Influence of Digital Finance on Household Leverage Ratio from the Perspective of Consumption Effect and Income Effect

Abstract

Household leverage ratio is an important factor affecting family stability. Digital finance has changed the means of payment and consumption frequency, but the relationship between digital finance and household leverage ratio is still unclear. The existence of household debt is defined as the existence of leverage. The higher the household debt, the greater the household leverage. Based on the matching data of the China Household Finance Survey (CHFS) 2019 and the China Digital Inclusive Finance Index, this paper studies the impact of digital finance on household leverage ratio and explores its mechanism theoretically and empirically. This research finds that digital finance can significantly promote the household leverage ratio and this conclusion is still valid after instrumental variable method and robustness test. The mechanism analysis shows that digital finance can promote household over-consumption and further expand household leverage ratio. Digital finance can also reduce household leverage ratio by increasing household income. The heterogeneity analysis suggests that the role of digital finance in expanding leverage ratio is stronger for urban areas and households with low educational level. For households with higher assets, digital finance helps to reduce leverage ratio. Therefore, the government should guide residents to consume rationally and give full play to the entrepreneurship-facilitating and income-increasing effect of digital finance. Meanwhile, the residents themselves should speed up the cultivation of digital financial literacy, which is of vital significance for lowering household leverage ratio and systemic financial risks. China’s development level of digital finance ranks among the top in the world. Studying the role of digital finance in China is helpful to provide experience reference for countries around the world.

1. Introduction

Since China implemented the supply-side reform in 2015, the growth rate of corporate and local government leverage ratio has declined, but the household leverage ratio has been rising [1]. According to the data of the People’s Bank of China, the household leverage ratio in the household sector increased from 52.2% in 2016 to 72.2% in 2021, already exceeding the warning line of 65% set by the International Monetary Fund. The sharp increase of the household leverage ratio increases the living burden of residents and the high pressure of repayment easily leads to financial risks. Since the outbreak of COVID-19 in 2019, the living pressure of residents in China has also risen. It is essential to find ways to improve the household leverage ratio.

There are many existing studies on household leverage ratio which can summarize its related causes and consequences. The following is an analysis of the causes from macro and micro perspectives. From the macro perspective, some scholars believe that the leverage ratio risk of China’s household sector should not be exaggerated even though the current leverage ratio is climbing rapidly. They believe that the income flux can bear the pressure of repaying the principal and interest, that the low-risk assets can cope with the liquidity crisis, and that the high-savings rate of residents has sufficient solvency [1]. The rapid rise in housing prices has sharply driven up the household leverage ratio in China, with each doubling of housing prices leading to a 39.2% increase in the household leverage ratio [2]. Income inequality significantly increases the household leverage ratio, and the leverage ratio of low-wealth households is mainly driven by housing debt [3]. Downward fluctuations in house prices and unemployment shocks can both increase leverage ratio [4]. Long-term low-interest rates and tax cuts can also increase household debt [5]. Some scholars use the panel data of 41 countries from 1966 to 2017 and find that high savings rates lead to high leverage ratio, but there is a “U” shape relationship between them [6]. Some scholars believe that the higher the urbanization level, the higher the household debt level [7]. The convenience of information technology provides technical support for the expansion of credit channels and further expands household debt [8]. The theoretical basis of household debt behavior is primarily based on life cycle theory and liquidity constraint hypothesis [9]. The influencing factors from the micro perspective are as follows: household leverage ratio decreases with the age of the householder [10]. Human capital level, educational level, and financial literacy are some factors that affect the household leverage ratio. People with high financial literacy can make better financial decisions and will not blindly borrow money to expand the household leverage ratio [11,12]. The increase of per capita income helps to improve the household leverage ratio [13]. In addition, demographic characteristics, subjective attitudes, and financial situations have different effects on households with different debt sizes [14]. Household leverage ratio is the ratio of household debt to total assets and is used to describe the level of household debt.

