A Study on the Utilization Rate and Influencing Factors of Small Agricultural Machinery: Evidence from 10 Hilly and Mountainous Provinces in China

Abstract

Hilly and mountainous areas are weak places for the development of agricultural mechanization in China. The way to improve the utilization rate of small agricultural machinery widely used in hilly and mountainous areas is of positive significance for optimizing resource allocation efficiency of agricultural production and ensuring food security supply. Taking microtillers as a representative tool, this study systematically analyzed the main factors affecting the utilization rate of small agricultural machines and its influencing mechanism. Then, based on the survey data of 4905 farmers in 100 counties in 10 hilly and mountainous provinces of China, empirical analysis was carried out by some econometric models, such as censored regression and the mediating effect model. Results show the following.: (1) Among farmers in hilly and mountainous areas, the average use time of each microtiller is 218.41 h per year. (2) Age, social identity, terrain conditions, crop types, land area, the number of microtillers, the number of large tractors, and the machinery purchase subsidy policy are the significant factors affecting the utilization rate of microtillers. (3) The increase of cultivated land area not only directly improves the utilization rate of microtillers, but also indirectly improves the utilization rate of microtillers due to the increase in quantity.

1. Introduction

Without agricultural modernization, national modernization is incomplete, incomprehensive, and unstable [1]. Agricultural modernization is also an important part of rural revitalization [2,3]. Agricultural machinery is a significant new factor that can break the low production cycle of traditional agriculture [4], and agricultural mechanization is the only way for agricultural capital to further realize agricultural modernization [5]. Promoting the level of agricultural mechanization is conducive to strengthening the production capacity of agricultural products [6], improving land use and management [7], ensuring food security supply [8,9], and developing the growth of green efficiency and agricultural innovation sustainability [10]. In developing countries in particular, the use of agricultural machinery can help solve production difficulties such as the shortage of fertilizer, labor, and water resources [11]. Since 2004, the agricultural mechanization in China has increased rapidly. However, the problems of low utilization of agricultural machinery and the low quality of agricultural mechanization development have come into focus. During the process of agricultural production, the phenomenon of idle agricultural machines in some places and the shortage of agricultural machines in other places exist at the same time [12]. At present, the agricultural mechanization rate in hilly and mountainous areas, where the cultivated land accounts for one-third of China’s cultivated land, is only about 50%, which is more than 30 percentage points lower than that in plain areas [13]. Due to the complex terrain and weak economic level of cultivated land in hilly and mountainous areas, small agricultural machines such as microtillers are widely popular because of their lightness, small size, and low price. As a representative of small agricultural machines, the number of microtillers in China reached 7.9667 million in 2020 [14]. Therefore, the research on the influencing factors of the utilization rate of microtillers widely used in hilly and mountainous areas cannot only help to improve the utilization rate of microtillers and promote the development of agricultural mechanization in hilly and mountainous areas, but also play a reference role in improving the idle situation of agricultural machinery in other regions or other types.

The existing studies on the utilization of agricultural machinery mainly focus on the following two aspects. One is to study how to dispatch agricultural machinery in a single farm from the perspective of operational research, so as to minimize the cost of agricultural machinery in use [15,16,17,18]. For example, Ismail and Ismail (2014) embedded the linear programming model into the software and developed a software system that can select the combination of wheat production machinery with minimized cost according to the farm area and the available time of each operation [19]. Najafi and Dastgerduei (2015) used mixed integer and linear programming methods to optimize the use of agricultural machinery in 80 farms in southern Iran, and found that there is room for optimization in the utilization of agricultural machinery [20]. By evaluating the mechanization status and benefit-to-cost ratio of rice cultivation in Malaysia, Muazu et al. (2015) drew a conclusion that spraying operations have the highest agricultural machinery utilization [21]. Cogato et al. (2020) focused on analyzing the land status and mechanical accessibility of Italian vineyards, aiming at improving machinery utilization and optimizing land planning in vineyards [22]. The other is to study the impact of specific factors on the utilization of agricultural machinery [23,24,25,26,27]. For example, Luo and Zhang (2016) incorporated time, space, weather, and road factors to target an agricultural machinery association in Anhui Province, China, so as to maximize its utilization rate of agricultural machinery [28]. Fischer et al. (2018) discussed the relationship between gender and the use of feed-cutting machines and found that some women rely on the male labor force due to low education and skill level, which reduces the utilization efficiency of agricultural machinery [29]. Based on the investigation of rice farms in Fukui Prefecture, Japan, Yagi and Hayashi (2021) combined the types of organization with machinery utilization [8]. Their findings indicated that compared with small-scale farms, large-scale farms have higher agricultural machinery utilization. Similar research findings have been shown in other studies [30,31].

Compared with the Occident, the agricultural operation scale of farmers in China is relatively small, and the agricultural machinery operation focuses on socialized services [12]. However, the above studies were all about the situation of a single farm or a specific area, and their conclusions were not fully applicable to the hilly and mountainous areas in China. Specifically, there were few existing research studies that systematically analyzed the utilization rate of a specific type of agricultural machinery and its influencing factors, and even fewer studies focused on the utilization rate in China’s hilly and mountainous areas. Based on the above analysis, this study raises the following questions: (1) What is the utilization rate of microtillers for farmers in hilly and mountainous areas of China? (2) What are the factors affecting the utilization rate of microtillers? How do these factors affect utilization rate? The aims of this study are: (1) to discuss the utilization rate of microtillers for farmers in hilly and mountainous areas of China; (2) to analyze the factors and the influence degree that affect the utilization rate of microtillers for these farmers; and (3) to provide some implications to promote the use of agricultural machinery in hilly and mountainous areas of China. In view of this, first, based on the mechanism analysis, this study took microtillers as the research object, using the survey data involving 10 hilly and mountainous provinces, 100 hilly and mountainous counties and 4905 farmers in China. Secondly, the empirical study on the utilization rate of microtillers and its influencing factors was carried out, including four aspects: the characteristics of individual and family, the characteristics of agricultural machinery equipment and agricultural machinery socialized services, the characteristics of agricultural production and operation, and the characteristics of financial support. Finally, according to the research results, this study put forward suggestions and implications to improve the utilization rate of agricultural machinery in hilly and mountainous areas of China.

