Does Agricultural Mechanization Help Farmers to Strengthen Sustainability and Protect Cultivated Land? Evidence from 2118 Households in 10 Provinces of China

Abstract

The protection of cultivated land is related to food security and sustainable agricultural development. Improving agricultural planting efficiency and reducing chemical inputs are important to promoting sustainability and protecting cultivated land, and agricultural mechanization plays an important role in this process. Based on the survey data of 2118 households in 10 provinces of China, we used the Oprobit and IV-Oprobit models to analyze the impact and mechanism of agricultural mechanization on the behaviors of farmers in achieving sustainability and protecting cultivated land. The results show that agricultural mechanization has a significant promotion effect on the behaviors of farmers, especially in motivating them to adopt higher levels of protective behaviors in terms of sustainable land cultivation. At the same time, the impacts of agricultural mechanization on the different production links were different. The promotion effect of the harvesting link on the sustainability protection behaviors of farmers was the most obvious, and the promotion effects of the tillage and sowing links were the least obvious. In addition, planting income and fertilizer input played a role in mediating between mechanization and cultivated land sustainability protection. Further analysis showed that agricultural mechanization can more effectively motivate farmers with full-time businesses or higher land concentrations to prioritize cultivated land sustainability. Therefore, it is necessary to pay attention to the role of agricultural mechanization in promoting sustainability, protecting cultivated land, and promoting innovative green agricultural machinery. Via mechanization, we can increase the incomes of farmers, reduce excessive fertilizer use, and specifically target full-time farmers engaged in agricultural production and key aspects of land sustainability protection to promote the construction of better agricultural machinery systems, as well as agricultural machinery research and innovation, thereby fully leveraging the ecological protection effects of agricultural mechanization.

1. Introduction

Cultivated land is the foundation of ensuring national food security and stable social and economic development, and strengthening sustainable production on cultivated land is of considerable significance to realizing the aim of “storing grain in land and grain in technology” and promoting green and sustainable agricultural development [1]. Improving the quality level of cultivated land requires the use of cultivated land resources to transition from the pursuit of quantity intensity to the pursuit of quality and efficiency, as well as the transition from extensive production to green development. By reducing fertilizer and pesticide inputs and taking technological progress as the driving force for growth, higher outputs can be obtained with less input and environmental pollution can be reduced, as well as the excessive consumption of cultivated land resources, improving their efficient utilization. In recent years, in order to promote sustainability in land cultivation, China has successively introduced and implemented various policies and actions at the macrosystem level, such as “comprehensive land consolidation”, “high-standard farmland construction”, and “cultivated land rest” [2]. However, in the face of the challenges confronting China, a country with a large population and little free land in the existing land system, the utilization of cultivated land is characterized by its obvious small scale and low efficiency, and serious problems such as the excessive application of agricultural chemicals, the super-intensive utilization of cultivated land resources, and the overdrawing of cultivated land fertility [3,4] hinder high-quality agricultural transformation and development. Therefore, it is of considerable practical significance to explore the influencing factors that restrict cultivated land sustainability and protection in China against the background of the multiple challenges regarding cultivated land quantity, quality, ecology, and efficiency. Through a review of the existing research, it was found that the cognition of farmers regarding cultivated land protection [5,6], agricultural land property rights [7], agricultural technology promotion [8], government regulation [9], and other factors has an important impact on sustainable land cultivation. As a specific representation of agricultural technology progress, the development and application of agricultural mechanization also deserves attention.

Agricultural mechanization is the process of using advanced and applicable agricultural machinery and equipment to improve production and operating conditions, and constant improvements in agricultural production technology offer economic and ecological benefits. With the rapid development of industrialization and urbanization in China, agricultural mechanization has become an important part of modern agricultural production [10,11]. At present, many agricultural production links have been mechanized, such as plowing, fertilization, pesticide application, and harvesting, which not only reduces the labor intensity of farmers but also improves the quality and output of agricultural production. In the early stage, scholars mainly discussed the economic benefits of agricultural mechanization from the perspective of labor substitution. For example, based on theoretical and empirical analysis, Wang et al. [12] and Wang et al. [13] found that agricultural machinery development can effectively replace agricultural labor and promote an increase in agricultural output. Similarly, Thompson et al. [14] took rice planting in California as an example and found that harvesting mechanization could reduce the use of labor by 92–97%, as well as the labor costs. Benin [15] believes that mechanization, as a typical labor-saving process, could improve labor productivity and increase crop outputs. In recent years, with the continuous development of intelligent and precise agricultural machinery, some scholars have begun to discuss the impact of agricultural mechanization on chemical inputs and agricultural ecology from the perspective of improving operation standardization and quality. For example, Xu Z. et al. [16] believe that improvements in agricultural machinery operation precision and technology have a spillover effect that can not only reduce labor input but also chemical fertilizer application, which is conducive to agricultural ecological environment protection. Similarly, CAI Rong et al. [10] took wheat planting in China as an example and found that mechanization, as a carrier of technological progress, could reduce irrational fertilizer use and improve fertilizer application efficiency. Based on fertilizer and pesticide use data from 30 provinces of China, Du M. et al. [17] comprehensively measured the agricultural eco-efficiency and found that agricultural mechanization could significantly improve the eco-efficiency and ecological environment. In addition, from the perspective of purchasing professional mechanization services, Zhang et al. [18] and Lu et al. [19] believe that socialized agricultural machinery services, such as farming, pesticide application, and fertilization, can realize the efficient utilization of chemicals, improve the quality of cultivated land, and promote green agricultural production through quantitative and precise operational control.

