Does Financial Inclusion Have an Impact on Chinese Farmers’ Incomes? A Perspective Based on Total Factor Productivity in Agriculture

Abstract

The development of inclusive finance brings opportunities for farmers’ income growth. Based on panel data from 30 provinces in China from 2014 to 2023, this study explores the relationship between financial inclusion and two dimensions of financial inclusion (the degree of inclusion and the service efficiency of the financial function) and farmers’ incomes by constructing a two-way fixed-effects model. Meanwhile, this study further discusses the impact of agricultural total factor productivity on this relationship using a mediated-effects model. The results show that financial inclusion, the degree of inclusion, and the service efficiency of the financial function significantly contribute to the growth of farmers’ income in both the short and long run. Meanwhile, farmers’ reliance on financial inclusion and the two dimensions of financial inclusion deepens with income, and this reliance is more pronounced in economically developed regions. Further, promoting agricultural total factor productivity is an important way for financial inclusion and the degree of inclusion to contribute to short- and long-term income growth for Chinese farmers.

1. Introduction

Agriculture is the basic industry for the survival of mankind, the countryside is an important area for carrying the population, and farmers are a potential variable for social stability. Therefore, the “three rural issues” (agriculture, rural areas, and farmers’ issues) are fundamental issues related to global human survival and social development. At present, solving the “Three Rural Issues” by developing and expanding the rural economy and promoting the growth of farmers’ incomes has become a hot topic and the central task of the “Three Rural Issues” work. China, with nearly 600 million farmers, has always regarded solving the “three rural issues” as the basis for safeguarding people’s livelihoods and sustainable development. Therefore, promoting rural economic development and farmers’ income growth in a sustained manner may provide ideas for solving China’s “three rural issues”. However, as China’s economic growth slows down and is affected by unexpected negative events such as the COVID-19 pandemic, farmers’ income growth is under greater pressure. Against this backdrop, the development of inclusive finance may bring new opportunities for Chinese farmers to grow their incomes. In 2013, China proposed to vigorously develop inclusive finance in the document “Decision of the Central Committee of the Communist Party of China on Several Major Issues Concerning Comprehensively Deepening Reform”. Since then, policy documents such as the “Plan for Promoting the Development of Inclusive Finance (2016–2020)” and the “Implementation Opinions of the State Council on Promoting the High-Quality Development of Inclusive Finance” have been successively issued based on the development trend of inclusive finance in previous years, guiding the sinking of inclusive finance into rural areas. The business scope of inclusive finance has also gradually expanded from deposits and loans to various products such as credit, insurance, and wealth management. This has made inclusive finance more comprehensive and complete, better able to meet individualized needs, and enhance the strength of financial services for the “three rural areas”. Therefore, how the results of inclusive finance can further promote the long-term and stable growth of farmers’ incomes is the original purpose of China’s development of inclusive finance, as well as an important way to alleviate relative poverty in rural areas and build an inclusive economic growth system.

In summary, under the impact of the macroeconomic slowdown and the COVID-19 pandemic, farmers face greater pressure to sustainably increase their incomes, and there is an urgent need to explore a new impetus to promote the sustained growth of farmers’ incomes. Inclusive finance, because of its inclusive nature and its function of providing financial services, may bring new opportunities for Chinese farmers to increase their incomes. The organization of this study is as follows: the first part of this study is the introduction, which presents the background and research framework of the impact of financial inclusion on farmers’ income. The second part is the literature review, which reviews the relevant studies by scholars and comparatively analyzes the shortcomings of existing studies and the strengths of this study. The third part is the research design, which puts forward the research hypotheses based on the theory of the impact of financial inclusion on farmers’ income, and introduces the data sources, models, and variable selection for studying the related issues. The fourth part is the results and discussion, which presents the results of the benchmark regression, heterogeneity analysis, and the regression of the mechanism, and explains the reasons. The fifth part is the endogeneity and robustness test, which is conducted using various methods to ensure the reliability of the results. The sixth part is the conclusion and recommendation, which puts forward relevant policy recommendations based on the above research results, explains the limitations of this study, and provides an outlook for future research. The specific research logic is shown in Figure 1 below.

2. Literature Review

Financial exclusion is a widespread problem in most countries and is a serious impediment to the development of rural economies. Therefore, alleviating the problem of financial exclusion and allowing poor groups, such as farmers, to enjoy the dividends brought by finance in the process of financial development is the fundamental reason for the emergence of inclusive finance. In this process, whether inclusive finance will have a substantial impact on farmers, the most objective and direct research direction is to explore the relationship between inclusive finance and farmers’ income. In this area, scholars’ research has gone through the following three stages.

The first stage is centered on the introduction of the concept of inclusive finance and the impact of inclusive finance on farmers’ income, as well as the theoretical foundation. In 2005, the United Nations formally put forward the concept of inclusive finance, pointing out that inclusive finance is a financial system that can provide effective services for all levels of society, especially for low-income people such as small and micro enterprises and farmers [1]. The formal introduction of the concept of inclusive finance laid the foundation for the theoretical framework of the impact of inclusive finance on farmers’ income. In the early days, scholars focused on exploring the impact of financial inclusion on farmers’ income from the perspective of improving financial inclusion infrastructure, including increasing the number of rural financial institution outlets [2,3] and laying ATMs [4]. With the development of inclusive finance, scholars have found that the impact of inclusive finance on farmers’ income is not limited to the level of infrastructure. The emergence of new inclusive financial products, such as microcredit loans and agricultural insurance, has provided impetus to the growth of farmers’ income. Scholars have also shifted their research perspective from infrastructure to products and services. In this process, there are two mainstream views among scholars. One is that inclusive finance can ease credit constraints and realize farmers’ income growth [5,6,7]. To reach out to rural areas, inclusive financial institutions often provide credit products with lower interest rates and longer cycles according to the needs of farmers [5], and the interest rate mechanism can facilitate the transfer of funds to farmers [6]. With financial support for their production and operation, farmers can increase their output and realize increased income [7]. Another is that inclusive finance can smooth out the risks faced by farmers and realize their increased income [8,9]. Financial inclusion expands the scope of financial coverage and allows financial resources to be reallocated [8], which not only effectively reduces social risks and increases employment opportunities for farmers but also breaks down the risks faced by farmers’ production and life and promotes the growth of farmers’ income [9].

Although the research on the impact of inclusive finance on farmers’ income has built up a theoretical framework at this stage, the theoretical framework is more general and only focuses on the impact of the availability of financial services on farmers’ income, with no specific impact path or impact mechanism. At the same time, the relevant research also only focuses on the theoretical level and lacks the support of data, which may lead to a situation in which theory and reality are contradictory as shown in

Table 1. First stage of development of research related to financial inclusion and farmers’ income.

Author	Key Points	Comparison
United Nations Publications (2006) [1]	Inclusive finance can effectively serve all segments of society, especially those with low incomes.	The concept of financial inclusion was explicitly mentioned for the first time, laying a theoretical foundation.
Rafique S. (2006) [2]; Fuller D. et al. (2006) [3]; Cook S. (2006) [4].	Improved financial inclusion infrastructure improves farmers’ incomes.	A Theoretical Framework for Financial Inclusion and Farmers’ Income from the Perspective of Infrastructure Improvement, but with limitations.
Basu P. (2006) [8]; Dev S.M. (2006) [9]; Sutton C.N. et al. (2007) [6]; Rand J. (2007) [5]; Sriram M.S. (2007) [7].	Inclusive finance can smooth out the risks farmers face to realize income gains. Inclusive finance can ease credit constraints to realize farmers’ income growth.	A Theoretical Framework for Financial Inclusion and Farmers’ Income from the perspective of financial products and services expands theoretical perspectives.

.

The second stage is the rise of measurement and empirical evidence of financial inclusion. With the deepening of research and the improvement of data availability, scholars have begun to utilize empirical methods to test the impact of financial inclusion on farmers’ incomes as well as to explore its impact mechanism in depth. In this process, the first thing scholars discuss is how to measure financial inclusion, including the selection of measurement indicators and the identification of measurement methods. In terms of the selection of measurement indicators, the early measurement of financial inclusion used a single indicator of financial deepening, mainly including the ratio of financial assets to physical assets [10,11], the ratio of money supply to GDP [12], and the ratio of private sector bank loans to GDP [13]. However, financial inclusion includes not only a single deepening of financial services, but also the integration of various financial services. Therefore, a single index system cannot measure financial inclusion well, and scholars have begun to expand their horizons to a variety of index systems. Sarma (2008) drew on the construction method of the HDI index and, for the first time, constructed the Inclusive Finance Composite Index (IFI) by using the three indicators: the bank penetration rate, the availability of banking services, and the efficiency of the service use [14]. The emergence of the IFI index lays the foundation for scholars to utilize a variety of indices to construct a comprehensive index of financial inclusion. Gupte et al. (2012), based on Sarma (2008), further considered the transaction costs of banking services and utilized the four indicators of bank penetration rate, bank service availability, service use efficiency, and transaction costs to measure the level of inclusive finance [15]. However, due to the shortcomings of Sarma (2008) in constructing the indicators, such as the relevant data in some countries being difficult to obtain, the measurement is the financial inclusion index at the national level. Scholars have begun to select financial inclusion measurement indices from more accessible indicators, more micro perspectives, and more comprehensive information. Mialou et al. (2017) constructed a new financial inclusion measurement index using the number of ATMs, the number of ODC branches, the total number of residents lending to ODCs, and the number of residents borrowing from ODCs [16]. With the detailed micro-household data, Zhang and Posso (2017) constructed a household financial inclusion index in the form of dummy variables to fill the gap in the measurement of household-level financial inclusion [17]. Although the methods of constructing financial inclusion indices vary greatly, the principle of unity among scholars is to select the method that can obtain the most information from the indicator system. The most common methods used by scholars to measure the level of financial inclusion are as follows: (a) Average Euclidean Distance Method: This method can reflect the spatial distance characteristics; however, it may conceal the complementarity between indicators and is complicated to calculate [14,18,19]. (b) Simple geometric mean: This method is simple to calculate but is less interpretable for indices constructed using multiple indicators [20]. (c) Factor analysis can reduce the dimension and reveal the underlying structure; moreover, it can be used to deal with high correlation between indicators. However, it has to satisfy the prerequisite assumptions, such as normality and sufficient correlation. Meanwhile, factor naming and interpretation rely on subjective judgment [21,22]. (d) Principal component analysis (PCA) is completely dependent on the data; no subjective assumptions are required, and the problem of multicollinearity can be effectively eliminated. However, the principal component meaning is difficult to interpret the index, and this method is sensitive to the data distribution; moreover, there is also the problem of omitting secondary information [23,24,25,26]. (e) Entropy value method: Determining weights based entirely on data avoids the influence of subjective factors [27,28]. It is robust to outliers and insensitive to data distribution. At the same time, the entropy value method can maximize the reflection of information and improve the efficiency of the use of information in indicators.