“Digital finance” generally refers to the aggregation of “Internet Finance”, defined by the People’s Bank of China, and “financial technology”, defined by the Financial Stability Board. It represents new financial business modes, such as financing, payment, and investment, that traditional financial institutions and Internet companies realize through advanced digital technology. The emergence of digital finance in China was initiated by the launch of Alipay in 2004, while the first year of digital finance was marked by the launch of Yu Ebao in 2013 [15]. The progress of third-party payment, online lending, digital insurance, and digital currency business has given rise to new financial institutions, shortened the space-time distance between financial supply and financial demand, lowered the threshold for residents to use financial services, and exerted a significant impact on real life. In recent years, scholars have conducted more research on the impact of digital finance in various fields. The first kind of research defines the concept and connotation of digital finance [15]. The second kind of research analyzes the influence of financial institutions implementing digital finance, including the transmission effect of monetary policy, reducing the number of employees in the financial industry, enhancing financial efficiency, and affecting financial stability [16,17,18,19]. The third kind of research focuses on the economic effect of digital finance, which primarily includes alleviating the financing constraints of energy enterprises, promoting green technology innovation, reducing the bankruptcy risk of enterprises, reducing environmental inequality, promoting sustainable development, stimulating the investment of small and micro enterprises, lowering urban environmental pollution, driving inclusive growth, and facilitating renewable energy consumption [20,21,22,23,24,25,26,27]. The fourth kind of research focuses on the influence of digital finance on micro-residents, which is strongly related to this present paper. Digital finance can increase residents’ income, boost consumption, enhance entrepreneurship, affect labor mobility, squeeze out traditional private loans, and alleviate relative poverty [28,29,30,31,32,33,34,35]. There have been many studies on the impact of digital finance on residents, but few studies have analyzed the relationship between digital finance and household leverage ratio.

This paper holds that the relationship between digital finance and household leverage ratio has a consumption effect and an income effect. Regarding the analysis of the consumption effect of digital finance, some studies have proven that digital finance can generate payment convenience, relax liquidity constraints [31], and speed up residents’ consumption decisions. Digital finance also reduces the uncertainty faced by households, releases consumer demand, and provides assistance to household consumption expansion [36]. Meanwhile, the growth of consumption further expands household debt [37,38]. In short, the income effect of digital finance may affect the household leverage ratio. Regarding the analysis of the income effect of digital finance, digital finance can narrow the income gap between urban and rural areas. With the advancement of digital finance, residents’ entrepreneurship and innovation have increased dramatically, which has also led to a significant increase in residents’ income [39,40]. Some studies also conclude that the increase in per capita income can improve the households’ consumption expectations, strengthen their willingness to borrow, and thus expand the household debt [13]. Besides, digital finance can promote low-income household leverage ratio to a greater extent [41]. In short, the income effect of digital finance may affect the household leverage ratio. In view of this, this paper aims to investigate the relationship between digital finance and household leverage ratio.

The marginal contributions of this paper are as follows. First, this paper analyzes the effect of digital finance on household leverage ratio theoretically and empirically. It finds that digital finance can expand household leverage ratio, which holds true even after overcoming endogenous problems. Second, from the perspective of consumption effect, it is found that digital finance can expand over-consumption and thus improve household leverage ratio. From the perspective of income effect, it is found that digital finance can increase income and thus reduce household leverage ratio. Third, it is found that digital finance has characteristic heterogeneity when it affects household leverage ratio. Wang et al. [38] found that digital finance affects household leverage ratio from the perspective of consumption and liquidity constraints. Compared with that, this paper also believes that digital finance can improve household leverage ratio.

The follow-up arrangement of this paper is as follows. The second part is data interpretation and model setting. The third part is empirical results. The fourth part is conclusions and suggestions.

2. Data Interpretation and Model Setting

2.1. Data Source and Variable Definition

The data of this paper comes from three aspects. The data of micro households comes from the CHFS2019, published by the China Household Finance Survey of Southwestern University of Finance and Economics. The data of digital finance comes from the China Digital Inclusive Finance Development Index released by the Institute of Digital Finance, Peking University. The data of city level comes from China City Statistical Yearbook. Considering the characteristics of digital financial users and private lending groups, the samples of residents younger than 18 and older than 80 are excluded. In addition, the samples with missing household key variables are also eliminated. This paper defines the main explained variables and control variables based on the nationwide large-scale micro-household survey data conducted by the China Household Finance Survey of Southwestern University of Finance and Economics. In order to study the relationship between digital finance and household leverage ratio, this paper conducts research based on the newly published CHFS questionnaire data, and can also use all CHFS data to further expand the research.

Household leverage ratio (Leverage). The explained variable in this paper is household leverage ratio, and the existence of household debt is defined as the existence of leverage. The higher the household debt, the greater the household leverage. Therefore, the ratio of household debt to total assets is used as a measure of household leverage ratio [3].

