Digital Financial Capability and Entrepreneurship in China: A Digital Economy Perspective

Abstract

The digital economy is the fastest-growing sector in the world today, and will also be the key development direction of China in the future. The level of digital financial capability (DFC) determines whether households can fully embrace the inclusivity of digital finance and seize the business opportunities presented by the digital economy. Hence, we utilized panel data from the China Household Finance Survey (CHFS) and built a probit model with two-way fixed effects. The results revealed that DFC has a positive impact on household entrepreneurship through both supply- and demand-side mechanisms. The results also revealed that DFC has not only catalyzed the evolution of digital business models but has also enhanced the digital business performance of enterprises. The analysis of heterogeneity confirms that DFC can promote entrepreneurship in rural settings and in the retail and restaurant sectors. This research has important theoretical contributions and policy suggestions for governments to help them pay attention to the inclusive role of DFC in developing the digital economy.

1. Introduction

Entrepreneurship not only injects vitality into economic growth but also serves as the bedrock of stable employment [1,2]. According to data from the State Administration for Market Regulation of China, between 2012 and 2022, the number of private enterprises in China surged from over 10 million to more than 49 million.

The digital economy has ushered in fresh opportunities for innovation and entrepreneurship. It has emerged as the fastest-growing sector both in China and globally. China’s digital economy, especially digital finance, has developed rapidly and is at the forefront of the world, which can be used for reference by other developing countries in the future. The digital economy’s evolution hinges on the Internet revolution and emerging digital technologies such as big data, the Internet of things (IoT), and artificial intelligence (AI). In the realm of e-commerce, as of June 2022, there were 24 million online retail platform stores in China, marking a year-on-year increase of 12.1%, which directly provided employment to over 70 million people. Hence, the government of China emphasizes the significance of digital economy development.

Digital finance constitutes a vital component of the digital economy’s progression, enhancing inclusivity within the financial system and extending development opportunities to the vast numbers previously marginalized by mainstream finance [3]. China’s low credit card penetration rate left the majority of residents without credit records and thus unable to secure loans. Digital finance has disrupted this landscape. The tendencies of digital finance user growth are shown in Figure 1. By 2022, China boasted 911 million digital payment users, accounting for 85.4% of total Internet users. These digital financial platforms collect user consumption data and employ big data and machine learning to evaluate their creditworthiness, establishing extensive databases of individual credit histories. Leveraging these records, digital financial platforms can offer microloans to entrepreneurs through online applications, simplifying the process and eliminating the need for collateral [4]. Figure 1 shows the growth in digital finance users and private enterprises in China, which have a similar tendency. This means that the two are likely to be related, which this paper tries to explain at the household level.

Based on the development of digital finance, the differences between users of digital finance are further highlighted. Digital financial capability is a new concept, which reflects the ability of individuals or households to use digital finance. Luo and Zeng defined DFC as the capability of households to use digital financial products and services [5]. The essence of DFC is the capability to allocate digital finance resources and meet their economic interests in the era of the digital economy. Luo and Zeng also verified the positive impact of DFC on household business ownership and business innovation.

Furthermore, Luo et al. discussed how DFC impacts entrepreneurial performance through sales, borrowing, and investment channels [6]. These preliminary studies have allowed a preliminary discussion on the concept of DFC and its impact, but they hardly reflect the true connotations of DFC. Especially in recent years, digital finance in China has experienced rapid growth, and research on entrepreneurship using cross-sectional data faces more endogenous problems. This paper attempts to re-measure DFC by using the panel data of a high-speed growth period of digital finance and discussing its impact on entrepreneurship.

Against this background and the literature gap, we intend to answer the following research questions in this paper.

RQ1: Does DFC have a positive impact on entrepreneurship in China? If there is a positive impact, from what mechanisms?

RQ2: Does DFC promote entrepreneurs to operate online?

RQ3: What is the heterogeneity of DFC’s impact on entrepreneurship?

To answer these research questions, this paper utilitises the China Household Finance Survey (CHFS) data to define and measure the DFCs of households. A probit model with two-way fixed effects is used to analyze the impact of DFC on entrepreneurship. Firstly, the positive impact of DFC on entrepreneurship is verified. Secondly, in terms of the impact mechanisms, DFC can enhance the accessibility of financial services and reduce credit constraints to households. It also expands the demand-side market, creates more opportunities, and improves aspiring entrepreneurs’ motivation and willingness. Thirdly, it confirmed that DFC promotes the development of digital operations without displacing traditional offline operations. Moreover, DFC has enhanced the online revenue of startups, proving the benefits of participating in the digital economy. Lastly, the analysis of heterogeneity confirms the inclusive role of DFC in promoting rural household entrepreneurship and entrepreneurship in the retail and restaurant sector.

2. Literature Review and Hypothesis

2.1. Literature Review

This paper mainly studies the impact of DFC on entrepreneurship and further discusses the mechanisms of DFC in promoting entrepreneurship. Therefore, this paper organizes the relevant literature from the following aspects.

