Agricultural Machinery Socialization Service Adoption, Risks, and Relative Poverty of Farmers

Abstract

Agricultural machinery socialization services are an important means of alleviating poverty and promoting agricultural modernization. Based on 2750 items of survey data from farmers in Henan Province, this paper empirically tests the impact and mechanism of agricultural machinery socialization service adoption on the relative poverty of farmers by using a binary logit model and mediation effect model. The results show that the adoption of agricultural machinery socialization services has a significant negative impact on the relative poverty of farmers. The reduction in natural risk plays an intermediary role in the impact of the adoption of agricultural machinery socialization services on the relative poverty of farmers. The size of the household labor force and the land operation scale affect the poverty reduction effect of agricultural machinery socialization services. It can be seen that in the relative poverty governance stage, we should improve the development of the agricultural machinery socialization service system, enhance the risk resistance of farmers, build a supporting system for farmers’ employment skills training, and encourage an orderly connection between the large-scale operation of agricultural land and the large-scale operation of services so as to fully realize the poverty reduction role of agricultural machinery socialization services.

1. Introduction

Poverty has always been a global problem that hinders the progress of human society. The United Nations Development Programme (UNDP) works in some 170 countries and territories, in line with the Sustainable Development Goals (SDGs), to help end poverty, reduce inequality and achieve inclusion, make societies more inclusive, and protect the planet. According to the latest Global Multidimensional Poverty Index (MPI) report [1], 25 countries, including China, have succeeded in halving the multidimensional Poverty Index in 15 years, confirming the feasibility of accelerating development. In 2020, China accomplished the task of eliminating absolute poverty [2]. Rural households on the edge of development still face a high level of poverty vulnerability, and are vulnerable to natural disasters, major diseases, and other risks of returning to or falling into poverty [3]. The extensive development of agricultural mechanization can effectively improve the ability of farmers to cope with risks [4,5,6] and prevent farmers from becoming poor due to disasters. However, factor endowments vary considerably between countries, as does the degree of mechanization development. In the United Kingdom, there are more than 6 million hectares of arable land, but less than 2% of the population is engaged in agricultural production. In order to develop modern agriculture, the British government attaches importance to the development of the agricultural machinery industry and agricultural mechanization. Similarly, in the United States, the agricultural land area is large, and the rapid development of industrialization and urbanization has caused a shortage of agricultural labor, so the country mainly relies on the expansion of agricultural mechanization services in order to avoid the abandonment of arable land. Japan is limited by land resource endowment—the process of agricultural production mainly adopts biotechnology to increase output, and mechanization only plays a secondary role [7]. In India, agricultural mechanization has also accelerated the replacement of human and cattle capital, with faster agricultural mechanization and a sharp increase in agricultural labor productivity [8]. In Asian countries such as China and India, small farmers dominate, yet lack the resources to purchase machinery; therefore, the development of agricultural services is particularly important [9]. Agricultural services are central to agricultural income growth, food security, and poverty reduction among farm households in developing countries [10]. Affected by the shortage of labor, the social services provided by agricultural machinery have gradually become a bridge for the organic connection between small farmers and modern agriculture [11]. China’s agricultural machinery socialization service refers to the various types of agricultural machinery operation services provided by organizations and households to other agricultural producers, such as machine ploughing, machine sowing, machine harvesting, irrigation and drainage, and plant protection, as well as the related paid services such as maintenance, supply, intermediation, and the leasing of agricultural machinery. After more than 40 years of development, a Chinese-style socialized agricultural machinery service organization has been formed (Figure 1), which can be divided into five types: machine-owning farmers, agricultural machinery service consortiums, grassroots agricultural machinery service institutions, agricultural machinery stock cooperatives, and agricultural machinery service intermediaries. Of these, grassroots agricultural machinery service institutions and agricultural machinery service intermediaries are organized by the government or agricultural machinery management departments, while the other organizations are formed by farmers on their own initiative [12,13]. By 2020, there were more than 40 million specialized agricultural machinery households in China, about 195,000 operation service organizations such as agricultural machinery cooperatives, nearly 20 million hm² of cross-regional operational areas of agricultural machinery, and 57 million hm² of operation service areas for agricultural machinery cooperatives. The social services associated with agricultural machinery profoundly affect agricultural and rural development. Therefore, exploring the impact of socialized agricultural machinery services on the relative poverty of farmers is of great significance for optimizing the development system of such services, thus helping to achieve poverty alleviation and build a long-term poverty alleviation mechanism.