After solving the problem of financial inclusion measurement, scholars have empirically verified the impact of financial inclusion on farmers’ income and further explored the impact path of financial inclusion on farmers’ income. Prior to this, scholars explored the path of impact mainly focused on the supply level, including the improvement of infrastructure and the innovation of inclusive financial products and services. With the depth of the study, scholars have found that the availability and use of financial inclusion does have an impact on farmers’ income; however, the ability to use it also affects the effectiveness of financial inclusion. At this stage, scholars began to explore the impact of financial inclusion on farmers’ income from the perspective of utilization capacity. The ability to use is mainly related to farmers’ own characteristics, and many studies that stand on the level of human capital [29,30,31], risk attitude [32], social network [33], and other dimensions to explore the impact of financial inclusion on farmers’ income have appeared in this stage. With the development of the digital economy, digital technology is rapidly spreading in the financial system, and traditional infrastructure and product services are transforming to digitalization. Scholars have also keenly captured the changes in finance and farmers’ lives brought about by digital technologies and have begun to explore the relationship between digital financial inclusion and farmers’ incomes. Many scholars have found that digital financial inclusion can break through time and space distances to allow farmers to enjoy financial services [31]. Farmers can use cell phones and computers to access financial services through mobile payments and digital financial platforms [34], which reduces the cost of accessing financial services and increases the probability of farmers using inclusive financial products, which, in turn, leads to income growth for farmers [31,35].

The above study empirically verified the impact of financial inclusion on farmers’ income and further analyzed the path of impact. At the same time, the impact of digital technology was considered to explore the relationship between digital financial inclusion and farmers’ income. However, scholars’ explorations at this stage are still deficient. Scholars focus more on the direct impact of financial inclusion on farmers’ income and less on the indirect path of financial inclusion acting on farmers’ income. At the same time, research has focused mainly on the overall impact of financial inclusion on farmers’ incomes, with less attention paid to differences such as groups and regions as shown in

Table 2. The second stage of the development of research related to financial inclusion and farmers’ income.

Author	Key Points	Comparison
Shetty N.K. (2008) [10]; Bebczuk R.N. (2008) [11]; Hol S. (2007) [12]; Gadanecz B. (2008) [13].	Measuring inclusive finance using the ratio of financial assets to physical assets, the ratio of money supply to GDP, and the ratio of private sector bank lending to GDP.	A single indicator cannot comprehensively measure financial inclusion.
Sarma M. (2008) [14]	For the first time, the Integrated Financial Inclusion Index (IFI) was constructed using three indicators: bank penetration, bank service availability, and service utilization efficiency.	The IFI index lays the foundation for scholars to utilize a variety of indicators to construct financial inclusion.
Gupte R. et al. (2012) [15]	Building on Sarma (2008), the transaction costs of banking services are further considered.	Expanded IFI index. However, the relevant data in the IFI constructed indicators are difficult to obtain in some countries, and the measurement is at the national level.
Mialou A. et al. (2017) [16]; Zhang Q. et al. (2019) [17].	Measuring the financial inclusion index using the number of ATMs, the number of branches of ODCs, the total number of residents lending to ODCs, and the number of residents borrowing from ODCs, dummy variable form.	Financial inclusion measures are selected from more accessible indicators, more micro perspectives, and more comprehensive information.
Sarma M. (2008) [14]; Apparicio P. (2008) [19]; Sarma M. (2012) [18].	Measuring financial inclusion using the mean Euclidean distance method.	Capable of capturing the spatial distance characterization of financial inclusion, but may mask complementarities between indicators and be computationally complex.
Arora R.U. (2012) [20]	Measuring financial inclusion using simple geometric averages.	Simple to calculate, but poorly interpreted for indices constructed using multiple indicators.
Aisaiti G. et al. (2019) [21]; Omar M.A. et al. (2020) [22].	Measuring financial inclusion using factor analysis.	It can be downscaled and reveal underlying structure, and can handle high correlation between indicators. But the prerequisite assumptions, such as normality and sufficient correlation, have to be satisfied. Meanwhile, factor naming and interpretation rely on subjective judgment.
Cámara N. (2014) [24]; Yorulmaz R. (2018) [23]; Dungey M. (2018) [25]; Le T.H. (2019) [26].	Measuring financial inclusion using principal component analysis.	Completely dependent on the data and without subjective assumptions, it can effectively eliminate the problem of multicollinearity. However, this method is sensitive to data distribution and also suffers from the problem of omitting secondary information.
Yanzhi T. (2021) [27]; Zhang B. (2021) [28].	Measuring financial inclusion using the entropy approach.	Determines weights based entirely on data, avoiding the influence of subjective factors. It is robust to outliers and insensitive to data distribution. At the same time, the entropy value method can maximize the reflection of information and improve the efficiency of the use of information in the indicators.
Zhou G. (2020) [29]; Fowowe B. (2020) [30]; Aisaiti G. (2019) [21]; Adegbite O.O. et al. (2020) [32]; Abor J.Y.et al.(2018) [33].	Exploring the impact of financial inclusion on farmers’ incomes from a capacity to use perspective.	Shifting the study of the impact of financial inclusion on farmers’ incomes from the direct to the indirect level.
Wang X. et al. (2020) [31]; Jiang L. et al. (2019) [34]; Mhlanga D. (2020) [35].	Digital financial inclusion can effectively contribute to farmers’ income growth.	Focusing on the innovative potential of technology-driven inclusive financial development, it reveals the pathways of inclusive finance to increase farmers’ incomes when augmented by digital technology.

.

The third stage is to explore the impact of financial inclusion on farmers’ income from a multidimensional perspective. With regard to the indirect impact of inclusive finance on farmers’ income, early scholars explored the mechanism of its impact from the perspective of farmers’ ability to use it (human capital, risk appetite, and social networks). However, the development of financial inclusion as a form of finance cannot be separated from the economic development of a country and a region. At the same time, farmers’ income is the sum of the benefits that farmers obtain in a specific period of time by participating in various economic activities such as production, processing, business, and resource allocation. Therefore, scholars have begun to focus on considering the impact of inclusive finance on farmers’ incomes from an economic perspective. Research on economic pathways mainly includes the level of regional economic development [36,37,38,39,40], industrial structure [36,38,41], non-farm employment [41,42], agricultural productivity [30,36,43], international trade [44], fiscal and monetary policies [37,45], urban and rural factor flows [41,42,46], and resource allocation efficiency [30,41,42,43,46]. With the development of the digital economy, scholars have further considered the impact of digital technology on the process of financial inclusion on farmers’ income on the basis of the above studies. The avenues of digital technology include digital account penetration [47,48], mobile payment frequency [49,50,51], online credit approval efficiency [47,49], digital insurance participation rate [52,53,54], bio-digital technology use [55,56,57], and the degree of business digitization [58,59]. While the findings of many scholars indicate the impact of digital technology on the process of financial inclusion on farmers’ income. However, some scholars also come to the conclusion of no effect. Further, scholars explore the effects of the digital divide [60,61] and the sophistication of digital infrastructure [62] in this process to explain why financial inclusion does not enhance farmers’ incomes through the development of digital technologies.

At this stage, the deepening of heterogeneity analysis reflects the transformation of scholars’ research perspective from “overall effect” to “group differentiation”, which we sort out from the research content and empirical methods. At the level of research content, scholars’ heterogeneity analysis has experienced the development from single heterogeneity to multidimensional heterogeneity. The issue of poverty reduction has always been a major problem plaguing human development. Therefore, at this stage, scholars mainly focus on the income level and explore the single rich-poor difference in the impact of financial inclusion on farmers’ income [63,64]. However, there are differences in the degree of sophistication of financial facilities, policies, and economic development in different regions, and environmental factors may have an impact on the effect of financial inclusion on farmers’ incomes [65,66,67]. At the same time, there are differences in the financial literacy, education, age, etc., of different farmers, and personal factors may also have an impact on the effect of financial inclusion on farmers’ income. Since then, scholars have further added environmental and personal factors in the process of research to analyze the differences in the impact of financial inclusion on farmers’ income [68,69]. In terms of research methodology, early on, differences in the impact of financial inclusion on farmers’ incomes across incomes, regions, and groups were mainly explored through group regression [20,70]. In determining the grouping criteria, it is often subjective and may lead to biased results. Moreover, grouping regression cannot handle continuous heterogeneity. Subsequently, scholars explored the introduction of interaction terms into heterogeneity studies [37,43,71]. Although the interaction term can avoid the subjectivity of grouping as well as deal with continuous heterogeneity, it cannot fully reveal the heterogeneity of the distribution, and there may also be the risk of multicollinearity. Therefore, because quantile regression can comprehensively reveal distributional heterogeneity and is robust to distributions with extreme values, such as income, it can accurately measure the differences in the impact of financial inclusion on different quantile levels [43,66,72]. Currently, the quantile regression method is used by a wide range of scholars to analyze the heterogeneity of inclusive finance on farmers’ income.

At this stage, the research on financial inclusion and farmers’ income has entered a multidimensional and multi-method research phase of direct impact, indirect impact, and heterogeneity. However, the vast majority of studies have examined the overall effect of financial inclusion on farmers’ income. Studies of the overall effect cannot provide insight into which features of financial inclusion actually affect farmers’ incomes. At the same time, existing scholars rarely focus their vision on total factor productivity in agriculture when studying how inclusive finance affects farmers’ incomes, as shown in

Table 3. The third stage of development of the research related to financial inclusion and farmers’ income.

Author	Key Points	Comparison
Fowowe B. (2020) [30]; Pomeroy R. et al. (2020) [40]; Huang Y. (2020) [46]; Peprah J.A. et al. (2021) [43]; Liu G. et al. (2020) [36]; Liu T. et al. (2021) [39]; Arshad M.U. et al. (2021) [45]; Li Y. et al. (2022) [37]; Ge H. et al. (2022) [38]; Yang B. et al. (2023) [42]; Wang W. et al. (2023) [44]; Zhu K. et al. (2024) [41].	Exploring the impact of inclusive finance on farmers’ incomes in terms of the level of regional economic development, industrial structure, non-farm employment, agricultural productivity, international trade, fiscal and monetary policies, urban–rural factor flows, and resource allocation efficiency.	The research horizon expands from focusing on the supply and demand side of financial inclusion to exploring the impact of financial inclusion on farmers’ incomes from various pathways in the economy.
Ezzahid E. et al. (2021) [50]; Liu Z. et al. (2021) [57];Hasan M.M. et al. (2022) [47]; Tay L.Y.et al.(2022) [48]; Qian H. (2022) [55]; Zhang L. et al. (2023) [51]; Yu W. et al. (2023) [52]; Xu S. et al. (2023) [53]; Liu Y. et al. (2023) [58]; Zhang C. et al. (2024) [49]; Qin Z. et al. (2024) [54]; Chen Y. et al. (2024) [56]; Wang Y. et al. (2024) [59].	Exploring the impact of inclusive finance on farmers’ incomes in terms of digital account penetration, mobile payment frequency, online credit approval efficiency, digital insurance participation rate, bio-digital technology use, and the degree of business digitization.	The indirect impact of financial inclusion on farmers’ incomes extends from the perspective of farmers’ ability to use it (human capital, risk appetite, and social networks) to the state of economic development of a country or a region.
Zhao H. et al. (2022) [60]; Xiong M. et al. (2022) [61]; Nutassey V.A. et al. (2025) [62].	The digital divide and the inadequacy of digital infrastructure have prevented the emergence of inclusive finance from realizing the effects of digital technology on farmers’ incomes.	Provide a digital divide and infrastructure development level explanation for the inability of financial inclusion to enhance farmers’ incomes through the development of digital technologies.
Mohan R. (2008) [63]; Swamy V. (2014) [64].	Income-level differences in the impact of financial inclusion on farmers’ incomes.	A single indicator dimension does not capture the differences in the impact of financial inclusion on farmers’ incomes in other dimensions.
Kumar A. et al. (2019) [65]; Afrin S. et al. (2017) [66]; Sanderson A. et al. (2018) [67]; Hussain S. et al. (2023) [68]; Ren J. et al. (2023) [69].	Examining the Heterogeneity of Financial Inclusion on Farmers’ Income from Environmental and Financial Literacy Dimensions.	Expanded dimensions of heterogeneity analysis of financial inclusion on farmers’ income.
Arora R.U. (2012) [20]; Cnaan R.A. (2012) [70].	Subgroup regression to explore the heterogeneity of financial inclusion on farmers’ income.	The criteria for determining grouping are often subjective and may lead to biased results. Moreover, it is not possible to deal with continuous heterogeneity.
Li Y. et al. (2022) [37]; Peprah, J.A. et al. (2021) [43]; Xia Y. et al. (2025) [71].	Interactivity modeling to explore the heterogeneity of financial inclusion on farmers’ income.	Compensates for the shortcomings of grouped regressions, but does not fully reveal the heterogeneity of the distribution, and there may also be a risk of multicollinearity.
Afrin S. et al. (2017) [66]; Peprah J.A. et al. (2021) [43]; Dirir, S.A. et al. (2022) [72].	Quantile regression modeling to explore the heterogeneity of financial inclusion on farmers’ income.	It can fully reveal distributional heterogeneity, is robust to distributions with extreme values such as income, and can accurately measure differences in the impact of financial inclusion across quartile levels.

