Financial Literacy, Borrowing Behavior and Rural Households’ Income: Evidence from the Collective Forest Area, China

Abstract

Since the introduction of a series of collective forest tenure reforms in China, diverse forestland mortgage financial products have been available even in rural areas. It is difficult for rural households to make appropriate financial decisions in order to increase their income due to a lack of financial knowledge and relevant skills. It is important to analyze the relationship between financial literacy, borrowing behavior, and rural household income. Based on the learning-by-doing theory, the credit constraint theory, and the data obtained from the survey of 460 households in five rural countries of Liaoning Province, alternative econometric models were used to estimate the “Financial Literacy-Borrowing Behavior-Household Income” transmission channel. The findings reveal that the financial literacy levels are positively associated with household income and that there is an inverted U relationship between them which is low on both sides and high in the middle. In addition, the financial literacy has a significantly positive effect on the farmers’ credit behavior, which in turn promotes their income growth. The results provide a new perspective on the study and a clear explanation of the role of financial literacy in improving the loan amountsavailable in China’s rural areas. The paper concludes with recommendations for policymakers to prioritize financial education that will promote and support credit constraint reduction in collective forest areas.

1. Introduction

Forestland tenure is a crucial component of rural land policy that influences the marketization, the diversification of the economy, and environmental protection throughout the world, especially in developing countries [1,2,3]. Ghana, Indonesia, the Philippines, Vietnam, and other countries have created legislation and devised policies concerning the definition and allocation of property rights to improve the management of forestland [4,5,6,7]. A new wave of reforms regarding collective forestland tenure was implemented in China, beginning in the early 2000s [8,9]. Forestland can be used as collateral for loans by households in this situation. With the progress of reforms in China’s rural land financial systems, there has been a proliferation of new financial products in recent years that cater to the collective forestlands, such as forestry cooperative guarantees, forest rights mortgages, and public interest pledges for forest compensation revenue rights [10,11]. However, there are differences in the responses of the collective forest area farmers to the forestland mortgage policies and the household welfare that they generate [12]. While some people actively apply for the forestland mortgage loan, others never apply for it even if they need it. Similarly, there can be significant differences in the economic benefits that people are able to achieve after obtaining the loans [13,14].

This begs the question of whether these differences are related to their levels of financial literacy. The collective forest areas often tend to be inaccessible and disaster-prone, with the people living in them being poorly educated in general and lagging behind those living in the plains in terms of awareness and financial literacy [15,16]. Farmers find their economic fortunes being increasingly exposed to and influenced by the dynamics of land financial markets. The less financially literate among them are not able to deal as effectively with the new dynamic environment, reducing their production efficiency and driving down their income [17]. According to the “Analysis Report on Consumer Financial Literacy Survey” released by the Bureau of Financial Consumer Protection of the People’s Bank of China in 2017, the correct rate of financial knowledge questions among Chinese rural consumers was 50.7%, 4.08% lower than that of urban consumers; the correct rate of identifying financial products or services in rural areas was 53.4%, 4.2% lower than that of urban consumers [18]. There is a significant disparity between the urban and rural areas in terms of financial knowledge and skills.

Existing studies have focused primarily on retirees, students, consumers, and investors. Previous studies have revealed that borrowing behavior is influenced by the level of financial literacy. As one may expect, individuals with low financial literacy tend to be limited in terms of financial knowledge, the understanding and application of skills, access to relevant information, and awareness of the policies and various financial procedures, which makes them prone to suffering from cognitive bias and causes them to underestimate investment risks and face self-rationing in credit [19,20,21,22]. In addition, the lack of financial literacy increases the likelihood of wrong financing decisions, leading to higher borrowing costs, limited liquidity of available funds, and a higher risk of sustaining investment losses, which thus adversely affects household income [23,24,25]. They have also revealed the relationship between financial literacy and household income. Firstly, financial literacy is directly correlated with household income; families with a higher level of financial literacy have a higher chance of increasing their household income [26].Secondly, different types of financial literacy, such as deposits literacy, risk literacy, and debt literacy, have positive effects on wealth [27]. While financial literacy and education are both correlated with saving behavior, higher education is positively associated with more wealth only in conjunction with financial literacy [28].

The extant literature has analyzed the strengthened relationship between forestland mortgage credit behavior and rural income from various perspectives. Xu et al. [29] reported a significant increase in income-based variables, such as the total household income and the share of non-farm income after borrowing money. Their study demonstrated the importance of access to credit in improving household income. The impact of different financing channels on farm household income varies significantly. Formal borrowing is more effective in boosting income as compared to informal borrowing [30]. Middle-income groups and those borrowing on a moderate scale have been reported to benefit more from borrowing as compared to other groups [31]. In addition to that, there is a positive relationship between credit supply and department economic growth [32,33]. Credit constraint status also affects modern agricultural technologies [34,35].

While these reports provide a rich theoretical basis and empirical reference for our study, there is further scope for more in-depth investigations. While most of the existing literature focuses on the influence of a single causal chain, either “financial literacy–borrowing behavior” or “borrowing behavior-household income”, very few publications have explored the inherent transmission channel, “financial literacy–borrowing behavior–household income”. The borrowing behavior of farmers is influenced by their level of financial literacy, which affects their overall income. Our study uses data from 460 households in the collective forest areas in Liaoning Province to empirically analyze the direct impact of financial literacy on household income and the mediating effect of borrowing behavior in the income transmission channel. The research sought to understand the relationship between financial literacy, borrowing behavior, and household income, which is crucial to being able to frame appropriate policies which can facilitate the spread of financial education in rural areas, improve financial literacy levels in the general populace, and reduce the constraints to obtaining rural household credit.The innovation may lie in the perspective of the study as most of the published studies are from the perspectives of national policies, financial institutions’ supply, and farmers’ household characteristics, but far fewer papers have been published on studies conducted from the perspective of the financial decision-making ability of the micro-entities on the demand-side, and these papers have included financial literacy, borrowing behavior, and household income in the same framework. The findings provide evidence to assist with the policies in developing countries that improve financial education, which can help rural households to reduce credit constraint.

The paper is organized into five sections, after this introductory section. Section 2 focuses on the theoretical analysis of the transmission channelsbetween financial literacy, borrowing behavior, and household income. Section 3 shows the research methods, including the data sources, model specification, and variables selection. Section 4 reports the empirical results and tests the endogenous problems. In the last section, the conclusions and discussion are presented, followed by some policy suggestions.