In general, the existing research not only focuses on the economic benefits of agricultural mechanization but is also beginning to pay attention to the ecological effects of agricultural mechanization, providing rich theoretical and methodological support for the relevant research in this paper. First, in the existing studies, there is a lack of analysis on the mechanism of how agricultural mechanization acts on cultivated land sustainability protection at the microlevel of farmers; most of them measure the cultivated land sustainability protection behaviors of farmers with a single measure, ignoring the diversity of the actual measures taken by farmers to protect cultivated land. Second, due to differences in the mechanical operating conditions and technical complexity in the different production links [16,19], mechanical application in the different links has different impacts on the cultivated land sustainability protection behaviors of farmers. However, the existing research lacks in-depth analyses of the influence of agricultural mechanization on the farmland sustainability protection behaviors of farmers from the perspective of the agricultural production sub-links. Third, there are differences in the adaptability and application effects of agricultural machinery under the different employment statuses and cultivated land conditions of farmers. In view of this, based on the survey data of 2118 farmers in 10 provinces of China, we comprehensively investigated the cultivated land sustainability protection behaviors of farmers based on 10 types of cultivated land protection measures. We comprehensively used the Oprobit model, an instrumental variable model, and a mediating effect model to systematically evaluate the impact and mechanism of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers. Furthermore, we assessed the effects of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers in different production links, management characteristics, and cultivated land conditions. The relevant research and findings in this paper have important reference value for deeply understanding the important role of agricultural mechanization development and application, building a differentiated policy support system to balance the relationship between agricultural mechanization and cultivated land sustainability protection, and alleviating the adverse impact of “agricultural labor outflow” on the quality of agricultural development.

2. Theoretical Analysis and Research Hypotheses

2.1. Analysis of the Impact of Agricultural Mechanization on the Cultivated Land Sustainability Protection Behaviors of Farmers

Compared with manual operations, mechanization is conducive to the standardization of agricultural production, improving the input use efficiency and reducing damage to the quality of cultivated land. The traditional artificial farming method requires a lot of manpower and time, and it causes a large degree of soil compaction, which easily causes the destruction of cultivated land [20]. However, agricultural mechanization can reduce the deep tillage of the soil and the degree of soil compaction through reasonable mechanical operations, thereby reducing the damage to cultivated land [21]. Moreover, agricultural mechanization contributes to improvements in soil quality. Through reasonable mechanical operation, organic materials such as straw can be fully returned to the field, improving the soil organic matter content and soil structure and increasing the soil fertility, thereby improving the soil quality [22]. In addition, agricultural machinery is also needed for the control and operation of wasteland reclamation, the transformation of medium- and low-yield fields and grassland, the development of sandy land, and the treatment of waste-sand salts and alkalis. Therefore, theoretically, agricultural mechanization encourages farmers to strengthen their cultivated land sustainability protection.

In addition, from the perspective of the agricultural production sub-links, considering the differences in the agronomic complexity and factor inputs, the difficulty of mechanical substitution, as well as other aspects in different production links [16], the agricultural mechanization in different links may have different impacts on the cultivated land sustainability protection behaviors of farmers. For example, harvesting mechanization can help realize the comprehensive utilization of straw and environmental protection treatment, improve soil quality, and achieve “quality improvement”, and the mechanization of pesticide spraying mainly achieves “pollution reduction” by improving disease and insect control efficiency and reducing pesticide use and soil pollution; however, the effects of the two on cultivated land sustainability protection are not the same. Based on the above analysis, in this paper, we propose hypothesis H1:

H1: 

Agricultural mechanization can encourage farmers to strengthen their cultivated land sustainability protection, and different mechanization links have different impacts on the cultivated land sustainability protection behaviors of farmers.

2.2. Analysis of the Mechanism of Agricultural Mechanization Affecting the Cultivated Land Sustainability Protection Behaviors of Farmers

(1) Analysis of the “income promotion” impact mechanism.

Their agricultural planting income is an important motivation and goal for farmers in the management of farmland, and protecting the sustainability of farmland is an indispensable element for sustainable agricultural development. Agricultural mechanization can promote the sustainability protection of the cultivated lands of farmers by increasing the planting returns. Li [23], Wang [12], and other scholars point out that agricultural mechanization is an inevitable choice for agriculture to cope with the Lewis turning point and the disappearance of the demographic dividend. Agricultural mechanization can not only replace human labor, save labor costs, and reduce labor intensity, but also improve operation efficiency, alleviate agricultural time constraints, and improve agricultural outputs. Therefore, agricultural mechanization can improve the agricultural planting incomes of farmers by reducing production costs [24] and improving their agricultural outputs [15]. The increase in planting income helps farmers to increase their investment in and management of farmland and to promote cultivated land sustainability protection. Thompson [14] points out that the increase in planting income can effectively improve the financial security of farmers so that they can purchase high-quality pesticides and fertilizers and apply green production technology, providing them with more funds to strengthen their cultivated land sustainability protection. Based on the above analysis, in this paper, we propose Hypothesis H2:

H2: 

Planting income plays an intermediary role in the agricultural mechanization impact on the cultivated land sustainability protection behaviors of farmers.

(2) Analysis of the “fertilizer reduction” impact mechanism

There is a close relationship between fertilizer application and cultivated land sustainability protection. In reality, farmers apply chemical fertilizers to increase their crop yields. However, the excessive use of fertilizers not only harms the soil environment and ecosystem but also increases agricultural production costs and risks. Agricultural mechanization can promote the sustainability protection of cultivated land for farmers by reducing the amount of chemical fertilizer applied. In cases of high labor costs, farmers using artificial fertilization generally reduce the fertilization frequency and increase the fertilizer amount applied each time, which often leads to excessive fertilizer application [25]. However, mechanized fertilization can avoid uneven and non-standard artificial fertilization and can implement quantitative and standardized fertilization control, which can effectively improve the chemical fertilizer application efficiency and promote chemical fertilizer application reduction [26]. Moreover, mechanical fertilization is conducive to the adoption of fertilizer-saving methods, such as lateral and deep fertilization, which can reduce fertilizer application by reducing fertilizer loss and improving utilization efficiency [16]. Therefore, agricultural mechanization can help farmers achieve cultivated-land sustainability protection by reducing the amount of chemical fertilizer applied and weakening the action of chemical fertilizers on soil pollution. Based on the above analysis, we propose Hypothesis H3:

H3: 

The fertilizer input plays an intermediary role in the impact of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers.