.

The innovations of this study: firstly, financial inclusion is categorized into two dimensions (degree of inclusion and efficiency of financial function services). The policy objective of financial inclusion is to make financial services available to all groups at a reasonable cost; the internal logic is to compensate for the profit-seeking nature of finance in order to correct market failures. It can be seen from the purpose and operational logic that inclusive finance has the dual characteristics of inclusion and finance. Based on this, this study analyzes in detail the impact of the two dimensions of financial inclusion (the degree of inclusion and service efficiency of the financial function) on farmers’ incomes at the theoretical level and verifies it at the empirical level using various statistical methods. Secondly, this study focuses its vision on agricultural total factor productivity. In fact, total factor productivity in agriculture is an important indicator of production efficiency. Financial capital (funds, credit, insurance, etc.) not only serves as a direct means of production but also provides security for production operations, which inevitably has an impact on production efficiency. Production efficiency measures the output under the unit cost, and the improvement of production efficiency brings about an increase in output or a reduction in costs, which inevitably translates into more income. Therefore, this paper is of value in exploring the mechanism of the role of agricultural total factor productivity in the process of financial inclusion affecting farmers’ income, as shown in

Table 4. Comparison of financial inclusion and farmers’ income at various stages of development and the strengths of this paper.

Stage	Research Topics	Comparison/Marginal Contribution
The first stage	The concept of inclusive finance introduction and the impact of inclusive finance on farmers’ income theoretical foundation.	Contribution: The research on the impact of financial inclusion on farmers’ income has built up a theoretical framework at this stage. Weaknesses: (a) The theoretical framework is more general, only staying in the impact of financial service availability on farmers’ income, without specific impact paths and impact mechanisms. (b) The relevant research also stays only at the theoretical level and lacks the support of data, which may lead to a situation in which theory and reality are contradictory.
The second stage	The rise of measurement and empirical evidence of financial inclusion.	Contributions: (a) The impact of financial inclusion on farmers’ income is empirically verified, and the path of impact is further analyzed. (b) Considered the impact of digital technology and explored the relationship between digital financial inclusion and farmers’ income. Weaknesses: (a) More focused on the direct impact of financial inclusion on farmers’ income, and less research on the indirect pathway of financial inclusion acting on farmers’ income. (b) The study mainly focuses on the overall impact of inclusive finance on farmers’ income, and pays less attention to differences such as groups and regions.
The third stage	A multidimensional perspective on the impact of financial inclusion on farmers’ incomes.	Contribution: Research on financial inclusion and farmers’ income has entered a phase of in-depth studies with multiple dimensions and methodologies, including direct impact, indirect impact, and heterogeneity. Weaknesses: (a) The vast majority of studies examine the overall impact of financial inclusion on farmers’ income. Studies of the overall impact cannot provide insight into which features of financial inclusion actually affect farmers’ incomes. (b) Existing scholars rarely focus their vision on total factor productivity in agriculture when studying how financial inclusion affects farmers’ incomes. (c) Studies have mostly focused on immediate impacts and have not considered long-term changes.
My study	The relationship between financial inclusion and the two dimensions of financial inclusion (the degree of inclusion and the service efficiency of the financial function) and farmers’ incomes is discussed, and the impact of total factor productivity in agriculture on this relationship is further discussed.	Contribution: (a) Financial inclusion is categorized into two dimensions (degree of inclusion and efficiency of financial functioning services). The impact of the two dimensions of financial inclusion (the degree of inclusion and the efficiency of financial function services) on farmers’ income is analyzed in detail at the theoretical level and verified at the empirical level using a variety of statistical methods. (b) Focusing the vision on total factor productivity in agriculture. Exploring the mechanism of total factor productivity in agriculture in the process of financial inclusion affecting farmers’ income. (c) Investigating the immediate and forward impact of financial inclusion and its two dimensions on farmers’ income. At the same time, examining the impact of financial inclusion and its two dimensions on farmers’ immediate and forward income through total factor productivity in agriculture.

.

3. Research Design

3.1. Theoretical Analysis

Due to the serious financial exclusion in the traditional financial market, it is difficult to provide adequate and complementary financial services for the development of the agricultural and rural economy, which restricts the development of the agricultural and rural economy and the growth of farmers’ income [73,74]. Inclusive finance mainly provides decentralized small farmers and microenterprises with comprehensive financial services such as payments, credit, wealth management, and insurance that have wide coverage, high levels of utility, and relatively low costs, thereby helping to improve the efficiency of rural resource allocation and promote the growth of farmers’ incomes.

From the perspective of an inclusive nature, on the one hand, the development of inclusive finance can effectively improve the availability of financial services. The development of inclusive finance can promote the increase in financial institutions’ outlets, strengthen the construction of financial institutions’ outlets [75], and allow financial institutions’ outlets to be extended to rural areas, especially poor rural areas, to expand the coverage of financial services. On the other hand, the development of inclusive finance promotes the realization of the goal of benefiting the people. In order to promote the development of inclusive finance, the central and local governments have granted special subsidies, refinancing concessions and tax breaks to inclusive financial institutions to reduce their operating costs. Inclusive financial institutions subsidize the reduced costs of their operations to the corresponding products, which ultimately benefit farmers. At the same time, the development of inclusive finance will also lead to an increase in the number of financial institutions in rural areas, increasing competition in the industry [76]. Increased competition in the industry further encourages inclusive financial institutions to improve their service efficiency and provide farmers with low-cost and high-quality financial services. In addition, the establishment of financial inclusion outlets in rural areas has brought farmers and financial services closer together, allowing farmers to handle financial services quickly and reducing their transaction and transportation costs.

From the perspective of financial function, firstly, the emergence of inclusive finance has expanded the coverage of basic payment services and provided farmers with low-cost payment and settlement services [77]. PFIs improve the efficiency and quality of payment and settlement services by optimizing the account opening process, constructing “aging” payment demonstration outlets, and providing door-to-door services, etc., and improve the construction of internal clearing networks to protect user information security and improve the efficiency of payment and settlement and fund clearing. The convenience of payment and settlement has eliminated the need for farmers to meet with their employers for cash wage settlements and has made it possible for farmers to obtain sales and subsidies through bank transfers and other means, thus improving the efficiency of farmers’ access to income.

Secondly, characteristics such as the difficulty of reasonably estimating the value of crops and the illiquidity of the market for rural self-constructed houses have left farmers without better assets to use as collateral. Financial institutions have had to raise the threshold for financial services to reduce risk, making it more difficult for farmers to obtain loans from formal financial institutions. At present, inclusive finance is still characterized by policy-based finance, which provides relatively diverse and below-market interest rate agricultural loans to specific farmers under the guidance of national policies to improve their income. Referring to the existing literature, such as Khatun Tahmina and Sadia Afroze (2016) [78], this paper constructs the Cobb–Douglas production function considering financial capital. Assuming constant returns to scale, the production function of the farmer’s household is a function of labor ($L$), total capital ($K$), total financial capital ($F$), and production efficiency ($A$). $\alpha$ is the contribution of physical capital to farmers’ income, $\beta$ is the contribution of financial capital to farmers’ income, and $1-\alpha -\beta$ represents the contribution of surplus labor to farmers’ income.

$$Y=A{K}^{\alpha }{F}^{\beta }{L}^{1-\alpha -\beta }$$

(1)

It is assumed that the size of a farmer’s decision to acquire financial capital is influenced by a combination of the cost of acquiring financial capital $C$, income from the previous period $Y_{t-1}$, risk appetite $\theta$, and other expenditures $\pi$, and it is assumed that the farmer spends all of his residual income on financial capital acquisition, taking into account risk and the desire to maximize their income.

$$Y_t=A_t{K_t}^{\alpha }{\left[{F_t}^{\beta }\left(C_t,Y_{t-1},{\theta }_t,{\pi }_t\right)\right]L}^{1-\alpha -\beta }$$

(2)

$$F_t={\theta }_t\left(Y_{t-1}-C_t-{\pi }_t\right)$$

(3)

$${\displaystyle \frac{\partial Y_t}{\partial C_t}}={-\beta A}_t{K_t}^{\alpha }{{\theta }_{t}L}^{1-\alpha -\beta }{\left(Y_{t-1}-C_t-{\pi }_t\right)}^{\beta -1}<0$$

(4)

From Equations (3) and (4), it can be seen that the decline in the cost of financial capital acquisition contributes to the scale of farmers’ financial capital investment, which, in turn, increases farmers’ income.

Furthermore, P&W Finance provides farmers with diversified financial products, such as “Government and Bank Insurance”, “P&W Anxian”, “YangGuang BiLeWu 1” and “YangGuang Jin P&W Day Open (90-day minimum holding) A”, which are low-threshold and low-risk financial products that help farmers allocate their idle funds in a scientific manner to realize the growth of their property income. Low-threshold and low-risk financial products, such as “Government and Bank Insurance”, “Puhui Anxian”, “Yangguangbi Lokwood No. 1”, and “Sunshine Puhui Rikai (90-day minimum holding) A”, have helped farmers to scientifically allocate their idle funds, thereby realizing the growth of their property income.