2. Theoretical Analysis and Research Hypothesis

2.1. Definitions

2.1.1. Financial Literacy

Different researchers and organizations have defined financial literacy in different ways. The concept of financial literacy can be divided into three categories:first of all, it can be described as basic knowledge regarding the fundamentals of finance, borrowing, saving, investing, and protection [36]. Secondly, financial literacy refers to a person’s ability to apply financial concepts that go beyond just the basic knowledge [37,38]. Finally, based on previous studies, the subjective financial tests can be added; these tend to measure people’sself-assessments of financial literacy [39,40].

Xu, L. and Zia, B. [41] have opined that the definitions of literacy frequently overlap. It is hard to put forward a uniform standard of research because the difference in content is related to the economic level.In low-income countries, the degree of financial outreach is limited, and the papers on financial literacy focus on access to relevant finance-related knowledge and the receptivity of the population towards it. In high-income countries, financial literacy is often viewed as being complementary to consumer protection.The capabilities to navigate a complex array of financial products are research priorities. The Liaoning Statistical Yearbook, published by the Liaoning Provincial Bureau in 2022, has shown that the GNP value of the research area is lower than the average GNP of the entire of Liaoning, China [42]. According to the second category of method, we have defined financial literacy as the knowledge of finance, as well as the application skills.

2.1.2. Borrowing Behavior

There are two primary aspects of borrowing behavior pertinent to the present study. The first is from the perspective of the borrowing process, which includesthe lending channels, access to credit, and the scale of borrowing volumes [43,44]. With regard to the type of lending channels, they can also be divided into formal or informal financial institutions [45,46]. Based on the existing literature and the actual situation in the collective forest areas, this paper defines borrowing behavior as the actual act of obtaining Forest Property Collateral Credit fundsand the amount borrowed from formal financial institutions such as the Rural Commercial Bank, the Postal Saving Bank of China, and other banks in rural areas. These formal financial institutions are the main credit channel for applying for the Forest Property Collateral Credit.

2.2. Theoretical Analysis of the Impact of Financial Literacy on Household Income

Academic research on the impact of financial literacy on household income mainly draws on the theory of thereturn on investment in education [47,48,49], which says that investment in financial education, such as school and on-the-job training, has a positive impact on financial literacy and generates a return on investment [50].However, this theory does not apply to the situation in China. The penetration of financial education is relatively low in China’s rural areas; the effect of interventions on financial literacy is somewhat weak; the farmers lack the training channels that enable comprehensive access to financial literacy; and the likelihood of return on investment in financial education is relatively low [51].

The development of financial literacy among farmers primarily occurs via the learning-by-doing mode, namely on-the-job learning or through skills and experience gained from real-life participation in various financial activities. Meanwhile, environmental factors may be other determinants, including their homes, the neighbors, and the famers in the same village [52]. Hence, this study was based on the learning theory, which suggests that knowledge accumulation not only relies on the surroundings, but it can also be gained from the experience of daily activities [53,54]. The learning-by-doing AK model and the Douglas production function in the short term are combined in this paper [55,56]. The farmer family’s economic output of the T period is given by Equation (1):

$$Y=K^{\alpha }{\left(AL\right)}^{\beta }{\overline{O}}^{\gamma }Y \alpha +\beta +\gamma =1$$

(1)

where $Y$ is the economic output of the farming household, $K$ is the capital input, including investment in relevant fixed assets, investment in the means of production, etc., $A$ is the financial knowledge of the farming household, L is the labor input of the farm household, and $\overline{\mathrm{O}}$ is the fixed input of forestland. To simplify the model, we assumed that the land resource endowment of farmers remains unchanged and does not consider the long-term production investment. In a market with perfect competition, with a given output condition and profit maximization as the ultimate objective, the profit is given by Equation (2):

$$TR=P{K}^{\alpha }{\left(AL\right)}^{\beta }{\overline{O}}^{\gamma }$$

(2)

$$C={\omega }_1+{\omega }_{2}K+F_A$$

(3)

$$\pi =P{K}^{\alpha }{\left(AL\right)}^{\beta }{\overline{O}}^{\gamma }-({\omega }_1+{\omega }_{2}K+F_A)$$

(4)

where TR represents the total returns, P is the price of the products, $C$ is the cost of the economic output, ${\omega }_1$ is the price of labor, i.e., the wage rate, ${\omega }_2$ is the cost of capital, i.e., the interest rate, $F_A$ represents the fixed indirect costs incurred in the process of accruing financial literacy, and $\pi$ is the profit. The first-order conditions for the profit maximization of the production are given by Equations (5) and (6):

$$\frac{\partial \pi }{\partial L}=P\beta A^{\beta }{K}^{\alpha }{L}^{-1+\beta }{\overline{O}}^{\gamma }-{\omega }_1=0$$

(5)

$$\frac{\partial \pi }{\partial K}=P\alpha A^{\beta }{K}^{-1+\alpha }{L}^{\beta }{\overline{O}}^{\gamma }-{\omega }_2=0$$

(6)

From Equations (5) and (6), we can conclude that the marginal output of labor input and the marginal output of capital input are proportional to the profit from the farming household production, and the magnitude of marginal returns is positively related to the level of financial literacy, i.e., in order to maximize profit, we need to determine how the cost of achieving a given output can be minimized. The conditions for cost minimization may be expressed by the following equations:

$$C\left(\omega ,Y\right)=mi{n}_{L,K,A}{\omega }_{1}L+{\omega }_{2}K+F_A$$

(7)

$$s.t.K^{\alpha }{\left(AL\right)}^{\beta }{\overline{O}}^{\gamma }=Ys.t.$$

(8)

The optimal solution can be found by calculating the first-order and second-order conditions for cost minimization separately, as shown below.

$$\frac{\partial C}{\partial A}=-\frac{L{\left(AL\right)}^{-1-\frac{\beta }{\alpha }}{\overline{O}}^{-\frac{\gamma }{\alpha }}{\omega }_2{Y}^{\frac{1}{\alpha }}\beta }{\alpha }$$

(9)

$$\frac{{\partial }^{2}C}{\partial A^2}=\frac{{\left(AL\right)}^{\frac{-\beta }{\alpha }}{\overline{O}}^{-\frac{\gamma }{\alpha }}{\omega }_2{Y}^{\frac{1}{\alpha }}\beta \left(\alpha +\beta \right)}{A^2{\alpha }^2}$$

(10)

We can conclude that when the conditions are such that ∂C/∂A < 0, an increase in the level of financial literacy decreases the production costs, while the conditions that cause ∂²C/∂A² > 0 result in financial literacy having a marginal decreasing effect on the rate of the reduction in production costs and the rate of the increase in production profits, resulting in the curve representing profits having an inverted U-shape—high in the middle and low on both sides.

