4)	(5)	(6)
	Resilience	Innovation Capability	Consumption Vitality
Degree of Digitalization of Inclusive Finance	0.036 ***	5.286 ***	0.608 ***
	(0.008)	(1.200)	(0.116)
Control Variables	YES	YES	YES
Time Fixed Effects	YES	YES	YES
Individual Fixed Effects	YES	YES	YES
Constant	−0.737	230.277 **	−11.899
	(0.613)	(105.092)	(8.913)
N	300	300	300

Note: (1). Standard errors are in parentheses; (2). ** p < 0.05, *** p < 0.01.

.

The digitization level has a significant positive effect on economic resilience, while the innovation capability and consumption vitality still play a significant mediating effect, which is in line with the conclusions of the regression results in this study.

5.6. Heterogeneity Analysis

Heterogeneity analysis based on advanced industrial levels. According to the previous analysis, we take the median of the advanced industrial level as the basis for grouping and performing regression analysis, the results of which are shown in columns 1 and 2 of

Table 8. Heterogeneity analysis results.

	(1)	(2)	(3)	(4)
	Resilience	Resilience	Resilience	Resilience
Digital Finance	0.044	0.096 **	0.116 ***	0.036
	(0.028)	(0.037)	(0.029)	(0.030)
Control Variables	YES	YES	YES	YES
Time Fixed Effects	YES	YES	YES	YES
Individual Fixed Effects	YES	YES	YES	YES
Constant	−1.424	0.082	−1.341	−0.066
	(0.942)	(0.410)	(0.904)	(0.719)
N	150	150	150	150

Note: (1). Standard errors are in parentheses; (2). ** p < 0.05, *** p < 0.01.

. It can be seen that the influence of digital finance on economic resilience is not significant when the advanced industrial level is low. Digital finance has a significant positive effect on economic resilience when the level of industrial upgrading is high.

A heterogeneity analysis based on the traditional financial development level. According to the previous analysis, we take the median of the traditional financial development level as the basis for grouping and performing regression analysis. The results are shown in columns 3 and 4 of Table 8. Digital finance has a significant positive effect on economic resilience in areas where the traditional financial development level is low; the effect is not significant, although it is positive in areas with high traditional financial development levels. This is also in line with the previous analysis, that is to say, digital finance can reach key areas and weak links that traditional finance cannot, thus significantly improving economic resilience. There are many key areas and weak links that traditional finance cannot reach when the traditional finance development level is low; consequently, digital finance has a significant promoting effect. There are relatively few key areas and weak links when the traditional financial development level is high, and as a result, the promoting effect of digital finance is relatively limited.

6. Further Analysis

Digital finance has a significant spatial spillover effect on adjacent areas because digital finance breaks through geographical restrictions with the help of information technologies such as the Internet of Things and big data and can promote the rapid flow of resources, including finance and technology, thereby influencing adjacent areas. In addition, we calculated using Moran’s I and found that there was a significant spatial effect on the development process of supply chains. Therefore, it is necessary to study whether digital finance can enhance the economic resilience of adjacent areas through spatial spillover effects.

Based on the above analysis, we construct a spatial Dubin model (SDM model):

$$Y_{it}={\phi }_{1}W{Y}_{it}+{\gamma }_1{X}_{it}+{\gamma }_{2}Control{s}_{it}+{\theta }_{1}W{X}_{it}+{\theta }_{2}WControl{s}_{it}+{\mu }_i+{\nu }_t+{\epsilon }_{it}$$

(7)

where $i$ represents the province, $t$ represents the year, $Y_{it}$ represents the economic resilience of the explained variable, $X_{it}$ represents the digital finance of the explanatory variable, $Controls$ represents the control variable, $u_i$ represents the individual fixed effect, $v_t$ represents the time fixed effect, ${\epsilon }_{it}$ represents the stochastic disturbance term, and $\phi$ is the spatial autocorrelation coefficient, $\theta$ is the coefficient of the spatial interaction term, and $W$ represents the spatial weight matrix.