The first is the general factors that affect entrepreneurship, which do not consider the influence of digital economy. At the micro level, the influence of an entrepreneur’s cognition on entrepreneurial behavior was first noticed [7,8,9]. The second aspect involves demographic characteristics, such as gender [10] and age [11]. Additionally, the resource endowments of entrepreneurs all have an impact on entrepreneurship, such as human capital level [12], work experience [13], and political relations [14]. At the macro level, macro variables such as property rights arrangement, regulation, and liquidity will also affect entrepreneurship [15,16,17].

The emergence of digital finance has profoundly affected the business environment in China. Even though China lags behind the United States in the development and cognition of digital economy, the development of digital finance in China has certain advantages [18]. The development of digital finance has stimulated scientific and technological innovation in the financial field, which better serves the long-tail market and reduces financial exclusion [19]. The existing literature also confirms that digital finance has improved the business environment in China [20] and promoted regional industrial upgrading through innovation and entrepreneurship [21]. Digital finance has improved the structure of human capital and the credit allocation in SMEs, so it has a positive impact on entrepreneurship [22,23].

However, based on the background of the development of digital finance, the differences in the ability of individuals or households to use digital finance should be paid attention to. Prasad et al. was the first to study DFC, which is defined as having relevant knowledge about digital payment and online banking [24]. Morgan et al. further proposed more comprehensive dimensions, including knowledge about digital finance, awareness of and control over digital finance risks, and knowledge about consumer rights and compensation procedures [25]. These advanced studies abroad emphasize individuals’ comprehension of digital finance and the use of knowledge in digital finance. In terms of the impact of DFC, Luo and Zeng verified the positive impact of DFC on household business ownership and business innovation [5]. Furthermore, Luo et al. discussed how DFC impacts entrepreneurial performance through sales, borrowing, and investment channels [6]. Meng found that DFC has a positive impact on Chinese household consumption by online business [26].

To summarise, the research related to entrepreneurship is very rich. With the explosive growth of digital finance, the effect of digital finance development on entrepreneurship has begun to receive attention. Even though the development of digital finance enhances the availability of regional financial services, there may be significant differences in the capability of individuals or households to use digital finance, which can be reflected in their DFCs. In addition, although the existing literature focuses on the role of digital finance in promoting entrepreneurship, against the background of the digital economy, there is still a lack of theoretical basis and empirical evidence on whether new enterprises participate in the digital economy and whether they present a different development trend from that in the past.

2.2. Hypotheses

2.2.1. DFC and Entrepreneurship

In summary, most of the literature suggests that financing can promote entrepreneurship. Financing significantly impacts the operation of startup enterprises [27]. For a long time, entrepreneurship has relied on traditional financial development. Small and medium-sized enterprises face difficulty and high costs when obtaining financing [28]. Therefore, equity financing, such as venture capital and angel investment, is important for startups to obtain financing [29]. In recent years, new financing instruments, such as microcredit and crowdfunding, have emerged to provide financing support for small and micro enterprises [30]. Finance can promote entrepreneurship by relaxing risk constraints and enhancing entrepreneurs’ adventurous spirit [31]. DFC refers to an individual or household’s ability to utilize digital finance to meet financial needs. It also serves as a manifestation of the maturity of digital finance at the micro level. This capability should have a positive impact on entrepreneurship. Based on this, research hypotheses proposed in this paper are listed below:

Hypothesis 1 (H1).

DFC has a positive effect on household entrepreneurship in developing countries.

Hypothesis 1a (H1a).

DFC has a positive effect on household entrepreneurship by enhancing financial access in developing countries.

In the digital age, the platform economy’s rise is one of the three landmark events of the “digital revolution”, which changed people’s production and lives [32]. Supply-side and demand-side users will interact with each other in the platform, which means that any increase in one side will result in an increase in the other [33,34,35]. On the demand side, digital finance enhances payment convenience for households and provides consumers with access to short-term credit for smoothing out inter-period financial needs, thereby releasing constrained liquidity for consumption [26]. The consumption market has expanded, creating more business opportunities and fostering household entrepreneurship. These emerging demands open up new opportunities for entrepreneurial families. Based on these, this paper proposes research hypothesis H1b:

Hypothesis 1b (H1b).

DFC has a positive effect on household entrepreneurship by expanding the consumption market in developing countries.

2.2.2. DFC and Online Entrepreneurship

The entry barriers are higher for the manufacturing sector, which must be operated offline [36]. Households can reduce operating costs through online operations. Otherwise, households who own a business benefit from the platform economy. The platform lowers search costs by matching buyers and sellers [37]. Trade efficiency is improved by the platform [38]. So, the enterprise value can be improved by platform innovation [39]. Therefore, households with better digital financial capabilities are generally more engaged in the platform economy by using digital finance, which can promote enterprise digitalization. Based on this, the research hypothesis H2 in this paper is proposed:

Hypothesis 2 (H2).

DFC has a positive effect on household online entrepreneurship in developing countries.

In order to verify these research hypotheses, this paper uses CHFS data from China, including the three years 2015, 2017, and 2019. Please refer to the next chapter for a detailed description of the data and research design.