The research is organized as follows. Section 2 puts forward the mechanism analysis and research hypotheses. Section 3 introduces the research methods. Section 4 shows the data sources, followed by descriptive analysis and single factor cross analysis. Section 5 conducts the empirical research by using econometric models such as censored regression and mediating effect model and discusses the research results. The final section summarizes the main findings, limitations, and suggestions.

2. Mechanism Analysis and Research Hypotheses

2.1. Influence of the Characteristics of Individual and Family on the Utilization Rate of Microtillers

According to the human capital theory, effective training and management, and giving full play to people’s subjective initiative can promote the achievement of goals [32]. Therefore, research on agricultural machinery operations needs to consider human factors [33,34]. For example, young men engage in complex and heavy agricultural machinery operations, whereas women and the elderly are more inclined to engage in simple agricultural activities due to their lesser physical strength and other reasons [33]. In addition, women prefer small agricultural machines that are easy to operate. However, some research findings indicate that the age and education level of farmers have no significant impact on the utilization rate of agricultural machinery [8]. By improving the operation ability of agricultural machinery, skill training can significantly improve farmers’ willingness to use agricultural machines and the efficiency of their operation [23,35]. In order to maximize the return on the input of production factors, when the number of household agricultural labor force increases, farmers tend to expand the scale of agricultural operations, which may increase the demand for agricultural machines. However, there is little research on the relationship between the number of family farming laborers and the utilization rate of agricultural machinery. Therefore, its impact on the utilization rate of agricultural machinery is uncertain. In addition, from the perspective of social capital, having social identity is conducive to strengthening interpersonal relationships [36]. Farmers used to be enterprise managers. On the one hand, this shows that they have a wide range of contacts and obtain more external information about the needs of agricultural machines. They can use self-purchased agricultural machines for socialized services, so as to improve the utilization rate of agricultural machinery. On the other hand, this indicates that they have strong management ability and are good at operation, which can make full use of the self-purchased agricultural machinery.

Based on this, this research puts forward the following hypothesis from the dimension of the characteristics of individual and family.

H1: The characteristics of individual and family have a significant impact on the utilization rate of microtillers (gender, age, education, skill training, number of family farming laborers, and social identity).

2.2. Influence of the Characteristics of Agricultural Machinery Equipment and Agricultural Machinery Socialized Services on the Utilization Rate of Microtillers

As a physical asset, agricultural machinery is the equipment used by farmers to produce agricultural products and replace labor. A major feature of physical assets is natural loss, that is, the service life of agricultural machinery is limited. As the years of service increases, the performance of agricultural machinery decreases [37]. The growth in the number of microtillers increases the utilization rate of microtillers. This is because when there is only one agricultural machine, it breaks down during the busy period, which has an impact on agricultural production. Farmers need to buy socialized services to solve the problem of insufficient machines, and then the utilization rate of these broken machines drop. However, if a large number of agricultural machines are purchased, even if a few machines fail, a sufficient number of self-purchased agricultural machines can still complete agricultural production. Considering the fact that the use of physical capital can increase other capital, when farmers have multiple types of agricultural machinery, based on the characteristics of cost minimization, it is necessary to consider the combination efficiency and arrange agricultural machinery resources rationally [38]. Therefore, when farmers use other types of power machines for operations, this may have an impact on the utilization rate of microtillers. In addition, from the perspective of utility maximization, in order to give full play to the role of agricultural machinery and obtain more benefits, some farmers provide agricultural machinery socialized services to others, thereby improving the utilization rate of agricultural machinery [35,39].

Based on this, this research puts forward the following hypothesis from the dimension of the characteristics of agricultural machinery equipment and agricultural machinery socialized services.

H2: The characteristics of agricultural machinery equipment and agricultural machinery socialized services have a significant impact on the utilization rate of microtillers (years of service, number of microtillers, number of small tractors, number of large tractors and socialized services).

2.3. Influence of the Characteristics of Agricultural Production and Operation on the Utilization Rate of Microtillers

Natural capital refers to the stock of natural resources and environmental services conducive to livelihoods [40]. The natural capital of farmers mainly comes from land. The lower the natural capital, the smaller the geographical advantage of farmers [41]. At the same time, the slope of cultivated land also limits the utilization rate of agricultural machinery [13,42,43,44]. The distance from the residence to the town may have a positive impact on the utilization rate of agricultural machinery [41]. This may be because the farther away from the township, the lower the opportunity cost of farmers and the more land resources, so they are willing to use microtillers for agricultural production.

In addition, the expansion of production scale and effective land concentration produce economies of scale. The scale of operation improves the utilization rate of fixed assets, reduces farmers’ agricultural production costs, and promotes agricultural machinery utilization [6,31,45]. Cultivated land area is the embodiment of natural capital value [41]. With the expansion of cultivated land area, on the one hand, agricultural machines purchased and used by farmers themselves are useful, and then the working area of a single agricultural machine increases. On the other hand, a larger area requires more agricultural machines. According to H2, the increase in the number of agricultural machines may also improve the utilization rate of agricultural machinery. In other words, the increase of operation scale not only brings the direct effect of improving the utilization rate of agricultural machinery, but also brings the indirect effect of increasing the utilization rate of agricultural machinery by the growth of the number of agricultural machines.

Based on this, this research we put forward the following hypotheses from the dimension of agricultural production and operation characteristics.

H3: The characteristics of agricultural production and operation have a significant impact on the utilization rate of microtillers (terrain, crop types, distance from residence to town, and cultivated land area).

H4: The cultivated land area not only has a positive impact on the utilization rate of microtillers, but also has an indirect effect on the utilization rate of microtillers through the mediator of the number of microtillers.

