Can Digital Financial Inclusion Promote the Sustainable Growth of Farmers’ Income?—An Empirical Analysis Based on Panel Data from 30 Provinces in China

Abstract

The issue of farmers’ income is a widespread concern in countries worldwide, and the Chinese government has always prioritized promoting the sustainable growth of farmers’ income. The development of digital financial inclusion provides a new opportunity for farmers’ income to achieve sustainable growth. With the implementation of digital financial inclusion, whether it can effectively promote farmers’ income growth deserves in-depth study. Based on the panel data from 30 provinces in China from 2011 to 2021, this study uses a dual fixed effects model to empirically analyse the impact of digital inclusive finance on farmers’ income and further analyses the impact of various dimensions of digital financial inclusion on farmers’ income. From a policy perspective, the DID (difference in differences) method is used to analyse, in general, the impact of the implementation of digital financial inclusion policies on farmers’ income and, in particular, the impact of such inclusion on farmers’ income from the perspective of income structure. The results of this study show that digital financial inclusion can significantly promote farmers’ income growth. The dimensions of the breadth of coverage and depth of use can significantly contribute to the increase in farmers’ income, whereas digitization has a negative effect on this increase. Furthermore, the DID results show that digital financial inclusion policy implementation has a significantly positive effect on farmers’ income growth, that is, it can significantly contribute to their wage income, can contribute to family operating income but at a low level of significance, and does not significantly contribute to their property income. Moreover, regional heterogeneity analysis demonstrates that the marginal contribution of digital financial inclusion to the growth of farmers’ income in the eastern region is less than that in the central and western regions. Therefore, the development of digital inclusive finance in rural areas should be vigorously promoted in order to provide high-quality financial services and achieve sustainable growth in farmers’ incomes.

1. Introduction

How to promote the sustainable growth of farmers’ incomes is a widespread problem faced by countries worldwide. The issue of increasing farmers’ income is a priority for the Chinese government. How to promote its sustainable increase is also a subject of concern for scholars. The level of farmers’ income is a significant indicator of the level of agricultural development [1]. In the context of the Chinese government’s implementation of its rural revitalization strategy [2], research on increasing farmers’ income and finding the motivating factor to improve the sustainable growth of this income is imperative. One of the key elements in terms of increasing farmers’ income is capital [3,4]. One of the essential elements of agricultural economic growth and farmers’ earnings increase is capital. Owing to the high threshold, high cost, and constraints on the depth and breadth of the service coverage of traditional finance, farmers’ entry into formal financial services is limited and costly [5,6,7]. Therefore, improving the quality of rural financial services, lowering the threshold and cost, promoting equal opportunities for farmers to access formal financial services, and empowering farmers to increase their income are the keys to solving this problem.

The United Nations first introduced the concept of financial inclusion in 2005 [8]. The World Bank defines financial inclusion as offering a comprehensive and unrestricted system of financial services to every social group to alleviate financial exclusion [9]. Its definition is based on the principles of commercial sustainability and equal opportunity and the provision of suitable and efficient financial services at an affordable price to the population that needs financial services, according to the Plan for Promoting Inclusive Financial Development (2016–2020), released by the China State Council in 2015. The main beneficiaries of inclusive finance include small and microbusiness owners, farmers, urban low-income populations, and impoverished individuals [10]. Obviously, we can foresee the future prospects of rural financial inclusion, which is expected to be alleviated, injecting momentum into rural development and farmers’ income. When traditional finance was unable to provide for the needs of equitable opportunity, fairness, justice, and sustainable social development, inclusive finance was developed. The difference between the two lies in the following: first, inclusive financial development requires expanding the breadth of coverage, emphasizing equal opportunity, and highlighting that the service groups include farmers, low-income groups, and other disadvantaged groups, which were excluded in the traditional financial system owing to the high costs and high threshold of traditional finance. Second, when gauging development indicators, traditional finance often calculates the ratio of the financial scale to the overall scale of the economy [11]. The penetration of services, such as the number of banking accounts held by adults in formal financial institutions, is used to gauge inclusive finance [12]. Third, unlike traditional finance, inclusive finance places greater emphasis on social justice and fairness in social development as well as sustainable development by using market principles and sustainable business development to offer financial services to financially disadvantaged groups at reasonable prices.

International research on inclusive finance has mainly focused on measuring its development level and impacts on economic development and social income distribution. Beck et al. were the first to propose constructing an evaluation index for inclusive financial development from two aspects: the availability and use of financial services [13]. Sarma used availability, usage, and utility to measure the ’financial inclusion index’ [14,15,16]. Gupte et al. improved the inclusive financial index system and measured India’s financial inclusion index in 2008 and 2009 using four key dimensions: outreach (penetration and accessibility), usage, transaction convenience, and transaction costs. Chinese scholars have also made some achievements in the study of inclusive financial measurement [17]. Yang Yan measured inclusive finance from the two dimensions of availability and use. Under the availability index, the availability of geographical latitude and the availability of the population dimension are established. The use of the deposit dimension and the use of the loan dimension are set up under the usage index [18]. Chen Shufu constructed an index system from five dimensions, including financial service coverage, permeability, convenience, service quality, and regional characteristics [19]. Liu Yiwen et al. constructed an inclusive financial index system for China from the four dimensions of financial service penetration, service availability, use utility, and affordability, considering China’s national conditions. Internationally, the impact of inclusive finance on social income distribution and economic development has also been studied [20]. Claessens highlighted the importance of finance to economic development, emphasizing that universal access to financial services is necessary and that financial development can promote increases in farmers’ income [21]. Kapoor stated that financial development is conducive to people’s contribution to economic growth and enables them to benefit from it. Chinese scholars have also made some achievements in studying the impacts of inclusive finance on economic development and income [22]. Through empirical analysis, Ma Yufei et al. found that inclusive finance is generally conducive to poverty alleviation and indirectly affects poverty alleviation by promoting inclusive growth [23]. Zhu Yiming’s empirical research showed that inclusive finance is conducive to increasing rural residents’ income, but the effects of poverty reduction and income growth vary significantly for different income groups. Economic growth is an important mechanism for inclusive financial development to promote poverty reduction and income growth [24]. Tan Yanzhi and Peng Qianrui used the spatial Durbin model to examine the impact mechanism of inclusive financial development on poverty alleviation. They found that the level of regional inclusive financial development plays a significant role in promoting poverty alleviation. The level of financial inclusion development in neighboring areas and spatial relationships work together to produce positive spatial spillover effects on poverty alleviation [25]. International and Chinese scholars’ efforts to measure the development level of inclusive finance and its impacts on income and economic development provide a theoretical foundation for the research presented in this paper.