3. Materials and Methods

3.1. Data Sources

The data utilized in this study are derived from the CHFSs conducted in 2015, 2017, and 2019. This survey collected micro-level information about households in China [40]. The CHFS encompasses 29 provinces and over 1400 communities in China. The survey includes initial data from about 40,000 households, most of which are part of tracking survey samples. These samples contain comprehensive information on the digital financial behaviors of households. The structure of the questionnaire can be acquired in Supplementary Material. Additionally, the survey extensively inquires about household business operations, including the type of business and its performance, thereby providing data support for this study. The second part of the data came from the Digital Financial Inclusion Index of China (DFIIC) published by the Digital Finance Research Center of Peking University. This index was jointly launched by the Digital Finance Research Center of Peking University and the Ant Financial Group, and it is the most representative index among the existing digital finance indexes in China [41]. We merged the DFIIC at the province level with the CHFS. After data cleaning and removal of missing values, a final sample of 90,992 valid cases was obtained, as shown in Figure 2.

3.2. Empirical Strategies

This paper establishes an empirical model of DFC and household entrepreneurship, which is a binary choice model with household entrepreneurship as the dependent variable. A dummy variable represents the entrepreneurial behavior of the household. The survey only observes whether the household starts a business. Therefore, a continuous latent variable exists behind the household’s decision to own a business. Because the household’s choice to start a business brings revenues and costs, the latent variable may represent the net revenues or net utility generated by the household’s decision to engage in business. When the household chooses to start a business, and the income exceeds the cost of starting a business, the latent variable is greater than 0 and the household opts for entrepreneurship; otherwise, the household chooses not to start a business. Therefore, in this paper, we employ the following probit model to estimate the impact of DFC on household entrepreneurship:

$$\begin{array}{c}{Entrepre}_{ijt}^{\mathrm{*}}={\beta }_0+{\beta }_1{DFC}_{it}+{\beta }_2{X}_{ijt}+{\mu }_j+{\tau }_t+{\epsilon }_{ijt}\\ {Entrepre}_{ijt}=1({Entrepre}_{ijt}^{\mathrm{*}}>0)\end{array}$$

(1)

Among them, ${Entrepre}_{ijt}^{\mathrm{*}}$ is the latent variable, ${Entrepre}_{ijt}$ is household entrepreneurship, which is 1 when the household manages an industrial and commercial project; otherwise, it is 0. ${DFC}_{it}$ refers to the household’s DFC, $X_{ijt}$ refers to the control variables, including the head of the household and the household and region characteristic variables. ${\mu }_j$ refers to the dummy variable of the region where the household is located, ${\tau }_t$ refers to the dummy variable of the year, and ${\epsilon }_{ijt}$ refers to the random error term. The ${\beta }_1$ coefficient measures the impact of digital financial capabilities on household entrepreneurship. Since the probit model is biased when individual effects are fixed, this paper only controls the fixed effects of provinces and years in the regression. Considering the correlation of families in the time series, the standard errors of the model are clustered at the household level.

3.3. Variables

3.3.1. Digital Financial Capability (DFC)

The primary independent variable in this paper is the household’s DFC. Based on previous research [5,6], and considering data availability, three key indicators were chosen to gauge the household’s DFC:

• “Digital payment” indicates whether the surveyed households utilized online banking or mobile payments, signifying their capacity for digital payments;
• “Digital financial management” denotes whether the surveyed households held digital financial management products, demonstrating their ability to engage in digital financial management and investments;
• “Digital credit” represents whether households accessed digital financial channels for borrowing when conducting operations or making purchases, including internet microloans, consumer finance, or P2P platforms.

Consequently, this paper employs the entropy weight method (EWM) to calculate household DFC using the selected indicators. The EWM is a widely used weighting method to calculate index according to the discriminating degree of indicators. The process of calculating the weight with the EWM can cover each value, which avoids the non-objectivity caused by the expert review. Lastly, the value of the household’s DFC ranges from 0 to 1.

3.3.2. Entrepreneurship

The dependent variable in this paper is entrepreneurship, which is measured by whether the household owns a business and represented as a dummy variable. Drawing upon prior literature [5,42], if the household is involved in industrial and commercial production and operational projects, including self-employed individuals or enterprises, entrepreneurship is denoted as 1; otherwise, it is represented as 0.

3.3.3. Control Variables

The control variables selected in this paper include ① characteristic variables of the head of the household, such as gender, age, education, marital status, political identity, health level, and risk attitude; ② household characteristic variables, such as household income, total assets, indebtedness, house ownership, household size, child ratio, and elderly ratio, and ③ regional characteristic variables, including the DFIIC for the province. Refer to Appendix A for the complete variable definitions. The result of VIF analysis is shown in

Table A2. VIF analysis.

Variable	VIF	1/VIF
DFC	1.32	0.759
Male	1.24	0.809
Age	2.33	0.429
Edu	1.61	0.620
Married	1.14	0.877
Party	1.12	0.891
Health	1.10	0.905
Risk	1.07	0.933
Ln_income	1.31	0.764
Ln_asset	1.74	0.575
If_debt	1.07	0.932
If_house	1.28	0.778
Scale	1.68	0.594
Child ratio	1.50	0.668
Elderly ratio	2.60	0.385
DFIIC	1.41	0.709
Mean VIF	1.44

.