2.4. Influence of the Characteristics of Financial Support on the Utilization Rate of Microtillers

The main beneficiaries of the agricultural machinery subsidy policy are large farmers, production cooperatives, and other production and operation organizations, some of which are used by themselves, while others provide socialized services [44,46]. The incentive theory shows that effective incentives mobilize the enthusiasm of farmers. Agricultural subsidy policy is an incentive for the government to promote agricultural development and increase welfare in the agricultural field [47], but it may also cause low resource allocation efficiency and economic benefits [8]. At the same time, the implementation effect of this policy is sustained because agricultural machinery has the characteristics of durable goods, that is, the number of machines may be increased in the year when the policy is launched, but the increase in machines in that year can be used for many years [48]. Agricultural machinery purchase subsidy policy can increase farmers’ demand for large agricultural machinery [49], whereas small farmers are not the main body of the purchase and use of agricultural machinery due to their small scale and limited purchasing power. In addition, driven by utility maximization, agricultural machinery insurance is conducive to farmers’ risk transfer and loss compensation. Compared with low-power agricultural machinery, farmers are more willing to buy high-power agricultural machinery insurance [50]. However, the impact of farmers’ purchase of agricultural machinery insurance on the utilization rate of agricultural machinery has not yet been conclusive.

Based on this, this research puts forward the following hypothesis from the dimension of the characteristics of financial support.

H5: The characteristics of financial support have a significant impact on the utilization rate of microtillers (agricultural machinery purchase subsidy policy and agricultural machinery insurance).

Based on the above analysis, this research constructs the mechanism analysis framework (Figure 1), and believes that the characteristics of individual and family, the characteristics of agricultural machinery equipment and agricultural machinery socialized services, the characteristics of agricultural production and operation and the characteristics of financial support have a significant impact on the utilization rate of microtillers. In addition, the scale of cultivated land may affect the utilization rate of microtillers through the increase in the number of microtillers.

3. Methods

According to the above theoretical analysis, this study first briefly analyzes the relationship between the influencing factors of each dimension and the utilization rate of agricultural machinery through the cross-analysis, and then constructs the influencing factors model and the mediating model of the utilization rate of microtillers to verify the research hypotheses. The research methods are as follows.

3.1. Cross-Analysis

In order to find the characteristics of farmers with a high utilization rate, the utilization rate of microtillers of monitored farmers (we use the hours used by each microtiller every year to represent the utilization rate [8,51]) and all variables are grouped and cross analyzed. Machinery utilization is often referred to as the actual operating time compared to the maximum total operating time in the annual operating window. We assume that the maximum working time of the samples is the same, so only the actual use time of the microtillers is discussed. To be specific, first we calculate the average utilization rate of microtillers of farmers in different groups of each variable. Then, divide the calculation result by the average utilization rate of microtillers of all farmers with microtillers. Finally, the deviation coefficient of the relative average value of each group of variables is obtained. The formula is as follows:

$$B_j=\frac{{\displaystyle {\sum }_{i=1}^n{C}_{ji}}}{n}÷U$$

(1)

where Bj is the deviation coefficient of the utilization rate of microtillers of a certain variable in group j, and n is the number of samples in this group.$C_{ji}$ is the utilization rate of the farmer i in group j, ${\displaystyle {\sum }_{i=1}^n{C}_{ji}}$ is the total utilization rate of the farmers in group j, and the value of i ranges from 1 to n; U is the average utilization rate of all farmers with microtillers. In order to facilitate intuitive analysis, it is set here that when the deviation coefficient is between 0.8 and 1.2, the difference between the average value of this group and the average value of all samples is considered to be insignificant. When the deviation coefficient is greater than 1.2 or less than 0.8, it is considered that the average value of this group is significantly higher or lower than that of all samples.

For a dataset, if the standard error is σ, the mean value is X, so 68.27% of the data is distributed in the X±σ range and 95% of the data is distributed in the X±1.98σ range. If σ = 0.1X (that is, the coefficient of variation is 0.1), then the data distributed between 0.8X and 1.2X (X±2σ) is greater than 95%. If data is outside this range, it deviates significantly from the mean value X. If σ = 0.3X (the coefficient of variation is large), then the data distributed between 0.8X and 1.2X (X±0.67σ) is still 49.72%, and the deviation degree between samples not in this range and the mean is large. Therefore, in order to quickly find possible significant differences, this paper uses an interval of 0.8–1.2 times before the formal econometric regression to preliminarily determine the significance of differences in the utilization rate of microtillers among samples in different groups.

3.2. Censored Regression

According to the data characteristics, we adopt the censored regression model. Generally, the Tobit model is selected for analysis by using MLE estimation [52]. The model is carried out under the condition that the dependent variable is continuous but subject to certain restrictions and is mostly used for regression analysis where the dependent variable has a zero value and other values are positive and continuous. The latent variable y_i* needs to be used when estimating the result, and y_i* satisfies the basic assumption of the linear model. The initial establishment of the Tobit model is

$$\begin{array}{l}{y}_i{}^{*}=\beta x_i+{\epsilon }_i\\ y_i=max(0,y_i{}^{*})=\left\{\begin{array}{l}{y}_i{}^{*},if y_i{}^{*}\ge 0\\ 0,if y_i{}^{*}<0\end{array}\right.\end{array}$$

(2)

where y_i is the average hours (hours/year) used by the farmer i per year for each microtiller, x_i is the influencing factor of the utilization rate of microtillers, β is the unknown parameter in the equation, and ε_i is the random error term and obeys N (0, δ²).