The “time” and “space” issues that financial services have faced have been resolved mainly by the progress of digital strategies, and financial inclusion has now reached the stage of digital development [26,27,28]. Digital financial inclusion (DFI) refers to the application of technology; that is, digital technology enables finance to serve more extensive groups and enhances financial inclusion. Disadvantaged populations, such as rural residents and small and micro companies, can have more opportunities to access formal financial services with reasonable costs, practical procedures, and increased efficiency under the guidance of market and sustainable principles. This situation helps lessen the degree of financial exclusion experienced by vulnerable groups. Its advantages include the following. First, by leveraging digital strategies, DFI may successfully address “last-mile financial services”. That is, impoverished farmers in distant places and other disadvantaged groups can obtain financial services that match their needs. Second, by leveraging technological and other advantages, DFI can lead to the creation of new financial products, encourage product diversification, better meet the needs of various financial entities, implement efficient resource allocation, and effectively spur development. The third advantage is the low-cost advantage, in which digital technology is used to expand economies of scale, greatly reducing the cost of services while enabling a wider range of groups, especially farmers in remote areas, to share the fruits of the development of fintech at an affordable price. Fourth, the emergence of DFI has improved data sharing and has the potential to employ fintech to increase the standard of financial management and oversight. As such, such finance can be “inclusive” while guaranteeing the orderly and healthy development of the financial market. Fifth, the evolution of DFI and its knowledge publicity can increase the financial literacy of financial subjects [29]. In addition, the publicity of financial knowledge, especially for remote areas, farmers, and other groups, can effectively increase the financial literacy of inhabitants. This is conducive to meeting the needs of their own financial services, promoting their own development, and preventing information barriers and capital and other problems associated with missed opportunities for development and constraints on their own development. Compared with traditional finance, DFI clearly offers the following benefits: broad coverage, low cost, equal opportunity, sharing, and convenience. Such finance increases the effectiveness of financial resource allocation, raising the standard for inclusive financial services available to farmers and low-income urban dwellers and encouraging an understanding of the “inclusive” aspect of inclusive finance [30,31].

International research on DFI and its impact on income and economic and social development has just begun. The main research results are as follows. Kumar et al. studied the development potential of the digital footprint generated by mobile phone use. Under the premise that the interests, privacy, security, and moral problems of consumption are solved, these data can be effectively used to develop inclusive finance [32]. Munyegera et al. highlighted that mobile money can solve the problem of Ugandan adults being unable to obtain formal financial services, especially in rural areas. It can promote the penetration and coverage of inclusive finance, facilitate remittances, and reduce the costs of financial transactions [33]. Durai et al. stated that DFI offers advantages such as affordability, safety, and convenience, ensuring access of low-income groups to financial services [34]. Marco López-Paredes and Andrea Carrillo-Andrade believe that various actions are sustainable under certain network structures. Chinese scholars have also made some achievements in the study of DFI and the impact of DFI on poverty reduction and income distribution [35]. Guo Feng et al. (2020) used the microdata of representative institutions in China to compile a Chinese DFI dataset as instrumental basic data [36]. Through empirical research, Song Xiaoling found that DFI can significantly narrow the urban–rural income gap [37]. Liu Jinyi et al. and others determined that DFI directly alleviated rural poverty through internet credit and insurance and indirectly alleviated rural poverty through individual and private employment [38]. In their research, Liang Shuanglu et al. concluded that DFI significantly narrowed the income gap between urban and rural residents through the threshold effect, poverty reduction effect, and exclusion effect [39]. Zhou Li et al. conducted an empirical analysis by constructing a household credit threshold model for urban and rural residents and using the MM (Machado and Mata) decomposition method based on quantile regression. They concluded that the development of DFI was conducive to narrowing the urban–rural income gap and the marginal effect was greater at low quantiles [40]. Wang Yongjing et al. used the spatial Durbin model to conclude that there were significant positive spatial correlations among China’s provincial DFI, new urbanization, and urban–rural income gap [41]. Based on an empirical analysis of China’s provincial data using a non-linear threshold regression model, Zhang He et al. confirmed that DFI can narrow the urban–rural income gap [42]. Wang Yingzi empirically examined the regional differences in the impact of DFI on the urban–rural income gap using multiple regression analysis and the TOPSIS comprehensive evaluation method [43]. International and Chinese scholars’ research on the development potential and advantages of DFI, as well as its effects on poverty, income, and economic and social development, has laid a theoretical foundation for the further discussion presented in this study.

DFI was first introduced in 2016 during the G20 Hangzhou Summit [44,45]. In 2019, the People’s Bank of China (PBC) published its Report on Chinese Rural Financial Services 2018, which noted that efforts have been made to develop DFI for rural areas and that closing the “digital divide” will require continued full reliance on digital technology [46]. The Outline of Digital Rural Development Strategy, released in 2019, defined ways to further innovate rural DFI, enhancing the conditions for its development, and offering rural residents convenient and excellent financial services [47]. With the implementation of guidelines and policies, rural DFI has been effectively developed. Its implementation has injected new kinetic energy into rural development. The income growth of farmers is an important indicator of their development. Researchers are particularly concerned with the connection between farmers’ income growth and DFI. Some academics contend that DFI increases rural inhabitants’ earnings and plays a crucial role in promoting rural income. It is widely believed that financial development helps reduce income inequality, emphasizing the importance of the financial system to low-income groups. In the long run, the higher the degree of financial development, the lower the degree of inequality. In addition to promoting growth, financial development also reduces inequality [13,48,49]. Other scholars believe that rural finance has not had a substantial effect on farmers’ earnings growth. It is widely believed that the financial inclusion index does not significantly affect economic growth. In general, the impact of rural financial development on farmers’ income is very limited [50,51,52]. Moreover, some scholars have noted that owing to the lack of digital technology and financial literacy, low-income groups and farmers who have no access to the internet are unable to increase their incomes through digitally inclusive financial services, resources are crowded out, and the numbers of employment opportunities for digitally disadvantaged groups are reduced, which has negative effects on the labour market [53,54]. According to studies on how DFI affects rural residents’ earnings growth, it does so by encouraging economic development and having an impact on entrepreneurial activity [55,56,57,58]. The current academic research on this topic draws the above conclusions.

