The Impact of Farmland Management Rights Mortgage Loan on the Agri-Food Industrial Agglomeration: Case of Hubei Province

Abstract

The Chinese government encourages rural economic entities to use farmland management rights as collateral for loans, which helps to alleviate multi-level financing needs in rural areas. Based on the panel data of counties in Hubei Province, this paper adopts the Difference-in-Differences (DID) and the intermediary effect model to evaluate the impact of farmland management rights mortgage loans (FMRML) on the agri-food industrial agglomeration (AIA) in China. The study found that the pilot policy has significantly promoted the AIA. Moreover, the regression results remain robust after conducting the placebo test and the Propensity Score Matching Difference-in-Differences (PSM-DID) model, which demonstrates that the improvement effect is stable and long-lasting. From the heterogeneity analysis, it can be seen that the policy of FMRML has a more significant effect on the AIA in mountainous and hilly areas. By further analysis of the mechanism of action, it can be concluded that the pilot policy promotes the AIA by enhancing agricultural specialized production. The main findings can provide information for policymakers in China. The recommendations we have summarized encompass gradually expanding the scope of the pilot policy of FMRML, advancing the institutionalization and legalization of the policy, and promoting agricultural production specialization.

1. Introduction

The agri-food industry, which encompasses the agriculture, industry, and service sectors, has emerged as a fundamental, strategic, and pivotal industry for economic and social development. China, being a prominent player in agri-food production, has consistently ranked among the world’s leading nations in terms of major agri-food output. Nevertheless, the processing and transformation capabilities of agri-food in China have consistently lagged behind those of developed countries [1]. As of 2021, the deep processing rate of grains, oils, fruits, beans, meat, eggs, aquatic products, and other commodities in China stood at only around 30%, significantly lower than the levels exceeding 70% in developed nations, indicating substantial untapped potential for improvement. Moreover, China’s agri-food enterprises are generally small in scale and scattered in a dispersed pattern, exhibiting a notable dispersion of industrial spatial layout and diminished production efficiency [2]. To address these challenges, China issued the National Rural Industry Development Plan (2020–2025)1 in July 2020, which emphasizes the imperative to promote the AIA. This entails fostering the establishment of agri-food industries in production areas and concentrating them in advantageous regions to create industrial clusters with favorable characteristics.

Finance undoubtedly plays a vital role in promoting the AIA [3]. However, due to China’s unique ownership system, assets such as farmland cannot be used as collateral for bank loans. Consequently, farmers have long faced the problems of no collateral and loan difficulties, which makes it impossible to effectively promote the AIA [4]. To alleviate financial constraints in rural areas, the Chinese government has implemented the policy of FMRML, adopting the method of FMRML to promote financial services for agriculture [5,6]. Since 2008, China has begun to issue guiding opinions on the policy of FMRML one after another, but due to conflicts with the Land Law, the implementation has only been carried out through pilot programs. The establishment of FMRML was first discussed in the Guiding Opinions on Further Strengthening the Adjustment of Credit Structure and Promoting the Stable and Rapid Development of the National Economy2 document, which was jointly released in March 2009 by the People’s Bank of China and the China Banking Regulatory Commission. In March 2016, the Interim Measures for Pilot Mortgage Loans for the Management Right of Contracted Rural Land3 were promulgated, which was the first time that government departments had standardized opinions on FMRML through interim measures. In 2018, the “Rural Land Contracting Law” was amended to include provisions allowing for land management rights to be mortgaged in accordance with the law, providing a legal basis for FMRML. A significant policy statement titled Opinions on Promoting the Organic Connection between Small Farmers and the Development of Modern Agriculture4 was released by the General Offices of the Communist Party of China Central Committee and the State Council in February 2019. This statement emphasized the importance of fully implementing the FMRML policy.

By the end of September 2018, the total outstanding balance of farmland mortgage loans in 232 pilot areas nationwide had reportedly reached CNY 52 billion, according to the Summary Report of The State Council on the Nationwide Pilot Project of Mortgage Loans for the Management Right of Rural Contracted Land and Farmers’ Housing Property Rights5. This figure reflects an increase of 76.3% year over year, and the fact that CNY 96.4 billion had been released proved the policy of providing financial support for agriculture was working. In 2021, the “Opinions on Financial Support for the Development of New Agricultural Management Entities” further proposed expanding the range of mortgages and pledges for new agricultural management entities, while prioritizing the role of finance in promoting the development of the agri-food industry. Therefore, it is important to investigate whether the implementation of the policy of FMRML is advantageous to the agglomeration development of rural industries and to ascertain its mechanism of action and heterogeneity in the context of fostering the development of rural industries. The investigation of the aforementioned difficulties will assist in not only determining the impact of the policy of FMRML but also in providing policy references for advancing the AIA.