Existing studies have focused on the impact of agricultural machinery socialization services on farmers’ income, and it is generally believed that these services can significantly promote the growth of farmers’ income [14,15]. On the one hand, with the help of technological improvements, social agricultural machinery services can effectively improve production efficiency, reduce costs [16], and promote an increase in farmers’ agricultural operating income, especially in terms of breeding and fertilization [17,18]. On the other hand, the effective replacement of an agricultural labor force by agricultural machinery socialization services can further release the demographic dividend and accelerate the non-agricultural transfer of the labor force while reducing labor employment costs, thus increasing the wage income of farmers [19,20,21,22]. With the gradual deepening of the research, some scholars began to notice the impact of the adoption of agricultural machinery socialization services on the relative poverty of farmers, pointing out that it can improve agricultural production efficiency, labor allocation efficiency, and factor transaction efficiency, thereby increasing the agricultural operating income and wage income of farmers and alleviating the relative economic poverty of farmers from the perspective of urban and rural comparisons [23]. At the same time, the agricultural machinery socialization service reduces the gap of agricultural operational and wage income by narrowing the endowment difference among farmers, thus alleviating the relative economic poverty of farmers from the perspective of intra-rural comparison [24].

In general, there have been discussions on the relationship between agricultural machinery socialization services and the relative poverty of farmers, which has laid a good foundation for the smooth development of this study. However, there are still three aspects that need to be further expanded. First, the depiction of the relative poverty of farmers is mostly confined to the single dimension of income, and the relative poverty of farmers is not measured from a multidimensional perspective. Second, the lack of investigation into the heterogeneity of agricultural machinery social services on the relative poverty of farmers is not conducive to identifying the external constraints of agricultural machinery socialization services in poverty reduction. Third, the lack of the perspective of risk in exploring the mechanism of agricultural machinery socialization services on the relative poverty of farmers is not conducive to optimizing the path system of agricultural machinery socialization service poverty reduction. Based on this and on the basis of theoretical analysis, this paper uses 2750 items of micro-survey data from farmers in Henan Province, and it uses a binary Logit model and intermediary effect model to empirically test the impact of the adoption of agricultural machinery socialization services on the relative poverty of farmers.

2. Theoretical Analysis and Research Hypothesis

2.1. Impact of Agricultural Machinery Socialization Service Adoption on the Relative Poverty of Farmers

The evolution of the division of labor is regarded as a key factor in determining long-term economic growth. One way to solve the dilemma of the weak income increase of farmers is to promote the agricultural division of labor. Agriculture “imports” the division of labor economy from industry through the purchase of machinery and the indirect production effect, which greatly extends the space of the specialized division of labor and the trade radius [25], thus generating an agricultural machinery socialization service and improving production efficiency [26]. The adoption of agricultural machinery socialization services can improve production efficiency, increase farmers’ agricultural operating income and wage income, and alleviate relative poverty. On the one hand, the replacement of an agricultural labor force with agricultural machinery means that farmers can begin to concentrate on the advantageous production links, improve the level of individual specialization, and improve the efficiency of agricultural production. On the other hand, the adoption of agricultural machinery socialization services by farmers adds mechanical elements to agricultural production, and the mechanical operation efficiency is generally higher than that of manual operations, thus improving agricultural production efficiency. At the same time, the adoption of agricultural machinery socialization services can accelerate the non-agricultural transfer of labor by promoting the improvement of labor allocation efficiency [25], increasing the wage income of farmers, and alleviating relative poverty. In addition, agricultural machinery socialization services will embed other production factors such as pesticides and fertilizers into the production process so that farmers can obtain other production factors more conveniently, reduce transaction costs, and alleviate relative poverty [27].

Based on this, this paper proposes research Hypothesis 1.

H1. 

Agricultural machinery socialization service adoption has a significant negative effect on relative poverty and is conducive to alleviating the relative poverty of farm households.