In summary, according to the current research, there are different conclusions regarding whether DFI can improve farmers’ income, but the advantages and development value of such finance are recognized. Moreover, there have not been many studies on the impact of DFI on farmers’ income, and few studies have considered a multi-dimensional perspective on DFI and the structure of farmers’ income. In the policy direction, specific research on the impact of such policy on farmers’ income has not been carried out in depth. Therefore, considering the above analysis, the possible contributions of this research are as follows. First, this research uses a longer data time span. By adopting provincial panel data from 2011 to 2021, important time points such as the time of implementation of DFI policies and the COVID-19 period are included, allowing for a more comprehensive and in-depth analysis of the topic. Second, the DID (difference in differences) method is used to study this topic, and an exogenous policy shock event is used to study the problem of an increase in farmers’ income. Third, from a research perspective, an analysis of the impact of DFI on farmers’ income from a policy perspective is included. In the empirical research, a multi-dimensional perspective of DFI and an analysis of farmers’ income structure are included.

The research contents of this paper are as follows. The first part is the introduction and literature review. The second part covers the theoretical analysis and hypothesis. The third part is the methods and materials, including model design, variable description, and data sources. The fourth part presents the empirical analysis, including descriptive statistics, benchmark regression, robustness test, and further heterogeneity analysis. The fifth part is the policy effect analysis and robustness test. The sixth part is the conclusion and suggestion.

2. Theoretical Analysis and Hypotheses

Leyshon and Thrift first proposed the financial exclusion theory, which describes how certain people are excluded from financial services, in 1993 [8]. Kempson and Whyley categorized financial exclusion into the following six types: geographic, appraisal, conditional, price, marketing, and self [59]. Geographic exclusion refers to the fact that financial institutions do not open branches in isolated locations for profitability reasons. In addition, residents of remote areas are excluded. Appraisal exclusion means that owing to farmers’ low income and low degree of risk tolerance, banks prefer to provide financial services to customers with high appraisal values. Conditional exclusion refers to groups, such as low-income groups like farmers, who are excluded because they do not meet the loan conditions set by financial institutions. Price exclusion refers to groups that are excluded because the cost of accessing financial services is too high. Marketing exclusion means that financial institutions are profit-oriented and that target customers are usually groups with better economic strength, whereas groups with worse economic strength are excluded from financial services. Self-exclusion refers to groups that automatically exclude themselves from financial services owing to a lack of financial knowledge. On the basis of an analysis of such theory, Chinese rural financial exclusion is widespread, and farmers are generally excluded from financial services. Lewis put forward the theory of a dual economic structure in 1954, pointing out the economic structure of traditional agriculture and modern industry in developing countries. Owing to backwards agricultural technology and low-level production efficiency, backwards agricultural development and a surplus rural labour force exist. Modern industrial technology is advanced, with high-level productivity and high wages, resulting in a wide urban–rural wealth gap [60]. Currently, China has a dual economic structure in which both urban and rural areas coexist. Farmers’ income has been at a low level, and the income gap in the country is enormous. US scholar Simon Smith Kuznets proposed the theory of an inverted U-shaped curve in the 1950s, pointing out that as the economy develops, the wealth gap initially expands and then narrows. However, since China’s reform, the speed of economic growth has been fast, the income gap has grown, and the income level of farmers remains low [61,62].

China has long experienced unbalanced urban–rural and regional economic conditions, and farmers’ income levels need to be improved. According to the above theory, DFI is naturally inclusive due to its low-threshold and low-cost features, which can effectively address financial exclusion. Such inclusion can overcome geographical restrictions and alleviate geographical exclusion by means of internet technology. Therefore, farmers in isolated places can easily enjoy financial services. Such inclusion can also successfully lower the operating costs of financial institutions, make banks willing to lend to farmers with low economic strength, and achieve true inclusiveness. This effectively alleviates the four exclusions of appraisal, conditional, price, and marketing. The development of the internet is conducive to the popularization of financial knowledge, improving farmers’ financial literacy and potentially alleviating farmers’ self-exclusion. Thus, the digital dividend can benefit more disadvantaged groups, including farmers, reduce the urban–rural divide, solve the problem of binding their own development due to financial constraints, enable farmers to obtain more opportunities for development, promote the employment and entrepreneurship of farmers, and increase farmers’ income [63,64]. Furthermore, DFI offers broad coverage, high efficiency, and a high level of sharing, making it possible to extend financial services to farmers. Such inclusion can provide affordable, formal, and effective financial services to farmers while simultaneously enhancing the effectiveness of communicating information, encouraging the flow of rural inputs for production, assisting in the growth of agriculture, and successfully increasing farmers’ income [65,66]. In addition, DFI using digital techniques and other financial technologies can allow for innovations in financial products to be realized, and the existence of various products and services can better fulfil the diverse demands of different groups [67]. Examples of such financial products and services are those that specifically target agricultural development, farmers’ groups, and green agriculture, which can effectively improve resource allocation efficiency and serve agriculture and farmers more accurately. Accordingly, the following hypothesis is proposed:

Hypothesis 1.

Digital inclusive finance can contribute to the growth of farmers’ income.

The theory of institutional innovation was developed by US economists D. North, Lance E. Davis, and Robert P. Thomus. The establishment of an appropriate personal incentive system is conducive to economic development. The key is that this system guarantees economic growth. When economic development stagnates, we should identify the reasons for such stagnation in the system and update the system to promote economic development [68,69]. Therefore, to encourage the Chinese rural economy, we should innovate the rural economic system. Under the framework of the economic system, pertinent policies should be formulated and implemented. The theory of financial constraints was proposed by economist Thomas Hellman, who argued that excessive financial liberalization exacerbates financial risk and that governments should take appropriate macro control interventions. The government’s macro control should grasp the principle of moderation, formulate reasonable policy regulations following the actual situation, and promote development. Therefore, because of the lower income level of farmers in China, difficulties, such as widespread financial exclusion, are faced in the country’s development. DFI provides an opportunity to address this problem. On this basis, the government should formulate policies to create rural DFI, solve the problems of financial exclusion and capital constraints for farmers’ development, promote farmers’ development, and increase farmers’ income. Therefore, the Chinese government has issued policies to create rural DFI.