In summary, financial literacy affects the farming household income in four ways:first of all, financial literacy affects the choice of labor allocation. Financial literacy enables farmers to be able to estimate costs and future returns on theirinvestments.These estimations can help farmers make decisions, such as whether to use their own labor or hired labor. Secondly, financial literacy helps to improve the utilization of capital inputs. Assuming a fixed level of capital input within a certain period, the level of financial literacy determines whether farmers can optimally plan the capital input according to their strengths and the market conditions. Thirdly, financial literacy promotes the optimal combination of economic output. In a scenario where the farming household has two products, different combinations of production factors are possible. The financial management skills that the farmers may have by being financially literate determine whether rational producers can quickly identify profit signals and efficiently allocate resources as the situation demands.Finally, financial literacy has a marginally diminishing effect on the increase in the profitability of farming household production. The positive impact of financial literacy is stronger on low-income farming households, and the effectiveness seems to diminish as the income gap decreases.

 Hypothesis 1.

Financial literacy can significantly boost farming household income, and there is a marginal decreasing effect on the rate of increase in household income.

2.3. Theoretical Analysis of the Effect of Financial Literacy on Farming Household Income Mediated by Borrowing Behavior

When farmers are unable to reach the optimal equilibrium point on the production curve, they need to increase labor input or productive capital input through external financing in order to achieve Pareto optimality. However, Chinese farmers have long faced severe creditconstraints, and the credit is hard to access [57]. According to the credit constraint theory, it mainly arises from adverse selection and the moral hazards caused by information asymmetry [58]. Financial literacy mitigates credit constraints through both direct and indirect channels. Farming households with higher levels of financial literacy can actively search for and understand information aboutfinancial products and are more familiar with credit policies, credit processes, credit rates, and other related information. This reduces the possibility of information transmission distortions between borrowers and lenders and alleviates the “no confidence in applying for credit” constraint caused by the misconception that formal credit is not available [59,60]. Furthermore, financial knowledge and application skills strongly influence the business power andentrepreneurial competency, as well as saving ability, which reduces the moral hazard to some extent and increases the likelihood of meeting the financial institution’s qualification criteria for the borrower [61,62,63]. Thus, an improvement in the level of financial literacy increases the access to credit.

It is assumed that farming households will invest all the loans they receive (I) in a forestland production to maximize their utility, thus creating conditions for an optimal mix of labor input (L) and productive capital input (K). The farmer may increase the efficiency of production by increasing the labor input (L) or expand the scale of production by increasing the productive capital input (K). External financing may also adjust the ratio of the labor input (L) to the productive capital input (K) to achieve Pareto optimality. In the actual economic operation, external financing may also lead to project failure [64]. Financial literacy will help alleviate this risk by improving the farmers’ investment planning, the ability to recognize market profit signals, and other business skills, thus reducing the possibility of “poverty due to credit” and increasing the positive effect of credit on income, as shown in Figure 1. This leads us to the following hypothesis 2, as given below.

 Hypothesis 2.

Financial literacy helps to increase access to credit, which in turn increases the farming household income.

3. Materials and Methods

3.1. Data Sources

The research data were sourced from five areas in Liaoning Province, namely Benxi County, Huairen County, Xinbin County, Qingyuan County, and Beipiao City, between October 2019 and January 2020. This area is the earliest in Liaoning province to complete the implementation of the collective forest rights system reforms. Its forest cover rate and forest accumulation are at the forefront of the country. The average area of collective forest rights confirmation is about 210,353.44 ha, which is a typical county with a lot of forests in China. The survey included questions about family resource endowments, production inputs and returns, credit behaviors, and financial literacy-related tests.

The main characteristics of these sample sets are as follows: the respondents are mainly aged between 36 and 65, and the proportion of farmers with junior high school education or above is 69.35%.The reform of the forest rights system has clarified the contractual management and the ownership rights of the collective forestland. A total of 91.96% of the sample households have obtained forest rights certificates and have the right to manage forestland.The proportion of the scale of forestland management larger than 18 ha was 25.01%. At 2–6 ha or 6–18 ha, it reached 23.04% and 24.78%, respectively. Only 27.17% of the sample of managed forest areas was less than 2 ha.In terms of the source of income, part-time businessesare no longer the main source of income for farmers, but instead, there is a trend toward specialization. A total of 22.83% of the sample farmers are engaged in pure agriculture such as rice, medicinal materials, aquaculture, fruit trees, and understory products, and 23.04% are engaged in tertiary industries such as farming, transportation, and forest product sales. The main characteristics of these sample sets are as detailed in

Table 1. Basic characteristics of individual interviewees.

Variable Name	Category	Frequency	Proportion (%)
Interviewee Age (years)	<35	11	2.61
	36–50	156	33.91
	51–65	233	50.65
	>66	59	12.83
Schooling years (years)	≤9	141	30.65
	10–12	238	51.74
	13–15	44	9.57
	≥16	37	8.04
Forest rights certificates issued	Not issued	37	8.04
	Partially issued	59	12.83
	Fully issued	364	79.13
The scale of forestland management (ha)	<2	125	27.17
	2–6	106	23.04
	6–18	114	24.78
	>18	115	25.01
The main source of income	Purely farming households	105	22.83
	Part-time farming households	82	17.83
	Non-agricultural part-time households	167	36.3
	Non-agricultural households	106	23.04

Note: Purely farming households, those whose non-agricultural income accounts for less than 10% of total household income; part-time farming households, those whose non-agricultural income accounts for 10–50% of total household income; non-agricultural part-time households, those whose non-agricultural income accounts for 50–90% of total household income; and non-agricultural households, those whose non-agricultural income accounts for more than 90% of total household income.

.

3.2. Econometric Model Specification

3.2.1. Baseline Regression

To verify the hypotheses above, this study constructs a mediating effect model, as shown below:

$$income={\alpha }_0+{\alpha }_{1}literacy+{\alpha }_2+e_1$$

(11)

$$credit=b_0+b_{1}literacy+b_2{X}_i+e_2$$

(12)

$$income=c_0+c_{1}literacy+c_{2}credit+c_3{X}_i+e_3$$

(13)

In Equations (11) to (13), income is the household income of the farmer, literacy is the financial literacy of the farmer, and credit represents the intermediary variable of forestland mortgage lending behavior. X_i represents a group of control variables, and e₁, e₂, and e₃ are random error terms. a_i, b_i, and c_i the parameter estimations, respectively. Equation (11) is used to verify the effect of financial literacy on household income, applying the ordinary least squares. Equation (12) is used to verify the effect of credit behavior on financial literacy. Due to the credit restrictions in rural areas, more than half of the respondents with credit cannot receive forest rightsmortgages; so, the Tobit model is adopted. Finally, we added the credit variable to Equation (13), observing the changes in the coefficient between financial literacy and household income.