However, due to the strong dependence on distribution, the Tobit model is not robust enough. If the likelihood function is incorrect, for example, there is a problem in that the disturbance term does not obey a normal distribution or has heteroscedasticity, the MLE estimation results become inconsistent, and a more robust censored least absolute deviation (CLAD) method can be used [53]. The CLAD method only requires the disturbance term to be independent and identically distributed and can obtain consistent estimates even in the case of nonnormality and heteroscedasticity. Moreover, under certain regularity conditions, the estimator obeys an asymptotically normal distribution. The censored data regression model is as follows:

$$y_i=\mathit{max}(0,\beta x_i+{\epsilon }_i).$$

(3)

The objective function of CLAD method is the sum of absolute deviation,

$$\underset{\beta }{\mathit{min}}{\displaystyle {\sum }_{i=1}^n|y_i-\mathit{max}(0,\beta x_i)}|$$

(4)

where if $\beta x_i+{\epsilon }_i\ge 0|$, then $y_i=\beta x_i+{\epsilon }_i$; Conversely, $y_i=0$.

3.3. Mediating Effect Model

In order to verify the hypothesis (H4) that cultivated land scale affects the utilization rate of microtillers through the mediating effect of the number of microtillers, we use the stepwise method of Baron and Kenny [54] to analyze the mediating effect. Next, because the Sobel method needs to be assumed to obey the normal distribution; in view of this, we choose the Bootstrap method, which is generally considered to be better, to directly test the coefficient product [55]. We construct the following mediating effect model,

$$Y\prime =cScale+e_1$$

(5)

$$M\prime =aScale+e_2$$

(6)

$$Y\prime =c\prime Scale+bM\prime +e_3,$$

(7)

where M’ is the number of microtillers, and Scale is the cultivated land area. The mediating effect test process is as follows. The first step is to test the total effect of Scale on Y’, that is, to test the significance of the coefficient c. The second step is to test the significance of the coefficient product, which is done indirectly by testing the coefficients a and b in turn. If both a and b are significant, go to the fourth step, and if at least one of them is not significant, go to the third step. In the third step, use the Bootstrap method to test ab directly. If ab is significant, continue to the fourth step, and if ab is not significant, stop. The fourth step is used to distinguish complete mediation or partial mediation. If c’ is significant, it is a partial mediation, and if c’ is not significant, it is a complete mediation.

4. Data Sources and Descriptive Analysis

4.1. Data Sources

The data used in this study comes from the sampling questionnaire survey conducted by the Department of Agricultural Mechanization of the Ministry of Agriculture and Rural Affairs of the People’s Republic of China from May to July 2020 on farmers in 10 provinces (Yunnan, Guizhou, Sichuan, Chongqing, Hunan, Hubei, Guangxi, Fujian, Shaanxi, and Shanxi). Figure 2 shows the survey location. Statisticians, who worked in county-level agricultural departments, helped farmers answer questions through the specially designed mobile application. In order to ensure the reliability of the data, before the formal survey, the questionnaire was preinvestigated and revised, and the statisticians in county-level agricultural departments were trained. During the formal survey, the questionnaire was processed, including the preliminary review by the statisticians of the county-level agricultural departments, the final review by the project team members of the Department of Agricultural Mechanization of the Ministry of Agriculture and Rural Affairs, the rejection and refilling of the problems found, the subsequent telephone sampling return visit, the cancellation of the unqualified questionnaire, and so on. A total of 5030 questionnaires were collected, of which 4905 were valid, and the effective rate was 97.51%. Among 4905 samples, 2467 monitored farmers had microtillers, with a total of 5797 microtillers.

4.2. Basic Information of Data

In this study, the utilization rate of microtiller is defined as the working hours per microtiller per year. According to the previous theory and research, the factors affecting the utilization rate of microtillers include the characteristics of individual and family (gender, age, education, skill training, number of family farming laborers, and social identity), the characteristics of agricultural machinery equipment and agricultural machinery socialized services (years of service, number of microtillers, number of small tractors, number of large tractors, and socialized services), the characteristics of agricultural production and operation (terrain, crop types, distance from residence to town, and cultivated land area), and the characteristics of financial support (agricultural machinery purchase subsidy policy and agricultural machinery insurance).Then, take them as independent variables. See

Table 1. Summary of variables.

Type of Variables	Name of Variables	Mean	Std. Dev.	Description of Variables
Dependent variable	Utilization rate of microtillers	218.41	270.74	Service time per microtiller per year (hours)
Independent variables
1. Characteristics of individual and family	Gender	0.89	0.31	Gender of head: Female = 0, Male = 1
	Age	45.97	9.22	Age of head: A ≤ 30 = 1, 30 < A ≤ 40 = 2, 40 < A ≤ 50 = 3, 50 < A ≤ 60 = 4, A > 60 = 5
	Education	3.41	0. 90	Education level of head: Illiteracy = 0, Primary school = 1, Junior high school = 2, Senior high school = 3, College and above = 4
	Skill training	0.89	0.32	No = 0, Yes = 1
	Number of family farming laborers	2.21	0.92	Total number of family members, mainly engaged in agriculture
	Social identity	0.08	0.27	Used to be a member of enterprise managers = 1, Other = 0
2. Characteristics of agricultural machinery equipment and agricultural machinery socialized services	Years of service	4.72	2.25	Average service life of a single microtiller owned by farmers (years)
	Number of microtillers	1.99	1.37	Number of microtillers still in use
	Number of small tractors	0.49	1.61	Number of tractors under 60 horsepower still in use
	Number of large tractors	0.37	1.17	Number of tractors over 60 horsepower still in use
	Socialized services	0.33	0.47	Does the farmer provide agricultural machinery socialized services to others: No = 0, Yes = 1
3. Characteristics of agricultural production and operation	Terrain	2.76	1.06	Shallow hill area = 1, Deep hill area = 2, Shallow mountain area = 3, Deep mountain area = 4
	Crop types	0.49	0.50	Non-grain crops = 0, Grain crops = 1
	Distance from residence to town	1.83	0.80	(km): D ≤ 5 = 1; 5 < D ≤ 10 = 2; 10 < D ≤ 20 = 3; 20 < D ≤ 50 = 4; D > 50 = 5
	Cultivated land area	3.71	2.02	Land area cultivated by farmers themselves(mu): L ≤ 10 = 1, 10 < L ≤ 30 = 2, 30 < L ≤ 50 = 3, 50 < L ≤ 100 = 4, 100 < L ≤ 200 = 5, 200 < L ≤ 500 = 6, 500 < L ≤ 1000 = 7, L > 1000 = 8
4. Characteristics of financial support	Agricultural machinery purchase subsidy policy	0.80	0.40	Does the farmer receive agricultural machinery purchase subsidy: No = 0, Yes = 1
	Agricultural machinery insurance	0.23	0.42	Does the farmer purchase agricultural machinery insurance: No = 0, Yes = 1

for specific variables. A total of 2467 farmers with microtillers have the utilization of 218.41 h per microtiller per year on average.