This paper mainly studies the impact of the policy of FMRML on the AIA in pilot areas of China. Driven by relevant policies, the country all over actively participates in the FMRML. However, due to regional diversity, differences in agricultural development levels, economic systems, and rural land management mode in each pilot area, the effects of the policy in each region may be different [7]. Hubei Province, which is rich in agricultural resources and has made significant progress in land system reform, is selected as a representative province for analysis in this study. As of September 2019, the amount of agriculture-related loans in the province totaled CNY 1.25 trillion, with a year-on-year growth rate of 10.7%, ranking fifth in China for growth rate [8,9,10]. Therefore, it provides a suitable case study to analyze the effects of the policy. To explore the impact of the pilot policy on the AIA, this study utilizes the DID and intermediary effect models, analyzing the sample data of 66 counties in Hubei Province from 2010 to 2019.

The remaining parts of this article are as follows: The Section 2 reviews the theoretical logic of the policy of FMRML promoting the AIA, revealing its internal mechanism and heterogeneity. The Section 3 uses the DID model to examine the impact of FMRML on the AIA. The Section 4 empirically analyzes the mechanism and heterogeneity of the policy. The Section 5 covers policy recommendations while summarizing the research findings.

2. Literature Review and Mechanism

2.1. Literature Review

2.1.1. Research on the AIA

As a key force to promote the revitalization of rural industries, the AIA is of self-evident importance. However, previous studies have primarily focused on the economic effects of such agglomeration [11]. Liu et al. (2022) [12] found that the AIA has significantly boosted agricultural economic growth. Similarly, Wu et al. (2020) [13] demonstrated that the AIA had a positive impact on agricultural energy efficiency through spatial econometric analysis. Furthermore, in addition to the research on the direct effects of the AIA mentioned above, Zeng et al. (2021) [14] discovered a positive spatial spillover effect of the AIA from the perspective of agglomeration externality. To effectively leverage the economic benefits of the AIA, some scholars have further summarized its influencing factors. Li et al. (2017) [15] believed that the improvement of transportation networks and the development of new global markets greatly promote agricultural clusters. Additionally, Miao et al. (2023) [16] found that institutional factors, natural factors, and trade conditions have an impact on the AIA, so agricultural market integration has practical significance for promoting the AIA.

2.1.2. Research on Finance and the AIA

Previous studies have extensively explored the relationships between finance and industrial development, with an emphasis on three aspects. Firstly, relaxing financial constraints is known to promote the growth of industrial productivity [17]. Secondly, diversifying risks in financial services contributes to the development of industrial planning by establishing risk-sharing mechanisms [18]. Thirdly, innovative financial products further promote the integration of finance and industry. [19]. In addition, some literature has further analyzed the relationship between the financial environment and the AIA from a credit perspective. Anshari et al. (2019) [20] believed that the AIA is facing a shortage of funds, and its agglomeration development requires more credit support. Moreover, several studies thoroughly discuss the impact of financial policies and the financial environment on the AIA. The findings suggest that the promulgation of financial policies and improvements in the financial environment are conducive to the development of the AIA [21].

The FMRML is a crucial policy to transfer rural land and promote rural financial development [22]. Previous research has examined the link between FMRML and rural finance. Yu et al. (2014) [23] studied the pilot program in Faku County, Liaoning Province, and discovered that FMRML improved the efficiency of rural land resources and further improved rural financial services through practical experience. From a legal standpoint, Zhou et al. (2021) [24] indicated that the policy of FMRML has functional value in resolving the financing difficulties associated with farmland mortgages. Luo et al. (2018) [25] proposed promoting the industrialization, capitalization, and commercialization of farmland property rights by utilizing the right to land contractual management as a financing tool.

To sum up, current research has mainly focused on either the policy of FMRML or the AIA, but there is limited research on the relationship between the two. Existing studies have the following deficiencies: Firstly, the economic effect of FMRML is mostly concentrated on the level of income increase of farmers [26], and neglects their impact on industrial development. Secondly, the mechanism and heterogeneity of the role played by the policy of FMRML in the AIA have not been thoroughly explored. In view of this, this paper combines the policy of FMRML and uses the DID model to examine how FMRML affects the AIA. The aim is to enrich the research content on the relationship between the policy of FMRML and the AIA, and to provide valuable reference materials for the development of rural finance and the AIA.

2.2. Mechanism and Research Hypothesis

2.2.1. The Impact of FMRML on the AIA: Comprehensive Effect

First of all, the policy of FMRML can directly impact the AIA by expanding the avenues for rural financial services and reducing financing constraints that hamper the development of this industry [27]. Prior to the introduction of this policy, mortgaged farmland property rights were insecure with competing claims. This situation forced financial institutions to adopt strategies like tightening credit supply, increasing interest rates, and implementing complex procedures due to the absence of effective collateral for farmers. These measures significantly restricted the potential of agricultural business entities to secure loans from financial institutions [28,29]. Following the issuance of the policy, the mortgage feature of farmland management rights was acknowledged at the policy level. This measure has effectively reduced information asymmetry between financial institutions and agricultural operators, weakened the adverse implications of adverse selection and moral hazards, enhanced the availability of credit for agricultural operators, and facilitated the extension of the agricultural industry from the primary sector to the secondary sector [30,31].