Digital Finance (Dfi). The core explanatory variable of this paper is the development of digital finance, which is measured by the Digital Inclusive Finance Development Index with reference to existing research practices [29]. The data comes from the Digital Inclusive Finance Index compiled by the Institute of Digital Finance, Peking University, and Ant Financial [42]. The second dimensions of this index consist of coverage (cov), use depth (USE), and digital service (dig). In this study, digital finance and its sub-indicators are logarithmic to reduce the influence of heteroscedasticity.

Mechanism variables. Over-consumption and income. With reference to relevant research, the difference between total household consumption and household income is used to measure over-consumption [43]. According to the general practice of previous studies, total household income is employed to measure income. These two variables are both logarithmic to reduce the influence of heteroscedasticity.

Control variables. In terms of control variables, this paper mainly controls the characteristic variables of householder, household, and region. Characteristic variables of householder: whether the householder is male (man), age (age), education years (eduyear), rural residence (rural), marriage (marriage), and health (health). To alleviate the problem of missing variables at the resident level, this paper also includes the square of the householder’s age (age2). Characteristic variables of household: risk attitude (risk), whether the place of residence is rural, logarithm of total household assets (asset), and logarithm of total household income (income). To distinguish the function of traditional finance from digital finance, this paper also includes the traditional finance at the regional level and measures it by the logarithmic value of the year-end loan balance of financial institutions in the province. Descriptive statistical results of main variables are shown in

Table 1. Descriptive statistics.

Variable	Sample Size	Mean	Standard Error	Minimum	Maximum
Leverage	33,415	0.192	2.19	0	194.617
Dfi	33,415	5.713	0.091	5.573	5.934
man	33,415	0.757	0.429	0	1
age	33,415	55.407	12.847	19	80
age2	33,415	3234.95	1393.426	361	6400
eduyear	33,415	9.17	4.002	0	22
rural	33,415	0.567	0.495	0	1
marriage	33,415	0.97	0.17	0	1
health	33,415	0.395	0.489	0	1
risk	33,415	0.054	0.226	0	1
asset	33,415	12.749	1.729	0	21.465
income	33,415	10.289	2.281	0	13.239
overcons	33,415	10.016	1.598	−0.693	13.225
finance	33,415	2.18	1.115	1.057	6.368

.

2.2. Estimation Method

Referring to previous research practices, this paper uses the following model for research:

$$Leverag{e}_{i,p}={\alpha }_0+{\alpha }_1\times Df{i}_p+{\alpha }_2\times Control{s}_{i,p}+{\epsilon }_{i,p}$$

(1)

In this model, $i$ represents the household. $p$ represents the province. $Leverag{e}_{i,p}$ represents the private borrowing needs of the household $i$ in the province $p$. $Df{i}_p$ represents the development level of digital finance in all households of the province $p$, which is measured by the China Digital Inclusive Finance Index. $Control{s}_{i,p}$ represents the control variables at the levels of householder, household, and region. ${\epsilon }_{i,p}$ is a random interference term.

3. Empirical Results

3.1. Benchmark Results

To analyze the influence of digital finance on household leverage ratio, this paper first makes benchmark regression.

Table 2. Benchmark regression results.

	(1)	(2)	(3)	(4)
	Leverage	Leverage	Leverage	Leverage
Dfi	−0.648 ***	−0.495 **	0.549 ***	0.503 ***
	(0.178)	(0.181)	(0.126)	(0.168)
man		−0.026	−0.009	−0.009
		(0.036)	(0.034)	(0.034)
age		0.001	0.022 **	0.022 **
		(0.006)	(0.008)	(0.008)
age2		−0.000	−0.000 ***	−0.0003 ***
		(0.000)	(0.000)	(0.000)
eduyear		−0.017 ***	0.009	0.009
		(0.005)	(0.006)	(0.006)
rural		−0.022	−0.147 ***	−0.146 ***
		(0.035)	(0.049)	(0.049)
marriage		−0.063	0.184	0.184
		(0.138)	(0.140)	(0.140)
health		−0.135 ***	−0.053 **	−0.053 **
		(0.029)	(0.020)	(0.020)
risk			0.149	0.148
			(0.090)	(0.090)
asset			−0.237 ***	−0.237 ***
			(0.039)	(0.039)
income			0.007 *	0.007 *
			(0.004)	(0.004)
finance				0.006
				(0.011)
_cons	3.895 ***	3.469 ***	−0.400	−0.146
	(1.024)	(1.009)	(0.624)	(0.913)
N	33,415	33,415	33,415	33,415
R2	0.001	0.003	0.026	0.026