In terms of provinces, Guizhou has the highest number of microtillers (1061) and utilization rate (278.98 h). The lowest number is in Shanxi, where the number and utilization rate of microtillers are only 172 and 152.98 h, respectively. In addition, 820 farmers with microtillers provide socialized services, accounting for 33.24% of all farmers with microtillers. Figure 3 shows the proportion of self-cultivation area and socialized service area in the total operating area. With the increase of the total operating area, the proportion of the area where farmers provide agricultural machinery socialized services to others has increased, from 47.77% to 77.73%.

4.3. Cross Analysis

To find the characteristics of farmers with high utilization rate, the utilization rate of microtillers and all variables were analyzed by grouping and cross-analysis. These results are presented in Appendix

Table A1. Results for cross-analysis.

Type of Variables	Name of Variables	Description of Variables	Number of Samples	Utilization Rate of Microtillers (Hours)
	Average utilization rate per microtiller per year for farmers with microtillers		2467	218.41
Characteristics of individual and family	Gender	Female	270	221.14
		Male	2197	218.07
	Age	≤30	138	201.57
		30–40	523	201.93
		40–50	1010	243.28
		50–60	678	202.69
		>60	118	188.59
	Education	Illiteracy	19	324.32
		Primary school	303	191.11
		Junior high school	1130	209.41
		Senior high school	674	226.96
		College and above	341	249.68
	Skill training	No	277	193.48
		Yes	2190	221.56
	Number of family farming laborers	≤2	1842	214.42
		3	365	225.27
		4	216	235.31
		>4	44	245.25
	Social identity	Other	2271	211.60
		Enterprise manager	196	297.22
Characteristics of agricultural machinery and socialized services	Years of service (years)	≤5	1366	224.99
		5–8	864	211.83
		>8	237	204.42
	Number of microtillers	≤3	2163	208.34
		>3	304	290.02
	Number of small tractors	≤3	2411	217.56
		>3	56	254.98
	Number of large tractors	≤3	2421	218.99
		>3	46	187.57
	Socialized services	No	1647	217.68
		Yes	820	219.87
Characteristics of Agricultural operation	Terrain	Shallow hill area	451	228.40
		Deep hill area	407	267.08
		Shallow mountain area	890	217.48
		Deep mountain area	719	185.72
	Crop types	Non-grain crops	1270	262.89
		Grain crops	1197	171.21
	Distance from residence to town (km)	≤5	1030	204.11
		5–10	828	220.82
		>10	609	239.30
	Cultivated land area (mu)	≤10	376	120.81
		10–30	546	170.85
		30–50	285	203.98
		50–100	355	235.24
		100–200	339	271.29
		200–500	335	267.60
		500–1000	133	297.93
		>1000	98	379.74
Characteristics of financial support	Agricultural machinery purchase subsidy policy	No	504	259.17
		Yes	1963	207.94
	Agricultural machinery insurance	No	1902	215.58
		Yes	565	227.91

.

4.3.1. Characteristics of Individual and Family

From the gender grouping of the monitored farmers (Figure 4), although the number of women using microtillers is less than that of men, the average microtiller utilization rate of women is slightly higher than that of men. This may be because female farmers prefer to choose simple and easy-to-operate microtillers instead of labor, compared to riding tractors that are complicated to operate and difficult to maintain. From the age grouping of the monitored farmers (Figure 5), both the sample size and the deviation coefficient peak between the ages of 41 and 50. The 41–50-year-old farmers mentioned here and later refer to the monitored farmers with microtillers who are older than or equal to 41 and younger than 50 years old. The reason may be that compared with farmers at a younger age, farmers aged 41–50 have more experience in using microtillers and have a higher utilization rate. However, with the further increase of age, the physical fitness decreases, resulting in the drop of the frequency of using microtillers. Based on this, we use “whether the age group is 41–50 years old or not” as a 0, 1 explanatory variable (new age variable/No = 0, Yes = 1) to replace the previous age variable, and incorporates it into the regression equation later. In terms of the education grouping of the monitored farmers (Figure 6), except for farmers who are not educated, with the rise of education level, the microtiller utilization rate of farmers ascends. From the skill training grouping of the monitored farmers (Figure 7), farmers who have participated in skill training have a higher utilization rate of microtillers than those who have not participated in training. The reason is that training that is related to agricultural machinery can help farmers increase agricultural machinery knowledge and improve agricultural machinery operation skills. Among the family farming laborers in the group of monitored farmers (Figure 8), the utilization rate of microtillers gradually increases with the ascent of the number of family farming laborers. For the social identity grouping of the monitored farmers (Figure 9), although only 8% of the sample now or once have the identity of enterprise management, the utilization rate is significantly higher than the average level, which is 0.36 times higher than the average value. This is because the farmers at the enterprise management level have a strong social network and can obtain more information about the needs of agricultural machinery, and find more service objects who are willing to purchase socialized services. Moreover, they have strong management ability which enables them to arrange and use agricultural machinery more reasonably.