3.3.4. Mechanisms Variables

The mechanism variables selected in this paper include the number of bank cards held by families, whether families hold credit cards, total household consumption, household online shopping consumption, whether families have business plans, and whether families take the initiative to start businesses. Refer to Appendix A for the complete variable definitions.

3.4. Descriptive Statistics

Table 1. Descriptive statistics of DFC constructing indicators.

DFC	All	2015	2017	2019
	Mean	Mean	Mean	Mean
Construction index:
Digital payment	0.3543	0.2089	0.3285	0.5607
Digital financial management	0.0716	0.0669	0.0696	0.0798
Digital credit	0.0026	0.0000	0.0002	0.0086

displays the descriptive statistics for the constructed DFC indicators. Over time, digital payment has become more prevalent. Concurrently, digital financial management and digital credit have been growing in China.

The descriptive statistics for the full sample studied in this paper are presented in

Table 2. Descriptive statistics of the full sample.

Variables	N	Mean	Std. Dev	Min	Max
Entrepreneurship	90,992	0.148	0.355	0	1
DFC	90,992	0.050	0.094	0	1
Male	90,992	0.590	0.492	0	1
Age	90,992	53.261	13.248	18	80
Edu	90,992	9.631	3.998	0	22
Married	90,992	0.868	0.339	0	1
Party	90,992	0.154	0.361	0	1
Health	90,992	3.109	1.044	1	5
Risk	90,992	2.631	1.560	1	5
Ln_income	90,992	10.475	1.973	0	16.311
Ln_asset	90,992	12.788	1.744	0	20.414
If_debt	90,992	0.378	0.485	0	1
If_house	90,992	0.910	0.286	0	1
Scale	90,992	3.317	1.573	1	20
Child ratio	90,992	0.104	0.157	0	0.833
Elderly ratio	90,992	0.278	0.376	0	1
DFIIC	90,992	238.844	53.197	145.930	377.734
Account	90,992	2.337	2.812	0	400.000
Credit	90,992	0.201	0.401	0	1
Ln_consump	90,992	10.732	0.875	5.53339	14.795
Ln_online	90,992	2.479	3.606	0	13.017
If_plan	90,992	0.090	0.286	0	1.000
Initiative	90,992	0.103	0.304	0	1
Time trend(year):
2015	90,992	0.352	0.478	0	1
2017	90,992	0.355	0.479	0	1
2019	90,992	0.293	0.455	0	1

Notes: The numbers of samples from 2015 to 2019 are 32,029, 32,302, and 26,660 respectively.

. Regarding the dependent variable, 14.8% of households opted to initiate their businesses. The average value of the independent variable “DFC” is 0.05. When we combine this with the statistical analysis of the indicators used to construct DFC in Table 1, it becomes evident that the DFCs among households in China remain relatively low.

4. Results

4.1. DFC and Household Entrepreneurship

Table 3. Impact of DFC on household entrepreneurship.

Entrepreneurship	(1)	(2)	(3)	(4)
	Probit	Probit	Probit	Probit
DFC	0.4238 ***	0.2763 ***	0.0777 ***	0.0773 ***
	(0.0120)	(0.0125)	(0.0121)	(0.0121)
Male		0.0030	0.0054 **	0.0053 **
		(0.0026)	(0.0025)	(0.0025)
Age		−0.0044 ***	−0.0026 ***	−0.0026 ***
		(0.0001)	(0.0001)	(0.0001)
Edu		−0.0022 ***	−0.0085 ***	−0.0085 ***
		(0.0004)	(0.0004)	(0.0004)
Married		0.0740 ***	0.0194 ***	0.0193 ***
		(0.0042)	(0.0042)	(0.0042)
Party		−0.0337 ***	−0.0389 ***	−0.0388 ***
		(0.0040)	(0.0038)	(0.0038)
Health		0.0082 ***	0.0029 **	0.0029 ***
		(0.0011)	(0.0011)	(0.0011)
Risk		0.0051 ***	0.0019 **	0.0018 **
		(0.0008)	(0.0008)	(0.0008)
Ln_income			0.0013 *	0.0013 *
			(0.0007)	(0.0007)
Ln_asset			0.0651 ***	0.0651 ***
			(0.0012)	(0.0012)
If_debt			0.0254 ***	0.0254 ***
			(0.0024)	(0.0024)
If_house			−0.1322 ***	−0.1321 ***
			(0.0050)	(0.0050)
Scale			0.0167 ***	0.0167 ***
			(0.0009)	(0.0009)
Child ratio			−0.0092	−0.0094
			(0.0088)	(0.0088)
Elderly ratio			−0.0999 ***	−0.1000 ***
			(0.0057)	(0.0057)
DFIIC				0.0600 ***
				(0.0221)
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	90,992	90,992	90,992	90,992
Pseudo R2	0.0367	0.0726	0.1571	0.1572

Notes: Cluster standard errors are given in brackets. *** p < 0.01, ** p < 0.05, * p < 0.1. Marginal effects of the probit model are reported. In the subsequent tables, this paper added all the control variables in column (4).

presents the primary findings regarding the influence of DFC on household entrepreneurship. After progressively incorporating and controlling for head characteristic variables, household characteristic variables, and regional characteristic variables, the marginal effect of DFC is 0.0773, and it is statistically significant at the 1% level (p < 0.01), as shown in column (4). This suggests that households with better digital financial capabilities are more inclined to choose entrepreneurship. Hence, H1 has been confirmed.