2.3. Heterogeneity Analysis of the Influence of Agricultural Mechanization on the Farmland Sustainability Protection Behaviors of Farmers

The impact of agricultural mechanization on cultivated land sustainability protection will be different under different concurrent agricultural employment statuses and land fragmentation levels. First of all, whether farmers engage in part-time or full-time agricultural production will directly affect the time and energy required for them to engage in agricultural production and management. For example, full-time farmers can devote more time and energy to agricultural production and management, and they can better balance the relationship between agricultural mechanization and cultivated land sustainability protection to more easily carry out detailed cultivated land sustainability protection work [27]. Conversely, part-time farmers regard agriculture as a sideline, invest insufficient energy in agricultural production and management, and have a limited ability to operate and manage agricultural machinery, which leads to their inability to give full play to the positive role of agricultural machinery in cultivated land sustainability protection. Second, different land fragmentation levels will affect the ability of agricultural mechanization to promote cultivated land sustainability protection behaviors in farmers. Land fragmentation refers to the phenomenon of small and highly dispersed farmland areas. The higher the degree of land fragmentation, the smaller and more complex the farmland area, the poorer the adaptability of agricultural machinery and equipment, and the more time and higher cost required for the mechanical operations, making it difficult for farmers to realize mechanized operations [28], resulting in a relatively low mechanization promotion effect on their farmland sustainability protection behaviors. Based on the above analysis, we propose Hypothesis H4 below. The theoretical framework of this study is illustrated in Figure 1.

H4: 

In agricultural production, the employment type of farmers (full- or part-time) and the level of land fragmentation can affect the ability of agricultural mechanization to promote cultivated land sustainability protection behavior in farmers.

3. Materials and Methods

3.1. Data Sources

The data used in this study were obtained from the China Rural Revitalization Comprehensive Survey (CRRS), organized by the Rural Development Institute of the Chinese Academy of Social Sciences in 2020, which covers agricultural production, rural development, and the lives and consumption behaviors of farmers. In order to obtain convincing samples, according to the social economy and the conditions of the agricultural and rural development and geographical locations in different regions of China, the research group selected 10 provinces (autonomous regions) as the research areas, including Guangdong, Zhejiang, and Shandong in eastern China; Anhui and Henan in central China; Guizhou, Sichuan, Shaanxi, and Ningxia in western China; and Heilongjiang in northeastern China. Specifically, in the selection of sample farmers, the research group adopted the stratified random sampling method. The steps were as follows: First, according to the per capita GDP level of the sample provinces, all the counties (cities and districts) were divided into five groups: low-level, medium–low-level, medium-level, medium–high-level, and high-level counties. Then, according to the same sampling method, the townships (towns) in the sample counties were divided into low, medium, and high groups according to the per capita GDP level, and one township (town) was randomly selected from each group; that is, three townships (towns) were selected from each county; then, the village economic development level was determined according to the guidance of the township (town) government, all the villages were divided into two groups with poor and good economic levels, and one village was randomly selected from each group; that is, two villages were selected from each township (town). Subsequently, based on the roster provided by the village committee, between 12 and 14 households were randomly selected from each village using the equidistant sampling method to conduct the survey. A total of 3833 sample data points were collected in this survey.

3.2. Research Methods

In this study, we used the Oprobit model to analyze the impact of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers, and we further used the mediating effect model to deeply analyze the impact mechanism. The reasons for selecting the two abovementioned models and their basic introductions are as follows:

(1) Oprobit model protection

Due to the treatment of the explained variable, the protection of cultivated land sustainability, as an ordered categorical variable, we used ordinal models for the quantitative analysis, among which the Oprobit and Ologit models are the two most commonly used. The Ologit model assumes that the random variable follows a logistic probability distribution, while the Oprobit model assumes that the random variable follows a normal distribution [29], which is more consistent with the characteristics of the research data in this study. Therefore, we adopted the Oprobit model for the fitting estimation. The Oprobit model is set as follows:

$${Y_i}^{\ast }=a_{0 }+a_1{Mechanization}_i+a_2{X}_i+{\epsilon }_i$$

(1)

In Equation (1), ${Y_i}^{\ast }$ represents the latent variable that cannot be estimated; ${Mechanization}_i$ represents agricultural mechanization; $a_{0 }$ and $a_1$ represent the intercept term and the coefficient to be estimated of explanatory variable $Mechanization$; $X_i$ represents the control variables, including the agricultural decisionmaker characteristics, household characteristics, operation characteristics, policy characteristics, and internet information acquisition characteristics; $a_2$ represents the control variable coefficient to be estimated; ${\epsilon }_i$ is the random disturbance term.

(2) Mediating-effect model

Agricultural mechanization can affect cultivated land sustainability protection by increasing the planting incomes of farmers and reducing their fertilizer inputs. In order to test the action mechanism of agricultural mechanization on cultivated land sustainability protection, we built the following mediating effect model with reference to the research of Wen Zhonglin et al. [30]:

$$M_i={\beta }_{0 }+{\beta }_1{Mechanization}_i+{\beta }_2{X}_i+{\epsilon }_i$$

(2)

$$Y_i=c_{0 }+c_1{Mechanization}_i+c_2{X}_i+c_3{M}_i+{\epsilon }_i$$

(3)

In Equations (2) and (3), $M_i$ is the mediating variable, which is the farmer’s planting income and fertilizer input, and $\beta$ and $c_{ }$ are the parameters to be estimated. The meanings of the remaining variables and symbols are consistent with those of Equation (1). It should be noted that the measurement software used in the data analysis process using the above methods in the article is Stata17.0.