4.3.2. Characteristics of Agricultural Machinery Equipment and Agricultural Machinery Socialized Services

From the average years of service grouping of agricultural machinery (Figure 10), the microtiller utilization rate within five years is the highest, which is 1.03 times as high as the average utilization rate. The reason is that with the increase of service life, there are some problems, such as mechanical wear, air corrosion, aging, and incorrect use and maintenance, which weaken the performance of microtillers and reduce the utilization rate of microtillers. From the number of microtillers owned by the monitored farmers (Figure 11), compared with farmers with three or fewer microtillers, farmers with more than three microtillers have a higher utilization rate of microtillers. This is because the higher the number of microtillers, the more likely it is to have a scale effect. When microtillers are used in production, even if a microtiller is broken, farmers with multiple microtillers do not need to buy additional socialized services, and the use of self-purchased microtillers is enough for agricultural activities. In terms of the number of small tractors (Figure 12), farmers with more than three small tractors have a higher utilization rate of microtillers than those with three or fewer small tractors. On the contrary, from the perspective of the number of large tractors (Figure 13), as the number of large tractors rises, the utilization rate of microtillers for farmers declines. From the socialized services grouping of the monitored farmers (Figure 14), the utilization rate of microtillers for farmers who provide mechanical farming services to others is slightly higher than that for farmers who do not provide services.

4.3.3. Characteristics of Agricultural Production and Operation

According to the land terrain grouping of the monitored farmers (Figure 15), the utilization rate of microtiller in deep hill area is the highest. The reason is that the terrain condition of the deep hill area is suitable for the use of small agricultural machinery for farming. The area of farmland that can be cultivated by machines in the shallow mountain area and the deep mountain area is too small, resulting in a low microtiller utilization rate. Moreover, there are more large plots in the shallow hill area, which mainly use large agricultural machines for farming, resulting in less use of microtillers. Based on this, we take “whether the land terrain is deep hill area or not” as a 0, 1 explanatory variable (new terrain variable/No = 0, Yes = 1), which are included in the regression equation later, to replace the previous terrain variable. From the crop types grouping of the monitored farmers (Figure 16), the utilization rate of microtillers for farmers who mainly focus on nongrain crops is higher than the average level. The main reason is that grain crops generally need large plots suitable for large and medium agricultural machinery, while cash crops and forest and fruit agricultural products use small plots suitable for microtillers. At the same time, cash crops such as vegetables can be planted several times a year, and thus the frequency and efficiency of using microtillers are high. According to the grouping of the distance from the residence to the town (Figure 17), the utilization rate of microtillers generally shows an upward trend. The reason is that the longer the distance, the fewer opportunities for farmers to go to cities and towns for part-time work, and the lower the opportunity cost for them to spend time using microtillers. From the cultivated land area grouping of the monitored farmers (Figure 18), as the cultivated land area increases, the utilization rate is on the rise. In the cultivated land of less than 10 μ and 10-30 μ, their utilization rates are significantly lower than the average, 0.55 times and 0.78 times, respectively. In 100-200 μ, 200-500 μ, 500-1000 μ and more than 1000 μ cultivated land, their utilization rates are significantly higher than the average level, 1.24 times, 1.23 times, 1.36 times and 1.74 times, respectively. The reason is that the expansion of farmland area generates economies of scale, and farmers improve the utilization rate of agricultural machinery in order to meet the needs of large-scale agricultural production.

4.3.4. Characteristics of Financial Support

From the agricultural machinery purchase subsidy policy grouping of the monitored farmers (Figure 19), 79.52% farmers obtain agricultural machinery purchase subsidies, but the utilization rate of farmers who do not have subsidies is higher than that of farmers who have subsidies. It is shown that the utilization rate of microtillers purchased spontaneously by farmers is high, and it is not driven by the subsidy policy. From the agricultural machinery insurance grouping of the monitored farmers (Figure 20), the sample number of farmers who purchase insurance is lower than that of farmers who do not purchase insurance. However, the utilization rate of microtillers of farmers who purchase agricultural machinery insurance is slightly higher than that of farmers who do not purchase insurance.

5. Empirical Results

Before the empirical analysis, the multicollinearity diagnosis was carried out, and it was found that the tolerance values of all variables are less than 5, so there was less possibility of multicollinearity between variables. In addition, based on the Tobit model, a conditional moment test was used to test the normality of disturbance term [56]. As shown in

Table 2. Results for the conditional moment test.

	Critical Values
CM	10%	5%	1%
1634.9	5.38	7.71	10.93

, the conditional moment statistic is 1634.9, which is much higher than the critical values. Therefore, the original assumption that the disturbance term obeys the normal distribution is strongly rejected. The CLAD method, which is more robust, thus, was adopted in this research.

Table 3. Estimation results for OLS, Tobit and CLAD.

Name of Variables		(1)	(2)	(3)
		OLS	Tobit	CLAD
Characteristics of individual and family	Gender	−2.196 (16.87)	−2.156 (16.87)	2.331 (8.271)
	Age	30.19 *** (10.78)	30.80 *** (10.78)	19.98 *** (5.279)
	Education	−14.40 ** (6.425)	−14.42 ** (6.422)	1.654 (3.153)
	Skill training	−15.03 (17.39)	−15.27 (17.39)	10.81 (8.890)
	Number of family farming laborers	4.428 (5.747)	4.412 (5.748)	−1.635 (2.874)
	Social identity	42.72 ** (20.45)	42.68** (20.45)	51.80 *** (9.970)
Characteristics of agricultural machinery equipment and agricultural machinery socialized services	Years of service	−2.025 (2.351)	−2.009 (2.352)	−0.433 (1.129)
	Number of microtillers	8.683 * (4.439)	8.741** (4.441)	7.102 *** (2.170)
	Number of small tractors	−0.394 (3.446)	−0.616 (3.446)	−0.580 (1.110)
	Number of large tractors	−15.19 *** (5.087)	−15.09 *** (5.085)	−4.834 ** (2.200)
	Socialized service	29.82 ** (12.74)	29.34 ** (12.75)	−1.933 (6.252)
Characteristics of agricultural production and operation	Terrain	50.81 *** (14.21)	51.21 *** (14.20)	31.11 *** (6.960)
	Crop types	−65.14 *** (11.44)	−65.94 *** (11.44)	−36.27 *** (5.617)
	Distance from residence to town	14.30 ** (6.581)	14.39 ** (6.578)	11.07 *** (3.251)
	Cultivated land area	26.46 *** (3.169)	26.52 *** (3.170)	20.19 *** (1.550)
Characteristics of financial support	Agricultural machinery purchase subsidy policy	−32.62 ** (13.58)	−31.54 ** (13.59)	−17.81 *** (6.705)
	Agricultural machinery insurance	7.210 (13.67)	7.098 (13.67)	−4.006 (6.647)
	Constant	168.6 *** (38.92)	167.1 *** (38.92)	36.08 * (19.34)
	Observations	2467	2467	2467
	Pseudo R-squared	0.09	0.01	0.07
	P value	0.00	0.00	0.00