It should be noted that the spatial weight matrix used in this study is a spatial distance matrix that comprehensively considers the influence of geographical distance, and the corresponding formula is constructed as follows:

$$W_{ij}=\left\{\begin{array}{cc}1/d_{ij}^2& i\ne j\hfill \\ 0& i=j\hfill \end{array}\right.$$

(8)

$d_{ij}$ represents the geographical distance of the province $i$ and $j$ calculated based on highway mileage.

Before conducting the regression analysis, we need to use Moran’s I to conduct global and local spatial autocorrelation tests for digital financial and economic resilience to verify whether the variables have a spatial correlation. According to the test results, the global spatial autocorrelation test shows that digital finance and economic resilience have a remarkable spatial positive autocorrelation under the spatial distance matrix as shown in

Table A1. The global Moran index of digital finance and economic resilience.

Year	Digital Finance		Resilience
	Moran Index	p	Moran Index	p
2011	0.361	0.000	0.342	0.000
2012	0.421	0.000	0.358	0.000
2013	0.392	0.000	0.346	0.000
2014	0.391	0.000	0.363	0.000
2015	0.339	0.000	0.400	0.000
2016	0.381	0.000	0.389	0.000
2017	0.389	0.000	0.393	0.000
2018	0.406	0.000	0.407	0.000
2019	0.417	0.000	0.397	0.000
2020	0.426	0.000	0.418	0.000

of Appendix A. In terms of the local spatial autocorrelation test, we test the digital finance and economic resilience in 2011 and 2020, respectively. The results show that the vast majority of provinces are located in the first or third quadrant, indicating significant spatial autocorrelation as shown in Figure A1, Figure A2, Figure A3 and Figure A4 of Appendix A. In conclusion, it is reasonable to choose a spatial econometric model for this study.

We conduct LM, Hausman, Wald, and LR tests to select a suitable spatial econometric model for empirical analysis as shown in

Table A2. LM test, Hausman test, Wald test, and LR test.

Test Type	Results	p
LM-error	6.951	0.008
Robust LM-error	30.662	0.000
LM-lag	44.507	0.000
Robust LM-lag	68.217	0.000
Hausman	493.060	0.000
Wald-error	67.010	0.000
Wald-lag	53.500	0.000
LR-error	60.600	0.000
LR-lag	54.090	0.000

of Appendix A. The results show that all pass a significance test of 1%. Therefore, the SDM model should be adopted for the analysis, and the results are shown in

Table 9. SDM model regression results and effect decomposition.

	(1)	(2)	(3)	(4)	(5)
	Principal Regression	Spatial Regression	Direct Effect	Indirect Effect	Total Effect
Digital Finance	0.074 ***	0.035 ***	0.077 ***	0.059 ***	0.136 ***
	(0.014)	(0.006)	(0.015)	(0.014)	(0.020)
Human Capital	0.136	1.806	0.180	2.163	2.344
	(0.842)	(2.005)	(0.834)	(2.582)	(2.930)
Foreign Investment	0.499 ***	0.553 **	0.536 ***	0.784 ***	1.320 ***
	(0.103)	(0.235)	(0.104)	(0.302)	(0.366)
Transportation Facilities	0.110 ***	−0.098 *	0.106 ***	−0.088	0.018
	(0.018)	(0.053)	(0.018)	(0.064)	(0.072)
Population Density	0.539	1.170	0.569	1.526	2.095
	(0.658)	(1.463)	(0.636)	(1.639)	(1.685)
Individual Effects	YES
Time Effects	YES
$\rho$	0.208 **
	(0.089)
R-squared	0.142
Log-L	890.276
N	300

Note: (1). Standard errors are in parentheses; (2). * p < 0.1, ** p < 0.05, *** p < 0.01.

.

First, the spatial coefficient $\rho$ of the SDM model is very significant, indicating that economic resilience shows a strong spatial dependence. Second, the main regression results in Column 1 show that digital finance has a significant effect on economic resilience. Moreover, according to the regression results in Column 2, the previous conclusions still hold when considering the spatial effect. Finally, we decompose the spatial spillover effect of the SDM into direct, indirect, and total effects, and the results are shown in columns 3, 4, and 5 of Table 9, respectively. It can be found that digital finance has a significant positive role in promoting economic resilience regardless of its direct effects, indirect effects, or total effects. This proves the previous hypothesis that digital finance can enhance the economic resilience of adjacent areas through spatial spillover effects.