Finally, inclusive finance can increase farmers’ resilience to risks and safeguard their capital. On the one hand, financial savings services are characterized by the preservation and enhancement of value. Financial savings services can pool idle funds and provide interest to savers. This low-risk way of preserving and increasing the value of assets is generally accepted by farmers, who have increased their savings. When life is shaken by risks, financial savings can have a cushioning effect and increase farmers’ resilience to risks. At the same time, the flow of funds to inclusive financial institutions can also enhance their credit capacity. Through credit funds, farmers can smooth out basic living expenses such as healthcare and education, and can also be used to bridge investment gaps and enhance the economic resilience of their households [79]. On the other hand, inclusive finance provides farmers with special insurance such as agricultural insurance, agricultural product value insurance, etc.; moreover, when unexpected events cause losses to farmers, the insurance can give compensation or payment, increase the transfer income of rural residents, and effectively reduce the harm caused by unexpected events. Accordingly, Hypothesis 1 and Hypothesis 2 are proposed.

Hypothesis 1.

The development of inclusive finance can effectively promote the growth of farmers’ income.

Hypothesis 2.

The increase in the degree of financial inclusion and the efficiency of financial functions and services can lead to the growth of farmers’ income.

Total factor productivity (TFP) is a concept that measures the efficiency of production and is expressed as the ratio of outputs to inputs in the process of economic activity. The relationship between financial inclusion, agricultural total factor productivity, and farmers’ incomes is further explored below.

First, in terms of easing credit constraints, agricultural production has a long cycle and is greatly affected by climate and other uncertainties; therefore, financial institutions have to face a certain degree of risk in providing credit funds for it [80]. At the same time, there are problems such as information asymmetry and a lack of collateral and credit guarantee between rural residents and financial institutions, which makes it more difficult to obtain financial support from formal financial institutions [81,82]. The emergence of inclusive finance alleviates the financial exclusion of low-income groups such as rural residents, reduces the cost of access to financial services for rural residents, and guarantees the timely availability of funds for agricultural production. With access to production capital, farmers are in a position to use more specialized techniques in the production process [71], increase their investment in new agricultural machinery and equipment, and learn more advanced biochemical and agricultural techniques [83]. The promotion and use of new agricultural machinery and equipment will help to increase labor productivity; the promotion and use of biochemical technologies will help to increase the efficiency of agricultural resource use; and the learning of advanced technologies will help to increase the level of human capital [84,85]. The enhancement of labor productivity, agricultural resource utilization efficiency, and human capital, on the one hand, reduces the marginal consumption rate of resources, improves the input–output ratio of all kinds of agricultural factors, and ultimately enhances the total factor productivity of agriculture, promotes the development of large-scale and intensive agricultural production [86], and realizes the growth of operating income. On the other hand, under the established output, it reduces the cost of labor capital and land and other resources that farmers need to pay to engage in agricultural production activities, promotes to a certain extent the transfer of land, releases a large number of surplus labor to transfer to non-agricultural industries, and raises the wage income of farmers.

Second, risk avoidance. Agricultural production is faced with greater uncertainty, in addition to pests and diseases, climate and environmental changes, and other factors; agricultural sales and market price risk also affect the production activities of farmers. Specialized financial inclusion agricultural insurance, such as disaster insurance and agricultural price insurance, designed for different needs, provides risk-bearing price compensation for disaster losses and sales losses, and further improves the ability of farmers to manage their production, thus contributing to the improvement of total factor productivity in agriculture. The increase in total factor productivity in agriculture means that the established capital inputs, the single yield growth, and the input–output ratio of capital increased. As the sales channels of agricultural products are more stable in the short term, in the case of market prices with insurance protection, it can promote the increase of sales income and realize the growth of operating income; accordingly, Hypothesis 3 is proposed.

Hypothesis 3.

Financial inclusion and its dimensions drive up farmers’ incomes by contributing to the growth of total factor productivity in agriculture.

3.2. Model Construction

This paper adopts the DEA–Malmquist index method to measure the level of total factor productivity in agriculture. First, the set of production possibilities of each decision unit in period $t$ is defined, assuming constant returns to scale and strong factor disposition; second, the Malmquist productivity index is defined for the period $t$ to $t+1$ using the output distance function based on the following technology:

$$\begin{gathered} {\mathrm{M}}_0^t={\mathrm{D}}_0^t\left(x_{t+1},y_{t+1}\right)/{\mathrm{D}}_0^t\left(x_t,y_t\right) \\ {\mathrm{M}}_0^{t+1}={\mathrm{D}}_0^{t+1}\left(x_{t+1},y_{t+1}\right)/{\mathrm{D}}_0^{t+1}\left(x_t,y_t\right) \end{gathered}$$

(5)

The Malmquist productivity index measured using Equation (5) above is geometrically averaged to obtain the Malmquist index for the independent decision unit from period $t$ to period $t+1$:

$$\begin{gathered} TFP={\mathrm{M}}_0\left(x_t,y_t,x_{t+1},y_{t+1}\right) \\ ={\displaystyle \frac{{\mathrm{D}}_0^t\left(x_{t+1},y_{t+1}\right)}{{\mathrm{D}}_0^t\left(x_t,y_t\right)}}\sqrt{\left[{\displaystyle \frac{{\mathrm{D}}_0^t\left(x_{t+1},y_{t+1}\right)\times {\mathrm{D}}_0^t\left(x_t,y_t\right)}{{\mathrm{D}}_0^{t+1}\left(x_{t+1},y_{t+1}\right)\times {\mathrm{D}}_0^{t+1}\left(x_t,y_t\right)}}\right]} \end{gathered}$$

(6)

where $x_t$ and $y_t$ are the input and output variables in period $t$, and $x_{t+1}$ and $y_{t+1}$ are the input and output variables in period $t+$1; ${\mathrm{D}}_0^t\left(x_t,y_t\right)$ and ${\mathrm{D}}_0^{t+1}\left(x_t,y_t\right)$ are the distance output functions derived from comparing the production point with the frontier technology in the same period; and ${\mathrm{D}}_0^t\left(x_{t+1},y_{t+1}\right)$ and ${\mathrm{D}}_0^{t+1}\left(x_{t+1},y_{t+1}\right)$ are the distance output functions derived from comparing the production point with the frontier technology in the mixing period.

In order to test the impact of financial inclusion on farmers’ income, this paper constructs the following benchmark model to develop empirical evidence:

$${IN}_{it}={\gamma }_0+{\gamma }_1{FIN}_{it}+{\gamma }_2{X}_{it}+{\lambda }_i+{\mu }_t+{\epsilon }_{it}$$

(7)

where ${IN}_{it}$ is the explained variable, denoting the farmer’s income in province $i$ in period $t$. ${FIN}_{it}$ is the explanatory variable, denoting the level of financial inclusion development in province $i$ in period $t$. $X_{it}$ is the control variable, and ${\gamma }_0$ is the constant term, ${\gamma }_1$ and ${\gamma }_2$ are the correlation coefficients, ${\lambda }_i$ and ${\mu }_t$ represent the area fixed effect and time fixed effect, respectively, and ${\epsilon }_{it}$ is the random error term.

In order to examine the mechanism of the impact of financial inclusion on farmers’ income, this paper refers to the mediation model proposed by Baron and Kenny (1986) [87] with the following expression:

$${TFP}_{it}={\alpha }_0+{\alpha }_1{FIN}_{it}+{\alpha }_2{X}_{it}+{\lambda }_i+{\mu }_t+{\epsilon }_{it}$$

(8)

$${IN}_{it}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{FIN}_{it}+{\mathsf{\beta }}_2{TFP}_{it}+{\mathsf{\beta }}_3{X}_{it}+{\lambda }_i+{\mu }_t+{\epsilon }_{it}$$

(9)

where ${TFP}_{it}$ denotes the total factor productivity of agriculture in province $i$ in period $t$.

3.3. Data and Variables

Raw data values are from the China Insurance Yearbook, China Statistical Yearbook, China Electronic Information Industry Statistical Yearbook, China Agricultural Machinery Industry Yearbook, China Water Conservancy Statistical Yearbook, China Agricultural Reclamation Statistical Yearbook, China Rural Statistical Yearbook, China Grain Yearbook, the official website of the Ministry of Industry and Information Technology of the People’s Republic of China, the official website of the People’s Bank of China, the official website of the China Insurance Regulatory Commission (CIRC), the official website of the State Grain Administration of China, and the official website of local finance departments, spanning the 2014–2023 period. In this case, missing data were filled in using linear interpolation. Specifically, firstly, for a variable with a missing value, the data are grouped by province; secondly, the missing values are filled in chronological order based on known neighboring values; and, finally, the original variable with a missing value is replaced with the filled variable to construct the panel data. Hong Kong, Macau, Taiwan, and Tibet are not included in the study due to the lack of partial sample data. In order to exclude the effect of extreme values, the two ends of the spectrum of farmers’ incomes were reduced by 1%.

• Explained variables

In this paper, the per capita disposable income of farmers is selected as a proxy indicator for measuring farmers’ income. In order to exclude the impact of price level changes on farmers’ disposable income, it is deflated with 2013 as the base period.

• Core explanatory variables

This paper combines the financial inclusion evaluation index system established by the Global Partnership for Financial Inclusion (GPFI) and constructs a total of 22 indicators from the two dimensions of the degree of inclusion and the efficiency of financial functions and services; moreover, it measures the level of financial inclusion through the entropy value method, as shown in

Table 5. Inclusive finance indicator measurement system.

Primary Indicators	Secondary Indicators	Tertiary Indicators	Description of Indicators
Inclusive finance	Degree of inclusion	Population density of banking and financial institutions	Number of outlets/resident population
		Density of workers in banking and financial institutions	Number of employees/resident population
		Geographic density of banking and financial institutions	Number of outlets/land area
		Geographic density of workers in banking and financial institutions	Number of practitioners/land area
		Population density of insurance-based financial institutions	Number of outlets/ resident population
		Population density of employees of insurance-based financial institutions	Number of employees/resident population
		Geographic density of insurance-based financial institutions	Number of outlets/land area
		Geographic density of employees of insurance-based financial institutions	Number of practitioners/land area
		Population density of microfinance companies	Number of outlets/resident population
		Geographic density of microfinance companies	Number of outlets/land area
		Insurance claims per capita	Expenditure on insurance claims/ resident population
		Loans per capita	Loan balance/resident population
		Deposits per capita	Savings balance/resident population
		Agricultural loan ratio	Balance of agricultural loans/all loans
	Financial function service efficiency	Various loan balances	Balance of all loans to financial institutions
		Various deposit balances	Balance of all deposits in financial institutions
		Loan level	Various loan balances/GDP
		Deposit level	Various deposit balances/GDP
		Non-performing loan ratio	Non-performing loans/all loan balances
		Insurance depth	Insurance premium income/GDP
		insurance density	Insurance premium income/ resident population
		Year-on-year rate of change in premium income	Value of change in insurance income/ previous year’s insurance income

.

• Mediating variables

The paper refers to Tipi and Rehber (2006) [82] and Nin-Pratt and Yu (2010) [81] and chooses the DEA–Malmquist index method to measure total factor productivity in agricultural production. The input indicators were selected as the agricultural labor force, land input, and mechanization degree, which were measured by the number of rural employees, the sown area of crops, and the total power of agricultural machinery, respectively; and the production indicator was selected as the value of agricultural output.