Note: * Statistical significance at the 10% level, **at the 5% level, ***at the 1% level.

shows the estimation results of OLS, Tobit, and CLAD. If Tobit satisfies the original assumption, there should be little difference between Tobit and CLAD [57]. However, it shows that the estimation results of Tobit and CLAD are quite different (Table 3), and the coefficient estimates and significance change (Table 2), which indicates that the estimation results of Tobit are biased. Therefore, the CLAD method was finally used for analysis.

5.1. Results for the Influencing Factors of the Utilization Rate of Microtillers Analysis

Farmers aged 41-50 years has a significant positive impact on the utilization rate of microtillers. This may be because, on the one hand, farmers aged 41-50 have more experience in using agricultural machinery than younger farmers, which leads to their higher utilization rate of agricultural machinery; On the other hand, compared with farmers over 50 years old, their physical quality is better, and it is convenient to operate microtillers with certain labor intensity. A similar finding was found in Liu and Liu’s (2022) study in eastern China [17]. Social identity has a significant positive impact on the utilization rate of microtillers. The reason is that, for one thing, farmers at the enterprise management level have high social capital and strong interpersonal relationships, which indicates that they can obtain more information about agricultural machinery needs and have more opportunities for socialized services; For another, they have strong ability in management and operation, which means that they can make full use of self-purchased machines.

The number of microtillers has a significant positive impact on the utilization rate, that is, the utilization rate increases by 7.102 h for each additional 1 unit of microtillers. With the purchase of a large number of machines, even if a few machines are broken, there is no need to buy socialized services, because enough self-purchased agricultural machines can still complete agricultural production. The number of large tractors has a significant negative impact on the utilization rate of microtillers, that is, for every increase in the number of large tractors, the utilization rate of microtillers decreases by 4.834 h. Large tractors are different from small agricultural machines such as microtillers, and there is a competitive relationship between them. Large agricultural machines are applicable to large plots of land. When farmers increase the number of large tractors for agricultural production, the utilization rate of microtillers that are different from the use conditions and functions of large agricultural machines reduces correspondingly.

In terms of land terrain, the deep hill area has a significant positive impact on the utilization rate of microtillers. It is shown that the land conditions in deep hill area are suitable for small agricultural machinery production, and the utilization rate of microtillers is high. For crop types, grain crops have a significant negative impact on the utilization rate of microtillers. This may be because large and medium tractors are generally used in grain production, whereas microtillers are more commonly used in the production of cash crops and forest and fruit agricultural products. The distance from the residence to the town has a significant positive impact on the utilization rate of microtillers. The reason is that the farther away from cities and towns, the fewer opportunities there are for farmers to go to cities and towns for part-time work, the lower the opportunity cost for them to spend time using microtillers, the stronger the investment attribute of microtillers, and the higher the utilization rate; On the contrary, the closer to the town, the more opportunities for farmers to engage in part-time work, and the more time they spend on part-time work, while the time for farmers to use microtillers is reduced. The cultivated land area has a significant positive effect on the utilization rate of microtillers. This is because the expansion of arable land has a scale effect, which makes the agricultural machinery purchased and used by farmers have room to play a role. The working area of a single agricultural machine increases, resulting in an increase in utilization. The study in Japan had a similar finding, that is, the scale effect offsets the low efficiency of the organization and operation of large farms, so as to improve the machinery utilization [8].

In terms of characteristics of financial support, the agricultural machinery purchase subsidy policy has a significant negative impact on the utilization rate of microtillers, that is, compared with the farmers without agricultural machinery purchase subsidies, the farmers who obtain agricultural machinery purchase subsidies have lower utilization rate of microtiller by 17.81 h. The reason is that, on the one hand, approximately 20.4% of the farmers who own microtillers do not receive subsidies for the purchase of agricultural machinery, and the utilization rate of their microtillers is significantly higher than the average, which shows that the utilization rate of farmers who buy spontaneously are high, rather than being driven by the subsidy policy. On the other hand, the objects of subsidy policy for the purchase of agricultural machinery tends to be large machines, and the impact of this policy is sustained, which indicates that its implementation effect takes time [48,58,59].

5.2. Results for the Mediating Effect of Cultivated Land Area and Number of Microtillers

To verify the hypothesis of mediating effect in H4, this research used stepwise method and Bootstrap method for regression.

Table 4. Results for mediating effect test with stepwise method.

	(1) Utilization Rate of Microtillers	(2) Number of Microtillers	(3) Utilization Rate of Microtillers
Cultivated land area	24.44 *** (1.261)	0.242 *** (0.0132)	20.19 *** (1.550)
Number of microtillers			7.102 *** (2.170)
Control variables	Introduced	Introduced	Introduced

Note: *** Statistical significance at the 1% level.

shows the stepwise regression results. The first step (model 1) took the utilization rate of microtillers as the dependent variable and the cultivated land area as the independent variable, the second step (model 2) took the number of microtillers as the dependent variable and the cultivated land area as the independent variable, and the third step (model 3) took the utilization rate of microtillers as the dependent variable and took the cultivated land area and the number of microtillers as the independent variables simultaneously. The regression coefficient of the cultivated land area in the second step and the regression coefficient of the number of microtillers in the third step are both significant, which indicates that the mediating effect is significant. That is to say, farmers with large cultivated land area not only directly improve the utilization rate of microtillers due to the increase of cultivated land area, but also indirectly promote the utilization rate of microtillers by increasing the number of microtillers and then reducing the failure rates. The results of Bootstrap retest are shown in

Table 5. Results for mediating effect test with Bootstrap method.