7. Conclusions and Implications

7.1. Conclusions

This study aimed to provide insights into the impact of digital finance on economic resilience. First, our findings consistently indicate that digital finance has become a key force in reducing information asymmetry, optimizing resource allocation, and alleviating financing constraints, which is of great significance for enhancing economic resilience. Second, by improving the convenience and efficiency of financial services, digital finance provides more financing opportunities for small and medium-sized enterprises and improves consumption vitality, thereby enhancing the economic resilience of the city. Third, a highly market-oriented city can enhance the role of digital finance in promoting economic resilience. Fourth, the enhancement effect of digital finance on economic resilience is more significant in areas with higher levels of industrial advancement or lower levels of traditional financial development. Fourth, digital finance significantly improved the economic resilience of geographically adjacent areas through spatial spillover effects. Fifth, the development of digital finance not only helps enhance economic resilience in one’s area but also promotes the improvement of economic resilience in geographically adjacent areas.

7.2. Implications

In this paper, we propose the following policy implications based on the above research conclusions: First, we should achieve efficient operation of the digital financial system to enhance economic resilience by increasing investment in the construction of digital financial infrastructure, continuously improving digital payment systems, financial data centers, network security systems, etc. In addition, we should improve the transparency and traceability of supply chains to enhance the overall level of economic security by actively encouraging fintech innovation and developing new fintech means, such as blockchain technology and smart contracts. Second, we should keep improving the marketization level, and the relevant departments should lower the market access threshold, simplify the approval process, improve legal transparency, advocate free development of the private economy, attract more types of enterprises to settle in, achieve the diversification and rationalization of industries, enhance the agglomeration of industries, and reduce external dependence on basic products. Third, we should enhance financial support for scientific research and technological innovation, encourage enterprises to carry out R&D activities, and improve their overall innovation capability. In the meantime, we can encourage the development of new forms of consumption and promote diversification of consumer markets to improve urban consumption vitality to ensure that both the supply and demand sides of the market are more stable, thereby improving the resilience and security of the entire economy. Fourth, we should implement differentiated development strategies, provide more support for areas where the advanced industrial level is low, and advocate the use of a variety of digital financial means to enhance local economic resilience and security. Meanwhile, for the areas where the traditional financial development level is low, we should give full play to the complementary role of digital finance and allocate more factor resources to the key areas and weak links in the industrial and supply chains. Fifth, we should actively establish inter-area cooperation and sharing mechanisms and strengthen knowledge transfer, technology exchange, and economic and trade exchanges through the establishment of a digital financial cooperation platform, thereby giving full play to the spatial spillover effect of digital finance and promoting the improvement of economic resilience and security level in adjacent areas. Sixth, Digital finance can significantly affect economic resilience by enhancing financial inclusion, reducing transaction costs, and fostering innovation. However, it also presents challenges in terms of regulation, security, and privacy. The government should formulate technical standards and business norms related to digital finance and improve the legal supervision system of digital finance. The government should accelerate the training of talents who understand both science and technology and finance, and improve the awareness and ability of employees to innovate in science and technology. The government should also encourage financial institutions to strengthen the independent innovation of key core technologies, improve the disaster preparedness capability of the system, improve the network security technical defense system, and eliminate the hidden dangers of information technology security loopholes.

7.3. Limitations and Future Research

The limitations of our analysis and future research are as follows. First, this study merely takes China as an example to provide policy suggestions for developing countries and does not discuss the situation of other countries. In future research, studies on other representative developing countries and comparative studies between developed and developing countries should be conducted. Second, it does not discuss how to deal with problems such as the digital divide and monopoly, which may be caused by the development of the digital economy. Finally, owing to data limitations, more micro-level data were not used, and discussions on some enterprise cases were not included. These limitations should be studied further in the future.