3.2.2. Unconditional Quantile Regression

Unconditional quantile regression (UQR) was used to construct a recentered influence function(RIF) to examine the differences in the impact of financial literacy on farmers at different income levels [65], using Equation (14).

$$RIF\left(q_{\tau },income,F_{income}\right)=q_{\tau }+\frac{\tau -1\left(income\le q_{\tau }\right)}{f_{income}\left(q_{\tau }\right)}$$

(14)

where $RIF\left(q_{\tau },income,F_{income}\right)$ denotes the RIF corresponding to the τ- quantile of F_income, $q_{\tau }$ is the unconditional quantile of income that satisfiesthe relation F_income = τ, and f_income(∙) is the density function of income.The bias effect of the unconditional quantile can be obtained using Equation (15).

$$UQPE\left(\tau \right)={{\displaystyle \int }}^{ }\frac{\partial E(RIF\left(q_{\tau },income,F_{income}\right)|X)}{\partial X}d{F}_{literacy}$$

(15)

Equation (15) represents the marginal effect of the unit advection transformations of financial literacy on the τ—the unconditional quantile of household income. In contrast to conditional quantile regressions (CQR),UQR focuses on the impact of the changes in financial literacy levels on the distribution of farming household incomes, thus reducing the need for the over-explanation of individual characteristics.

3.3. VariablesSelection and Analysis

3.3.1. Dependent Variable: Farming Household Income

This study takes farming household annual income as a dependent variable in assessing the transmission effect of financial literacy and credit behavior, which contains the net agricultural income and non-agricultural income.

3.3.2. Core Explanatory Variables: Financial Literacy

The variable of financial literacy includes both financial knowledge and application skills. In this context, financial knowledge includes knowledge about savings, currency, and credit. The variables, investment planning, financial policy familiarity, and the use of financial products are classified as application skills. Appropriate weights were assigned to financial literacy to overcome the lack of equal weighting of the financial literacy scores, and the factor scores for financial literacy were obtained through factor analysis. First, we determined whether or not the conditions for the applicability of the factor analysis were met. The KMO test value was calculated to be 0.751; the Cronbach’s alpha was 0.703; and the spherical test was significant at the 1% level, confirming that the financial literacy variables were suitable for factor analysis. The common factors were retained based on the principle that the characteristic root is greater than one, and the contributions of the common factors were used as weights to obtain the final financial literacy factor scores. (See

Table 2. Factor analysis results and reliability and validity tests of financial literacy measures.

Classification	Variable Name	Financial Literacy Question Setting	Factor Load	KMO Test Results
Knowledge	Savings knowledge	Will the deposit rate be calculated?	0.68	0.80
		Will the deposit rate be calculated correctly?	0.56	0.75
	Currency knowledge	Will inflation be calculated?	0.78	0.79
		Will inflation be calculated correctly?	0.73	0.73
	Credit knowledge	Will credit interest be calculated?	0.84	0.71
		Will credit interest be calculated correctly?	0.88	0.68
Application skills	Investment planning	Is production planning performed well in advance of production investments?	0.53	0.82
	Financial policy familiarity	Are you familiar with relevant production investment policies?	0.56	0.86
	Use of financial products	Do you take out personal accident insurance or property insurance to prevent accidents?	0.60	0.65
		Should bank financial products be purchased to obtain additional income?	0.52	0.71
		Should forest insurance be purchased to preventforest disaster?	0.47	0.73

).The descriptive statistics for the financial literacy indicators and application skills indicators are presented in

Table A1. Descriptive statistics for financial literacy indicators.

Variable Name	Financial Literacy Topic Setting	Correct Rate (%)	Awareness Rate (%)
Savings knowledge	Suppose you have 100 yuan now, and the bank’s annual interest rate is 4%. If you save this money for 5 years, what is the total amount you will get after combining the principal and interest after 5 years? (1 = less than 120 yuan, 2 = equal to 120 yuan, 3 = greater than 120 yuan, 4 = cannot calculate)	11.96	27.83
Currency knowledge	Suppose you have 100 yuan now, the bank’s annual interest rate is 5%, and the annual inflation rate is 3%. Compare the amount of things you could buy with your money if you had put this money in the bank a year ago. (1 = more than a year ago, 2 = as much as a year ago, 3 = less than a year ago, 4 = cannot calculate)	10.87	24.35
Credit knowledge	Suppose you go to the bank for a loan of 10,000 yuan and the annual interest rate of the loan is 7%, how much is the annual interest amount? (1 = less than 700, 2 = equal to 700, 3 = greater than 700, 4 = cannot calculate)	36.3	48.48

Note: The financial literacy test was assessed by using two dummy variables to represent the score for each question. The first dummy variable was used to assess whether the question was answered correctly by assigning a value of 1 to correct answers and 0 otherwise. The second dummy variable was used to assess the farmers’ perception of whether they were sufficiently informed about the topic by assigning a value of 0 if they answered, “don’t know” and 1 otherwise.

and

Table A2. Descriptive statistics for application skills indicators.

Variable Name	Applied Competence Question Setting	Average Value	Standard Deviation
Production planning	Do you have any production plans to maintain your current level of production, expand your production, or add production items in the coming years? (No = 0, Yes = 1)	0.6	0.49
Financial policy familiarity	How well do you know about rural financial policies such as forestry mortgage policies and agricultural insurance policies? (1 = not at all, 2 = not well, 3 = fairly well, 4 = very well, 5 = very well)	1.95	1.28
Using financial products	Do you take out personal accident insurance or property insurance to prevent accidents? (No = 0, Yes = 1)	0.37	0.48
	Do you buy bank financial products for extra income? (No = 0, Yes = 1)	0.08	0.28
	Should you purchase forest insurance to obtain additional income? (No = 0, Yes = 1)	0.18	0.38

.

3.3.3. Mediating Variable: Forestland Mortgage Borrowing Behavior

The loan amount available is the main variable in the model to verify the causal chain between financial literacy and household income. It is defined as the average amount of forestland mortgage funds borrowed from formal financial institutions. We chose this measurement not only because most existing studieshave adopted this variable to reflect the credit decision-making behavior of farmers, but because banks constitute the main channel for forestland mortgage loans in rural areas in China.