		Observed Coef.	Bootstrap Std. Err.	Normal-Based [95% Conf. Interval]
Cultivated land area and Number of microtillers	Direct effect	25.56 ***	3.71	18.29	32.83
	Indirect effect	2.76 ***	1.38	0.05	5.47

Note: *** Statistical significance at the 1% level.

. The direct effect and indirect effect pass the significance test at the level of 1%, but the indirect effect is far less than the direct effect as a whole. Existing studies have found that the expansion of cultivated land area promote the improvement of agricultural machinery utilization rate [6,8,31,45]. Our research confirmed this conclusion and further found that the cultivated land area also indirectly affects the utilization rate through the number of microtillers.

6. Conclusions

Based on the survey data of 4905 sample farmers in 100 hilly and mountainous counties in 10 hilly and mountainous provinces, China, this research systematically analyzed the utilization rate and influencing factors of small agricultural machinery represented by microtillers from four aspects: the characteristics of individual and family, the characteristics of agricultural machinery equipment and agricultural machinery socialized services, the characteristics of agricultural production and operation, and the characteristics of financial support. The main findings and suggestions are as follows, so as to have an in-depth understanding of the utilization rate of small agricultural machinery.

First, among farmers with microtillers in hilly and mountainous areas, the average annual use time of each microtiller is 218.41 h. Secondly, in terms of the characteristics of individual and family, farmers aged 41-50 have a higher utilization rate of microtillers than farmers in other age groups, because their experience in using agricultural machinery is better than that of younger farmers and their physical fitness is better than that of older farmers. Returning farmers with experience in enterprise management not only have a wide network of contacts and many opportunities for socialized services, but also are good at management and can make full use of self-purchased agricultural machinery. Thirdly, for the characteristics of agricultural machinery equipment and agricultural machinery socialized services, there is a scale effect in the utilization rate of microtillers, that is, for each additional microtiller, the utilization rate increases by 8.452 h. At the same time, there is a competitive relationship between large tractors and microtillers. In other words, the number of large tractors has a significant negative impact on the utilization rate of microtillers. Fourth, from the characteristics of agricultural production and operation, because deep hill area is not suitable for large agricultural machinery and is more suitable for small agricultural machinery than mountain area, the land condition of deep hill area has a significant positive impact on the utilization rate of farmers’ microtillers. Farmers who grow grain crops rarely use microtillers. The farther the residence is from the town, the higher the utilization rate of microtillers. In addition, the scale of cultivated land not only has a direct positive impact on the utilization rate of microtillers due to the increase of operation opportunities, but also has an indirect positive impact through the number of microtillers. Finally, regarding the characteristics of financial support, it is found that compared with the farmers who do not obtain agricultural machinery purchase subsidies, the utilization rate of microtillers for farmers with subsidies is 17.81 h lower. In other words, the agricultural machinery purchase subsidy policy has a significant negative impact on the utilization rate of microtillers.

It is worth noting that this research also has some limitations. First, the study ignores the land fragmentation factor. The degree of land fragmentation has an inhibitory effect on farmers’ purchase behavior, and the degree of land fragmentation in hilly and mountainous areas is relatively high. Therefore, the land fragmentation may also have an impact on the utilization rate of agricultural machinery. That is to say, the higher the degree of cultivated land fragmentation in hilly and mountainous areas, the lower the operation efficiency of agricultural machinery. Secondly, only the impact of grain crops is investigated. In this study, crop types are divided into grain crops and nongrain crops, and there is a lack of discussion on oil crops, sugar crops, melon and fruit crops and other crops. Thirdly, the study ignores the impact of food security. Agricultural production is an important concern for food security [60]. The improvement of agricultural mechanization is conducive to improving the production capacity of agricultural products and ensuring the safe supply of food [6]. Therefore, we can analyze the relationship between food security and agricultural machinery utilization in the future. Finally, some influencing factors may be omitted when using cross-sectional data, mainly due to the difficulty and workload of data collection. These may be issues for future research.

With regard to suggestions, combined with the research findings, this study puts forward the following points in order to promote the development of agricultural mechanization and the improvement of the agricultural machinery utilization in hilly and mountainous areas in China. First, the government should establish some technical exchange platforms for agricultural machinery operation, especially focusing on the experience exchange of agricultural machinery operators aged 41-50. Exchange platforms, such as training courses and WeChat groups, strengthen the communication between young agricultural operators and agricultural operators aged 41-50, and young agricultural operators can make up for the lack of experience through operation skill training and communication, which can improve the machinery utilization. Secondly, considering the finding that the terrain conditions severely limit the efficiency of agricultural machinery in hilly and mountainous areas, the government should strengthen the suitable-for-mechanization transformation of cultivated land, promote farmland plots to be small to large, curved to straight, steep to flat, and interconnected, improve the access and operation conditions of agricultural machinery, expand the scale and concentration of cultivated land, and improve the adaptability level of agricultural machinery. Thirdly, we can improve the subsidy policy for the purchase of agricultural machinery. In recent years, the Chinese government has promoted the development of agricultural mechanization through agricultural machinery subsidies. However, with the increase in the number of agricultural machines, the promotion effect of agricultural machinery subsidies on the machinery utilization has been restrained. Therefore, it is necessary to adjust the subsidy policy in time according to the actual situation of the region. In other words, the subsidies should be reduced for the agricultural machinery with high ownership, so as to prevent the excessive subsidies from reducing the marginal utilization rate of agricultural machinery for farmers.