* p < 0.10, ** p < 0.05, *** p < 0.01. The standard error of clustering to the province is in brackets.

reports the results of benchmark regression. Column (1) shows the results without including control variables. Column (2) shows the results after including control variables at the level of householder. Column (3) shows the results after further including control variables at the level of household, and column (4) shows the complete regression results. Comparing the results in columns (1) and (2) and columns (3) and (4), it can be seen that there is a significant error in the results when there are no control variables at the levels of household and region, indicating that the major problem of missing variables will occur when the control variables at the levels of household and region are not controlled. In this paper, the most complete and accurate column (4) is used for analysis. The coefficient of digital finance (Dfi) to household leverage ratio (Leverage) is 0.503, which is significant at the level of 1%. This indicates that digital finance can significantly improve the household leverage ratio and expand the household debt level. Digital finance provides households with convenient access to funds, such as Ant Credit Pay and Ant Cash Now, which makes households’ debt level easily increase. From the results of the control variables, the relationship between age (age) and household leverage ratio (Leverage) is a significantly inverted “U” shape, with the inflection point being approximately 38 years old. This suggests that the increase in age usually expands the leverage ratio before residents reach the age of 38 years old. After reaching 38 years old, residents decrease the household leverage ratio [10] with the increase of age. The relationship between the rural household (rural) and leverage ratio (Leverage) is negatively significant, indicating that residents with rural household usually have a lower household leverage ratio.

3.2. Robustness Test

Although the control variables of residents, households, and regions have been controlled as much as possible, there are still endogenous concerns such as missing variables and measurement errors. Therefore, this paper employs the 2SLS model for analysis. In addition, the leverage ratio is a truncated variable, so a Tobit model is applied for processing.

The instrumental variable used in this paper is the distance from each province to Zhejiang Province [31]. The underlying data of digital finance comes from the Ant Financial Group, which is located in Hangzhou, Zhejiang Province. Therefore, the level of digital finance in Zhejiang Province is supposed to be relatively high. A longer distance from Zhejiang Province suggests more backward development of digital finance, thus satisfying the correlation. The distance between the residents of each province and Zhejiang does not affect the leverage ratio, so it also satisfies the exogeneity. Column (1) of

Table 3. Robustness test results.

	2SLS Model	Tobit Model
	(1)	(2)
	Leverage	Leverage
Dfi	0.564 **	0.503 ***
	(0.230)	(0.168)
Controls	√	√
K-Paap rk LM statistic	11.650 [0.0006]
K-Paap rk Wald F statistic	30.247 [0.000]
N	33,415	33,415
R2	0.026

** p < 0.05, *** p < 0.01. The standard error of clustering to the province is in brackets. √ represents the meaning of yes, indicating control.

reports the results of using the instrumental variable method. It can be observed that the impact of digital finance (Dfi) on household leverage ratio (Leverage) is still significantly positive, indicating that the conclusion of this paper is still valid after endogeneity treatment. In addition, the K-P rk LM test and the K-P rk Wald F test verify that the selection of instrumental variables is reasonable and effective. In this paper, the model is also changed to check whether the benchmark model is robust. Column (2) of Table 3 displays the results of the Tobit model. The coefficient of digital finance (Dfi) to household leverage ratio (Leverage) is 0.503, which is significant at the level of 1%. This is consistent with the benchmark results, indicating that the model does not affect the robustness of the results in this paper. All the above results verify that the results of this paper are robust.

3.3. Test Results of Sub-Indexes

To further analyze the detailed influence of digital finance on household leverage ratio, this paper makes an analysis of the sub-indexes of digital finance.

Table 4. Test results of sub-indexes of digital finance.