2.2. Heterogeneous Impact of the Adoption of Agricultural Machinery Socialization Services on the Relative Poverty of Farmers

In theory, after the introduction of agricultural machinery socialization services, the original agricultural labor force will be released to non-agricultural industries, thereby increasing the non-agricultural income of farmers [19] and alleviating relative poverty. However, the labor release effect of agricultural machinery socialization services will be affected by the size of the household labor force, and there are differences. Generally speaking, the greater the number of family members involved in farm labor, the more obvious the labor release effect of the agricultural machinery socialization service will be, and, correspondingly, the more farmers can increase their non-agricultural income. For farmers with a small family labor force, although the agricultural machinery socialization service can replace part of the labor force in production links, they are still constrained by the non-mechanization of some production links as well as family care, and their labor force transfer degree is lower. Here, most of the labor exists in the form of part-time employment. In addition, the poverty reduction effect of agricultural machinery socialization services is also affected by the scale of land management. The larger the scale of land management, the lower the service cost, and the more obvious the income increase for the farmers [28]. On the contrary, the smaller the scale of land management and the higher the degree of fragmentation, the higher the service cost, thus reducing the poverty reduction effect of agricultural machinery socialization services.

Based on this, this paper proposes research Hypothesis 2.

H2. 

The effect of agricultural machinery socialization service adoption on the relative poverty of rural households is heterogeneous under the characteristics of different household labor force sizes and land management scales.

2.3. Role of Risk in the Impact of the Adoption of Agricultural Machinery Socialization Services on the Relative Poverty of Farmers

Agricultural production risk is an important reason why farmers become poor and find it difficult to escape poverty [29]. The adoption of agricultural machinery socialization services by farmers can effectively avoid agricultural production and operation risks [30], reduce the risk probability, and alleviate relative poverty. Firstly, the adoption of social agricultural machinery services can effectively deal with the harm caused by natural disasters and other emergencies that affect agricultural production, ensure agricultural output, stabilize planting income [31], reduce the loss of agricultural operating income, and alleviate the relative poverty of farmers. Secondly, the introduction of agricultural machinery socialization services effectively replaces the labor of hired workers [32], avoids the employment risk caused by the difficulties involved in hiring workers, makes the agricultural operation process measurable, reduces supervision costs [33,34], moral hazards, and adverse selection, and avoids income damage. Finally, the adoption of agricultural machinery replaces self-purchased agricultural machinery, avoiding investment risks. Agricultural machinery is characterized by high asset specificity, and farmers are prone to high precipitation costs when purchasing agricultural machinery; however, adopting agricultural machinery socialization services removes the need to bear the risk of locking up the machinery [35] and at the same time can obtain the benefits brought by mechanized operations and alleviate relative poverty (Figure 2).

Based on this, this paper proposes research Hypothesis 3.

H3. 

Agricultural machinery socialization service adoption alleviates the relative poverty of farm households by reducing natural risks.

3. Materials and Methods

3.1. Data Sources

The data in this paper were taken from the sample survey data of farmers in six counties of Henan Province in 2017 (Figure 3). Henan Province is a major agricultural province in China. According to data reported in the seventh national census, Henan’s resident rural population accounts for 46.79% of the resident population, which is higher than the national average of 36.11%. According to the data regarding China’s grain output in 2021, Henan’s total grain output is 65.442 million tons, ranking second in China. Therefore, the selection of Henan as a sample area is representative. According to indicators such as the village per capita net income, geographical location, wheat sown area, and per capita disposable income of rural residents, a total of 4000 rural households were investigated, and 3914 valid samples were obtained in accordance with the principle of random sampling. After data cleaning, 2750 rural household samples were finally selected for research according to the indicators required in this paper. Among the six counties in this sample, the terrain is complex and varied, with plains and mountains dominating in Anyang County, mountains and hills dominating in Xin’an and Qi counties, plains dominating in Wuyang County, plains and depressions dominating in Shangcai County, and inclined plains and gentle slopes dominating the topography of Zhengyang County, which covers the geographic features of south, north, and central Henan Province, as well as providing a reference for the study of other cultivated areas in China. Therefore, the development of agricultural machinery socialization services in Henan Province is representative. At the same time, Henan Province is an underdeveloped area; the overall economic level of rural residents is low, and its relative poverty is typical. Therefore, this paper selects the rural residents of Henan Province as the research object to explore the impact of the adoption of agricultural machinery socialization services on the relative poverty of farmers.

3.2. Variable Selection

3.2.1. Dependent Variables

Scholars represented by Amartya Sen define multidimensional poverty as the deprivation of basic human viable ability [36]. Based on the ideas of Amartya Sen combined with domestic and foreign scholars’ construction and measurement methods for multidimensional poverty indicators [37,38,39,40], and taking into account existing survey data, a measurement index system of the relative poverty of rural households is constructed (see

Table 1. Identification index system of relative poverty of farmers.