• Control variables

In this paper, urbanization, farmers’ investment, the use of agricultural machinery, the use of fertilizer, and the level of economic development are selected as control variables.

Urbanization is mainly manifested in the transfer of population to cities, the expansion of the size of cities, and the development of the urban economy, and its impact on farmers’ income is manifested in the following aspects. First, the transfer of farmers to cities promotes land mobility to realize increased income for farmers. For farmers transferring into the land, the improvement of land transfer efficiency will enable them to quickly obtain land capital, realize the efficient integration and utilization of land, and obtain the benefits of economies of scale. For farmers transferring out of the land, they can not only obtain rental income, but also discard the land bound to a higher-quality employment platform. Second, the expansion of urban scale and urban economic development improves the employment environment for farmers, promoting the expansion of the secondary and tertiary industries, providing farmers with a large number of non-farming jobs, and the growth of farmers’ wage income. In this study, urbanization is measured using the proportion of urban population to total population in each province.

Farmers’ investment in agricultural machinery and equipment, farmland and farming infrastructure, and forestry production facilities can effectively improve production efficiency, and non-farm production infrastructure inputs can expand farmers’ sources of income. In this study, farmers’ investment is measured by the ratio of farmers’ fixed asset investment after deducting residential investment to agricultural GDP.

Increased efficiency in the use of farm machinery and fertilizers can reduce farmers’ labor costs, avoid nutrient losses, and ultimately increase productivity and bring about an increase in farmers’ incomes. In this study, the use of farm machinery is measured by the total power of farm machinery per hectare; the use of fertilizer is measured by the amount of agricultural fertilizer applied per hectare.

Stabilized and high-quality economic development can reduce residents’ precautionary savings and increase their consumption expenditures. Agricultural products, as necessities of life, will see their consumption demand increase under conditions of stable economic development. The growth of consumption demand for agricultural products promotes the sustainable development of agriculture and the growth of farmers’ income. The stable and high-quality development of the economy also provides farmers with a favorable employment environment, more employment opportunities, and increases farmers’ income. In this study, the level of economic development is measured using gross product, which is the sum of the market value of final goods and services produced, as shown in

Table 6. Results of descriptive statistics.

VarName	VarSymbol	Obs	Mean	Median	SD	Min	Max
Farmers’ Income	IN	300	1.604	1.471	0.644	0.696	3.841
Financial Inclusion	FIN	300	0.106	0.076	0.082	0.025	0.392
Degree of Inclusion	FPN	300	0.084	0.058	0.083	0.006	0.840
Financial Function Service Efficiency	FGN	300	0.114	0.071	0.105	0.021	0.457
Agricultural Total Factor Productivity	TFP	300	1.641	1.395	0.882	6.926	0.293
Urbanization	UR	300	0.624	0.610	0.112	0.402	0.941
Farmers’ Investment	FTZ	300	5.141	4.524	3.199	1.864	17.543
Agricultural Machinery Use	NJ	300	0.423	0.354	0.219	0.136	1.152
Fertilizer Use	HF	300	1.929	0.410	4.783	0.014	31.872
Level of Economic Development	GDP	300	0.116	0.056	0.204	0.002	1.643

.

4. Results and Discussion

4.1. The Total Effect of Financial Inclusion Affecting Farmers’ Income

Table 7. Regression results of financial inclusion and farmers’ income.

VAR	OLS	RE	FE	FE	FE
	(1)	(2)	(3)	(4)	(5)
FIN	3.242 ***	3.017 ***	2.986 ***	3.534 ***	1.226 ***
	(0.401)	(0.546)	(0.583)	(0.335)	(0.348)
UR	2.269 ***	6.282 ***	7.617 ***	1.373 ***	−1.053 **
	(0.272)	(0.372)	(0.394)	(0.237)	(0.432)
FTZ	−0.019 *	−0.034 ***	−0.030 ***	0.006	0.016 ***
	(0.010)	(0.008)	(0.007)	(0.009)	(0.005)
NJ	−0.070	−0.458 ***	−0.343 **	0.156	−0.106
	(0.124)	(0.140)	(0.139)	(0.105)	(0.078)
HF	−0.003	−0.002	−0.011	−0.009 *	−0.001
	(0.006)	(0.007)	(0.007)	(0.005)	(0.004)
GDP	−0.337 **	0.068	0.107	−0.390 ***	0.043
	(0.139)	(0.089)	(0.083)	(0.119)	(0.048)
Cons	0.017	−2.269 ***	−3.151 ***	0.341 **	2.093 ***
	(0.173)	(0.264)	(0.267)	(0.146)	(0.275)
Province fixed effects			yes	no	yes
Year fixed effects			no	yes	yes
N	300	300	300	300	300
${\mathrm{R}}^2$	0.474		0.909	0.651	0.974

Note: ***, **, * indicate 1%, 5%, and 10% significance levels, respectively, and the following regressions have the same results.

presents the results of the regression using the OLS model, random effects model, individual fixed effects model, time fixed effects model, and two-way fixed effects model. The results show that the estimated coefficients of financial inclusion are 3.242, 3.3017, 2.986, 3.534, and 1.226, respectively, and all of them are significant at a 1% confidence level, which indicates that the development of financial inclusion can significantly increase the level of farmers’ income.

Inclusive finance still has the characteristics of policy-based finance [88], and it is able to provide comprehensive financial services with wide coverage and relatively low cost for disadvantaged groups in the process of financial services. Payment and settlement functions can improve the efficiency of farmers’ access to income. Credit services can alleviate the plight of rural capital mismatch, improve farmers’ operational efficiency, and bring about income growth. Financial management services can activate idle funds and bring value-added benefits; secondly, they can optimize the structure of farmers’ incomes and keep them growing steadily over the long term. Insurance and other services can provide protection for farmers’ lives and operations, and hedge against the losses caused by risks. Hypothesis 1 is, therefore, verified.

It has been demonstrated earlier that inclusive financial development can significantly contribute to farmers’ income; so, is this impact short-term or long-term? In the process, does the degree of inclusion and financial functions have an impact on farmers’ incomes? If there is an impact, is it short-term or long-term? In order to explore the above questions, this paper analyzes the level of farmers’ income in the current period, the next period, and the next two periods by the level of inclusive financial development, the degree of inclusion, and the service efficiency of financial functions, the results of which are shown in

Table 8. Financial inclusion and the long-term effects of its dimensions.

VAR	T		T + 1			T + 2
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
FIN			1.085 ***			1.005 ***
			(0.349)			(0.335)
FPN	0.403 ***			0.329 ***			0.295 **
	(0.131)			(0.123)			(0.114)
FGN		0.950 **			1.018 *			1.127 *
		(0.478)			(0.550)			(0.609)
control variable	yes	yes	yes	yes	yes	yes	yes	yes
Cons	2.224 ***	2.160 ***	1.981 ***	2.104 ***	2.007 ***	1.900 ***	2.019 ***	1.919 ***
	(0.271)	(0.283)	(0.283)	(0.279)	(0.294)	(0.293)	(0.288)	(0.306)
Province fixed effects	yes	yes	yes	yes	yes	yes	yes	yes
Year fixed effects	yes	yes	yes	yes	yes	yes	yes	yes
N	300	300	270	270	270	240	240	240
${\mathrm{R}}^2$	0.973	0.973	0.977	0.977	0.977	0.982	0.981	0.981

Note: ***, **, * indicate 1%, 5%, and 10% significance levels, respectively, and the following regressions have the same results.

.

From Models (3) and (6), it can be seen that the estimated coefficient of financial inclusion is significantly positive at the 1% confidence level when farmers’ income in the next period and the next two periods are used as explanatory variables, indicating that there is a long-term income-generating effect of the level of financial inclusion development on farmers’ income. From Models (1) and (2), it can be seen that both the level of financial inclusion and the efficiency of financial function services can effectively promote the growth of farmers’ income, the confidence level of both of them is 1% and 5% respectively, and the effect of the level of financial inclusion is more significant than the efficiency of financial function services. In Models (4), (5), (7), and (8), when farmers’ incomes in the next period and the next two periods are used as explanatory variables, the estimated coefficients of the degree of inclusion and the efficiency of financial function services are all significantly positive. Among them, the estimated coefficients in Models (4) and (5) are 0.329 and 1.018, respectively, with a gradual decrease in significance; the estimated coefficients in Models (7) and (8) are 0.295 and 1.127, respectively, and the confidence level also decreases from 1% to 10%. This situation indicates that the promotion of farmers’ income by the dimensions of financial inclusion has a long-term effect, and the effect of the degree of inclusion is stronger.

Financial inclusion has the dual attributes of an inclusive nature and a financial function. Deepening the degree of inclusion can break the dilemma of “financial exclusion” in rural areas, promote the low-cost reach of financial services to farmers, and establish the basis for immediate as well as long-term participation of farmers in financial markets. Meanwhile, the financial function allows farmers to access financial services such as payment and settlement, credit and loans, and financial management and insurance. The payment and settlement function allows funds to flow efficiently in the process of farmers’ financing, production, sales, and financial management, which reduces farmers’ short-term transaction costs and improves the long-term turnover efficiency of funds. The credit system can be set up to realize credit tracking and rating, which can not only reduce the financing cost of farmers in time, but also cover the capital gap in the long cycle of agricultural production. Financial products such as insurance can provide long-term risk protection for farmers’ operations. Therefore, in both the short and long term, the development of inclusive finance, the deepening of the degree of inclusion, and the improvement of the service efficiency of the financial function can lead to an increase in farmers’ incomes.

Further, the degree of inclusion focuses on reshaping the relationship between farmers and the financial system. The degree of inclusion integrates farmers into the modern financial system by expanding outlets and lowering barriers to entry. Therefore, the degree of inclusion is the basis for farmers to utilize the service efficiency of the financial function to work. Only on the basis of farmers’ access to financial services can payment and settlement, credit loans, and financial products play a role in increasing income. Through the above analysis, we can see why the degree of inclusion has a stronger effect on farmers’ income than the service efficiency of the financial function.

4.2. Heterogeneity Analysis

The development of inclusive finance has contributed to the sustained growth of farmers’ incomes, and, in the process, the extent to which farmers with different incomes have benefited from inclusive finance may vary. For this reason, this paper uses a quantile regression model to explore the income increase in inclusive finance for farmers with different incomes, as shown in

Table 9. Impact of financial inclusion and its dimensions on farmers at different income levels.