3.3. Variable Selection

3.3.1. Explanatory Variable

Agricultural production involves a series of processes, such as tillage, planting, and harvesting, and farmers undertake a variety of measures to protect the quality of their cultivated land in practical agricultural production. In order to comprehensively and systematically measure the cultivated land sustainability protection behaviors of farmers, we based this study on the Action Plan for Cultivated Land Sustainability protection and Improvement issued by the Ministry of Agriculture and Rural Affairs of China, which emphasizes that the focus of cultivated land sustainability protection is “soil improvement, fertility enhancement, water and fertilizer conservation, and pollution control and restoration.” From the four aspects of cultivated land sustainability protection—“soil improvement”, “fertility enhancement”, “water and fertilizer conservation”, and “pollution control and repair”—two–three indicators were selected for comprehensive investigation, specifically (1) “rotation tillage”, “fallow tillage”, and the “integration of tillage and sowing” in soil improvement; (2) “organic fertilizer application” and “straw return” in terms of soil fertility; (3) the “application of water-saving irrigation technology” and the “integration of tillage and application” in water and fertilizer conservation; and (4) the “reduced application of chemical fertilizer”, “reduced application of pesticide”, and “recovery of pesticide packaging” in pollution control and repair, drawing on the research findings of Cheng C. [31] and Li B. [32]. The cultivated land sustainability protection levels were measured according to the numbers of the above 10 protection measures adopted by farmers; that is, the greater the number of protection measures adopted by farmers, the higher the degree of cultivated land sustainability protection.

According to the results (Figure 2), only 57 farmers did not adopt any cultivated land sustainability protection measures, accounting for 2.69%. More farmers adopted two–three cultivated land sustainability protection measures, accounting for 51.84%. Fewer farmers adopted seven–eight cultivated land sustainability protection measures, accounting for 0.75%. Although we selected 10 measures to measure the cultivated land sustainability protection behaviors, none of the sample farmers adopted more than 8 measures at the same time; thus, the value of this variable ranges from 0 to 8. In addition, in terms of the adoption of different cultivated land sustainability protection measures, the adoption rates of “returning straw to the field” and “pesticide packaging recycling” were the highest at 81.07% and 55.48%, respectively.

3.3.2. Explanatory Variables

According to the previous theoretical analysis, mechanization can promote the protection of cultivated land sustainability. However, the impacts of the different links will vary due to factors such as the mechanized operation conditions and technical complexity. Based on this, we divided the mechanization degree into the whole-link mechanization degree and the sub-link mechanization degree. The mechanization degree in the whole link is measured as the average proportion of the sample farmers in the plowing, sowing, fertilization, spraying, and harvesting links. In the sub-link analysis, according to the requirements of the complete agricultural production process and the indispensable operations for crop growth and yields, and considering that, in order to promote the sustainability protection of cultivated land, farmers usually integrate the two production links of cultivated land and sowing, we analyzed the four links of plowing and sowing, fertilization, spraying, and harvesting, and we selected the proportion of mechanized operation of the sample farmers in each link to measure its mechanization degree.

3.3.3. Mediating Variables

Based on the previous theoretical analysis, we believe that the level of agricultural mechanization encourages farmers to strengthen their cultivated land sustainability protection mainly by increasing their planting incomes and reducing their fertilizer inputs. Therefore, in this study, we selected planting income and fertilizer input as the mediating variables for testing and analysis, among which the planting income is represented by the annual net planting incomes of the farmers, and the fertilizer input is represented by the total amount of fertilizer input per unit of cultivated land area.

3.3.4. Control Variables

There are many complex factors affecting the cultivated land sustainability protection behaviors of farmers, and previous studies [7,27,33] have shown that their personal characteristics, household management characteristics, policy characteristics, and internet information acquisition characteristics have important impacts on their cultivated land sustainability protection behaviors. According to the theory of farmer behavior and referring to the above research, we selected eight variables from five aspects as the control variables, including the following: (1) in terms of the characteristics of agricultural decisionmakers, two variables were selected: “age” and “education level”; (2) in terms of family characteristics, the “total family population” variable was selected; (3) in terms of the operation characteristics, three variables were selected: the “land scale”, “agricultural disaster”, and the “distance from home of the plot”; (4) in terms of policy characteristics, the “government subsidy” variable was selected; (5) in terms of the internet information acquisition characteristics, the “network convenience” variable was selected. The definitions and descriptive statistics of the variables in this study are presented in

Table 1. Variable definitions and descriptive statistics.

Type of Variable		Name of Variable	Variable Definitions and Assignments	Mean	St. Dev.
Explained Variable		Cultivated land sustainability protection	Number of cultivated land sustainability protection measures adopted by farmers in the planting process	2.848	1.387
Explanatory Variable		Agricultural mechanization	Average proportion of mechanical operations in plowing, sowing, spraying, fertilization, and harvesting	0.453	0.308
Mediating Variable		Farmers’ planting income	Annual net planting incomes of rural households (CNY ten thousand)	1.129	2.962
		Fertilizer input	Total amount of fertilizer input per unit area of cultivated land (kg)	55.652	30.574
Control Variables	Characteristics of Farmers	Age	Actual age of agricultural decisionmakers (years)	54.873	10.866
		Level of education	1—primary school and below; 2—junior high school and technical secondary school; 3—senior high school and vocational high school; 4—junior college and above	1.763	0.719
	Characteristics of agricultural decisionmakers	Total household population	Total rural household population (persons)	4.129	1.542
	Characteristics of operations	Size of land	Total operating area (mu)	31.196	88.789
		Agricultural disaster	Disaster or not: 0 = no; 1 = yes	0.340	0.474
		Distance between the plot and the residence	Average distance between maximum three plots and residence (km)	1.168	2.866
	Policy characteristics	Government subsidy	Whether farmers have received government subsidies for land sustainability protection: 0 = no; 1 = yes	0.107	0.309
	Internet information acquisition characteristics	Network convenience	1 = poor, frequent internet disconnections; 2 = good, occasional internet disconnections; 3 = very good, internet is rarely disconnected.	2.223	0.766

.