Secondly, the policy of FMRML helps facilitate the transfer of rural land and promotes the development of the AIA. The policy primarily caters to various large-scale agricultural management entities that have long-term financing requirements, which are generated through land management rights transfer. These entities can utilize loans to expand their operations further, while simultaneously promoting the development of agriculture towards scalability, industrialization, and modernization [32,33]. Additionally, the provision of loans assists in releasing the labor force of small-scale farmers, thereby meeting the labor-intensive employment needs of the AIA [2].

2.2.2. The Impact of FMRML on the AIA: Mechanism Analysis

Due to the impact of “more people with less cultivated land” and “fragmentation” in China’s agriculture, farmers tend to opt for diversified planting as a means to avoid risk and maximize the value of cultivated land. However, this model does not contribute to enhancing the level of agricultural specialization [34]. The differentiated planting mode severely limits the overall yield of farmland and is incapable of meeting the quality requirements of agri-food materials, thereby hindering the AIA. The expansion of land parcel scale resulting from the implementation of the policy of FMRML can encourage farmers to shift from multi-plot and differentiated crop planting to single-plot and specialized planting, leading to the formation of high-quality, specialized raw material production bases. This effect helps to effectively improve the quantity and quality of agri-food supply [35].

2.2.3. The Impact of FMRML on the AIA: Heterogeneity Analysis

There are heterogeneities in the effects of FMRML on the AIA in different topographical regions. The plain areas with flat terrain and relatively concentrated distribution of farmland are more suitable for the development of the AIA. However, due to the advantageous resource endowments in the plain areas, the agri-food industries have already achieved a certain scale of development in the early stages, making it difficult for the policy of FMRML to show prominent effects [2,15]. In addition, the hilly and mountainous regions have complex terrain, fragmented distribution of farmland, low land value, and limited development of the agri-food industry, making it more urgent to break the financial constraints. Therefore, the policy of FMRML that releases funds, land, and labor factors may be more conducive to the AIA in non-plain areas.

The mechanism of the effect of the policy of FMRML on the AIA is shown in Figure 1. Based on the above theoretical analysis, this paper puts forward the following research hypotheses:

Hypothesis 1 (H1).

The policy of FMRML can promote the AIA.

Hypothesis 2 (H2).

The policy of FMRML can promote the AIA by improving the specialization of agricultural production.

Hypothesis 3 (H3).

The impact of the policy of FMRML on the AIA varies significantly with different landforms.

3. Research Methods and Data

3.1. Overview of the Study Area

Located in Central China, Hubei Province, with a subtropical monsoon humid climate, is the core node of the Yangtze River economic belt. It shares borders with Anhui Province to the east, Chongqing City to the west, Shaanxi Province to the northwest, Jiangxi Province, Hunan Province, and Henan Province to the south and north, respectively. Its latitude ranges from 29°01′53″ to 33°6′47″ north latitude, while its longitude ranges from 108°21′42″ to 116°07′50″ east longitude. Hubei Province covers a total area of 18.59 × 10⁴ km². Hubei is a low-lying region with a flat landscape in the center that is surrounded by mountains on three sides, to the east, west, and north. With an average elevation of 363 m, it is a little incomplete basin that is open to the south. The terrain is complex and diverse, with distribution in plains, hills, and mountains, mainly consisting of hills and mountains, accounting for 80% of the landform in Hubei Province [9,10]. Cultivated land accounts for about 26% of the total area of the province and is one of the most important agricultural production bases in China. The abundant agricultural resources determine that Hubei Province has the conditions to carry out the policy of FMRML.

3.2. Identification Strategy

This paper aims to examine whether the policy of FMRML can promote the AIA. To answer this question, it is necessary to consider the influence of various factors such as “policy effects” and “time effects” on the AIA in the pilot counties. Therefore, to isolate the net effect of the policy, the primary concern in the empirical analysis is how to eliminate the interference of other factors such as “time effects” on the regression results. Traditional econometric methods for addressing this issue mainly involve two approaches: one approach involves including dummy variables for years in the model for regression analysis, while the other approach involves incorporating policy dummy variables into the model for analysis. The first method confounds the impact of time trends within the year dummy variables, leading to measurement results that are not purely reflective of policy effects. The second method may overlook other factors that influence the AIA, giving rise to endogeneity issues. Based on this, the pilot policy of FMRML that was put into place in 2016 is viewed as a quasi-natural experiment in this research, and the results are evaluated using the DID model [36,37]. The sample data of 66 counties (cities and districts) in Hubei Province studied in this paper include eight pilot counties: Zhongxiang City, Yiling District, Suixian County, Nanzhang County, Daye City, Gong’an County, Wuxue City, and Yunmeng County. Referring to the practice of Li et al. (2021) [38], this article selects eight pilot counties (cities and districts) as treatment groups, while the rest are control groups. The DID model can compare the effect of policy implementation from two aspects: time effects and policy effects. The effect of the policy on the AIA can be seen by comparing the actual location entropy index changes of the pilot counties (cities or districts) and other non-pilot counties (cities or districts) following the implementation of the policy.