The implementation of this DFI policy has largely accelerated rural digital financial infrastructure construction and its services. In addition, the availability, utilization, and quality of rural finance should be improved. This policy has provided accurate and effective financing for economic development in rural areas, helping revitalize the countryside and thus driving up farmers’ income growth. On the one hand, in accordance with the China Rural Financial Services Report 2018, released by the PBC in 2019, the rural financial credit system has been increasingly improved, credit files have been established for 2.61 million microenterprises and 184 million rural households, and in rural areas, everyone has a bank account, there are ATMs in townships and villages, there are POSs in villages, and rural digital financial products have been continuously enriched and upgraded. Finance has improved with the help of digital techniques, effectively reducing the degree of information mismatch, thus also reducing the cost of risk control and transaction expenses [46]. On the other hand, the key to development lies in the capital, which has been vigorously promoted by the implementation of the government’s policy on rural DFI, alleviating the capital constraints faced by rural economic development, farmers’ employment, and entrepreneurship for a long period [70,71]. Furthermore, farmers’ levels of financial literacy have increased owing to the digital inclusive financial policy for the countryside as well as the publicizing and promoting of digital inclusive financial knowledge in these areas. This situation has improved farmers’ ability to use DFI products more effectively, thus allowing them to share in the benefits of development. Moreover, this policy uses digital technology to overcome financial exclusion in rural areas and improve fund utilization [72,73]. The advantages of equal opportunity and the sustainable development of DFI should be fully utilized. In summary, DFI policies have a positive impact on farmers’ income growth by promoting the accessibility of financial services, promoting digital platforms, supporting farmers’ entrepreneurship, and improving farmers’ financial literacy. The theoretical framework of this paper is shown in Figure 1.Accordingly, the following hypothesis is proposed:

Hypothesis 2.

The implementation of digital inclusive finance policies as a macro policy tool significantly enhances the effectiveness of digital financial services in promoting farmers’ income.

3. Methods and Materials

3.1. Model Construction

In this study, a dual fixed effect model is constructed to estimate the impact of DFI on farmers’ income as follows:

Y_it = a0 + β₁IFI_it + β₂CV_it + u_i + λ_t + ε_it

(1)

where i represents the province, t represents the year, Y_it denotes farmers’ income, IFI_it denotes DFI, and CV_it is a set of control variables, including farmers’ education level (EDU), industrial structure (IS), rural investment level (INV), agricultural technology level (TEC), government fiscal expenditure (GE), and economic development level (GDP). β₁ and β₂ are the regression coefficients of the explanatory variables and the control variables, respectively. u_i is the individual effect, λ_t is the time effect, and ε_it is the random disturbance term.

To further analyse the impact of exogenous policy shocks—specifically, the DFI policy implementation in rural China on Chinese farmers’ income—the DID (difference in differences) method is adopted as follows:

Y_it = a₀ + β₁ (Treat_i * Post_t) + β2CV_it + u_i + λ_t + ε_it

(2)

where i represents the province, t represents the year, Y_it represents farmer’s income, Treat_i is a grouped dummy variable (treatment group = 1, control group = 0), Post_t is a staged dummy variable (after policy implementation = 1, before policy implementation = 0), and Treat_i * Post_t is the interaction term, which is the effect of the treatment group after policy implementation. β1 is the regression coefficient of the interaction term, CV_it is a set of control variables, β2 is the regression coefficient of the control variables, u_i is the individual effect, λ_t is the time effect, and ε_it is the random disturbance term. This paper uses Stata 17.0 software for empirical analysis.

3.2. Description of Variables

The definitions of the variables of this paper are shown in

Table 1. Definition of variables.

Variable Type	Variable Name	Variable Symbol	Variable Construction
Explained variable
	Level of farmers’ income	Y	Farmers‘ per capita disposable income, taken as a logarithm
Explanatory variable
	Digital financial inclusion	DFI	Digital financial inclusion Index, taken as a logarithm
Control variables
	Educational attainment of farmers (rural human capital)	EDU	Average education level of the rural population
	Industrial structure	IS	Proportion of the primary industry’s added value to the region’s GDP
	Rural investment level (investment rate)	INV	Ratio of rural fixed asset investment to the total rural population (rural per capita investment) represents the level of investment
	Agricultural technology level (agricultural mechanisation rate)	TEC	Overall power of agricultural machinery per hectare
	Government expenditure	GE	Proportion of agriculture, forestry, and water expenditure in the local general public budget expenditure
	Economic development level	GDP	Measured by calculating every province’s logarithm of per capita GDP

.

3.2.1. Explained Variable: Farmers’ Income (Y)

Farmers’ income is measured by the per capita disposable income of farmers and is taken as a logarithmic value, spanning the period of 2011–2021. The National Bureau of Statistics of China (NBS of China) used the net income of rural residents to measure the level of farmers’ incomes in 2011 and 2012. However, in 2013, it began to change the statistical caliber, measuring the level of farmers’ incomes in terms of the per capita disposable income of rural residents. Since the method of calculation has not changed significantly, the level of farmers’ income in 2011–2012 has been replaced by the net income of rural residents.

3.2.2. Explanatory Variable: Digital Financial Inclusion (DFI)

The DFI index, released by the Institute of Digital Finance (IDF) of Peking University, was used as the explanatory variable and is logarithmically transformed. This index also measures the DFI from coverage breadth (COV), use depth (USE), and digitized degree (DIG) [36]. The time span is 2011–2021.

3.2.3. Control Variables

The control variables selected are as follows: (1) Economic development level (ED): Measured by the per capita GDP, taking the logarithmic value; (2) Farmers’ educational attainment (EDU): Expressed by the average education level of the rural population, specifically calculated using the number of years of schooling as a weight (elementary school 6 years, junior high school 9 years, secondary and senior high school 12 years, and tertiary education and above 16 years); (3) Industrial structure (IS): Expressed by the proportion of the primary industry’s added value to the region’s GDP; (4) Rural investment level (INV): Measured by the per capita investment in rural; (5) Agricultural technology level (TEC): Measured by the rate of agricultural mechanization; (6) Government fiscal expenditure (GE). Expressed by the ratio of agricultural, forestry, and water expenditures in the local general public budget expenditure.