4. Results and Analysis

4.1. The Impact of Agricultural Mechanization on the Farmland Sustainability Protection Behaviors of Farmers

In order to ensure the accuracy of the estimated results of the Oprobit model, we first used the variance inflation (VIF) coefficient method to conduct the multicollinearity test. According to the VIF test results, the maximum VIF value of the explanatory variables is 1.16, the mean value is 1.08, and the results are all less than 5.

Table 2. Estimation results of the Oprobit model.

Variables	(1) Direct Effect	(2) Marginal Effect
		No Protection	Low Protection Level	High Protection Level
Agricultural mechanization	0.5657 *** (0.0782)	−0.0268 *** (0.0063)	−0.1247 *** (0.0261)	0.1515 *** (0.0314)
Age	−0.0010 (0.0022)	−0.0002 (0.0002)	−0.0007 (0.0007)	0.0009 (0.0009)
Level of education	0.0872 *** (0.0322)	−0.0083 *** (0.0025)	−0.0385 *** (0.0108)	0.0467 *** (0.0131)
Total household population	0.0072 (0.0147)	−0.0004 0.0011	−0.0021 (0.0049)	0.0025 (0.0060)
Size of land	−0.0001 (0.0003)	1.73 × 10−6 (0.0000)	8.07 × 10−6 (0.0001)	−9.81 × 10−6 (0.0001)
Agricultural disaster	−0.0993 ** (0.0486)	0.0078 ** (0.0036)	0.0364 ** (0.0163)	−0.0442 ** (0.0198)
Distance between the plot and the residence	−0.0155 * (0.0080)	0.0006 (0.0006)	0.0026 (0.0027)	−0.0031 (0.0033)
Government subsidy	0.7322 *** (0.0733)	−0.0418 *** (0.0070)	−0.1947 *** (0.0239)	0.2365 *** (0.0287)
Network convenience	0.0914 *** (0.0310)	−0.0066 *** (0.0023)	−0.0309 *** (0.0100)	0.0375 *** (0.0122)
Number of samples	2118
LR chi2	179.09
Pseudo-R2	0.0244

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

reports the results of the impact of agricultural mechanization on the farmland sustainability protection behaviors of farmers. Among them, according to the regression results of the direct influence effect in Column (1), agricultural mechanization has a significantly positive impact on the cultivated land sustainability protection behaviors of farmers, passing the significance test at the 1% confidence level, indicating that agricultural mechanization can encourage farmers to strengthen their cultivated land sustainability protection. Similar to this result, Xu Z. et al. [16] and Du M. et al. [17] also found that agricultural mechanization has a promoting effect on agricultural ecological protection.

Furthermore, according to the regression results of the marginal impact effect in Column (2), when the other conditions remain unchanged, the probabilities of unprotected and low-level-protected cultivated land sustainability decrease by 2.68% and 12.47% for every unit increase in the agricultural machinery application level, respectively. However, the probability of adopting high-level protection increases by 15.15% (referring to Liu et al. [8] and Qi et al. [33]) when farmers adopt zero cultivated land sustainability protection measures, meaning that the farmers do not carry out any cultivated land sustainability protection. The adoption of one–three protection measures by farmers (three is the average value of the research samples) means that the farmers carry out low-level cultivated land sustainability protection. When four or more measures are taken, this means that the farmers carry out high levels of cultivated land sustainability protection, indicating that agricultural mechanization development and application will encourage farmers not only to strengthen their cultivated land sustainability protection but also to achieve higher levels of cultivated land sustainability protection behaviors.

Among the control variables, the education level of the farmers, the agricultural disaster situations, the distances between the plots and the residences, whether there are government subsidies, and the network convenience of farmers have significant impacts on their farmland sustainability protection behaviors. Specifically, the higher the education levels of farmers, the more inclined they are toward cultivated land sustainability protection, possibly because farmers with higher education levels pay more attention to green agricultural production and cultivated land protection. The agricultural disaster situation has a significant negative impact on the cultivated land sustainability protection behaviors of farmers, probably because natural disasters, such as drought, aggravate cultivated land sustainability deterioration. The distance of the plot from the residence has a significantly negative effect on the cultivated land sustainability protection by farmers, possibly because the farther the plot is from home, the higher the cost of protection, and the less inclined farmers are toward it. The presence of a government subsidy has a positive and significant impact on the sustainability protection of cultivated land, indicating that subsidy policies such as rotation and fallow can stimulate the enthusiasm of farmers to protect their cultivated land. The convenience of internet information acquisition has a positive and significant impact on the cultivated land sustainability protection of farmers, possibly because cultivated land sustainability protection involves the application of new knowledge and new technology.

4.2. Results of the Influence of Mechanization in Different Links on the Cultivated Land Sustainability Protection Behaviors of Farmers

In order to further explore the differences in the impact of mechanical operations on the different links on the cultivated land sustainability protection behaviors of farmers, we further examined them based on the perspective of agricultural production sub-links.

Table 3. Estimation results of mechanization in different production sub-links based on Oprobit model.