3.2.1. Baseline Regression Models

To identify the impact of FMRML on the AIA, the models of this paper are set as follows:

$${Landloan}_{it}={\beta }_0+{\beta }_{1}DI{D}_{it}+{\beta }_2{X}_{it}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(1)

In Equation (1), ${Landloan}_{it}$ is the explained variable, which is measured by location quotient index; $DI{D}_{it}$ is the core explanatory variable, $DI{D}_{it}$ = $treatmen{t}_i\times tim{e}_t$. During the sample period, if County i is the pilot county, then ${treatment}_i$ = 1, otherwise ${treatment}_i$ = 0. When t ≥ 2016, ${time}_t$ = 1, otherwise ${time}_t$ = 0. Among them, for the treatment group from 2016 to 2019, $DI{D}_{\mathrm{it}}$ = 1, otherwise $DI{D}_{it}$ = 0. Subscripts i and t represent county i and year t, respectively; $X_{it}$ is the control variable; ${\gamma }_t$ represents time-fixed effect; ${\mu }_i$ represents the regional-fixed effect. Since there are only 66 counties (cities or districts) in the sample data and they are all in Hubei Province, this paper only controls the city level; ${\epsilon }_{it}$ represents stochastic error. This article focuses on the coefficient ${\beta }_1$ of the core explanatory variable, which is the net effect of the policy of FMRML on the AIA.

3.2.2. Parallel Trends Testing

The fundamental tenet of using the DID model is to satisfy the assumption of parallel trends, which states that the incidence of the AIA should not have significantly changed between the control and treatment groups over time prior to the implementation of the policy, and that the change tends should be parallel. To test whether the parallel trends assumption is met between pilot and non-pilot counties, the following model is constructed:

$${Landloan}_{it}={\beta }_0+{\beta }_{1}pr{e}_{-6}+{\beta }_{2}pr{e}_{-5}+\dots +{\beta }_{10}pos{t}_3+{\beta }_k{X}_{it}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(2)

In Equation (2), current represents the current period of policy establishment, pre refers to the early stage of policy establishment, subscript −6 represents the sixth year or more before the counties (cities or districts) are established as the pilot, post represents after counties (cities and districts) are established as the pilot, and subscript 3 represents the third year or above after the pilot is established.

3.2.3. Mechanism Test Model

To explore whether the policy indirectly promotes the AIA through agricultural production specialization, based on the aforementioned policy effect identification and the mediating effect test method suggested by Wen et al. (2014) [39], the following test equation is created in this study:

$$M_{it}={\alpha }_0+{\alpha }_{1}DI{D}_{it}+{\alpha }_2{X}_{it}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(3)

$$AI{A}_{it}={\varphi }_0+{\varphi }_{1}DI{D}_{it}+{\varphi }_2{M}_{it}+{\varphi }_3{X}_{it}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(4)

Among them, M is the mediator variable, represented by agricultural production specialization. The total effect of policies is ${\beta }_1$ as shown in Equation (1), the direct effect is ${\varphi }_1$, and the mediation effects of variable M is ${\alpha }_1$ and ${\varphi }_2$. The testing steps of the mediator model are as follows: First, based on Equation (1), it is verified that the policy of FMRML has a promoting effect on the AIA. Then, Equation (3) is regressed. If parameter ${\alpha }_1$ is significantly positive, it indicates that the policy of FMRML has a promoting effect on the mechanism variable. Finally, Equation (4) is regressed. If parameters ${\varphi }_1$ and ${\varphi }_2$ are significantly positive, it indicates that the mechanism variable is an important impact channel for the policy of FMRML to promote the AIA.

4. Variables and Data

4.1. Selection of Variables

4.1.1. Explained Variables

At present, the measurement methods of industrial agglomeration can be divided into the following types: spatial Gini coefficient [40], EG index [41], Herfindahl index [42], and location quotient [43], etc. Among them, the location quotient is a widely used clustering identification method and analysis indicator, which can reflect regional scale differences and also reflect the degree of specialized clustering in a certain industrial sector. The AIA in counties (cities or districts) of Hubei Province is therefore measured in this study using the location quotient. The following is the calculating formula:

$$AI{A}_{it}=\frac{q_{it}}{{\displaystyle \sum p_{it}}}/\frac{Q_t}{P_t}$$

(5)

In Equation (5), $q_{it}$ represents the output value of the AIA in Hubei Province’s county i (city or district) in the year t, $p_{it}$ represents the gross domestic product of Hubei Province’s county i (city or district) in the year t, $Q_t$ represents the AIA output value in year t of Hubei Province, $P_t$ represents the gross domestic product of Hubei Province in year t. If the ratio is greater than 1, it indicates that the county (city or district) has the phenomenon of the AIA, and the larger the ratio is, the higher the agglomeration degree in the county (city or district).