3.3. Data Sources

This paper selects the panel data from 30 provinces in mainland China (due to the deficiency of Tibetan data, it is not used as a sample for the empirical investigation) from 2011 to 2021 as the sample. The data are mainly from the NBS of China, the China Statistical Yearbook, and the China Rural Statistical Yearbook, which are collated and calculated.

4. Results and Discussion

4.1. Descriptive Statistics

Table 2. Summary Statistics.

VarName	Obs	Mean	SD	Min	Median	Max
InY	330	9.410	0.414	8.361	9.403	10.559
lnDFI	330	5.283	0.669	2.909	5.474	6.129
lnINV	330	7.344	0.556	4.423	7.443	8.491
lnGDP	330	10.831	0.451	9.682	10.792	12.142
lnTEC	330	1.812	0.351	0.923	1.786	2.630
IS	330	0.098	0.053	0.002	0.098	0.258
EDU	330	7.816	0.613	5.861	7.859	9.910
GE	330	0.114	0.033	0.041	0.114	0.204

shows the descriptive statistics of the variables. “InY” reflects the imbalance in farmers’ income levels across provinces, and “InDFI” reflects the imbalance in DFI levels across provinces. The standard deviations (SDs) of these two variables are 0.414 and 0.669, respectively. This finding indicates that the degree of dispersion of the two variables is large, and that the degree of dispersion of DFI is even greater. This suggests that the development of DFI across Chinese provinces is more unbalanced. This disparity can be attributed to China’s vast territory and unbalanced area development due to geographical factors, resource endowments, and other reasons. The implementation of the Western Development and Poverty Alleviation programmes has relatively narrowed the development gap across regions. In particular, the success of poverty alleviation has eliminated absolute rural poverty, and thus, farmers’ income disparities have relatively narrowed [74,75]. However, rural DFI has only recently emerged, and the infrastructure and financial products and services in the central and western regions are very backwards compared with those in the developed eastern region [76,77].

4.2. Analysis of Benchmark Model Estimation Results

To estimate the impact of DFI on farmers’ income, a dual fixed effects model is utilized.

Table 3. Benchmark regression results.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	lnY	lnY	lnY	lnY	lnY	lnY	lnY
lnDFI	0.073 ***	0.039 ***	0.040 ***	0.041 ***	0.040 ***	0.040 ***	0.039 ***
	(6.81)	(4.68)	(4.80)	(4.93)	(4.86)	(4.85)	(4.46)
lnGDP		0.252 ***	0.247 ***	0.236 ***	0.238 ***	0.238 ***	0.240 ***
		(15.42)	(14.72)	(14.08)	(13.97)	(13.81)	(13.70)
lnTEC			0.008	0.006	0.005	0.005	0.005
			(1.23)	(0.84)	(0.82)	(0.81)	(0.73)
lnINV				0.016 ***	0.016 ***	0.015 ***	0.015 ***
				(3.51)	(3.46)	(3.41)	(3.31)
IS					0.077	0.075	0.061
					(0.79)	(0.76)	(0.61)
EDU						0.001	0.002
						(0.28)	(0.30)
GE							0.048
							(0.59)
_cons	8.640 ***	6.139 ***	6.170 ***	6.177 ***	6.145 ***	6.140 ***	6.126 ***
	(220.51)	(37.27)	(37.06)	(37.82)	(36.55)	(36.28)	(35.77)
Provincial fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	330	330	330	330	330	330	330
r2	0.996	0.998	0.998	0.998	0.998	0.998	0.998

t statistics in parentheses. *** p < 0.01.

shows the results. Column (1) presents only the core explanatory and explained variable regression results. Columns (2) through (7) progressively control for other variables that affect farmers’ income. The core explanatory variable regression coefficients are all significant and positive at the 1% level. Specifically, it can be seen that without including any control variables, column (1) shows a 1% increase in DFI and a 0.073% increase in farmers’ income. Column (7) adds a series of control variables and shows that 1% growth in DFI increases farmers’ income by 0.039%. It shows that DFI has a significant positive effect on farmers’ income growth. These findings demonstrate that DFI can notably contribute to the increase in farmers’ income. Thus, Hypothesis 1 is verified.

For all control variables, the regression coefficients are positive. The regression coefficients of the rural investment level and economic level are significantly positive at the 1% level. This finding indicates that with economic growth, farmers’ income has increased. Moreover, the above findings indicate that investment has an impact on agricultural growth, thus increasing farmers’ income. The agricultural technology level has increased farmers’ income, indicating that the use of technology in agriculture has increased the efficiency of agriculture and encouraged agricultural growth; thus, farmers’ income has increased accordingly. A reasonable industrial structure promotes an increase in farmers’ income, indicating that the rational layout of the industrial structure and effective allocation of factor resources promote agricultural development, thus promoting farmers’ income growth. The rural education level promotes farmers’ income growth, indicating that the utilization of knowledge in production development is conducive to increasing farmers’ income. The government expenditure promotes farmers’ income growth, demonstrating that an increase in government expenditure on agriculture promotes agricultural development and is conducive to farmers’ income growth.

4.3. Robustness Analysis

4.3.1. Variable Lagged One Period

Taking into account the possibility of a delay in the impact of DFI on farmers’ earnings, a regression analysis is conducted with the DFI level (LlnDFI) lagged by one period. The findings are shown in Column (1) of

Table 4. Robust Analysis.

	(1)	(2)	(3)
	lnY	lnY	lnY
LlnDFI	0.026 ***
	(3.18)
lnDFI		0.077 ***	0.045 ***
		(3.21)	(3.57)
lnGDP	0.254 ***	0.245 ***	0.219 ***
	(13.51)	(12.61)	(12.18)
lnTEC	0.010	0.011	0.002
	(1.48)	(1.57)	(0.31)
lnINV	0.016 ***	0.017 ***	0.032 ***
	(3.41)	(3.66)	(5.58)
IS	0.167	0.159	−0.007
	(1.42)	(1.38)	(−0.07)
EDU	−0.001	0.000	−0.005
	(−0.17)	(0.08)	(−0.72)
GE	−0.020	−0.015	0.110
	(−0.24)	(−0.19)	(1.25)
_cons	6.121 ***	5.936 ***	6.163 ***
	(32.21)	(30.86)	(35.57)
Provincial fixed effect	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes
N	300	300	253
r2	0.998	0.998	0.998

t statistics in parentheses. *** p < 0.01.