Variables	Plowing and Sowing	Spraying Pesticides	Applying Fertilizers	Harvesting
Agriculture mechanization	0.1587 ** (0.0643)	0.3458 *** (0.0530)	0.2429 *** (0.0571)	0.5076 *** (0.0490)
Control variables	Controlled	Controlled	Controlled	Controlled
Number of samples	2118	2118	2118	2118
LR chi2	130.83	167.36	142.81	232.40
Pseudo-R2	0.0178	0.0228	0.0195	0.0317

Note: ** and *** represent significance at 5% and 1% levels, respectively.

reports the estimated results of the impact of mechanization in the different links on the farmland sustainability protection behaviors of farmers. On the whole, agricultural mechanization in the different links has a significantly positive impact on the cultivated land sustainability protection behaviors of farmers; that is, agricultural mechanization promotes cultivated land sustainability protection, which is consistent with the conclusions of the whole-link analysis above, indicating that the estimated results are robust. Furthermore, the promotion effect of mechanization is the most obvious in harvesting, followed by those of pesticide spraying and fertilizer application, and the promotion effects of tillage and sowing are the least obvious. As analyzed above, “returning straw to the field” and “pesticide packaging recycling” in the harvesting process are the two cultivated land sustainability protection measures with high adoption rates among the surveyed farmers. However, mechanical operations in tillage and sowing are easily constrained by the plot fragmentation level, which may limit the role of agricultural mechanization, resulting in a relatively small promotion effect of the mechanical operations in tillage and sowing on the cultivated land sustainability protection behaviors of farmers. Accordingly, Hypothesis H1 is verified.

4.3. Endogeneity Discussion and Treatment

Agricultural mechanization and cultivated land sustainability protection are both the results of the decisions of farmers, and they may have endogenous problems. Specifically, while agricultural mechanization promotes the cultivated land sustainability protection behaviors of farmers, the improvement in the cultivated land sustainability protection levels will also adversely affect the mechanization application levels. Although important control variables that affect the farmland sustainability protection behaviors of farmers were included in the model as much as possible, endogeneity may still exist due to missing variables. Therefore, based on the Oprobit model regression, we introduced the instrumental variable regression method to correct the potential endogeneity problem.

In terms of the instrumental variable selection, referring to the practice of Zhu J. [24], the village-level agricultural mechanization (excluding the household) was used as the agricultural mechanization instrumental variable. Village-level mechanization directly affects the mechanization application decisions of individual farmers. However, the application of agricultural mechanization by other farmers in the same village did not have a direct impact on the cultivated land sustainability protection of the interviewed farmers. Theoretically, the instrumental variables selected in this study met the requirements of correlation and exogeneity and are valid. In terms of model selection, considering that the explained variable of cultivated land sustainability protection is an ordered categorical variable, in this study, we referred to the conditional mixed process (CMP) method proposed by Roodman [34] to construct an IV-Oprobit model for two-stage estimation. The validity test results of the model (

Table 4. Endogeneity test results based on IV-Oprobit model.

Variable	(3) IV-Oprobit
	The First Stage: Agricultural Mechanization	The Second Stage: Cultivated Land Sustainability Protection
Agriculture mechanization		0.9676 *** (0.1228)
Agricultural mechanization at village level	5.7264 *** (0.1283)
Control variables	Controlled	Controlled
LR chi2	2617.74 ***
atanhrho_12	−0.1088 *** (0.0285)
Number of samples	2118

Note: *** represent significance at 1% level.

) show that the two-stage estimation results are significant, the endogeneity test parameter (atanhrho_12) is significantly different from 0, indicating that there is an endogeneity problem, and the two-stage estimation results using the CMP method are more accurate. Specifically, according to the second-stage estimation CMP results, when endogeneity is considered, agricultural mechanization still has a significantly positive impact on the cultivated land sustainability protection behaviors of farmers, which is significant at the level of 1%, verifying the robustness of the conclusion. Compared with the estimated results in Table 2, the regression coefficients of the key explanatory variables of the IV-Oprobit model increase significantly, indicating that, with the test analysis without considering the endogeneity problem, it is easy to underestimate the impact of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers, which further reflects that agricultural mechanization development and application can encourage farmers to strengthen their cultivated land sustainability protection. Thus, H1 is further verified.

4.4. Analysis of Action Mechanism

Agricultural mechanization can significantly promote the cultivated land sustainability protection behaviors of farmers. In order to deeply analyze and verify the agricultural mechanization action mechanism affecting the cultivated land sustainability protection behaviors of farmers, the planting income and fertilizer input were selected as the intermediary variables.

Table 5. Mechanism test results based on Oprobit model and OLS.

Variables	Income Enhancement Effect		Fertilizer Reduction Effect
	(4) OLS Planting Incomes of Farmers	(5) Oprobit Cultivated Land Sustainability Protection	(6) OLS Fertilizer Input	(7) Oprobit Cultivated land Sustainability Protection
Agriculture mechanization	1.0048 *** (0.2098)	0.5458 *** (0.0786)
Planting income		0.0205 ** (0.0081)
Mechanization in fertilization process			−3.2700 * (1.6970)	0.2385 *** (0.0572)
Fertilizer input				−0.0014 * (0.0007)
Control variables	Controlled	Controlled	Controlled	Controlled
F	34.93	—	2.25	—
LR chi2	—	185.57	—	146.63
Pseudo-R2	—	0.0253	—	0.0200
Number of samples	2118	2118	2118	2118

Note: *, **, and *** represent significance at 10%, 5%, and 1% levels, respectively.

presents the OLS regression results of Model (4) of the agricultural mechanization effects on the agricultural production income. Li et al. [23] also found that agricultural mechanization can significantly promote the growth of the agricultural operating income against the background of a reduced agricultural labor force. Model (5) is an Oprobit model regression analysis that includes both agricultural mechanization and the farming household planting income. The results show that the agricultural mechanization and farming household planting income coefficients are significantly positive, but the agricultural mechanization coefficient is smaller than the estimated results in Model (1) of Table 2, indicating that the planting incomes of farmers play a partial intermediary role in the impact of agricultural mechanization on their cultivated land sustainability protection.