4.1.2. Core Explanatory Variables

In this essay, ${treatmen}_i$ is taken as the policy grouping variable indicating whether the county (city or district) belongs to the pilot county (city or district), ${time}_t$ is taken as the policy time variable, and ${Landloan}_{it}$, the cross term between ${treatmen}_i$ and ${time}_t$, is taken as the core explanatory variable. Specifically, ${treatmen}_i$ of pilot county (city or district) is set as 1, while ${treatmen}_i$ of non-pilot county (city or district) is set as 0. The time dummy variable ${time}_t$ before and after the implementation of the policy is set as 0 or 1, respectively.

4.1.3. Mediation Variables

In this paper, the specialization level of agricultural production (Spe) is selected as the mediation variable, which reflects the degree of intensification and specialization of agricultural production [44], fully reflecting the advantages of agricultural industry layout, and further confirming whether the policy of FMRML has achieved the expected effect. The maximizing index, which is determined by the ratio of the planting area of the largest proportion of crops to the total planting area of crops, calculates the specialization of agricultural production [45].

4.1.4. Control Variables

In addition to the policy of FMRML, which will affect the AIA, there are other factors affecting it. Therefore, it is necessary to control the interference of these exogenous factors. This study chooses the following control variables: (1) economic development level (Lninome), using the logarithmic value of the average disposable income of urban residents to measure the level of local economic development; (2) industrialization degree (Industry); industrialization not only promotes the intensive of agri-food industry, but also promotes the optimization of regional distribution of agri-food. The ratio of the gross product of secondary industry to the regional gross domestic product of each region is used to measure the degree of industrialization of the region; (3) degree of external openness (Open) is the ratio of total export volume and regional gross domestic product of each region selected to represent the degree of openness of each region in which export volume is converted into CNY according to the exchange rate of the current year, and opening up has a certain promotion effect on the upgrading of industrial structure; (4) macro tax burden (Tax) is measured by selecting the ratio of general public finance budget revenue to the gross domestic product in each region and dividing it by the resident population; the tax burden level determines the key factor of whether a company can settle in the region, thereby further promoting the formation and development of local agri-food industry clusters; (5) mechanization degree (Mach) is the ratio of the total power of agricultural machinery to the total planting area of crops used to indicate the degree of mechanization [46,47,48,49].

4.2. Data Sources and Descriptive Statistics

Since the China County Statistical Yearbook only covers 66 counties (cities or districts) in Hubei Province, in order to make the research more rigorous and scientific, this paper selects panel data of 66 counties (cities or districts) in Hubei Province from 2010 to 2019 to empirically analyze the driving factors of the AIA. The related indicators are mainly from the China County Statistical Yearbook, Hubei Statistical Yearbook, Hubei Rural Statistical Yearbook, the statistical yearbooks of various cities in Hubei Province, and the annual Statistical Bulletin of National Economic and Social Development in each county (city or district) of Hubei Province. Linear interpolation is employed to fill in missing data, and simultaneous smoothing is performed.

Table 1. Variable definitions and descriptive statistics.

Variable Name	Variable Abbreviation	Observation	Average Value	Standard Deviation
Agri-food industrial agglomeration	AIA	660	0.9214	0.5821
Farmland management rights mortgage policy	Landloan	660	0.0485	0.2150
Specialization level of agricultural production	Spe	660	0.2557	0.1472
Economic development level	LnIncome	660	9.9386	0.3447
Degree of external openness	Open	660	0.0505	0.0986
Industrialization degree	Indus	660	0.4289	0.1089
Macro tax burden	Tax	660	12.2271	8.7263
Mechanization degree	Mach	660	4.7500	1.8303

displays the descriptive statistics for the variables incorporated into the model.

5. Empirical Results and Analysis

5.1. Difference-in-Differences Analysis

5.1.1. Parallel Trend Assumption Test

One of the essential prerequisites for the DID model is compliance with the parallel trend assumption, which requires excluding the impact of policies or factors unrelated to the policy of FMRML on the AIA in the pilot area. To address this, this paper uses the parallel trend test to draw the average growth trend of the AIA in pilot counties and non-pilot counties from 2010 to 2019. From Figure 2, the trend of the AIA in pilot and non-pilot counties before the implementation of the policy in 2016 is roughly similar. However, the trend of the AIA between pilot and non-pilot counties has changed significantly since the policy’s adoption. Therefore, the effect of the policy on the AIA complies with the parallel trend assumption and also indicates the policy’s unmistakable beneficial effect on the AIA.

5.1.2. Baseline Regression Results

Table 2. Baseline regression results.