Using the mean financial literacy score as a cut-off, farmers with a score below the mean value of 3.78 were classified as part of the low financial literacy group, while the rest were classified as part of the high financial literacy group (see

Table A3. Financial literacy questionnaire scores of interviewees.

Questionnaire Score	0	1	2	3	4	5	6	7	8	9	10	11	12
Sample size (pcs)	37	67	80	59	53	42	40	28	29	11	9	4	1
Share (%)	8.04	14.57	17.39	12.83	11.52	9.13	8.7	6.09	6.3	2.39	1.96	0.87	0.22

).

Table 3. Borrowing behavior of farm households with different financial literacy.

	Demand for Forestland Mortgage Loan		Access to Forestland Mortgage Loan
	Existence of Credit Demand	Credit Amount Needed (CNY Million)	Access to Credit	Loan Amount Available (CNY Million)
The low financial literacy group (N = 256)	0.24	5.93	0.10	2.55
The high financial literacy group (N = 204)	0.62	108.18	0.30	54.57
Difference between groups	0.38 ***	102.26 ***	0.20 ***	52.02 ***

Note: *** represents significance at the 1% statistical levels; the values in the table are average; the difference between different groups was compared by independent sample t-test.

shows a comparison of the borrowing behavior of farmers in groups corresponding to different financial literacy levels. It shows that the demand for forestland mortgage loans is significantly higher than the access to them, which indicates severe constraints in the availability of mortgage loans through formal channels. Secondly, the indicators show that access to credit is significantly higher for farmers in the high financial literacy level group than those of the low financial literacy level group. It follows that financial literacy is likely to be one of the important effect factors in the alleviationof the farmers’ credit constraints.

3.3.4. Control Variables

In order to comprehensively measure the impact of the level of financial literacy on household income and the mediating effect of borrowing behavior, as many control variables as possible were selected based on the previous literature. These variables were divided into three categories: each individual’s characteristics, household characteristics, and regional characteristics. The individual characteristics included gender, age, and health status. The household characteristics included the highest education level of the household members, the working population ratio, whether there were village cadres in the household, the cost of human interaction, the area of forestland operated, and whether they had joined a professional cooperative. The regional characteristics included the areas where they were located and whether they had been issued forest rights certificates. The definitions and descriptive statistics for each variable are shown in

Table 4. Definition of key variables and descriptive statistics.

	Variable Name	Definition	Mean	SD
Dependent variables	Farming household income	Average annual household income from 2017 to 2019 (taken as logarithm)	10.99	1.32
Core explanatory variables	Financial literacy	Final values obtained by factor analysis	−3.81 × 10−10	0.33
	Knowledge	Final values obtained by factor analysis	−1.90 × 10−9	0.50
	Application skills	Final values obtained by factor analysis	−4.00 × 10−9	0.55
Mediating variable	Loan amount available	Logarithm of the average loan amount from 2017 to 2019	0.54	1.37
Control variables	Gender	Male = 0; Female = 1	0.09	0.29
	Age	Age of respondent (years)	54.18	10.06
	Health status	Unhealthy = 0; Average = 1; Healthy = 2	1.75	0.56
	Highest educational level of household members	Years of education of the most educated member of the household	10.33	4.04
	Percentage of the labor force	Household labor force size/total household size	0.68	0.30
	Village officials identity	Village officials in the household: No = 0; Yes = 1	0.16	0.36
	Cost of human interaction	Expenses used to maintain social relationships, such as marriages, and funerals in 2019 (CNY million)	1.88	2.29
	Area of forestland operated	Logarithm of the scale of woodland operations(mu) in 2019	8.96	1.28
	Professional cooperatives	Participation in professional cooperatives: No = 0; Yes = 1	0.05	0.22
	Location	Chaoyang = 0; Fushun = 1; Benxi = 2	1.21	0.71
	Forest rights certificate	Not issued = 0; Partially issued = 1; Fully issued= 2	1.71	0.61

.

4. Results and Analysis

4.1. Baseline Regression

4.1.1. Regression Results of Financial Literacy on Farm Household Income

Regressions 1, 2, and 3 in

Table 5. The effect of financial literacy on farm household income.

	OLS (1)	OLS (2)	OLS (3)
Financial literacy	0.280 *** (0.893)	——	——
Knowledge	——	0.168 ** (0.099)	——
Application skills	——	——	0.256 *** (0.055)
Gender	0.366 (0.163)	0.389 (0.164)	0.352 (0.161)
Age	−0.023 *** (0.005)	−0.024 *** (0.005)	−0.025 *** (0.005)
Health status	0.283 ** (0.088)	0.295 ** (0.088)	0.265 ** (0.087)
Highest educational level of household	0.132 *** (0.027)	0.136 *** (0.028)	0.120 *** (0.027)
Percentage of the labor force	0.335 ** (0.155)	0.347 ** (0.156)	0.293 * (0.153)
Village officialsidentity	0.232 (0.123)	0.256 (0.124)	0.207 (0.121)
Cost of human interaction	0.166 *** (0.228)	0.174 *** (0.023)	0.151 *** (0.023)
Area of forestland operated	0.155 *** (0.042)	0.172 *** (0.041)	0.137 *** (0.041)
Professional cooperatives	−0.234 (0.217)	−0.185 (0.218)	−0.336 (0.216)
Location	0.022 (0.069)	0.018 (0.069)	0.030 (0.068)
Forest rights certificates	0.154 * (0.080)	0.157 * (0.081)	0.140 * (0.080)
Constant	8.89 *** (0.518)	8.727 *** (0.518)	9.306 *** (0.524)
Observations	460	460	460
R2	0.454	0.445	0.468
Fvalue/Waldchi2	30.94	29.90	32.70

Note: ***, **, and * represent significance at the 1%, 5%, and 10% statistical levels, respectively; robust standard errors are in brackets.

report the regression results of the effects of financial literacy, financial knowledge, and application skills, respectively, on the farming household income, using ordinary least squares (OLS). The regression results show that after accounting for the individual characteristics, household characteristics, and regional characteristics, financial literacy is significantly positive at the 1% level, and there is a positive effect of the financial literacy level on an increase in income, with the parameter estimated to be 0.280. The financial knowledge level passes the significance test at the 5% level, and there is a positive effect on farm household income. Application skills passed the significance test at the 1% level, with a higher regression coefficient than the value of the knowledge regression coefficient and a marginal effect of 0.256. While improvement in the financial knowledge levels has a positive impact on farm household income, knowledge application had a stronger effect in comparison. Research hypothesis 1 was partially tested, and financial literacy was found to have a positive direct effect on farming household income.