, which shows that at the 1% level, the regression coefficient is significantly positive, which aligns with the benchmark regression results. The research model and results are thus robust.

4.3.2. Instrumental Variable Approach

To resolve the endogeneity issues that can be caused by reverse causation, missing variables, etc., this study tests the robustness of the benchmark regression results via the instrumental variable approach. In accordance with the methods of Qunhui Huang et al. (2019) [78], the 1984 postal and telecommunications historical data of each province serve as an instrumental variable for DFI. The level of traditional regional communication technology has historically been the basis for the development of the internet, and regions with better telecommunication infrastructures influence the development and application of internet techniques through the communication technique level, residents’ usage habits, and other factors. As the economy has grown, the use of conventional telecommunications tools has progressively declined. Therefore, the instrumental variable requirement is met. Notably, the 1984 postal and telecommunication historical data are cross-sectional; to make them suitable for panel data, please refer to Nunn and Qian’s approach (2014) [79]. Introducing a time-varying variable, i.e., constructing an interaction term between the number of rural broadband access subscribers in the previous year corresponding to the sample and the number of installed rural telephones in 1984 in each province as an instrumental variable for the composite index of DFI in that year. Moreover, we consider the time lag effects of DFI on farmers’ earnings. A one-period lag of the DFI composite index is also employed as an instrumental variable. As displayed in Column (2) of Table 4, at the 1% level, the DFI regression coefficient is significantly positive. The research model and robustness of the results are further verified.

4.3.3. Elimination of Outlier Variables

Owing to the significant differences in economic levels across regions, outliers that affect the regression results may exist. Therefore, in this work, data from four municipalities with higher economic levels, namely, Beijing, Shanghai, Tianjin, and Chongqing, and three provinces, namely, Guangdong, Zhejiang, and Jiangsu, are excluded, and a regression to validate the robustness of the study is conducted. As shown in Column (3) of Table 4, at the 1% level, the DFI regression coefficient is still significantly positive. The research model and results are thus robust.

4.4. Impact of Various Dimensions of Digital Financial Inclusion on Farmers’ Income

The comprehensive index of DFI is composed of the following three dimensions: coverage breadth, use depth, and digitization level. Therefore, what is the impact of these three dimensions on farmers’ income? The answer to this question is of great reference value for adjusting digital inclusive financial policy. To encourage the increase in farmers’ income, it is necessary to analyse which aspects promote farmers’ income and which dimensions are the main driving forces. A dual fixed effects model is used to estimate the impact of the above three dimensions on farmers’ income.

Table 5. Benchmark regression results from the perspective of three dimensions of digital financial inclusion.

	(1)	(2)	(3)
	ly	ly	ly
InCOV	0.018 ***
	(5.03)
InUSA		0.033 ***
		(5.33)
InDIG			−0.032 ***
			(−6.15)
EDU	0.001	0.001	−0.000
	(0.14)	(0.16)	(−0.07)
InGDP	0.232 ***	0.257 ***	0.260 ***
	(13.11)	(15.74)	(16.18)
InINV	0.016 ***	0.016 ***	0.018 ***
	(3.51)	(3.54)	(4.11)
IS	0.032	0.019	0.044
	(0.32)	(0.19)	(0.45)
lnTEC	0.005	0.005	0.005
	(0.74)	(0.76)	(0.76)
GE	0.040	0.011	0.045
	(0.50)	(0.13)	(0.58)
_cons	6.295 ***	5.968 ***	6.172 ***
	(35.53)	(35.73)	(37.06)
Provincial fixed effect	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes
N	330	330	330
r2	0.998	0.998	0.998

t statistics in parentheses. *** p < 0.01.

, Columns (1) and (2), show that the regression coefficients of coverage breadth and use depth are significantly positive at the 1% level, indicating that these two indicators significantly increase farmers’ income. Moreover, the regression coefficient of use depth is greater than that of coverage breadth, showing that use depth has the most positive effect on farmers’ income among the three dimensions. The regression coefficient of the digitization level is negative, demonstrating that the digitization level has a negative effect on farmers’ income. The possible reasons for this are that the cultural level of farmers is relatively low, their ability to accept new things is relatively weak, and their ability to use new products combined with digital technology and finance has not yet been achieved. It is evident that the greater the degree of digitization is, the fewer opportunities the digital dividend provides to farmers to increase their income, and the smaller the role it plays in increasing farmers’ income. The development processes of these factors are negatively correlated, resulting in a greater level of digitization, and thus a greater decrease in farmers’ income. Therefore, we should adjust the policy of digital development according to the situation of rural areas, optimize the development of digitalization through technological innovation, enable the development of rural areas, and increase farmers’ income. Moreover, we should improve the education level of farmers, popularize financial knowledge among farmers, and provide practical guidance for the use of DFI.

4.5. Heterogeneity Analysis

Owing to the significant differences in the level of imbalance of the regional economy, some heterogeneity with respect to the DFI development level is found due to regional differences. There are three regions in the country according to the NBS of China. Specifically, the eastern, central, and western regions are used to investigate the degree of heterogeneity in the region.

Table 6. Regional heterogeneity analysis.

	(1)	(2)	(3)
	lnY East	lnY Central	lnY West
lnDFI	0.048 **	0.187 ***	0.038 ***
	(2.43)	(3.07)	(4.01)
EDU	0.009	−0.042 **	0.010 *
	(1.49)	(−2.46)	(1.71)
lnGDP	0.133 ***	0.196 ***	0.147 ***
	(3.30)	(4.00)	(8.29)
lnINV	0.004	0.039 ***	−0.008
	(0.66)	(3.44)	(−1.49)
IS	−0.623 ***	0.750	0.169
	(−2.70)	(1.66)	(1.51)
lnTEC	0.015	0.019*	0.037 ***
	(1.31)	(1.90)	(3.33)
GE	−0.135	0.192	−0.059
	(−0.82)	(0.80)	(−0.86)
_cons	7.532 ***	6.084 ***	6.922 ***
	(17.28)	(8.36)	(44.06)
Provincial fixed effect	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes
N	110	66	121
r2	0.999	0.999	1.000