Area	Variant	Timing	Obs	Q(10)	Q(25)	Q(50)	Q(75)	Q(90)
full sample	FIN	T	300	1.247 ***	1.201 ***	1.676 ***	3.406 ***	6.135 ***
				(0.358)	(0.406)	(0.499)	(0.641)	(0.993)
		T + 1	270	1.076 **	1.448 ***	1.961 ***	4.104 ***	6.719 ***
				(0.473)	(0.437)	(0.502)	(0.742)	(1.100)
		T + 2	240	1.213 **	1.605 ***	2.107 ***	3.942 ***	5.864 ***
				(0.548)	(0.470)	(0.529)	(0.719)	(1.168)
	FPN	T	300	0.772 **	1.730 ***	3.045 ***	5.400 ***	9.118 ***
				(0.350)	(0.372)	(0.452)	(0.429)	(0.760)
		T + 1	270	0.996 ***	1.455 ***	2.778 ***	5.867 ***	9.525 ***
				(0.354)	(0.361)	(0.457)	(0.485)	(0.960)
		T + 2	240	1.020 **	1.246 ***	1.916 ***	7.069 ***	10.563 ***
				(0.431)	(0.424)	(0.455)	(0.569)	(0.939)
	FGN	T	300	0.553 *	0.357	0.714 **	1.444 ***	3.549 ***
				(0.296)	(0.303)	(0.353)	(0.497)	(0.955)
		T + 1	270	0.740 **	0.425	0.813 **	1.958 ***	3.635 ***
				(0.362)	(0.361)	(0.389)	(0.538)	(0.915)
		T + 2	240	0.866 **	0.499	0.718 *	2.426 ***	4.285 ***
				(0.373)	(0.384)	(0.413)	(0.509)	(1.079)

Note: ***, **, * indicate 1%, 5%, and 10% significance levels, respectively, and the following regressions have the same results.

.

In both the short and long run, financial inclusion significantly promotes farmers’ income at different income levels, and the estimated coefficient of financial inclusion tends to increase with the increase in income level. From 0.1 to 0.9 quantile points, the estimated coefficients of inclusive finance on income in the current period are 1.247, 1.201, 1.676, 3.406, and 6.135, respectively; the estimated coefficients of inclusive finance on income in the next period are 1.076, 1.448, 1.961, 4.104, and 6.719, respectively; and the estimated coefficients of inclusive finance on income in the next two periods are 1.213, 1.605, 2.107, 3.942, and 5.864, respectively. The above results indicate that the impact of the development of inclusive finance on the sustained growth of farmers’ incomes increases as farmers’ incomes increase.

Rising income levels bring about an improved living environment for farmers, who, in addition to sustaining their daily lives, have more capacity and financial resources to expand their knowledge and awareness, and are more likely to be exposed to inclusive finance and more willing to accept inclusive financial services. Therefore, as income levels rise, farmers can continue to benefit from the development of inclusive finance.

In terms of the dimensions of financial inclusion, the degree of inclusion significantly contributes to the sustained growth of farmers’ incomes at different quantile levels and the estimated coefficients of the degree of inclusion show an increasing trend, although the financial function services efficiency does not have a significant impact on farmers’ incomes at the 0.25 quantile; however, the financial function services efficiency is significantly positively correlated with the current and future incomes of farmers at the other quantile levels. This suggests a deepening dependence of farmers’ incomes on the degree of inclusion and the financial function services’ efficiency, both in the short and long term.

Based on the average GDP level of each province from 2014 to 2023, the median is used as the cutoff point to divide the provinces into economically developed and less developed regions to further examine the regional heterogeneity of the impact of financial inclusion on farmers’ income, and the results are shown in

Table 10. Regional differences in the impact of financial inclusion and its dimensions on farmers at different income levels.

Area	Variant	Timing	Obs	Q (10)	Q (25)	Q (50)	Q (75)	Q (90)
economically developed area	FIN	T	150	3.724 ***	3.977 ***	3.998 ***	5.040 ***	5.677 ***
				(0.391)	(0.491)	(0.652)	(0.794)	(1.112)
		T + 1	135	4.291 ***	4.124 ***	4.558 ***	6.157 ***	6.040 ***
				(0.412)	(0.458)	(0.676)	(0.860)	(1.156)
		T + 2	120	4.724 ***	4.604 ***	5.177 ***	6.538 ***	6.359 ***
				(0.450)	(0.547)	(0.749)	(0.704)	(1.172)
	FPN	T	150	2.738 ***	3.130 ***	5.104 ***	4.975 ***	8.199 ***
				(0.505)	(0.716)	(0.514)	(0.798)	(1.092)
		T + 1	135	3.201 ***	2.609 ***	5.536 ***	5.531 ***	8.265 ***
				(0.427)	(0.772)	(0.526)	(0.901)	(1.182)
		T + 2	120	3.057 ***	2.538 ***	6.083 ***	5.950 ***	8.860 ***
				(0.707)	(0.806)	(0.620)	(1.027)	(1.362)
	FGN	T	150	2.595 ***	2.500 ***	3.225 ***	3.721 ***	4.215 ***
				(0.399)	(0.384)	(0.564)	(0.720)	(0.824)
		T + 1	135	2.960 ***	3.038 ***	3.176 ***	3.857 ***	3.586 ***
				(0.357)	(0.420)	(0.554)	(0.799)	(1.002)
		T + 2	120	3.376 ***	3.292 ***	3.338 ***	4.812 ***	3.607 ***
				(0.378)	(0.434)	(0.604)	(0.895)	(1.369)
economically less developed area	FIN	T	150	0.351	0.197	−0.073	−0.434	0.314
				(0.531)	(0.535)	(0.756)	(0.887)	(0.859)
		T + 1	135	0.376	0.096	0.531	0.241	0.400
				(0.545)	(0.589)	(0.766)	(0.812)	(0.916)
		T + 2	120	0.539	0.474	0.995	−0.260	−0.453
				(0.503)	(0.558)	(0.761)	(0.907)	(1.237)
	FPN	T	150	0.709 *	1.436 ***	1.989 ***	3.444 ***	2.335 ***
				(0.365)	(0.433)	(0.624)	(0.718)	(0.521)
		T + 1	135	1.016 **	1.534 ***	1.315 **	3.916 ***	2.602 ***
				(0.399)	(0.441)	(0.526)	(0.671)	(0.682)
		T + 2	120	0.984 ***	1.425 ***	1.962 ***	0.834	3.481 ***
				(0.350)	(0.394)	(0.518)	(0.822)	(0.779)
	FGN	T	150	0.074	−0.206	−0.321	−0.307	−0.530
				(0.364)	(0.414)	(0.547)	(0.621)	(0.551)
		T + 1	135	−0.017	−0.055	−0.007	−0.302	−0.198
				(0.426)	(0.400)	(0.557)	(0.551)	(0.679)
		T + 2	120	−0.148	−0.138	0.095	−0.316	−0.683
				(0.419)	(0.427)	(0.551)	(0.607)	(0.761)

Note: ***, **, * indicate 1%, 5%, and 10% significance levels, respectively, and the following regressions have the same results.

.

In terms of financial inclusion, in economically developed regions, financial inclusion has a significant positive impact on farmers’ income in both the short and long term. However, in less economically developed regions, the impact of financial inclusion on farmers’ income is not significant. In terms of the dimensions of financial inclusion, in both economically developed and underdeveloped regions, the degree of financial inclusion in general significantly contributes to the growth of farmers’ short-term and long-term incomes; however, the services’ efficiency of the financial function significantly contributes to farmers’ short-term and long-term incomes only in economically developed regions.

Farmers in less economically developed areas tend to operate on a smaller scale [89], and small-scale operators have relatively lower capital needs and more conservative business philosophies; therefore, their demand for financial services is also relatively low. Insufficient market demand cannot effectively stimulate the enthusiasm of inclusive financial institutions. On the one hand, it will lead to financial outlets that cannot effectively reach the farmers in economically underdeveloped areas; on the other hand, it will cause inclusive financial institutions to be reluctant to conduct in-depth market research, which reduces the accuracy of the match between the financial products and the needs of farmers. Insufficient financial demand and a lack of supply, ultimately, lead to the development of inclusive finance in economically underdeveloped regions lagging behind; the financial function fails to give full play to its service’s efficiency, and farmers are unable to raise their income level through inclusive finance.

4.3. Transmission Mechanisms of the Impact of Financial Inclusion on Farmers’ Incomes

According to the previous theoretical analysis, this paper verifies the mediating effect of agricultural total factor productivity in the process of inclusive finance acting on farmers’ income, and the results are shown in

Table 11. Transmission mechanism of the impact of financial inclusion on farmers’ income: total factor productivity in agriculture.

	T		T + 1	T + 2
VAR	TFP	IN	IN	IN
	(1)	(2)	(3)	(4)
TFP		0.019 *	0.025 ***	0.016
		(0.010)	(0.010)	(0.010)
FIN	3.781 *	1.154 ***	0.954 ***	0.897 ***
	(2.239)	(0.348)	(0.348)	(0.340)
control variables	yes	yes	yes	yes
Cons	−2.549	2.141 ***	2.089 ***	2.020 ***
	(1.766)	(0.274)	(0.282)	(0.300)
N	300	300	270	240
${\mathrm{R}}^2$	0.420	0.974	0.978	0.982

Note: ***, and * indicate 1%, and 10% significance levels, respectively, and the following regressions have the same results.

.

In Model (1), the estimated coefficient of financial inclusion on agricultural total factor productivity is significantly positive, indicating that financial inclusion contributes to the increase in agricultural total factor productivity. In Model (2), the estimated coefficients of agricultural total factor productivity and inclusive finance are 0.019 and 1.154, which passed the significance test at 10% and 1% confidence level, respectively. Combining the results of Model (1) and Model (2), it can be known that there is a mediating effect of agricultural total factor productivity in the process of the impact of inclusive finance on farmers’ income, and that the increase in the level of inclusive finance development will promote the growth of agricultural total factor productivity, which, in turn, will bring about the growth of farmers’ income.

It has been verified above that agricultural total factor productivity has a mediating effect in the process of financial inclusion, promoting farmers’ short-term income growth. Then, does agricultural total factor productivity also have a mediating effect in the process of financial inclusion promoting farmers’ long-term income growth? Models (3) and (4) answer the above question. Although the estimated coefficients of agricultural total factor productivity in Model (4) are not significant, the estimated coefficients of agricultural total factor productivity and inclusive finance in Model (3) are 0.025 and 0.954, respectively, which are both significant at the 1% confidence level, indicating that inclusive finance can raise the long-term income level of farmers by promoting the increase in agricultural total factor productivity.

By providing comprehensive financial services to farmers, inclusive financial institutions have cracked farmers’ financing problems, dispersed business risks, and improved production management capabilities, directly reducing farmers’ operating costs and achieving short-term income increases; moreover, they have optimized the allocative efficiency of capital, labor, and human capital allocation, and have achieved the modernization and large-scale development of agricultural production, bringing about a long-term increase in farmers’ operating incomes.

This paper further explores the mediating effect of agricultural total factor productivity in the process of influencing farmers’ income in each dimension of financial inclusion, and the results are shown in

Table 12. Transmission mechanisms of the dimensions of financial inclusion affecting farmers’ income: total factor productivity in agriculture.

VAR	T		T + 1	T + 2	T		T + 1	T + 2
	TFP	IN	IN	IN	TFP	IN	IN	IN
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
TFP		0.019 **	0.026 ***	0.018 *		0.022 **	0.027 ***	0.019 *
		(0.010)	(0.010)	(0.010)		(0.010)	(0.010)	(0.010)
FPN	1.478 *	0.375 ***	0.287 **	0.263 **
	(0.835)	(0.131)	(0.123)	(0.115)
FGN					1.638	0.914 *	0.889	0.947
					(3.039)	(0.475)	(0.543)	(0.613)
control variables	yes	yes	yes	yes	yes	yes	yes	yes
Cons	−2.197	2.266 ***	2.201 ***	2.135 ***	−2.132	2.206 ***	2.121 ***	2.063 ***
	(1.729)	(0.270)	(0.277)	(0.294)	(1.797)	(0.281)	(0.292)	(0.313)
N	300	300	270	240	300	300	270	240
${\mathrm{R}}^2$	0.421	0.974	0.978	0.982	0.414	0.973	0.977	0.981

Note: ***, **, * indicate 1%, 5%, and 10% significance levels, respectively, and the following regressions have the same results.