Variable	(1)	(2)
	AIA
Landloan	0.131 **	0.140 **
	(0.0623)	(0.0617)
Income		0.601 ***
		(0.2247)
Dus		0.684 *
		(0.3773)
Tax		−0.0087 ***
		(0.0026)
Fac		0.0108
		(0.0149)
Open		0.0026
		(0.1507)
Constant		−4.952 **
		(2.0984)
Time effects	YES	YES
Regional effects	YES	YES
R2	−0.1119	−0.0754
N	660	660

Note: ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively, and the figures in brackets represent the standard errors.

presents the baseline regression results of the impact of the policy of FMRML on the AIA. Under the fixed effects of control time and region, column (1) separately examines the linear relationship between the policy and the AIA. It is discovered that the estimated coefficient of the policy is 0.131, significant at the 5% level. After adding control variables such as economic development level, degree of external openness, industrialization degree, macro tax burden, and mechanization degree to column (2), according to the regression results, the core explanatory variable’s coefficient is equal to 0.140 and passes the 5% significance test. This indicates that the implementation of the policy can greatly aid in the AIA. Compared to counties without the policy, the implementation of the policy has increased the AIA in the county by 14%, and H1 is confirmed. In addition, the coefficient of economic development level and industrialization degree in column (2) is significantly positive, indicating that the improvement of economic development level and industrialization degree can effectively promote the AIA. The coefficient of the macro tax burden is significantly negative, indicating that the rise of the macro tax burden will reduce the AIA by affecting the profits of agri-food enterprises.

5.1.3. Analysis of Policy Dynamic Effects

After the implementation of the policy of FMRML in the pilot counties (cities or districts) in 2016, the changes in the objective environment and the actual degree of policy implementation may be different, which may also affect the dynamic consequences of the policy. Therefore, this paper plots the policy dynamic effects of the policy of FMRML over time. As seen in Figure 3, the regression results demonstrate that there is no significant difference in the AIA in the various regions before the launch of the policy, as indicated by the coefficient of the AIA being less than 0 and not significant. However, following the implementation of the policy, the total coefficient of the time policy interaction term is noticeably positive, demonstrating that the policy does indeed have a promoting impact on the AIA. In addition, the overall effectiveness coefficient of the policy in 2019 is not significant. The possible reason is that in the third year after the implementation of the pilot policy, the policy begins to be fully implemented, and before that, the non-pilot areas absorb the successful experience of the pilot areas in the same jurisdiction and begin to imitate the mode of FMRML, which led to the spillover of the policy effect and thus covered up the policy effect of the pilot areas.

5.2. Robustness Check

To further exclude time changes and sample characteristics from interfering with the effect of the policy of FMRML on the AIA, this paper further adopts the placebo test. The counties (cities or districts) with the same number of treatment groups are randomly selected from the total sample as the pilot counties, and the remaining counties (cities or districts) are taken as the control group. Equation (1) is re-estimated, and the above stochastic process is set for 500 repetitions. Figure 4 displays the distribution of regression coefficients and associated p-values. The estimated coefficients are normally distributed around 0, and the true value of the regression coefficient of −0.171 (the vertical dotted line on the left of Figure 4) is the outlier of the placebo test, indicating that the experimental estimation results are unlikely to be obtained accidentally, which means that the randomly selected samples did not impact the AIA. This further proves that the regression results of the treatment and control groups in the baseline regression are robust.

5.3. PSM-DID Analysis

In order to avoid endogenous issues caused by the non-randomness of policy implementation and eliminate selective bias caused by systematic errors, this paper also reevaluates the impact of the policy of FMRML using the PSM-DID model. To minimize the significant differences between the treatment group and the control group before the pilot policy, this paper selects economic development level, degree of external openness, industrialization degree, macro tax burden, and mechanization degree (specific variable definitions are the same as the control variable) as matching variables, and uses the logit model for propensity score matching. In order to reduce the significant differences between the treatment and control groups prior to the pilot, counties (or cities or districts) from the non-pilot counties (or cities or districts) with similar propensity scores are chosen as the treatment group to match, thus avoiding selection bias resulting from non-random events of the policy and its resulting endogeneity issues. The results in Figure 5 show that the deviation between the treatment group and the control group is significantly reduced, which means a good matching effect. On the basis of PSM, based on the regenerated samples, the DID model is used to identify the impact of the policy on the AIA. As shown in column (1) of

Table 3. Matching regression results of PSM-DID.

Variable	(1) PSM-DID	(2) Baseline Regression Results
Landloan	0.122 *	0.140 **
	(0.0708)	(0.0617)
Constant	−4.917 *	−4.952 **
	(2.8694)	(2.0984)
Time effects	YES	YES
Regional effects	YES	YES
R2	−0.1292	−0.0754
N	412	660

Note: ① ** and * denote significance at the 5% and 10% levels, respectively; ② figures in brackets are the standard errors.

, after matching with PSM, the sign and significance of the interaction coefficient are basically consistent with the baseline results, further confirming that the policy aids in enhancing the AIA.

5.4. Heterogeneity Analysis

Based on the prior empirical and theoretical investigations, we infer that the policy of FMRML can promote the AIA. However, does this promotion effect differ depending on the topography? Therefore, this paper divides the pilot areas into three types according to their topography: plains, hills, and mountains.

Table 4. Heterogeneity analysis.