At the same time, we assessed the mediating effect of chemical fertilizer reduction in agricultural mechanization. Table 5 presents the OLS regression results of Model (6) of agricultural mechanization regarding the chemical fertilizer input. Zhu Jianjun et al. [25] also found that mechanized fertilization can avoid uneven and non-standard artificial fertilization, effectively improve the chemical fertilizer application efficiency, and promote reductions in chemical fertilizer application. Model (7) is the regression analysis of the Oprobit model with both the agricultural mechanization and chemical fertilizer inputs. The results show that agricultural mechanization positively promotes the cultivated land sustainability protection behaviors of farmers at the significance level of 1%, while the chemical fertilizer input significantly negatively affects them. Moreover, the agricultural mechanization coefficient in the fertilizer application link in Model (7) is also smaller than the estimated result in Table 3 (0.2429), indicating that the input of chemical fertilizer plays a partial intermediary role in the impact of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers.

4.5. Heterogeneous Effect Analysis

In order to gain a deeper understanding of the impact of agricultural mechanization on the farmland sustainability protection behaviors of farmers, we further considered the problem of the heterogeneous effects of farmers in different management states and farmland conditions. According to the previous theoretical analysis, whether a farmer works full-time or part-time will have an impact on their agricultural production input decision. Therefore, the impact of agricultural mechanization on farmers with different agricultural production characteristics may be different. Moreover, different land fragmentation levels will have different effects on the cost and convenience of agricultural mechanization operations. Therefore, there may be differences in the agricultural mechanization effects on the cultivated land sustainability protection behaviors of farmers under different land fragmentation levels. Accordingly, we focused on two heterogeneity tests: one based on the different agricultural production statuses of farmers, and the other based on different land fragmentation levels.

4.5.1. Heterogeneous Effect of Full-Time Agricultural Production

To investigate whether the impact of agricultural mechanization on farmland sustainability protection is related to the full-time agricultural production of farmers, the sample was divided into two groups based on their full-time or part-time agricultural production statuses. We used the Oprobit model for the grouped-sample regression, and the results of this heterogeneity test are shown in

Table 6. Regression results of heterogeneity analysis.

Variables	Full Time or Not		Land Fragmentation Level
	(1) Yes	(2) No	(3) Low	(4) High
Agriculture mechanization	0.6757 *** (0.0996)	0.4422 *** (0.1303)	0.7253 *** (0.0937)	0.2111 (0.1459)
Control variables	Controlled	Controlled	Controlled	Controlled
LR chi2	122.46	64.99	153.25	47.00
Pseudo-R2	0.0262	0.0245	0.0291	0.0227
Number of samples	1351	767	1521	597

Note:*** represent significance at 1% level.

. Columns (1) and (2) show that the agricultural mechanization coefficients are significantly positive, which suggests that whether farmers work full-time or part-time, agricultural mechanization can promote farmland sustainability protection.

In addition, the coefficient (0.6757) in Column (1) is greater than that (0.4422) in Column (2), indicating that for full-time farmers, agricultural mechanization has a greater positive impact on their cultivated land sustainability protection, verifying Hypothesis H4. In other words, full-time agricultural production will strengthen the promotion effect of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers.

4.5.2. Heterogeneity Tests Based on Land Fragmentation Level

In order to investigate whether the impact of agricultural mechanization on cultivated land sustainability protection is related to the degree of land fragmentation, we used the number of plots owned by farmers to measure the land fragmentation degree. Specifically, the degree of land fragmentation was grouped according to the mean value of the number of parcels, and the group below the mean value (four parcels) was defined as the low-level land fragmentation group, while the group above the mean value was defined as the high-level group. We used the Oprobit model to conduct the sub-sample regression, and Table 6 reports the results of this heterogeneity test. Columns (3)–(4) show that the agricultural mechanization coefficients are all positive; however, at low land fragmentation, the agricultural mechanization impact on the cultivated land sustainability protection behaviors of farmers passes the significance test at the level of 1%, indicating that the agricultural mechanization impact on the cultivated land sustainability protection behaviors of farmers is different at different land fragmentation degrees, verifying Hypothesis H4. In other words, a higher degree of land concentration will strengthen the agricultural mechanization promotion effect on the cultivated land sustainability protection behaviors of farmers.

5. Discussion

The promotion effect of agricultural mechanization on cultivated land sustainability protection is significant, and we calculated it via the Oprobit model and IV-Oprobit model based on the survey data of 2118 farmers in 10 provinces of China, verifying the research hypotheses. Agricultural mechanization not only improves agricultural production efficiency, saves labor costs, and improves agricultural outputs, but also encourages regional farmers to strengthen their cultivated land sustainability protection, which is an important supplement to the evaluation of the effect of applying agricultural machinery.

There is a positive relationship between agricultural mechanization and cultivated land sustainability protection. This is similar to the findings of Xu Zhigang et al. [16] and Du Meiling et al. [17], which show that agricultural mechanization helps to promote high-quality grain production and improve agricultural eco-efficiency. However, high-quality agricultural development and agricultural ecological construction are relatively broad concepts, and there are still great deficiencies in the analysis of the impact of agricultural mechanization from the perspective of improving cultivated land sustainability. Compared with previous studies, the innovation of this study lies in the following three aspects: First, from the four dimensions of “soil improvement, fertility enhancement, water and fertilizer conservation, and pollution control and remediation”, a quantitative evaluation system was constructed to measure the cultivated land sustainability protection levels of farmers, which provides a new perspective for theoretical research on high-quality agricultural development. Second, based on the perspective of the agricultural production sub-links, we measured the agricultural mechanization levels in multiple links and tested the heterogeneous mechanical operation effects in them on the cultivated land sustainability protection behaviors of farmers to more comprehensively evaluate the micro-impact of agricultural mechanization. Third, the adaptability of agricultural machinery application was fully considered, and the differences in the effects of agricultural mechanization on the farmland sustainability protection behaviors of farmers under different cultivated land conditions and concurrent employment statuses were tested to more precisely and reasonably examine the relationship between them.