System	Index	Relative Poverty Deprivation Critical Criteria
Relative poverty identification index system of rural households	Income level	The per capita household income is lower than the per capita disposable income of urban residents in the county 50% of income is considered relative poverty
	Level of education	The highest education is considered relatively poor below junior high school
	Social trust	People who think they are not trustworthy when dealing with them are considered to be relatively poor
		People who score less than 5 for trust in the village community are considered relatively poor (10 overall)
		People who score less than 5 for trust in the town government are considered relatively poor (10 overall)
	Cognitive attitudes to new things	People who are less active in accepting new things are seen as relatively poor
	Employment contracts	Family members who migrate to work without a written employment contract are considered relatively poor
	Pension insurance	People who do not buy pension insurance are considered relatively poor
	Health insurance	People who do not buy health insurance are considered relatively poor
	Cooperative participation	People who do not join cooperatives are considered relatively poor

) according to the United Nations Development Programme’s (UNDP) equal weight method, that is, the equal weight division of poverty indicators. In this index system, if four or more indicators are deprived, farmers are considered to be in a relatively poor state and assigned a value of 1; otherwise, they are assigned a value of 0.

3.2.2. Core Independent Variable

Land preparation and harvesting services are the agricultural machinery socialization services generally adopted by farmers [41,42], which can better reflect the adoption of agricultural machinery socialization services. Therefore, whether an agricultural machinery social service is adopted in land preparation or harvest is selected as the proxy variable for the adoption of agricultural machinery socialization services.

3.2.3. Mediator Variables

Based on the above theoretical analysis, it can be seen that the adoption of agricultural machinery socialization services can reduce the impact of natural risk, employment risk, and agricultural machinery investment risk, and alleviate the relative poverty of farmers. With reference to the existing studies [43,44] and taking into account the existing data indicators, the natural risk, employment risk, and agricultural machinery investment risk were respectively described by selecting the reduction ratio, labor disputes, and self-purchase of agricultural machinery.

3.2.4. Control Variables

Referring to the study of Zhai Yujia et al. [45], the study further selected individual characteristics, family characteristics, and regional characteristics as control variables to be included in the analysis. Individual characteristics include whether individuals are party members or village cadres; household characteristics include the size of the household labor force, contracted land area, and deposit balance; and the regional characteristics include village topography, village traffic conditions, village economic development level, distance from the town center, and distance from the county seat. The definitions of the variables and their descriptive statistical results are shown in

Table 2. Definition, measurement, and descriptive statistical analysis of variables (N = 2750).

Variable Type	Variable Name		Variable Measures	Average Value	Standard Deviation
Dependent variables	Relative poverty of rural households		0 = No, 1 = Yes	0.468	0.499
Core independent variable	Agricultural machinery socialization service adoption		0 = No, 1 = Yes	0.873	0.332
Mediator variables	Risk	Natural risk	%	12.645	20.372
		Employment risk	0 = No, 1 = Yes	0.004	0.068
		Agricultural machinery investment risk	0 = No, 1 = Yes	0.141	0.348
Control variables	Whether party members or not		0 = No, 1 = Yes	0.082	0.275
	Whether village cadres or not		0 = No, 1 = Yes	0.054	0.227
	Size of household labor force		person	2.849	1.441
	Contracted land area		Hm2	0.554	0.397
	Deposit balance		0 = None; 1 = CNY 10,000 and below; 2 = CNY 10,000–50,000; 3 = CNY 50,000–100,000; 4 = more than CNY 100,000	1.557	0.824
	Village terrain		1 = Mountain area; 2 = Hill; 3 = Plain	2.892	0.330
	Village traffic conditions		1 = Very poor; 2 = Poor 3 = General; 4 = Good; 5 = Very good	3.146	0.927
	Level of economic development of the village		1 = Very poor; 2 = Poor 3 = General; 4 = Good; 5 = Very good	2.872	0.698
	Distance from town center		Kilometer	4.115	3.401
	Distance from the county seat		Kilometer	20.796	12.052

.

3.3. Data Characteristics

The cross-analysis of agricultural machinery socialization service adoption and the relative poverty of farmers shows that 349 households, accounting for 12.7%, have not adopted agricultural machinery socialization services, and their average probability of relative poverty is 0.650. The number of farmers who adopted agricultural machinery socialization services was 2401, accounting for 87.3%, and their average probability of relative poverty is 0.441. As can be seen from

Table 3. Cross analysis of agricultural machinery socialization service adoption and relative poverty of farmers (N = 2750).