. In Models (1) and (5), the estimated coefficients of the degree of financial inclusion and the efficiency of financial function services on agricultural total factor productivity are 1.478 and 1.638, respectively, and the degree of financial inclusion is significantly positive; in Model (2), the estimated coefficients of the degree of agricultural total factor productivity and the degree of financial inclusion are 0.019 and 0.375, respectively, which are significant at the confidence level of 5 percent and 1 percent; in Model (6), the estimated coefficients of agricultural total factor productivity and financial function service efficiency are 0.022 and 0.914, respectively, and both are significantly positive. The above regression results indicate that, in the process of the degree of inclusion to promote the growth of farmers’ income, the total factor productivity of agriculture can play a mediating effect; however, the effect is not obvious in the efficiency of financial function services. From the perspective of farmers’ long-term income, the estimated coefficients of agricultural total factor productivity and the degree of inclusion in Models (3) and (4) are 0.026, 0.287, 0.018, and 0.263, respectively. The confidence level and the marginal effect, although decreased, are significantly positive, indicating that the deepening of the scope of financial coverage and the degree of benefit to the people can effectively improve agricultural total factor productivity, which, in turn, promotes the sustained growth of farmers’ income. This effect is weakened with the passage of time. In Models (7) and (8), the estimated coefficients of agricultural total factor productivity are significantly positive; however, the estimated coefficients of financial function service efficiency are not significant.

There is a difference in the effect of agricultural total factor productivity on the role of the two dimensions of financial inclusion in driving farmers’ income growth. This discrepancy essentially reflects the paradox of the applicability of financial services to agricultural operations. The degree of inclusion allows for higher non-performing loan ratios, the exploration of non-traditional collateral, and the optimization of product design with policy intervention. Because of this, farmers and other disadvantaged groups can use inclusive finance to access credit facilities that match their production cycles and agricultural insurance that effectively hedges their business risks, thereby enhancing agricultural total factor productivity and realizing increased incomes for farmers. The efficiency of financial services reflects the financial attributes of inclusive finance. Finance is profit-seeking, and the long cycle, high risk, and low profit characteristics of the agricultural sector make agricultural production face natural “financial exclusion”. Therefore, although the degree of financial inclusion allows financial services to reach farmers, the imperfect credit evaluation system and guarantee market, as well as the limitations of farmers’ financial literacy and cognition, all constrain the effect of financial function service efficiency on total factor productivity in agriculture, which, in turn, affects the income-generating effect of farmers. This result also highlights the direction for the further development of inclusive finance. The government should not only further guide inclusive finance to sink into the countryside but also build a perfect credit system and continuously improve the financial literacy of farmers, so that the efficiency of financial functions and services in the process of agricultural production can effectively improve the total factor productivity of agriculture, which, in turn, will bring about an increase in the income of farmers.

5. Endogeneity and Robustness Tests

Sustained income growth can accumulate capital for farmers, and capital enhancement allows farmers to improve their living environment and have more opportunities to access inclusive finance, thus contributing to the development of inclusive finance. In order to avoid the endogeneity problems caused by the aforementioned reverse causality between financial inclusion and farmers’ income, omitted variables, and measurement error of financial inclusion, this paper adopts the average level of financial inclusion in the lagged period except for this province multiplied by the number of banks in 2005 as the instrumental variable for financial inclusion. The reasons for choosing this instrumental variable are as follows: first, the development of inclusive finance is a continuous process, the present level of development is gradually formed on the basis of the past, and the number of banks can reflect the development of finance, which will inevitably have an impact on the local level of inclusive finance. Secondly, in 2005, China’s financial development was more backward, the level of financial inclusion was low, and the financial data in 2005 did not have an impact on farmers’ income after 2014. Therefore, the selection of instrumental variables in this paper meets the requirements of relevance and exogeneity. In addition, 2SLS, GMM, and LIML regression models are also used in this paper for robustness testing. The purpose of choosing GMM and LIML regression models for testing, in addition to the 2SLS regression model, is to avoid the influence of weak instrumental variables and heteroskedasticity, etc., on the results, and to compare the 2SLS results in order to test the robustness.

This study explains the principle of the endogeneity treatment of the GMM model in detail. The principle of the GMM model to perform an endogeneity test is to estimate and test through conditional moments to achieve the following two objectives: first, to determine whether it is necessary to use instrumental variables to deal with endogeneity; second, to determine whether the selection of instrumental variables meets the conditions. For the first objective, the GMM model utilizes the results of the endogeneity test to make judgments. For the second objective, the GMM model utilizes the weak instrumental variables test (first-stage F-value) to determine whether the instrumental variables are strongly correlated with the endogenous variables. When the number of instrumental variables is greater than the number of endogenous variables, the results of the Hansen J test can be used to determine whether the instrumental variables are independent of the error term. Since the number of instrumental variables is equal to the number of endogenous variables in this study, the Hansen J test cannot be used, and we have already clarified exogeneity through a theoretical analysis in the previous section. However, in this process, the first and second stage error terms’ autocorrelation may have an impact on the results; therefore, we need to utilize the Breusch–Godfrey test to check whether there is autocorrelation between the first and second stage error terms.

As shown in

Table 13. Endogeneity test.

VAR	2SLS	GMM	LIML
	(1)	(2)	(3)
First-stage regressions	FIN	FIN	FIN
IVI_FIN	0.531 ***	0.531 ***	0.531 ***
	(0.142)	(0.142)	(0.142)
F-value	13.97	13.97	13.97
First-order autocorrelation p-value		0.164
Second-order autocorrelation p-value		0.379
Second-stage regressions	IN	IN	IN
FIN	0.457 ***	0.457 ***	0.457 ***
	(0.155)	(0.155)	(0.155)
Control variable	yes	yes	yes
Province fixed effects	yes	yes	yes
Year fixed effects	yes	yes	yes
Endogeneity test	0.0001	0.0001	0.0001
First-order autocorrelation p-value		0.395
Second-order autocorrelation p-value		0.363
N	270	270	270
${\mathrm{R}}^2$	−0.648	−0.648	−0.648

Note: *** indicates 1% significance level, and the following regressions have the same results.

, firstly, the endogeneity test rejected the original hypothesis at a 1% significance level in the regression results of the 2SLS model, GMM model, and LIML model. This indicates that there is an endogeneity problem that needs to be dealt with using instrumental variables. Secondly, the F-value of the first stage in the regression results of the 2SLS model, GMM model, and LIML model is greater than 10, which indicates that the selected instrumental variables satisfy the condition of endogeneity and are strongly correlated with the endogenous variables. Moreover, the exogeneity of instrumental variables has been explained in the previous section. Therefore, the selection of instrumental variables is reasonable.

Further, the first-stage Breusch–Godfrey test for first-order autocorrelation p-value of 0.164 and second-order autocorrelation p-value of 0.379 in the GMM model are greater than 0.05; thus, the original hypothesis cannot be rejected. This indicates that there is no autocorrelation in the first stage. The second-stage Breusch–Godfrey test first-order autocorrelation p-value is 0.395, and the second-order autocorrelation p-value is 0.363. Both are greater than 0.05, and, again, the original hypothesis cannot be rejected. This indicates that there is no autocorrelation in the second-stage error term as well. In conclusion, the results of the GMM test are robust.

The results of the second stage of the 2SLS, GMM, and LIML models show that the coefficient of financial inclusion is positive and significant at the 1% confidence level when endogeneity is taken into account, suggesting that the development of financial inclusion has become an effective driver of farmers’ incomes.

To further ensure the reliability of the results, this paper conducts a robustness test in the following ways.

First, the explanatory variables are replaced with the sum of wage income and business income. Data from the China Bureau of Statistics show that, by the end of 2023, wage and business incomes accounted for more than 70% of farmers’ incomes.

Second, the data for the four provinces of Xinjiang, Inner Mongolia, Ningxia, and Guangxi are deleted to avoid the impact of provincial autonomous regions due to different policies, and so on.

Third, the impact of financial inclusion and other policies on farmers’ income may have a lag, and financial inclusion lags one and two periods to replace the explanatory variables for testing.

Fourth, with the development of the digital economy, digital technology is more and more widely used in farmers’ production and life, and at the same time, there is some impact of risk shocks on farmers’ income; therefore, this study will add digital technology and risk to the control variables for robustness testing.

Possible impacts of digital technology on farmers’ income include the following: first, digital technology improves the macro employment environment, so that farmers can keep abreast of employment information, and promotes farmers towards better employment platforms; second, digital technology is deeply integrated with the industrial chain, and the business processes such as production, processing, circulation, and sales are moving towards informatization and intelligence, promoting the growth of farmers’ business income; third, digital technology has become an important tool for farmers to participate in the rural property rights market, leasing market, and investment market, increasing farmers’ property income. In this study, we refer to the methods of Li Zhenghui et al. (2024) [90], Leng Xuan (2022) [91], and Liu Bo et al. (2023) [92], and utilize the entropy value method to measure digital technology from the two dimensions of digital foundation development and digital network development, respectively, as shown in Table 10.

The possible impact of risk on farmers’ incomes includes the following: The agricultural sector is characterized by a high level of risk, which may have a negative impact on agricultural operations. Capital is an important factor of production that affects farmers’ business decisions and operational efficiency; the growth characteristics of plants and animals make agricultural operations vulnerable to natural factors such as weather, disease, pests, and biological invasions; and the conversion of agricultural products into income is affected by factors such as price fluctuations and market structure. In this study, we refer to Zhenghui Li et al. (2024) [93], Saqib Shahab E et al. (2016) [94], and Ullah Raza et al. (2016) [95], and utilize the entropy value method to measure the risk indicators in terms of financial risk, natural risk, and market risk faced by farmers, as shown in

Table 14. Digital technology and risk measurement indicator system.

Primary Indicators	Secondary Indicators	Tertiary Indicators	Description of Indicators
Digital technology	Digital foundation	Rural delivery line density	Rural delivery routes/area size
		Long-haul fiber optic cable line density	Fiber optic cable length/area size
		Density of cell phone base stations	Number of cell phone base stations/area size
	Digital network	Rural broadband Internet penetration	Number of rural broadband Internet subscribers/number of resident population
		Rural cell phone penetration	Mobile telephones per 100 inhabitants in rural areas
		Rural computer penetration rate	Computers per 100 inhabitants in rural areas
Risk	Financial risk	Formal loan repayment burden ratio	Loans to farmers from financial institutions/ Gross disposable income of farmers
		Contingency funding coverage	Available cash/necessary expenditures
	Natural risk	Coefficient of variation of precipitation	Standard deviation of precipitation in the past five years/average precipitation in the past five years
		Density of climate disasters	Area of agricultural land affected by severe weather Area of agricultural land
		Density of biological disasters	Area of agricultural land affected by pests and diseases/Area of agricultural land
	Market risk	Price coefficient of variation	Number of rural broadband Internet subscribers/number of resident population
		Market concentration index	Standard deviation of wholesale prices of agricultural commodities in the past five years/Average wholesale prices of agricultural commodities in the past five years

.