Regarding the control variables, male heads of households are more likely to own their own businesses. The impact of age on entrepreneurship is significantly negative. Married heads of households are also more likely to own a business. The impact of education level and party membership on the entrepreneurship is significantly negative. A high level of health can also encourage households to choose entrepreneurship. Risk-prone householders are more likely to start a business and are willing to undertake more entrepreneurial risks, which is consistent with conclusions in the existing literature.

Regarding household-related control variables, household income and asset levels positively encourage household entrepreneurship choices because they provide families with more entrepreneurial resources. Household debt shows a positive correlation with entrepreneurship. However, house ownership may reduce the likelihood of entrepreneurship. A larger household size promotes household entrepreneurship because it provides a sufficient labor force. In contrast, elderly members within a household are more likely to inhibit household entrepreneurship.

At the regional level, the development of digital inclusive finance positively impacts household entrepreneurship.

4.2. Influence Mechanisms

According to theoretical analysis, digital finance can enhance households’ access to financial services and alleviate credit constraints. It can also expand the consumer market and promote household entrepreneurship from both the supply and demand sides. The test results are presented in

Table 4. Mechanisms of DFC affecting household entrepreneurship.

Influence Mechanisms	Account	Credit	Ln_consump	Ln_online	If_plan	Initiative
	(1)	(2)	(3)	(4)	(5)	(6)
	FE	Probit	FE	FE	Probit	Probit
DFC	6.0869 ***	0.4966 ***	0.8159 ***	11.0315 ***	0.1462 ***	0.1644 ***
	(0.1421)	(0.0132)	(0.0264)	(0.1693)	(0.0098)	(0.0449)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes	Yes	Yes
N	90,992	90,992	90,992	90,992	90,992	13,475
Pseudo R2	0.2181	0.2783	0.4677	0.4151	0.1088	0.0601

*** p < 0.01.

to empirically test the mechanisms described above.

From the supply side, DFC can increase access to financial services and ease credit constraints for entrepreneurial households. The number of bank cards households hold can somewhat reflect their access to financial services. Furthermore, whether a household possesses a credit card reflects its borrowing eligibility and associated financial status. Columns (1) and (2) in Table 4 report the estimation results, demonstrating that DFC enhances households’ access to financial services and significantly reduces the probability of credit constraints.

From the demand side, DFC can boost household consumption, expanding the consumer market, increasing entrepreneurial opportunities, and further promoting household entrepreneurship. This paper utilizes household consumption expenditure and online shopping expenditure data, applying logarithmic transformations to assess the promotional effect of DFC. The estimation results are presented in columns (3) and (4) in Table 4. DFC promotes household consumption, with a more pronounced impact on online consumption.

The estimation results in Table 4, column (5) indicate that DFC significantly increases the probability of households planning to start a business and their willingness to do so. Additionally, for families that have already initiated businesses, the questionnaire explores their entrepreneurial motivations in detail. The estimation results in Table 4, column (6) demonstrate that DFC substantially promotes households to launch businesses. This implies that improved DFCs reduce startup costs for households, augment startup opportunities, and enhance household willingness to engage in entrepreneurial activities proactively, thus verifying research hypotheses H1a and H1b.

4.3. Robustness Checks

4.3.1. Endogenous Discussion

Additionally, the household’s DFC in the model is treated as exogenous. Although other factors have been controlled for to the best extent possible in the model, there may still be endogeneity issues with the DFC variable. On the one hand, the presence of unobservable factors, such as household social resources, beliefs, or macroeconomic changes, may lead to biased estimates due to the problem of missing variables. On the other hand, the reverse causal relationship between DFC and household entrepreneurship cannot be disregarded. Some self-employed families may use mobile payment to enhance the convenience of transactions during their operations, thereby deepening their use of digital financial instruments and improving their DFC.

This paper employs two-way fixed effects and instrumental variables to address endogeneity issues. First, the two-way fixed-effect model can mitigate the influence of unobservable factors not included in the model, thus reducing estimation errors. Second, the instrumental variable method is used to further mitigate endogeneity and enhance the estimation results’ reliability. Drawing inspiration from previous researchers’ work, this paper employs “whether families use mobile phones” as the instrumental variable [5,42]. As families engage in digital finance through mobile phones, their use of mobile phones is related to their DFC. Furthermore, the use of mobile phones by families does not directly impact household entrepreneurial behavior. Therefore, the instrumental variable is theoretically effective.

The estimated results are presented in

Table 5. Endogenous discussion: IV and household FE.