Agricultural Machinery Socialization Service Adoption	Relative Poverty of Rural Households		Frequency	Percentage
	Mean	Mean Difference
Not adopted	0.650	0.209 ***	349	12.70
Adopted	0.441		2401	87.30

Note: *** indicates significance at the 1% statistical levels.

, there is a significant difference between the mean values of the relative poverty of farmers, indicating that the adoption of agricultural machinery socialization services has a negative relationship with the relative poverty of farmers, which reduces the probability of farmers’ relative poverty. Of course, in order to confirm whether the adoption of agricultural machinery socialization services can effectively alleviate the relative poverty of farmers, more rigorous quantitative analysis is needed, which will be verified in the following section.

3.4. Model Settings

In this study, the relative poverty status of farmers can be divided into two types: “in relative poverty” and “not in relative poverty”. This is a binary choice problem, so this study intends to use the binary logit model to analyze the relative poverty of farmers. The basic form of the binary logit model is as follows:

$$p_i=F(y_i=1|X_i)=\frac{1}{1+e^{-y_i}}$$

(1)

In Equation (1), $y_i$ represents whether farmers are relatively poor, where $y_i$ = 1 represents that farmers are in a state of relative poverty and $y_i$ = 0 represents that farmers are not in a state of relative poverty; $p_i$ represents the probability of farmers being relatively poor; $X_i$ represents the adoption of agricultural machinery socialization services; and $y_i$ is a linear combination of variables $X_i$, namely,

$$y_i=a_0+a_1{X}_i+\beta D_i^{\prime }$$

(2)

In Equation (2), $D_i^{\prime }$ represents the control variables in terms of individual characteristics, family characteristics, and regional characteristics; $a_1$ and $\beta$ are the coefficients to be estimated.

By transforming Equations (1) and (2), the logit model expressed in the occurrence ratio can be obtained as follows:

$$logit(p_i)=ln\frac{p_i(y_i=1)}{1-p_i(y_i=1)}=a_0+a_1{X}_i+\beta D_i^{\prime }+{\zeta }_i$$

(3)

In Equation (3), $a_0$ is a constant term and ${\zeta }_i$ is a random disturbance term.

In order to explore the impact mechanism of the adoption of agricultural machinery socialization services on the relative poverty of farmers, the following intermediary effect model is constructed by referring to the research of Wen Zhonglin and Ye Baojuan [46]:

$$Y=c_0+c_{1}X+{\displaystyle \sum _{n=1}{c}_{2n}}{D}_{ni}+{\epsilon }_1$$

(4)

$$M=a_0+a_{1}X+{\displaystyle \sum _{n=1}{a}_{2n}}{D}_{ni}+{\epsilon }_2$$

(5)

$$Y=b_0+c^{\prime }X+b_{1}M+{\displaystyle \sum _{n=1}{b}_{2n}}{D}_{ni}+{\epsilon }_3$$

(6)

Here, $Y$, $X$, and $M$, respectively, represent the relative poverty of farmers, the adoption of agricultural machinery social services, and the intermediary variables (natural risk, employment risk, agricultural machinery investment risk). $D_{ni}$ represents control variables in terms of individual characteristics, family characteristics, and regional characteristics; $a_0$, $b_0$, $c_0$ are the constant terms; $a_1$, $b_1$, $c^{\prime }$, $c_1$, $a_{2n}$, $b_{2n}$ and $c_{2n}$ are the coefficients to be estimated; and ${\epsilon }_1$, ${\epsilon }_2$, and ${\epsilon }_3$ are error terms and follow a normal distribution.

4. Results

4.1. Baseline Regression Results

It can be seen from

Table 4. Benchmark regression results of agricultural machinery socialization service adoption on relative poverty of farmers (N = 2750).

Variable Name	(1)	(2)	Marginal Effects
	Relative Poverty of Rural Households	Relative Poverty of Rural Households	Relative Poverty of Rural Households
Agricultural machinery socialization service adoption	−0.856 ***	−0.858 ***	−0.203 ***
	(0.119)	(0.123)	(0.028)
Whether party members or not		−0.607 ***	−0.143 ***
		(0.172)	(0.040)
Whether village cadres or not		−0.607 ***	−0.143 ***
		(0.215)	(0.050)
Size of household labor force		−0.093 ***	−0.022 ***
		(0.028)	(0.006)
Contracted land area		−0.010	−0.002
		(0.007)	(0.001)
Deposit balance		−0.256 ***	−0.060 ***
		(0.049)	(0.011)
Village terrain		0.095	0.022
		(0.123)	(0.029)
Village traffic conditions		−0.005	−0.001
		(0.043)	(0.010)
Level of economic development of the village		−0.041	−0.009
		(0.058)	(0.013)
Distance from town center		0.026 **	0.006 **
		(0.012)	(0.002)
Distance from the county seat		0.007 **	0.001 **
		(0.003)	(0.001)
Region	Under control	Under control	Under control
Pseudo R2	0.014	0.039	-
Sample size	2750	2750	2750