	(1)	(2)	(3)
	Leverage	Leverage	Leverage
cov	0.746 ***
	(0.190)
use		0.366 **
		(0.143)
dig			0.554
			(0.388)
Controls	√	√	√
N	33,415	33,415	33,415
R2	0.026	0.026	0.026

** p < 0.05, *** p < 0.01. The standard error of clustering to the province is in brackets. √ represents the meaning of yes, indicating control.

presents the results of using different sub-indexes of digital finance. It can be observed that the coverage of digital finance (cov) has the greatest (=0.746) and the most significant impact on the household leverage ratio (Leverage), and the use depth of digital finance has the second greatest (=0.366) impact on the household leverage ratio (Leverage), with a significance level of only 5%. The digital service level of digital finance (dig) does not significantly affect the household leverage ratio (Leverage), so there is no comparative significance. The above results indicate that the coverage and depth of digital finance provide convenient financial services for households and expand the household leverage ratio, but the improvement of digital service level does not expand the household leverage ratio.

3.4. Mechanism Analysis

To test the influence of digital finance’s consumption effect and income effect on the household leverage ratio, this paper constructs the following models (2) and (3) for analysis on the basis of relevant research.

$$Me{d}_{i,p}={\beta }_0+{\beta }_1\times Df{i}_p+{\beta }_2\times Control{s}_{i,p}+{\epsilon }_{i,p}$$

(2)

$$Leverag{e}_{i,p}={\gamma }_0+{\gamma }_1\times Df{i}_p+{\gamma }_2\times Me{d}_{i,p}+{\gamma }_3\times Control{s}_{i,p}+{\epsilon }_{i,p}$$

(3)

Among them, $Me{d}_{i,p}$ is the mechanism variable of this paper. The proxy variables representing consumption effect and income effect are measured by over-consumption (overcons) and income (income) respectively. The definitions of other variables are the same as in model (1). First, attention needs to be paid to the coefficient ${\beta }_1$ in model (2). If ${\beta }_1$ is significant, it means that digital finance (Dfi) affects mechanism variable (Med), which can be further analyzed. If ${\gamma }_2$ is significant in model (3), it means that the mechanism variable (Med) has an effect. If ${\gamma }_2$ is not significant, it means that there is no mechanism effect.

This paper first analyzes the effect of over-consumption on digital finance’s effect on household leverage ratio, and the results are shown in columns (1) and (2) of

Table 5. Mechanism Analysis results.

	(1)	(2)	(3)	(4)
	Overcons	Leverage	Income	Leverage
Dfi	0.802 **	0.441 **	1.147 ***	−0.307
	(0.337)	(0.161)	(0.412)	(0.249)
overcons		0.078 ***
		(0.015)
income				−0.019 ***
				(0.004)
Controls	√	√	√	√
N	33,415	33,415	33,415	33,415
R2	0.155	0.028	0.133	0.003

** p < 0.05, *** p < 0.01. The standard error of clustering to the province is in brackets. √ represents the meaning of yes, indicating control.

. The coefficient of digital finance (Dfi) to household over-consumption (overcons) is 0.802, which is significant at the level of 5%. It indicates that digital finance can significantly improve household over-consumption. Digital finance provides convenient payment methods and a wide variety of online shopping options, which tremendously stimulates residents’ consumption. In column (2), the coefficient of household over-consumption to household leverage ratio is 0.078, which is significant at the level of 1%. It indicates that household over-consumption significantly increases household leverage ratio. Combined with the results in columns (1) and (2), it can be easily concluded that digital finance improves household leverage ratio by promoting household over-consumption.

Furthermore, this paper also analyzes the effect of income growth on digital finance’s effect on household leverage ratio, and the results are displayed in columns (3) and (4) of Table 5. It can be seen from the results in column (3) that the coefficient of digital finance (Dfi) to household income is 1.147, which is significant at the level of 1%. It indicates that digital finance can significantly increase household income. Digital finance boosts the economic level, promotes residents’ entrepreneurial behavior, and increases household income [40]. In column (2), the coefficient of household income growth (income) to household leverage ratio (Leverage) is −0.019, which is significant at the level of 1%. It indicates that household income growth significantly reduces household leverage ratio. Combined with the results in columns (3) and (4), it can be easily concluded that digital finance reduces household leverage by promoting household income growth.

Based on the above results, it can be found that digital finance has a consumption effect and an income effect when it affects household leverage ratio. The consumption effect intensifies the rise of household leverage ratio, and the income effect weakens the rise of household leverage ratio.