As shown in Table 5, age has a significant negative influence on farming household income. One possible reason is that older people have a limited source of income. Health status has a significantly positive impact on the farming household income at a 5% level. Higher educational levels of the household members are more likely to have a positive effect on income. For each additional year of schooling, the farming household income is increased by 0.132. The percentage of the labor force carries a positive coefficient, with parameter estimates of 0.335. The cost of human interaction, the area of forestland operated, and forest rights certificates were also found to have a positive correlation with farming household income, indicating that the improvement of collective forestland tenure policies will help forestry farmers increase their income.

4.1.2. Financial Literacy on Farm Households of Different Income Levels

The regression results in

Table 6. Unconditional quantile regressions of the effect of financial literacy on farm household income.

	q15	q30	Q45	Q60	Q75	Q90
Financial literacy	0.389 *	0.408 **	0.405 **	0.510 ***	0.333	0.254
	(0.230)	(0.201)	(0.186)	(0.194)	(0.236)	(0.405)
Gender	0.560 *	0.646 ***	0.395 **	0.179	0.293 *	0.498 ***
	(0.317)	(0.242)	(0.192)	(0.187)	(0.196)	(0.196)
Age	−0.044 ***	−0.029 ***	−0.025 ***	−0.016 **	−0.017 *	0.006
	(0.008)	(0.007)	(0.006)	(0.006)	(0.072)	(0.012)
Health status	0.545 **	0.265 **	0.342 ***	0.263 ***	0.182 **	0.209 *
	(0.198)	(0.134)	(0.110)	(0.093)	(0.099)	(0.111)
Highest educational level of household	0.107 ***	0.111 ***	0.137 ***	0.128 ***	0.093 ***	0.139 **
	(0.031)	(0.033)	(0.032)	(0.035)	(0.142)	(0.069)
Percentage of the labor force	0.434	0.211	0.234	0.147	−0.056	0.021
	(0.307)	(0.211)	(0.173)	(0.179)	(0.198)	(0.316)
Village officials identity	0.264	0.359 **	0.289 **	0.210	0.340 *	0.281
	(0.165)	(0.143)	(0.173)	(0.158)	(0.198)	(0.345)
The cost of human interaction	0.061 ***	0.099 ***	0.116 ***	0.167 ***	0.180 ***	0.394 ***
	(0.024)	(0.026)	(0.028)	(0.034)	(0.035)	(0.072)
Area of forestland operated	−0.007	0.026	0.132 ***	0.201 *	0.333 **	0.561 ***
	(0.067)	(0.055)	(0.051)	(0.052)	(0.059)	(0.094)
Professional cooperatives	−0.002	−0.068	−0.108	−0.082	−0.056	−0.601
	(0.244)	(0.265)	(0.268)	(0.296)	(0.404)	(0.747)
Location	0.224	0.103	−0.0571	−0.053	−0.031	−0.464 **
	(0.122)	(0.094)	(0.0816)	(0.083)	(0.094)	(0.147)
Forest rights certificates	0.0779	0.194 *	0.136 *	0.098	0.087	0.032
	(0.122)	(0.109)	(0.093)	(0.093)	(0.099)	(0.163)
Constant	10.031 ***	9.929 ***	9.216 ***	8.74 ***	8.44 ***	8.08 ***
	(0.542)	(0.706)	(0.627)	(0.643)	(0.760)	(1.20)
Observations	460	460	460	460	460	460
R2	0.236	0.253	0.325	0.333	0.305	0.294

Notes: ***, **, and * represent significance at the 1%, 5%, and 10% statistical levels, respectively; robust standard errors are in brackets.

show that the coefficient of the effect of financial literacy on farm household income tends to decrease at higher quantiles. When the household income level is at the 10th percentile, financial literacy has a positive relationship with farm household income with a coefficient of 0.389 and passes the significance test at the 10% level. When the household income level is at the 30th and 45th percentiles, the coefficients of financial literacy on farm household income rise to 0.408 and 0.405, respectively. When at the 60th percentile, financial literacy has the highest coefficient of the factor on rural household income levels. At the 75th percentile, the coefficient of financial literacy on household income begins to weaken and does not pass the significance level test. At the 90th percentile, the coefficient of financial literacy on household income is the lowest.

This shows that the curve representing the level of financial literacy on the growth of farming household income has an inverted U-shape that is low towards the extreme ends and high in the middle. When people from the low-income farming group acquire financial knowledge and integrate it, there is a significant increase in their level of financial literacy, resulting in a significant improvement in their income.On testing hypothesis 1, we found a marginally decreasing effect on the rate of increase in household income.

While the financial literacy level of the high-income farming group is relatively higher in general, other factors such as health, literacy, forestland area, and social capital have significantly more influence on the group’s income, in addition to the influence of external variables such as the cost of production and policy changes. An inspection of the gender variable yields some interesting observations. Gender is a higher coefficient than other variables at the 30th and the 90th percentile, which is related to the larger number of males compared to females among the respondents.

4.1.3. The Effect of Financial Literacy on Farming Household Income Mediated by Borrowing Behavior

In order to verify the existence of the transmission path between financial literacy, borrowing behavior, and farm household income, this study used a mediation model to estimate the channel behind the impact of financial literacy on household income. The core explanatory variables are financial literacy, knowledge, and application skills, respectively, as presented in columns (1), (4), and (7) in Table 6.

Columns (2), (5), and (8) in

Table 7. Testing on the mediating effect between financial literacy and land transfer.