Entrepreneurship	Stage One	Stage Two
	(1)	(2)	(3)	(4)
	FE	IVProbit	Household FE	Household FE + 2sls
Have a mobile phone	0.0096 ***
	(0.0011)
DFC		1.5159 ***	0.0749 ***	0.8694 **
		(0.1624)	(0.0201)	(0.4070)
Controls	Yes	Yes	Yes	Yes
Household FE	No	No	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	90,992	90,992	59,037	59,037
Pseudo R2	0.653	0.0751	0.754	0.754
KP WALD-F Statistic	-	1672.55 ***	-	125.681 ***
DWH Test	-	134.99 ***	-	3.996 **

*** p < 0.01, ** p < 0.05.

. A weak instrument test was conducted, confirming the absence of weak instrument problems. All models showed that DFC has a positive impact on entrepreneurship. Consequently, with the endogeneity issue resolved, this paper’s conclusion remains valid.

4.3.2. Dealing with Reverse Causal Errors

The issue of reverse causality persists in the model. During their operations, startup households may choose to actively employ digital financial tools to enhance payment efficiency or do so passively in response to customer payment demands, consequently improving their DFC. To address this, the definition of entrepreneurship is modified, focusing solely on households that used mobile phones before embarking on their entrepreneurial journeys.

The estimation results are presented in column (1) of

Table 6. Dealing with reverse causal errors.

Entrepreneurship	(1)	(2)
	Probit	Probit
DFC	0.0773 ***	0.0514 **
	(0.0121)	(0.0248)
Controls	Yes	Yes
Province FE	Yes	Yes
Time FE	Yes	Yes
N	57,266	33,413
Pseudo R2	0.1093	0.1519

*** p < 0.01, ** p < 0.05.

, revealing a reduction in the estimation coefficient compared to the benchmark regression, though it remains statistically significant. Furthermore, the DFC of households is lagged by one period to mitigate estimation errors attributable to partial reverse causality. The results are displayed in column (2) of Table 6, and they consistently support the robustness of the findings. It is evident that even after considering and addressing the issue of reverse causal bias, the conclusion regarding the impact of DFC on household entrepreneurship remains valid.

4.3.3. Change of Independent Variables

This paper examines the robustness of the conclusions by employing alternative independent variable methods.

First, DFC is constructed using the simple addition method and the estimation results are presented in the first column of

Table 7. Replaced independent variable.

Entrepreneurship	(1)	(2)	(3)	(3)
	Probit	Probit	Probit	Probit
DFC	0.0262 ***	0.0476 ***	0.0195 ***	0.0327 ***
	(0.0021)	(0.0028)	(0.0022)	(0.0027)
Controls	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	90,992	90,992	90,992	13,475
Pseudo R2	0.1588	0.1607	0.1578	0.1588

*** p < 0.01.

, indicating a significant positive coefficient. Next, samples with a non-zero DFC score are defined as possessing financial capability (assigned a value of 1), while those with a score of 0 are categorized as lacking financial capability (assigned a value of 0). A dummy variable representing household DFC is then utilized for re-estimation. The results are displayed in column (2) of Table 7, demonstrating that the conclusion remains stable.

Lastly, the same DFC measure is employed, but the construction method of the indicator is altered. Household DFC is reconstructed using principal component and factor analysis, respectively. The estimation results are presented in columns (3) to (4) of Table 7, and they continue to indicate a significantly positive impact. This confirms the robustness of the earlier conclusion.

4.4. DFC and Online Entrepreneurship

4.4.1. Online Operating

Considering the household’s type of entrepreneurship, DFC may have different impacts on household participation in offline and online entrepreneurship. In the questionnaire, household entrepreneurship types are differentiated as “offline entrepreneurship”, “online entrepreneurship”, and “online and offline integrated entrepreneurship”. The types of household entrepreneurship are distinguished and examined in this section to investigate this issue. If the household engages in any of these types, the corresponding dummy variable is set to 1; otherwise, it remains at 0.

Table 8. Impact of DFC on household entrepreneurship types.

Business Type	Offline	Online	Integrating	Digitalization
	(1)	(2)	(3)	(4)
	Probit	Probit	Probit	Probit
DFC	0.0099	0.0131 ***	0.0170 ***	0.1972 ***
	(0.0117)	(0.0019)	(0.0022)	(0.0201)
Controls	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	90,992	90,992	90,992	13,475
Pseudo R2	0.1335	0.1390	0.1843	0.1304

*** p < 0.01.

presents the impact of DFC on households’ participation in offline entrepreneurship, online entrepreneurship, and integrated entrepreneurship, respectively. The estimation results reveal that the effect of DFC on households’ participation in offline entrepreneurship is not significant. However, it promotes households’ participation in online entrepreneurship and dual-line integrated entrepreneurship.

4.4.2. Online Revenue

Moreover, does DFC positively impact the business performance of entrepreneurial households? This section employs a fixed-effect model to select a sample of households that have initiated entrepreneurial endeavors. The objective is to estimate the operating income derived from household entrepreneurial projects and assess the influence of DFC on the operational performance of these households. The estimation results are presented in

Table 9. Impact of DFC on household entrepreneurial performance.