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

displays the results. At the 1% level, both the western and central regions’ regression coefficients for DFI are notably positive. However, they are notably positive at the 5% level in the eastern region. This finding indicates that farmers’ incomes can increase in all three regions with DFI, but those increases in the central and western regions are very significant, whereas those in the eastern region are not as significant. Because the eastern area has a relatively high economic level, due to the law of diminishing marginal utility, the marginal utility of DFI for increasing farmers’ income is diminishing. In the central and western regions, the economic level is relatively backwards, the urban–rural gap is wider, and the gap between regions is also wider. In terms of time, the circulation of production factors between urban and rural regions requires much time, leading to low-level production efficiency. Space, geography, transportation conditions, resource endowment, and other factors result in the poor circulation of production factors. Digital technology can overcome the limitations of the above unfavourable conditions, breaking the information and production factor circulation barriers in terms of time and space. This situation is conducive to compensating for the deficiencies of the economically underdeveloped and undeveloped areas of the western and central regions of the country and has an obvious empowering effect. Thus, this situation plays a more obvious role in economic development and in driving farmers to increase their income. Therefore, the increase in farmers’ income in the central and western regions is more significantly impacted by DFI than in the eastern region.

5. The Policy Effect of Digital Inclusive Finance—An Empirical Analysis Based on the DID Method

5.1. Analysis of Benchmark Model Estimation Results

Because China has experienced urban–rural dualism for a long period [80], DFI policies have been implemented and popularized in the countryside. According to the reform measures and practical work reported in the China Rural Financial Services Report 2018, released by the PBC in 2019, “In terms of financial inclusion, we have vigorously promoted the construction of financial infrastructure and DFI” and proposed that the next reform measures “should fully rely on digital technique and enhance the popularization of digital finance knowledge” [46]. The introduction of this national-level policy is a natural experiment. Therefore, this study uses it as an exogenous policy impact event. The DID approach is used to study the effects of digital inclusive financial policies before and after their implementation in countryside regions and their effects on farmers’ income to test hypothesis 2 proposed above. Owing to the different levels of digital technology use in each area, the effects of digital inclusive financial policies differ. The impact of policies in regions with weak digital infrastructure and relatively weak financial development is relatively greater, i.e., the response to policy implementation is more sensitive in these regions compared to other regions. Therefore, the experimental and control groups are constructed from the perspective of differences in the effects of DFI policies. Specifically, drawing on Campello and Larrain (2016) and Song Min et al. (2021) [81,82], experimental and control groups are constructed on the basis of the heterogeneity of regional responses to this policy. On the basis of the annual median, the DFI digitization level is divided into high and low groups, that is, a control group and a treatment group. Drawing on Song Min et al. (2021) [82], regions with DFI digitization levels below the annual median comprise the treatment group, taking a value of 1. Regions with values higher than the annual median comprise the control group, taking a value of 0. The study sample period is 2018 and later, which takes a value of 1; otherwise, it takes a value of 0. As shown in

Table 7. Panel dual fixed effect DID regression results.

	(1)	(2)	(3)	(4)
	InY	InY1	InY2	InY3
did	0.012 ***	0.080 ***	0.033 *	0.022
	(3.52)	(3.60)	(1.77)	(0.39)
EDU	−0.002	0.031	−0.023	0.016
	(−0.30)	(0.92)	(−0.82)	(0.19)
InGDP	0.270 ***	−0.063	0.352 ***	0.710 ***
	(16.15)	(−0.60)	(4.01)	(2.68)
InINV	0.016 ***	0.029	0.061 **	0.078
	(3.41)	(1.01)	(2.50)	(1.06)
IS	0.062	−1.848 ***	1.040 *	−6.331 ***
	(0.61)	(−2.89)	(1.94)	(−3.92)
InTEC	0.001	0.035	0.061 *	−0.137
	(0.14)	(0.83)	(1.73)	(−1.29)
GE	0.049	1.021 *	−0.641	2.624 **
	(0.58)	(1.94)	(−1.45)	(1.97)
_cons	5.977 ***	8.172 ***	3.912 ***	−2.130
	(34.83)	(7.59)	(4.33)	(−0.78)
Provincial fixed effect	Yes	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes	Yes
N	330	330	330	330
r2	0.998	0.923	0.907	0.595

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

, Column (1), at the 1% level, the DID coefficient is substantially positive. This finding demonstrates that when there are greater policy shocks, the increase in farmers’ income is more obvious than when there are fewer policy shocks. This finding shows that DFI significantly promotes farmers’ income growth.

This study further uses the DID method to analyse the impact of the implementation of digital inclusive financial policies on farmers’ income from the perspective of farmers’ income structure. The income of farmers is divided according to its source; moreover, considering that transfer income depends mainly on government behaviour, farmers’ income structure is studied according to wage income, family operating income, and property income, measured with the corresponding per capita logarithm. The regression results are displayed in Columns (2), (3), and (4) of Table 7. The explained variable in Column (2) is wage income, and the DID coefficient is significantly positive at the 1% level, showing that the increase in farmers’ wage income is more obvious in areas with greater policy impact and that the development of DFI can significantly increase their wage income. The explained variable in Column (3) is household operating income, and the DID coefficient is significantly positive at the 10% level; that is, the significance level is low. This finding shows that the impact of DFI on the increase of farmers’ household operating income is not very significant in areas with large policy shocks. The explained variable in Column (4) is property income, and the DID coefficient is not significant, demonstrating that DFI has no significant effect on the increase in farmers’ property income. The possible reasons for this finding are as follows. First, the implementation of DFI policies has alleviated rural financial exclusion and eased the financial constraints for rural economic development and farmers’ entrepreneurship, increasing farmers ’access to credit. Thus, driving farmers’ non-agricultural employment and increasing their wage income. Since China is still dominated by traditional agriculture, the growth of agricultural income is limited, and the impact of DFI on the increase in agricultural income is not significant. It is believed that with the development of modern agriculture, the impact of the implementation of DFI policies on farmers’ agricultural income growth will continue to increase. Third, farmers’ property income consists mainly of rent and bank interest. Agriculture in China is still dominated by traditional agriculture, and there is a large urban–rural wealth gap. Farmers have very limited resources to participate in social production activities through movable and immovable properties and thus earn rents and interest; thus, DFI does not have a significant effect on the growth of farmers’ property income.