		Model 1			Model 2			Model 3
	OLS (1) Income	Tobit (2) Loan Amount Available	OLS (3) Income	OLS (4) Income	Tobit (5) Loan Amount Available	OLS (6) Income	OLS (7) Income	Tobit (8) Loan Amount Available	OLS (9) Income
Financial literacy	0.280 *** (0.893)	1.032 ** (0.350)	0.169 * (0.097)	——	——	——	——	——	——
Knowledge	——	——	——	0.168 ** (0.099)	0.581 * (0.331)	0.080 (0.093)	——	——	——
Application skills	——	——	——	——	——	——	0.256 *** (0.055)	0.979 *** (0.272)	0.188 ** (0.073)
Loan amount available	——	——	0.480 *** (0.062)	——	——	0.491 *** (0.062)	——	——	0.473 ***
									(0.062)
Gender	0.366 (0.163)	0.650 (0.739)	0.336** (0.154)	0.389	0.734	0.346 **	0.352	0.776	0.332 **
				(0.164)	(0.746)	(0.154)	(0.161)	(0.762)	(0.153)
Age	−0.023 *** (0.005)	−0.008 (0.021)	−0.025 *** (0.005)	−0.024 *** (0.005)	−0.013 (0.021)	−0.026 *** (0.005)	−0.025 *** (0.005)	−0.013 (0.021)	−0.025 *** (0.005)
Health status	0.283 ** (0.088)	0.285 (0.401)	0.271 ** (0.083)	0.295 ** (0.088)	0.332 (0.402)	0.281 *** (0.083)	0.265 ** (0.087)	0.306 (0.415)	0.257 ** (0.083)
Highest educational level of household	0.132 *** (0.027)	0.034 (0.098)	0.132 *** (0.026)	0.136 *** (0.028)	0.053 (0.098)	0.134 *** (0.026)	0.120 *** (0.027)	−0.0073 (0.099)	0.126 *** (0.026)
Percentage of the labor force	0.335 ** (0.155)	0.612 (0.624)	0.300 ** (0.146)	0.347 ** (0.156)	0.652 (0.631)	0.306 ** (0.146)	0.293 * (0.153)	0.488 (0.628)	0.280 * (0.146)
Village officialsidentity	0.232 (0.123)	0.281 (0.413)	0.189 (0.116)	0.256 (0.124)	0.406 (0.414)	0.209* (0.116)	0.207 (0.121)	0.193 (0.414)	0.164 (0.117)
The cost of human interaction	0.166 *** (0.228)	0.208 ** (0.064)	0.126 *** (0.022)	0.174 *** (0.023)	0.229 *** (0.065)	0.130 *** (0.022)	0.151 *** (0.023)	0.180 ** (0.065)	0.119 *** (0.022)
Area of forestland operated	0.155 *** (0.042)	0.755 *** (0.171)	0.087 ** (0.040)	0.172 *** (0.041)	0.810 *** (0.174)	0.093 ** (0.040)	0.137 *** (0.041)	0.768 *** (0.172)	0.086 ** (0.039)
Professional cooperatives	−0.234 (0.217)	0.792 (0.616)	−0.357 * (0.205)	−0.185 (0.218)	0.949 (0.620)	−0.338 * (0.205)	−0.336 (0.216)	0.584 (0.622)	−0.392 * (0.205)
Location	0.022	−0.409	0.0643	0.018	−0.413	0.062	0.030	−0.401	0.0706
	(0.069)	(0.260)	(0.065)	(0.069)	(0.262)	(0.065)	(0.068)	(0.262)	(0.0647)
Forest rights certificates	0.154 * (0.080)	0.345 (0.332)	0.139 * (0.076)	0.157 * (0.081)	0.378 (0.337)	0.139 * (0.076)	0.140 * (0.080)	0.315 (0.334)	0.134 * (0.076)
Constant	8.89 *** (0.518)	−10.48 *** (2.238)	9.527 *** (0.494)	8.727 *** (0.518)	−11.03 *** (2.277)	9.472 *** (0.495)	9.306 *** (0.524)	−10.03 *** (2.253)	9.660 *** (0.497)
Observations	460	460	460	460	460	460	460	460	460
R2/Pseudo R2	0.454	0.158	0.516	0.445	0.150	0.514	0.468	0.1652	0.5204
Fvalue	30.94	——	36.65	29.90	——		32.70	——	37.22

Note: ***, **, and * represent significance at the 1%, 5%, and 10% statistical levels, respectively; robust standard errors are in brackets.

show that the marginal effects of financial literacy, financial knowledge, and application skills on loan amount availability are significantly positive. This indicates that an increase in the level of financial literacy has a positive effect on the income of farmers in the collective forest area. Both knowledge and application skills significantly influence credit behavior. The correlation of application skills was found to be statistically significant at the 1% level. With a regression coefficient of 0.979, application skills had a stronger marginal effect than knowledge. Individuals with a higher level of financial literacy are better able to understand credit policy and related aspects such as mortgage loan procedures, interest rates, etc., which enables them to be confident and more willing to apply for credit. Furthermore, financial literacy generally results in better financial management skills and credit ratings, which in turn makes qualification for loans more likely [66].

In columns (3), (6), and (9) in Table 7, the loan amount available isadded to the baseline model. Columns (3) and (9) show that when the credit behavior variable is controlled, the regression coefficients of financial literacy and application skillsdecline and are no longer significantat the 1% level. Column (6) fails to pass the significance test. Financial knowledge cannot be highlycorrelated with income. The results reveal that the level of financial literacy increases the loan amount available and has an improvement on household income. Thus, hypothesis 2 is supported.

The amount of loan to access has a positive correlation with household income and is significant at a 1% level, as seen in columns (3), (6), and (9), implying that forestland mortgage loans can enhance the farmers’ income and improve household welfare.We found that credit funds were invested in forest produce or non-agricultural products by most of the people who had access to the credit in our field research. While productivity mightnot bedetermined by access to credit, it is natural thatthe farmer who is under credit constraints will makedifferent financial choicesto the onenot subject to such constraints. Hence, households with access to credit need to strive to reduce credit restrictions and add financial credit services [67,68].

4.2. Endogeneity Test

The simple linear regression model cannot avoid the endogenous problems caused by missing variables and the reverse causality between financial literacy and household income, which can lead to biased regression results. We follow Morgan, P.J. and Long, T.Q. (2020) [53] in using “the average financial literacy of other farmers in the village (area-financial)”as the instrumental variable and adopted the two-stage least square (2SLS) and the IV Tobit. We have taken two measures to examine whether or not this assumption is reasonable or not: (i) theendogenous test, the Durbin score, the Wu–Hausman F value, and the Wald test with a p-value of less than 1%, which indicates that the instrumental variable passes the endogenous test, and the (ii) weak instrumental variable test. The F values in the first stage are all below the 15%level value of 8.96.This indicates that there is no weak instrumental variable. So, the instrumental variable is reasonable (please see

Table 8. Results of endogenous analysis.