Operating Performance	Operating Revenue	Online Operating Revenue
	(1)	(2)
	FE	FE
DFC	0.9262 ***	2.6548 ***
	(0.3474)	(0.3560)
Controls	Yes	Yes
Province FE	Yes	Yes
Time FE	Yes	Yes
N	13,475	13,475
R2	0.059	0.071

*** p < 0.01.

. The findings indicate that DFC exerts a significant positive effect on household operating income and online operating income, with a more pronounced impact on online operating income than overall operating income.

Consequently, the research hypothesis H2 has been substantiated.

4.5. Heterogeneity

4.5.1. Heterogeneity of Urban and Rural

Table 10. Heterogeneity: urban and rural.

Entrepreneurship	Entrepreneurship	Offline	Online	Integrating
	(1)	(2)	(3)	(4)
	Probit	Probit	Probit	Probit
Panel A: Urban household
DFC	0.0685 ***	−0.0044	0.0165 ***	0.0195 ***
	(0.0142)	(0.0137)	(0.0026)	(0.0029)
Controls	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	63,340	63,340	63,340	63,340
Pseudo R2	0.1606	0.1410	0.1306	0.1682
Panel B: Rural household
DFC	0.2173 ***	0.1590 ***	0.0062 **	0.0146 ***
	(0.0269)	(0.0244)	(0.0031)	(0.0034)
Controls	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	27,652	27,653	27,654	27,655
Pseudo R2	0.1874	0.1708	0.1336	0.2715

*** p < 0.01, ** p < 0.05.

analyzes the impact of DFC on entrepreneurship in urban and rural households. The estimation results indicate that, overall, the promotional effect of DFC on rural household entrepreneurship is greater than that of urban households. Digital finance has ameliorated inequality in opportunities, particularly for rural households, demonstrating its inclusive nature. Enhanced DFC enables them to better access the inclusive benefits offered by digital finance.

Furthermore, a heterogeneity analysis was conducted based on the entrepreneurial types of households, with the estimation results presented in columns (2) to (4) of Table 10. DFC has a greater impact on promoting offline entrepreneurship among rural households but exhibits no significant effect on urban households’ entrepreneurship. This phenomenon may be attributed to the higher cost of starting offline businesses in urban areas and the prevalence of digital finance. Moreover, for online entrepreneurship and integrated online and offline entrepreneurship, the promotion effect of DFC on urban households is stronger than on rural households. This difference may reflect the faster development of new digital economy formats in urban areas compared to rural regions, potentially introducing new disparities in opportunities.

4.5.2. Heterogeneity of Different Industries

In this section, the industry-specific impact of DFC on household entrepreneurship is analyzed. The estimated results are presented in

Table 11. Heterogeneity: different industries.

Entrepreneurship	Secondary Industry	Retail and Catering	IT	Others
	(1)	(2)	(3)	(4)
	Probit	Probit	Probit	Probit
DFC	−0.0020	0.0789 ***	0.0051 ***	0.0289 ***
	(0.0057)	(0.0113)	(0.0016)	(0.0083)
Controls	Yes	Yes	Yes	Yes
Province FE	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes
N	90,992	90,993	90,994	90,995
Pseudo R2	0.1044	0.0816	0.1361	0.1366

*** p < 0.01.

. The effect of DFC on households’ engagement in secondary industry entrepreneurship is not statistically significant. Notably, DFC has the most pronounced impact on promoting entrepreneurship in the retail and catering industries. This suggests that DFC plays a more substantial role in stimulating entrepreneurship related to consumer activities.

5. Discussion

The objective of this study was to research the impact of DFC on entrepreneurship and its influencing mechanisms. Furthermore, the impact of DFC on online entrepreneurship was researched.

RQ1 was answered by H1, H1a, and H1b. The DFC of a household was defined using the entropy method in this paper. The positive impact of DFC on entrepreneurship in China was proven, which is consistent with the existing literature [5,6]. However, we measured DFC in a different way from Luo and Zeng [5]. They regarded financial literacy as an integral part of DFC, but we think it will exaggerate the impact of DFC on entrepreneurship because households can improve their financial literacy through traditional finance. We measured DFC in the way that Kumar et al. defined digital financial experience, which can obtain a more accurate measurement [43].

This paper analyzed the influencing mechanisms from both the supply side and demand side. On the supply side, DFC can enhance the accessibility of financial services and reduce credit constraints for households that own a business. Compared with Luo et al. [6], we have increased the inspection of mechanisms on the demand side, proving that DFC also expands the consumer market, creates more opportunities, and improves the motivation and willingness of aspiring entrepreneurs. This provides new empirical evidence for the development of a platform economy.

RQ2 was answered by H2. This paper is among the first to study the impact of DFC on online entrepreneurship. The results show that DFC promotes the development of online operations without displacing traditional offline operations, increasing digital operating revenue. However, the existing literature has confirmed that experience and education in digital technology encourages entrepreneurs to start an online business and has a positive impact on entrepreneurs’ innovation [44,45,46], which can support our results.