Farmers are the main body of agricultural production, and their production behaviors are crucial to land resource management and protection [32]. As revealed by the research results in this paper, the development and application of agricultural mechanization can provide farmers with more time and economic resources to pay attention to the sustainability protection of their cultivated land, motivating farmers to adopt higher levels of cultivated land sustainability protection behaviors by increasing their agricultural incomes. For example, agricultural mechanization improves the yields and quality of agricultural products, improves the market competitiveness of agricultural products, and enables farmers to obtain higher agricultural incomes, thereby providing economic motivation for farmers to actively participate in the quality management of their cultivated land. Moreover, the development and application of agricultural mechanization makes farmland management more refined and standardized, which can help farmers achieve more efficient farming and cultivated land sustainability protection. For example, mechanized fertilization can achieve the accurate application of chemical fertilizer and improve the utilization rate, reduce the excessive application of chemical fertilizer and the risk of soil pollution, and help farmers better achieve cultivated land sustainability protection. In particular, for farmers engaged in full-time agricultural production or with high land concentrations, the protection and promotion roles of cultivated land sustainability played by agricultural mechanization are more obvious.

Although the development and application of agricultural mechanization play an important role in promoting cultivated land sustainability protection, there are also some potential challenges. First, for the use of some heavy machinery, the air permeability and water seepage properties of the soil may be affected, resulting in soil structure changes and soil quality deterioration. Second, the waste generated during mechanization operations may have a certain negative impact on the ecological environment of cultivated land. Therefore, while promoting agricultural mechanization, government management departments and farmers should enhance the awareness of cultivated land sustainability and environmental protection, reasonably choose the mode and equipment of their mechanization operations, and ensure the coordinated development of cultivated land sustainability protection and agricultural mechanization.

6. Conclusions and Suggestions

6.1. Conclusions

Cultivated land sustainability is the foundation of sustainable development and plays a crucial role in ensuring food security, preserving the ecological environment, and improving the incomes of farmers. Mechanization, as a high-efficiency input factor in modern agriculture, has gradually become an important tool and means to promote the protection of cultivated land sustainability. In this study, based on the theoretical analysis and field survey data of 2118 farmers in 10 provinces in China, we systematically examined the impact of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers and its mechanism, and we further assessed the differences in the impact under different production links, different employment statuses of the farmers, and different land fragmentation levels. The main conclusions are as follows:

(1) Agricultural mechanization has a significantly positive impact on the cultivated land sustainability protection behaviors of farmers, and for every unit increase in the agricultural machinery application level, the probabilities of unprotected and low-level-protected cultivated land sustainability decrease by 2.68% and 12.47%, respectively, while the probability of adopting high-level protection increases by 15.15%. This shows that the development and application of agricultural mechanization will encourage farmers not only to strengthen their cultivated land sustainability protection but also to achieve higher levels of cultivated land sustainability protection behaviors.

(2) The analysis of the influence of mechanization in the different links shows that the promotion effect of mechanization is the most obvious in harvesting, followed by in pesticide spraying and fertilizer application, while it is the least obvious in plowing and sowing.

(3) The planting incomes and chemical fertilizer inputs of farmers play partial intermediary roles in the impact of agricultural mechanization on their cultivated land sustainability protection.

(4) The impacts of agricultural mechanization on the cultivated land sustainability protection behaviors of farmers are significantly different due to their different management statuses and land fragmentation levels. Agricultural mechanization has a greater impact on the cultivated land sustainability protection behaviors of full-time farmers than part-time farmers. However, land fragmentation limits the effect of agricultural mechanization; that is, a higher degree of land concentration can strengthen the agricultural mechanization promotion effect on the cultivated land sustainability protection behaviors of farmers.

6.2. Recommendations

(1) In view of the cultivated land sustainability protection and promotion effect of agricultural mechanization, the development and supply of agricultural mechanization should be further accelerated. While continuously improving the supply “quantity” of agricultural machinery operation, it is also necessary to pay attention to the role of agricultural mechanization in promoting the cultivated land sustainability protection by farmers and to the improvement in the “quality” of the mechanical operation supply. For example, the government should further promote and improve the agricultural machinery operation subsidy policy, increase the subsidy for the green agricultural machinery “R&D—production—purchase—operation”, and improve the ecological protection promotion effect of agricultural machinery operation.

(2) In view of the important intermediary role played by the planting income and fertilizer input, the income-increasing and fertilizer-reducing effects of agricultural mechanization should be further improved. Guided by improving the efficiency of the mechanical operation output and fertilizer application, we should further strengthen the joint research between machinery manufacturing enterprises and scientific research institutions, as well as promote the green technology innovation of agricultural machinery.

(3) In view of the differences in the impact of agricultural mechanization in the different production links, operation characteristics, and cultivated land conditions, precise and differentiated safeguard measures should be formulated. First, the publicity of cultivated land sustainability protection policies should be increased, and farmers should be encouraged to increase their rates of “returning straw to the field” and “pesticide packaging recycling” mechanical operations in the harvesting link. Second, policy support for full-time farmers for machinery use should be strengthened to promote the standardization, development, and quality of agricultural production. Third, the replacement and integration of planting plots or contiguous circulation, reductions in the degree of land fragmentation, and the creation of conditions for giving full play to the ecological protection and promotion benefits of agricultural mechanization should be encouraged.