		Model 1			Model 2			Model 3
	2SLS (1) Income	IV Tobit (2) Loan Amount Available	2SLS (3) Income	2SLS (4) Income	IV Tobit (5) Loan Amount Available	2SLS (6) Income	2SLS (7) Income	IV Tobit (8) Loan Amount Available	2SLS (9) Income
Financial literacy	2.140 ***	2.866 **	1.900 ***	——	——	——	——	——	——
	(0.319)	(0.876)	(0.316)
Knowledge	——	——	——	2.203 *** (0.348)	2.984** (0.935)	1.920 *** (0.334)	——	——	——
Application skills	——	——	——	——	——	——	3.021 *** (0.669)	4.260 ** (1.383)	2.874 *** (0.733)
Loan amount available	——	——	0.301 *** (0.086)	——	——	0.345 *** (0.087)	——	——	0.310 *** (0.087)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	9.712 *** (0.688)	−9.210 *** (2.262)	10.06 *** (0.645)	8.924 *** (0.714)	−10.27 *** (2.300)	9.426 *** (0.667)	12.30 *** (1.277)	−5.599 * (2.861)	12.32 *** (1.217)
Financial area	1.303 *** (0.138)	1.303 *** (0.135)	1.252 *** (0.139)	1.266 *** (0.143)	1.266 *** (0.141)	1.239 *** (0.142)	0.923 *** (0.193)	0. 923 *** (0.192)	0.828 *** (0.198)
Durbin(score)	71.274 ***	——	61.113 ***	78.446	——	66.143	69.417	——	56.951
Wu-Hausman	81.775 ***	——	68.177 ***	91.697	——	74.732	79.267	——	62.879
Wald value	——	5.87 ***	——	——	8.79 ***	——	——	8.27 ***	——
The first F	89.756	20.75	21.70	80.274	12.180	11.93	23.196	17.290	16.61
The first F at15%	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Note: ***, **, and * represent significance at the 1%, 5%, and 10% statistical levels, respectively; robust standard errors are in brackets.

).

The results show that the correlation coefficient of financial literacy on farming household income increased by 1.86 compared to the OLS, indicating that the original OLS underestimated the role of financial literacy. While financial knowledge and application skills were both significant at a 1% level, application skills contributed more to household income compared to financial knowledge, with a difference of 0.818 between the two coefficients. The key coefficient of the amount of credit available was 0.301, which was slightly less than 0.480, that of the baseline model in column 3, Table 8. From the estimated results, the conclusion is the same as that of Table 8. This shows that the estimation results above are robust and reliable.

5. Conclusions and Discussion

5.1. Conclusions

The purpose of collective forestland tenure reform is not only to encourage farmers to better manage forestland but also to increase the credit availability for them. As the reforms of China’s forest tenure progress and the rural financial market becomes increasingly sophisticated, it is important for farming households to make effective financial decisions and make optimal use of the gradually increasing opportunities to participate in the financial market. Using the survey data from 5 counties, 17 villages, and 460 farmers in the collective forest area of Liaoning province in China, this study empirically tests the association between financial literacy and household income and explores how borrowing behavior works as a mediator in this relationship.

The findings include the following three points: financial literacy has a positive impact on farm household income. The possible explanation is that good financial literacy helps the farmer to estimate costs and predict future returns on investments and, hence, enables the farmer to plan the capital inputs optimally based on their strengths and the market conditions. On the other hand, there are differences in the regression coefficients of the effects of financial knowledge on the low-income and high-income groups. The marginal effect of financial literacy is smaller for households with more wealth, thus demonstrating that financial literacy can significantly reduce household wealth inequality in the collective forest area of China. In addition, the sum of money that could be received as a forestland mortgage loan has mediating effects between financial literacy and the farming household income.

Based on the above findings, several policy implications can be drawn. More financial education should be provided for farmers, especially those of low-income households. As argued previously, financial literacy is very important and is one way to help farmers enhance access to credit and improve their income. Financial literacy knowledge and application skills popularization policies need to be implemented. Importance needs to be attached to application skills training in financial literacy education programs, such as financial policy familiarity and the use of financial products, which are more beneficial when exerting the role of financial literacy in credit availability. Moreover, compulsory schooling ought to set financial education courses for children. The popularization of financial education is the key goal of increasing the level of financial literacy.

5.2. Discussion

These results of this study have some similarities to and differences with the previous literature. From the perspective of the positive relationship of financial literacy with household income, the result is consistent with most of the findings, namely families with a higher level of financial literacy have a higher chance of increasing their household income. Rich adults have higher financial literacy than poor adults in major advanced economy countries and major emerging economies [69]. The possible reason is the following: whether or not one is financially literate and to what degree has a significant impact on that person’s ability to make appropriate portfolio choices and financial decisions by improving their understanding of the various financial assets and their ability to compare and choose from the various options available [70,71].

In terms of financial literacy on farm households of different income levels, the result is in line with Lusardi, A., Michaud, P. C., and Mitchell, O. S. [72], who examined the relationship between financial knowledge and the income gap, demonstrating that there are differences in the coefficients of the effects of financial knowledge on low-income and high-income groups. The marginal effect of financial literacy is smaller for households with more wealth.

At the same time, there are some similar studies. Morgan, P.J. and Long, T.Q. [53] construct “financial inclusion scores”, including credit products and yield, finding that financial literacy is associated with credit products and has effect on savings. Adetunji, O. M. and David-West, O. [73] analyze the effects of financial literacy on savings and financial inclusion by the survey of two relatively low-income Asian countries. Their results showed that both financial literacy and general education levels have a strong correlation with savings behavior and financial inclusion, respectively. Although the scopes of these investigations are not totally identical with that of this paper, the conclusions are similar.

Our study suffers some limitations. First, the content of financial literacy does not cover farmers’ understanding of financial products such as financial rights protection awareness and financial risk awareness and their subjective evaluation of their financial literacy. Future research will focus on the farmers’ understanding of the financial market and awareness of the level of their own financial literacy, to measure farmers’ financial literacy level more accurately and carefully. Second, household income, forestland mortgage borrowing behavior, and household income belong to a long-term and dynamic process, and the use of panel data which have two dimensions (time series and cross-section) helps to better reflect the changes in the mediating effect of borrowing behavior on the impact of financial literacy on farming household income before and after changes in rural land credit policies. Future sample tracking studies will be conducted to reflect the dynamic process of the “financial literacy–borrowing behavior–household income” transmission path. Third, we have not considered the “financial literacy–borrowing behavior–household income” adverse causal chain, namely that household income does influence financial literacy; the effect of borrowing behavior on financial literacy and higher income on financial literacy will be explored in future research. In the future research, we will explore the use of structural equation configuration to solve this reverse factor problem. Fourth, the data sources are only from household surveys, and the behavior of the financial institutions remains to be discussed. Finally, in addition to the amount ofloan available, the borrowing behavior also includes whether the demand is satisfied, the lending channels, and the borrowing structure, which will be emphasized in future research.