Table 15. Robustness test.

VAR	Replacement of Explanatory Variables	Replacement Samples	Replacement of Explanatory Variables		Add Control Variables of Digital Technology	Add Control Variables of Risk	Add Control Variables of Digital Technology and Risk
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
FIN	0.665 **	1.068 ***			0.838 ***	1.252 ***	0.914 ***
	(0.267)	(0.372)			(0.285)	(0.334)	(0.282)
L.FIN			0.859 **
			(0.339)
L2.FIN				0.816 ***
				(0.310)
Control variable	yes	yes	yes	yes	yes	yes	yes
Province fixed effects	yes	yes	yes	yes	yes	yes	yes
Year fixed effects	yes	yes	yes	yes	yes	yes	yes
Cons	1.921 ***	2.035 ***	2.122 ***	2.157***	1.805 ***	1.672 ***	1.651 ***
	(0.295)	(0.423)	(0.394)	(0.396)	(0.314)	(0.374)	(0.313)
N	300	260	270	240	300	300	300
${\mathrm{R}}^2$	0.975	0.973	0.978	0.984	0.979	0.971	0.979

Note: ***, and ** indicate 1%, and 5% significance levels, respectively, and the following regressions have the same results.

shows the test results based on the four methods mentioned above. In Table 15, the estimated coefficients of financial inclusion in Models (1) to (7) are significantly positive, and the signs of the coefficients are consistent with the signs of Models (1) to (5) in Table 7, indicating that the promotion effect of financial inclusion on farmers’ income is robust and reliable.

6. Conclusions and Suggestions

At present, inclusive finance is developing strongly in China’s rural areas and has become an indispensable part of farmers’ productive lives. Deepening the development of inclusive finance and allowing the fruits of inclusive financial development to drive the long-term and stable growth of farmers’ incomes are key to alleviating relative poverty in order to build an inclusive economic growth system. Using panel data from 30 provinces in China from 2014 to 2023, this study explores the impact of financial inclusion on farmers’ income and the transmission mechanism through agricultural total factor productivity from both theoretical and empirical perspectives through two-way fixed-effects models and mediation models, and draws the following conclusions. Firstly, the development of inclusive finance and both the dimensions of financial inclusion and the efficiency of financial functions and services can significantly increase the income level of farmers and have long-term income-generating effects. Among them, the degree of financial inclusion has a stronger effect than the other dimensions of financial inclusion. Secondly, in both the short and long term, as farmers’ income increases, their dependence on financial inclusion and its dimensions deepens, and the effect is stronger in economically developed regions than in less developed regions. Thirdly, an increase in the level of inclusive financial development will promote the growth of total factor productivity in agriculture, which, in turn, can lead to the growth of farmers’ income in both the short and long term. The deepening of the financial coverage and the degree of benefit to the people can effectively improve the total factor productivity of agriculture, and then promote the sustained growth of farmers’ income, which is weakened with the passage of time. Based on the above conclusions, this paper proposes the following suggestions.

In response to the finding that the degree of financial inclusion has a stronger effect on farmers’ income than the efficiency of the financial function, the study makes two recommendations.

The first recommendation is to expand the scope of financial inclusion. The purpose of expanding the scope of “universal” is to make financial inclusion more accessible to a wider range of people. First, digital technology should be fully utilized to expand the coverage of inclusive financial infrastructure. Digital technology has the characteristics of borderless dissemination, two-way dissemination, and decentralized dissemination, and the application of digital technology can effectively realize the coverage of financial inclusion. On the one hand, it is necessary to combine the characteristics of local economic development, accelerate the promotion of digital economic development and the construction of digital villages, and realize the digital transformation of rural infrastructure and the digital transformation of farmers’ lives, so as to lay a good foundation for the development of digital inclusive finance. On the other hand, inclusive financial institutions should actively apply digital technology, realize the deep integration of digital technology and inclusive finance, and develop digital inclusive finance. At the same time, digitally inclusive financial products (mobile payment, digital banking, smart outlets, etc.) are being promoted to remote areas, such as rural areas, through cooperation with the government or the formation of specialized product promotion departments. Second, the identity threshold of inclusive finance should be eliminated so that farmers can access financial services conveniently. According to the relevant regulations of inclusive financial institutions, farmers must first open an account with an inclusive financial institution in order to access financial services and purchase inclusive financial products (agricultural insurance, agricultural funds, and agricultural credit). Farmers purchasing some special inclusive financial products (e.g., credit products) also need to verify the user’s credit information. These restrictions do not facilitate farmers’ access to inclusive financial services. Therefore, the process of opening an account should be simplified so that farmers can open an account online through face recognition, electronic signatures, remote fingerprint records, etc., to reasonably reduce the resistance of farmers to using inclusive financial products. At the same time, inclusive financial institutions should utilize information that is easy to collect and can confirm farmers’ creditworthiness. For example, collecting electricity bill payment records and relevant information should be used to establish credit files for farmers who do not have them. Furthermore, it is necessary to strengthen the relevance of services and products and provide relevant financial services according to farmers’ needs. Expanding the scope of inclusive finance is essentially a matter of pulling demand and realizing supply, or providing products and creating demand. Either way, relevant products must be designed with farmers’ characteristics in order to expand the market for inclusive finance in rural areas. Therefore, inclusive financial institutions should focus on farmers and provide inclusive financial services such as microcredit, policy credit, agricultural insurance against pests, agricultural insurance against market price fluctuations, and sound financial management products, in accordance with the characteristics of the needs of farmers in different regions.

The second recommendation is the scientific implementation of the degree of “benefit” of financial inclusion. The extent of the “benefits” to be realized is that after farmers have access to inclusive finance, they can fully utilize it to gain benefits. The government should play a leading and coordinating role in guiding inclusive finance to sink into rural areas. On the one hand, it should increase financial subsidies to fully address the difficulties encountered by inclusive financial institutions in conducting their business. At the level of direct support, the government can take the form of subsidizing loan interest, allocating loans at preferential policy rates, and injecting financial capital to provide financial subsidies, loan support, and share operating costs. Indirect support level: The government can utilize governmental financing institutions and set up special funds (financing credit enhancement fund, agricultural credit guarantee fund) to share the operating risks of financial institutions, as well as provide tax incentives to P&P financial institutions by reducing or exempting part of the value-added tax (VAT) and income tax and incentivize PFIs to actively conduct business, innovative pilots, and healthy competition by setting up industry-specific rewards (annual scale of business awards, annual incremental business awards, and special awards for science and technology finance, etc.). On the other hand, the government should play the role of a supervisor and resolutely resist violations by P&I financial institutions. In terms of the main body of supervision, a three-tier supervision system should be formed with the government as the formulator, the CBI as the implementer, and the disciplinary departments of P&I financial institutions as the enforcer. In terms of the content of supervision, it is necessary to strengthen the supervision of the market access, risk control, information disclosure, and other day-to-day affairs of P&I financial institutions, as well as to establish a mechanism for the public to make complaints and reports, so as to form a comprehensive supervisory system.

Based on the conclusion that farmers in low-income and economically underdeveloped areas have benefited from the development of inclusive finance, in addition to the need to continuously increase the efforts of inclusive finance in terms of both “inclusion” and “benefits”, this study also puts forward the following two recommendations.

The first recommendation is that inclusive financial institutions should develop inclusive financial services according to the characteristics of farmers in low-income and economically less developed regions. For farmers in low-income and economically underdeveloped regions, the way to increase their income is often through traditional agriculture. Therefore, inclusive financial institutions should provide financial services that take into account the characteristics of traditional agriculture. On the one hand, inclusive financial institutions can combine land management rights and farmers’ credit profiles. P&I financial institutions can take land contract rights as a mandatory condition for credit evaluation and build a credit evaluation system by combining creditworthiness, the payment of utility bills, and online shopping. On the other hand, farmers in low-income and economically underdeveloped areas are predominantly agricultural, making them often lack tradable and valuable collateral. Based on this, inclusive financial institutions can use farmers’ forest rights, fishing boats, and land management rights as collateral, addressing the situation of farmers’ lack of valuable collateral.

The second recommendation is to improve the financial literacy of farmers in low-income and economically underdeveloped areas so that they can enjoy the dividends of inclusive finance. First, a tiered education system should be formed to achieve differentiated teaching. On the one hand, to improve the financial literacy of farmers, it is necessary to cultivate rural “financial leaders”. By prioritizing the cultivation of key people with convening power and leadership, such as village committees, village cadres, and the heads of cooperatives, farmers will be motivated to learn financial knowledge. On the other hand, financial knowledge is specialized and needs to be disseminated in a reasonable way. Government departments can set up “financial classrooms” during the busy farming season, explaining the application process of agricultural credit and the claim process of agricultural insurance, and analyzing financial products and other financial knowledge in accordance with the needs of farmers’ production. Secondly, we should make good use of digital technology to expand the dissemination of financial knowledge through digital means. Inclusive financial institutions can take advantage of the current trend of short videos to explain financial stories and answer financial problems by producing short videos and live broadcasts. At the same time, village committees can set up “financial literacy improvement groups” to produce financial knowledge graphics and explanation courses from time to time in the group chat.

The deepening of financial coverage and benefits to the people can effectively increase total factor productivity in agriculture, thus promoting the sustained growth of farmers’ income, while the effect of financial function service efficiency is not significant. This result also points out the direction for the further development of inclusive finance. First of all, we should optimize financial products to meet the needs of agricultural total factor productivity. Inclusive financial institutions should provide credit and insurance products covering the whole industrial chain of “planting, processing, circulation, marketing” to meet the financial needs of farmers from production to sales, prevent pests and diseases, climate hazards, and market fluctuations, and protect farmers’ stable production. Secondly, it is necessary to promote accurate data services for agricultural production. PFIs should set up an agricultural data platform to datamaterialize field crops, climate environment, and soil, and establish a data analysis platform as a reference basis for PFIs to assess risks and provide loan funds. Further, the organization of agricultural production should be optimized. The Government and inclusive financial institutions should promote the development model of “cooperatives + farmers + finance”, so that the products and services provided by inclusive financial institutions can be rapidly promoted among their members under the impetus of cooperatives, thus further safeguarding farmers’ production and operation.

In 2005, the United Nations had already put forward the concept of financial inclusion. And it was only in November 2013 that China formally proposed it. Compared to other countries with the same economic volume, China’s late start in financial inclusion has meant that most of the data related to financial inclusion stays at the macro level. Macrodata are often difficult to portray the individual characteristics of micro subjects (farmers). Therefore, our next step is to follow the research team to visit the field. Currently, we have visited Jiangxi, Hunan, Anhui, Hubei, Jiangsu, Zhejiang, Sichuan, Chongqing, and other regions following the research group to investigate the local farmers’ income situation and the development of inclusive finance. This study has laid a foundation for our work, and our next research work will further improve the microdata of financial inclusion and strive to reflect the impact of financial inclusion on farmers’ income in a real and detailed way.