Variable	(1) Plain	(2) Hilly	(3) Mountain
Landloan	0.0663	0.248 ***	0.144 *
	(0.1929)	(0.0895)	(0.0844)
Constant	−26.71 ***	−3.153	−3.927 *
	(9.6947)	(3.4910)	(2.3113)
Control variables	YES	YES	YES
Time effects	YES	YES	YES
Regional effects	YES	YES	YES
R2	−0.0770	−0.0355	−0.0347
N	130	180	350

Note: ① *** and * denote significance at the 1% and 10% levels, respectively; ② figures in brackets are the standard errors; ③ according to the classification standard of the 2012 China County (City) Social and Economic Statistical Yearbook, the plain region is composed of 13 counties and cities, namely, Zhijiang, Jianli, Huangmei, Tianmen, Yunmeng, Jiangling, Jiayu, Hanchuan, Shishou, Xiantao, Gong’an, Honghu, and Qianjiang; the hilly region includes 18 counties, cities, and districts in Daye, Dangyang, Xiangzhou, Laohekou, Zaoyang, Yicheng, Jingshan, Shayang, Zhongxiang, Xiaochang, Dawu, Yingcheng, Anlu, Songzi, Tuanfeng, Xishui, Qichun and Wuxue; the mountain region includes 35 counties, cities, and districts in Yangxin, Zhuxi, Yuan’an, Wufeng, Baokang, Macheng, Chibi, Lichuan, Xianfeng, Yunyang, Fang, Xingshan, Yidu, Hong’an, Tongcheng, Suixian, Jianshi, Laifeng, Yunxi, Danjiangkou, Zigui, Nanzhang, Luotian, Chongyang, Guangshui, Badong, Hefeng, Zhushan, Yiling, Changyang, Yingshan, Tongshan, Enshi, Xuan’en, and Gucheng.

displays the heterogeneity regression results of topography. The findings demonstrate that the core explanatory variable coefficient in both hills and mountains is significantly positive, while the core explanatory variable coefficient in the plain areas is not significant. It indicates that the effect of the policy on promoting the AIA is more significant in non-plain areas than in plain areas. The main possible reasons are that the economic development level and the degree of industrialization and mechanization in the plain areas are significantly higher than those in non-plain areas, and the AIA is already relatively high. As a result, the promoting effect of the policy of FMRML on the AIA in the plain areas is in the stage of diminishing marginal effect, resulting in an insignificant impact. On the contrary, in non-plain areas, the policy’s effect on promoting the AIA is in the stage of increasing marginal returns, making its impact more significant.

5.5. Analysis of the Influence Mechanism

5.5.1. Mechanism Analysis

In order to test the role of agricultural production specialization in the mechanism of the policy of FMRML in promoting the AIA, the mediation effect variable is used for regression, and

Table 5. Test results of the influence mechanism.

Variable	(1) AIA	(2) Spe	(3) AIA
Landloan	0.140 **	0.0115 *	0.123 **
	(0.0617)	(0.0059)	(0.0612)
Spe			1.525 ***
			(0.4275)
Constant	−4.952 **	0.480 **	−5.684 ***
	(2.0984)	(0.2020)	(2.0876)
Control variables	YES	YES	YES
Time effects	YES	YES	YES
Regional effects	YES	YES	YES
R2	−0.0754	0.0317	−0.0541
N	660	660	660

Note: ① ***, **, * denote significance at the 1%, 5%, and 10% levels, respectively; ② figures in brackets are the standard errors.

displays the precise outcomes. The coefficient value of the policy is considerably positive at the 10% level of statistical significance, according to the regression results in column (2), suggesting that the policy could increase the level of agricultural production specialization. The main reason is that, on the one hand, the policy of FMRML promotes productive lending. Based on the assumption of rational actors, to pursue optimal production efficiency, agricultural entities allocate more credit funds to agricultural production as the accessibility to credit increases. This includes upgrading agri-food techniques, purchasing agri-food equipment, and improving their cropping structure, thereby enhancing the specialization of agricultural production. On the other hand, the policy of FMRML promotes productive consumption. By alleviating the capital constraints of agricultural entities, this policy stimulates productive consumption, while also increasing the willingness to transfer land. This leads to the consolidation of fragmented farmland, reduction of fallow farmland, and the promotion of scale in agricultural production, thereby driving the development of agricultural production specialization. In addition, as indicated by column (3) of Table 5, after adding agricultural production specialization as the mediation variable, the coefficients of the policy and the AIA are significantly positive at the 5% and 1% levels, which means that agricultural production specialization is an important way for the policy to promote the AIA.

5.5.2. Endogeneity Issues

Next, when examining the relationship between agricultural production specialization and the AIA, potential endogeneity issues mainly arise from reverse causation [50]. That is, regions with relatively lower levels of the AIA may have less demand for agri-food, which in turn impedes the development of agricultural production specialization. Therefore, this article incorporates one-period lags of mediation variables as instrumental variables to mitigate endogeneity issues arising from reverse causation. On the one hand, agricultural production specialization exhibits some dynamic continuity, where the previous period’s level of agricultural production specialization lays the foundation for the present period’s development. This satisfies the requirement for relevance. Through affecting the present period’s agricultural production specialization, the previous period’s agricultural production specialization, on the other hand, will affect the present level of the AIA, thus meeting the condition of exclusiveness. As a result, this study employs the lagged period of agricultural production specialization as an instrumental variable and utilizes the instrumental variable approach through a two-stage regression, as shown in

Table 6. Results of the instrumental variable method.