3.5. Heterogeneity Analysis

Due to the vast territory of China, the development levels of digital finance vary from place to place, so the impact of digital finance on different areas may be different. Moreover, there will be differences in assets and educational level between households, which may lead to different impacts of digital finance on different groups of people. Therefore, this paper carries out group-based tests according to location (urban or rural areas), household assets level, and educational level of residents. The dummy variable of household assets level is divided according to the median of household assets, with high assets above the median and low assets for the rest. The dummy variable of educational level is divided according to the residents’ educational background, with low educational level below senior high school and high educational level for the rest.

Table 6. Heterogeneity Analysis results.

	(1) Urban Areas	(2) Rural Areas	(3) High Assets	(4) Low Assets	(5) Low Educational Level	(6) High Educational Level
	Leverage	Leverage	Leverage	Leverage	Leverage	Leverage
Dfi	0.641 ***	0.299	−0.068 **	−0.101	0.610 **	0.244 *
	(0.212)	(0.266)	(0.028)	(0.397)	(0.240)	(0.135)
Controls	√	√	√	√	√	√
N	21,366	12,049	16,707	16,708	22,232	11,183
R2	0.024	0.037	0.082	0.038	0.032	0.013

* p < 0.10, ** p < 0.05, *** p < 0.01. The standard error of clustering to the province is in brackets. √ represents the meaning of yes, indicating control.

below demonstrates the corresponding results.

According to the results in columns (1) and (2) of Table 6, digital finance (Dfi) can significantly improve the household leverage ratio (Leverage) in urban areas, but has no significant effect on the household leverage ratio (Leverage) in rural areas. On the one hand, the development level of digital finance in urban areas is often high, so the impact of digital finance is more significant. On the other hand, the financial demand in urban areas is often higher, and people will more frequently participate in the financial market, buy stocks, borrow, and spend money. To sum up, the role of digital finance in promoting household leverage ratio is more significant in urban areas.

According to the results in columns (3) and (4) of Table 6, digital finance (Dfi) can significantly reduce the leverage of households with high assets, but has no significant effect on the leverage of households with low assets. The possible reason is that the debt caused by digital finance is not big enough for households with high assets compared with households with low assets. Meanwhile, digital finance can provide more means of wealth maintenance and appreciation for households with high assets, so the leverage ratio easily declines, while households with low assets have no such channels.

According to the results in columns (5) and (6) of Table 6, digital finance (Dfi) can significantly promote the leverage of households with low educational level and households with high educational level, but digital finance has a greater and more significant impact on the leverage of households with low educational level. When human capital level is high, households can often make better financial decisions, thus reducing the possibility of expanding the household leverage ratio. Therefore, it is in line with the theoretical reality that digital finance has a greater impact on the leverage ratio of households with low educational level [11].

4. Conclusions and Suggestions

Based on the matching data of the China Household Finance Survey (CHFS) 2019 and the China Digital Inclusive Finance Index, this paper studies the impact of digital finance on household leverage ratio theoretically and empirically, and investigates its mechanism and heterogeneity. The empirical results demonstrate that digital finance can significantly improve the household leverage ratio. In addition, the conclusion is still valid after the robustness test such as the instrumental variable method and model changing. The mechanism analysis results show that digital finance has a consumption effect and an income effect on household leverage ratio. Digital finance increases household leverage ratio by expanding household over-consumption, while decreasing household leverage ratio by increasing household income. The heterogeneity analysis results indicate that digital finance plays a greater role in expanding household leverage ratio for urban areas and households with low educational level, while reducing household leverage ratio for households with high assets. Therefore, the government should guide residents to consume rationally, and give full play to the entrepreneurship-facilitating and income-increasing effect of digital finance. Meanwhile, the residents themselves should speed up the cultivation of digital financial literacy, which is of vital significance for lowering household leverage ratio and systemic financial risks.

Based on the above results, this paper proposes the following suggestions. First, it is necessary to cultivate the digital financial literacy of households. The improvement of residents’ well-being helps to maintain social stability, while the rising household leverage ratio is not conducive to household harmony and residents’ happiness. Therefore, cultivating households’ awareness of digital finance can help households make rational and effective use of digital finance, thus bringing benefits to household development. Second, households should not blindly expand consumption. Digital finance provides product convenience and payment convenience, but consumers should strengthen their shopping rationality, which is helpful to maintain household welfare. Third, the government should guide digital finance to support residents’ entrepreneurship and increase their income. Moreover, it is of vital importance to give full play to digital finance’s abilities of finance support and entrepreneurship facilitation, actively support entrepreneurial activities of households, and empower residents’ income growth, thus reducing the household leverage ratio.