This suggests that the positive impact of DFC on household entrepreneurship primarily occurs when households engage in the digital economy. The development of digital finance has fostered innovation and entrepreneurship in the digital economy, showcasing a trend of coordinated development between the two. Furthermore, it is worth noting that the promotional effect of DFC on household online business income significantly surpasses that on overall business income. Participation in the digital economy emerges as a more effective avenue for households to launch businesses, increase income, and accumulate wealth, which can prove the belief that digital finance can promote wealth accumulation [47].

These results affirm that DFC has stimulated household participation in the digital economy within the realm of entrepreneurship. It not only fosters digital innovation in business models but also yields higher revenue through digitalization. The benefits of participating in the digital economy are evident for enterprises.

RQ3 was answered by the heterogeneity analysis. The heterogeneity analysis confirmed the inclusive role of DFC in promoting rural household entrepreneurship and entrepreneurship in the retail and restaurant industries. The gap between urban and rural areas can be narrowed down by digital finance [48]. Existing literature has proven that DFC has a greater impact on entrepreneurship in rural areas [5,6], which is consistent with our results. Furthermore, we found that DFC is more inclined to promote urban households to start online businesses, but DFCs tend to promote rural households to start traditional businesses. This is partly contrary to the conclusions of research conducted in Germany [49], which argued that “platform-oriented firms” are relatively located in urban regions, but in rural regions “digital manufacturers” are relatively often found. We believe that this difference may come from the country’s development level. In China, the entrepreneurial projects that rural households can undertake are usually backward, although these projects may narrow the gap between them and urban households. In China, the development of the digital economy and platform economy provides new development opportunities for excluded groups, for whom e-commerce is the main way to participate.

5.1. Theoretical Contribution

This paper has provided an important contribution to the available literature on DFC and entrepreneurship. First, it defined DFC across three facets: digital payment, digital financial management, and digital credit, offering a fresh micro perspective on the comprehension of digital finance. The main work has proved the effect of DFC on entrepreneurship. Second, it substantiated the mechanisms through which DFC influences entrepreneurship, elucidating the dynamics of the supply and demand sides. Third, from the perspective of digital economy, this paper confirmed the positive impact of DFC on enterprise digitalization. Fourth, we analyzed the gap between urban and rural areas in terms of the influence of DFC on entrepreneurship. Fifth, we analyzed the influence of DFC on entrepreneurship in different industries. This paper found that DFC has a greater impact on entrepreneurial behavior in the retail and catering industries, which may be beneficial to improving the employment of residents. In addition, the influence of DFC on the secondary industry is not significant.

5.2. Limitations and Future Recommendations

This paper offers essential recommendations but has certain limitations that should be considered in future research. Firstly, this paper focused only on the digital financial usage behavior of households to measure DFC. This is because of the limitations of the questionnaire, which lacks dimensions such as knowledge and attitudes. More dimensions should be considered as indicators of DFC in the future, such as basic knowledge and skill, practical know-how, consumer awareness, etc. [50,51]. Secondly, the latest data from the CHFS have not been released yet because of COVID-19, which limits our conclusions to before the COVID-19 shock. We suggest that the 2021 CHFS be added to the sample when available to research the effect of COVID-19 on the economy. Last, but not least, this paper’s sample contains only Chinese households, which limits the findings to China.

6. Conclusions

This paper defines and measures households’ DFCs and analyzes the influence and mechanisms behind this capability on entrepreneurship. This paper arrives at the following research conclusions, with policy suggestions given in this regard.

Firstly, it measures DFC through three dimensions: digital payment, digital financial management, and digital credit, verifying the positive impact of DFC on entrepreneurship with the probit model with two-way fixed effects. It is essential to continue recognizing the vital role of digital finance in fostering digital economic growth. However, it should be noted that DFC is the basis for residents to use digital finance, and the government should strengthen the cultivation of digital skills for residents.

Secondly, in terms of the impact mechanism, DFC can enhance the accessibility of financial services and reduce credit constraints on households. It also expands the demand-side market, creates more opportunities, and improves aspiring entrepreneurs’ motivation and willingness. The government should encourage digital finance innovation to support SMEs in financing. For SME operators, they should pay attention to the financing channels of digital finance and grasp the potential opportunities of platform economy development.

Thirdly, it distinguishes enterprises’ engagement in digital operations, confirming that DFC promotes the development of digital operational models without displacing traditional offline operations. Moreover, DFCs have enhanced the business performance of startups, with increased digital operation income surpassing overall revenue. The benefits of participating in the digital economy are evident for enterprises. In order to improve the online business environment, other developing countries can learn from China’s successful experience, such as by paying attention to the construction of a digital infrastructure and improving the popularity of digital finance. Alternatively, they can reduce taxes on e-commerce to encourage small businesses and self-employed individuals to operate online.

Lastly, from an urban and rural perspective, across various industries, the analysis of heterogeneity confirms the inclusive role of DFC in promoting rural household entrepreneurship and entrepreneurship in the retail and restaurant industries. This research suggests that the government needs to pay attention to improving the sustainable development ability of vulnerable groups, especially rural residents and people with limited labor skills. DFC is the decisive factor allowing them to benefit from the growth of digital economy and share prosperity, which requires relevant policy support.