Variable	(1) Spe	(2) AIA
L.Spe	0.833 ***
	(0.0319)
Spe		1.467 **
		(0.5854)
Control variables	YES	YES
Time effects	YES	YES
Regional effects	YES	YES
R2	0.9884
Anderson LM statistic	338.559 ***
Cragg–Donald Wald F statistic	681.250
N	594

Note: ① *** and ** denote significance at the 1% and 5% levels, respectively; ② figures in brackets are the standard errors.

. Under the first phase of regression, the estimated coefficient of IV in column (1) is positive and significant at 1% level, and corresponds to the expected estimation results. Furthermore, to assess the effectiveness of the instrumental variable, tests for unidentifiable and weak instrumental variables are conducted. The results indicate that the Anderson LM test yields a p-value of 0.0000, indicating no underidentification issue. The Cragg–Donald Wald F statistic is substantially higher than the critical value (approximately 16.38) suggested by Stock et al. (2002) [51] at the 10% level, rejecting the hypothesis of weak instrumental variables. Since only a single instrumental variable is employed, the exact identification condition is met, obviating the need for an exogeneity test. Hence, selecting the lagged one-period agricultural production specialization as the instrumental variable is both reasonable and effective. Column (2) presents the estimation results of the second stage, which reveal a positive and statistically significant coefficient for agricultural production specialization, consistent with the findings of the mechanism test. These results substantiate that even after addressing endogeneity through instrumental variable approaches, the promotion effect of agricultural production specialization on the AIA remains significant.

6. Conclusions and Suggestions

6.1. Conclusions

This paper takes the policy of FMRML implemented in China in 2016 as a quasi-natural experiment. In this paper, using panel data from 66 counties (cities or districts) in Hubei Province from 2010 to 2019, the DID model and mediation effect model are used to empirically test the impact of the policy on the AIA. The theoretical mechanisms of the policy on the AIA are also explored. The main conclusions are as follows: (1) From the perspective of effect, the baseline regression results show that the policy can significantly promote the AIA, which is also supported by parallel trend tests and placebo tests. The above conclusions validate research hypothesis H1. (2) From the trend point of view, although the effect of the policy is significant, the long-term effect is gradually weakened, mainly due to the spillover effect of the policy. (3) In terms of the influence mechanism research, the policy of FMRML can promote the AIA through agricultural production specialization, such as increasing agri-food inputs, innovating agri-food technologies, upgrading agri-food equipment, etc., which verifies the research hypothesis H2 in this article. Moreover, the mediating effect of agricultural production specialization is further demonstrated by eliminating endogenous problems through instrumental variables estimation. (4) According to an analysis of regional heterogeneity, the promoting effect of the policy on the AIA has regional heterogeneity. Compared with the plain area, the agglomeration degree is more obvious in non-plain areas, mainly because the promotion effect of FMRML on the AIA in non-plain areas is in the marginal effect increasing stage, so the impact is more significant. These findings validate hypothesis H3.

6.2. Policy Recommendations

Based on the research conclusions of this paper, the following recommendations for policy are drawn: (1) It is recommended to gradually expand the scope of the pilot policy of FMRML. Our research demonstrates a significant positive relationship between the policy of FMRML and the AIA, which not only alleviates the financing constraints of farmers who mortgage farmland management rights but also promotes the factors of the labor force, land, and capital to tilt towards the agri-food industry and promotes the AIA. These findings can inform the reference for the country to enrich the methods of rural pledge guarantee. (2) Promoting the institutionalization and legalization of the policy of FMRML is crucial. According to the research conclusion, it is observed that the long-term effectiveness of the policy shows a gradual weakening trend. To ensure comprehensive implementation of the policy and achieve favorable medium to long-term outcomes, it is recommended to institutionalize and legalize the policy, guaranteeing its efficacy. (3) To accurately implement policies based on time, location, and household is necessary for the effective implementation of the policy of FMRML. While promoting the AIA as a whole, the effectiveness of this policy varies in different regions due to the differences in their own conditions. Therefore, the policy of FMRML needs to be implemented accurately according to the actual development situation. (4) Promoting agricultural production specialization is recommended. This study shows that the specialization of agricultural production plays a key role in the AIA, and the policy of FMRML mainly promotes the specialization of local planting patterns and the optimization and development of industrial structures by stimulating the increase of local agricultural loans. In the process of rural revitalization, the development of the AIA is an essential component of agricultural modernization. Therefore, we should advocate for farmers to make rational utilization of agricultural financial policies to promote specialized agricultural production and further expand the use value of FMRML.