5.2. Robustness Analysis

5.2.1. The Parallel Trend Test

A parallel trends test is performed because the parallel trends assumption is a prerequisite to ensure that the above DID estimation results satisfy unbiasedness. Figure 2 displays the results. Prior to the policy’s implementation, there was no significant difference between the treatment and control groups. The differences between the two groups appear following the policy’s implementation. This finding demonstrates the effectiveness of the policy’s execution and thus passes the abovementioned test. Notably, in the first year following its implementation, the policy does not have an immediate effect due to the time lag of the policy implementation effect.

5.2.2. Placebo Test

Since the findings of this study may be influenced by other unobservable factors, this study draws on La Ferrara et al. (2012) and Li et al. (2016) [83,84] to construct an experiment by randomly screening the provinces where the DFI policy is implemented and randomly generating the time of policy implementation. In Table 7, Column (1), the DID benchmark regression is performed according to the random sample, and the above process is repeated 1000 times. The results obtained are shown in Figure 3. Under random processing, the estimated coefficient is presented as a normal distribution with a mean of 0 on the left side of the vertical dotted line; that is, the real estimated coefficient of the DID model is 0.012. This finding shows that there is no problem with missing important variables in the model setting and that the impact of DFI on farmers’ income is not caused by other unobservable factors. That is, the impact of the benchmark regression analysis is indeed the result of policy implementation in this study. The placebo test is passed.

6. Conclusions and Suggestions

On the basis of the provincial data from 30 provinces from 2011 to 2021, the effect of DFI on farmers’ income is analysed in this study using a dual fixed effects model and the DID method. The conclusions are as follows: (1) Increases in farmers’ income are significantly promoted by DFI. On the basis of 11 years of panel data, this inclusion is sustainable. Using the three methods of robustness testing, namely, the one-period lag of the variables, the instrumental variable method, and the elimination of outlier variables, this conclusion still holds. This finding demonstrates the robustness of the study results. Further analysis shows that coverage breadth and use depth significantly contribute to the increase in farmers’ income, whereas the digitization level has a negative impact. (2) Additionally, this study uses the rural DFI policy as an external shock event to examine the effect of the policy shock on farmers’ income via the DID approach. The findings demonstrate that policy implementation has a major positive effect on farmers’ income, which also passes the parallel trends test and placebo test. Further analysis from the perspective of farmers’ income structure reveals that DFI significantly contributes to an increase in farmers’ wage income and can contribute to an increase in farmers’ household business income, but the effect is not very significant; however, such inclusion cannot significantly contribute to an increase in farmers’ property income. (3) According to a regional heterogeneity analysis, in the three areas considered, farmers’ income growth is boosted by DFI. However, the influence in the eastern region is not as significant as that in the central and western regions. With respect to long-term unbalanced regional development, the eastern area has more advanced social and economic levels than the central and western areas. The marginal utility of DFI for farmers’ income growth is decreasing in the eastern region.

Based on the above conclusions, the following recommendations are proposed: (1) Rural DFI should continue to be developed. The rural digital infrastructure, considering local features and regions with weak digital infrastructure, should be strengthened; moreover, resources should be given a relative tilt, investment and development efforts should be increased, and precise policies should be implemented. For example, to strengthen the digital infrastructure in the central and western regions, policy implementation should be adapted to local conditions and be moderately tilted according to the actual needs of the situation. The detailed implementation plan adjusts the rules according to the regional characteristics. Moreover, digital inclusive financial knowledge should be prioritized in such a way that farmers and other groups with relatively low cultural levels can easily understand and accept, and farmers’ financial literacy should be improved. The digital dimension of DFI in rural areas should be adjusted according to the traits of rural areas, and the development policy of the digital dimension should be adjusted in a direction that is suitable for rural areas and for farmers. In accordance with the situation of farmers’ property income, the government should establish corresponding policies and provide resources to increase farmers’ amount of property. Moreover, farmers can popularize the use of movable and immovable property to increase income-related financial knowledge. This should be synchronized with policies on DFI to promote each other. (2) The function of the “invisible hand” of the market for allocating resources should continue to be enhanced, and simultaneously, an excellent and fair business environment should be established. In terms of resource allocation, DFI can more effectively fulfill its enabling role through the principles of efficiency and fairness. To remain competitively fair, the government should also exercise effective “tangible hand” policies via macro control. Through taxation, transfer payments, and other means, the income of the low-income groups’ should be increased. Through guiding the power of morality, the role of a third distribution as a useful supplement to primary distribution and redistribution should be promoted so that vulnerable groups, such as farmers, can share in the fruits of social development. (3) The establishment of systems that enable “issues concerning agriculture, the countryside, and farmers” should be established to create DFI, offer institutional protection for this development, and enhance systems at both the macro and micro levels. Owing to China’s long-standing unbalanced regional development, imbalanced urban and rural evolution, and socioeconomic development being a systematic project, the corresponding system construction is also unbalanced, especially in less developed regions, where the system set up is relatively weak and cannot guarantee the smooth implementation of advanced new projects. Therefore, each region should establish and enhance the corresponding system for the promotion and implementation of this project according to the regional situation, implement it in a practical and detailed way, and establish and improve the appropriate system according to the characteristics of the region and the needs of the project. (4) A rural credit system should be established and then continuously improved, and financial supervision should be strengthened. Digital techniques should be used to promote a rural credit system, establish and perfect credit subject assessment, optimize the rural credit environment, strengthen financial supervision, prevent and manage financial risk, promote the sustainable growth of the rural economy and society, and vigorously promote DFI for the countryside. In addition, while digital technology is used to build a perfect credit system and share credit information, it is necessary to safeguard customers’ right to privacy in terms of their credit information. (5) Specifically, high-quality DFI products and services should be offered to rural areas. China is a vast country, and each region has its own characteristics in terms of resource endowment, geographic environment, human history, etc. Combined with the reality of unbalanced development and the development of regional specialties of agriculture, there are differences in the financial products and services demanded by each region and each type of development project. Therefore, according to the characteristics of each region’s agriculture and the continuous development of different types of projects, the financial supply side should continue to refine the design of financial products. Demand and supply should continue to be balanced, and the advantages of DFI should be maximized to increase farmers’ income.

This paper specifically studies the impact of DFI on farmers’ income and enriches relevant theories. It is necessary for the government to continue to use policy tools at the macro level to provide protection for the development of regional DFI to enable farmers to increase their income. The research results of this paper can also provide a reference for the related development of other countries in the world. Future research can use municipal or county-level data to analyse these issues, increase sample size, and conduct more in-depth research.