Note: **, and *** indicate significance at the 5% and 1% statistical levels, respectively, with standard errors in parentheses.

that the adoption of agricultural machinery socialization services has a significant negative impact on the relative poverty of farmers and is conducive to alleviating the relative poverty of farmers, which is consistent with the theoretical analysis above. From the perspective of the mitigation degree, if other control variables remain unchanged, the relative poverty incidence of farmers who adopt agricultural machinery socialization services will be reduced by 20.3%. The possible reasons for this are that the adoption of agricultural machinery socialization services can reduce risk, reduce the probability of farmers’ income damage, improve economic conditions, and then alleviate the relative poverty of farmers. In this regard, we will provide further proof in the examination of the intermediary mechanism in the following section.

Among the control variables, whether individuals are party members or not, whether they are village cadres or not, the size of the household labor force, and the balance of savings have a significant negative impact on the relative poverty of rural households, while the distance from the town center and the distance from the county seat have a significant positive impact on the relative poverty of rural households. The main reason is that farmers whose households are headed by party members or village cadres have higher social status, show stronger political capital, and generally have better economic conditions and self-development ability; thus, the probability of falling into relative poverty is lower. Generally speaking, the larger the size of the household labor force, the better the economic situation, and the lower the probability of falling into relative poverty. The balance of household deposits not only represents the existing economic conditions of the farmers, but also reflects their development ability to a certain extent. Rural households with considerable household savings can obtain social opportunities by improving their education level, improving their economic level, and enhancing their self-development ability. Moreover, rural households with more significant household savings can provide financial support for the establishment of protective security and can avoid falling into the health poverty trap by purchasing pension insurance and medical insurance, thus alleviating relative poverty. In addition, the distance from the town center and the distance from the county seat may cause farmers to lose some development opportunities. The living environment and village traffic conditions in remote areas are relatively poor, which leads to the low economic level of farmers and a lack of endogenous development ability, which may lead to a decline into relative poverty.

4.2. Robustness Test

In order to further investigate whether the baseline regression results are robust, in the following we use the method of changing samples for re-regression and adopt the research group’s 2018 survey data of 756 rural households in Xinfeng County, Shaoguan City, Guangdong Province (Figure 4). This region is dominated by mountainous and hilly terrain and rice cultivation. The current situation of agricultural machinery socialization services in Xinfeng County, Shaoguan City, Guangdong Province is representative of the development of agricultural machinery socialization services in southern mountainous areas to a certain extent, and it forms a comparative study on the development of agricultural machinery socialization services in Henan Province, which is dominated by plain terrain and wheat cultivation. Moreover, the economic level of Xinfeng County in Shaoguan City, Guangdong Province is still in the development stage, and the overall economic status of rural residents is not good; thus, its relative poverty is representative to a certain extent. As can be seen from

Table 5. Robust regression results of agricultural machinery socialization service adoption on relative poverty of farmers (N = 756).

Variable Name	(1)	(2)	Marginal Effects
	Relative Poverty of Rural Households	Relative Poverty of Rural Households	Relative Poverty of Rural Households
Agricultural machinery socialization service adoption	−1.826 ***	−1.705 ***	−0.319 ***
	(0.177)	(0.192)	(0.028)
Whether party members or not		−1.044 ***	−0.195 ***
		(0.296)	(0.053)
Whether village cadres or not		−0.753 **	−0.141 **
		(0.359)	(0.066)
Size of household labor force		−0.107	−0.020
		(0.069)	(0.012)
Contracted land area		0.002	0.001
		(0.005)	(0.001)
Deposit balance		−0.142	−0.026
		(0.089)	(0.016)
Village terrain		0.452 ***	0.084 ***
		(0.163)	(0.030)
Village traffic conditions		0.059	0.011
		(0.201)	(0.037)
Level of economic development of the village		0.084	0.015
		(0.185)	(0.034)
Distance from town center		0.032 **	0.006 **
		(0.013)	(0.002)
Distance from the county seat		−0.006	−0.001
		(0.004)	(0.001)
Region	Under control	Under control	Under control
Pseudo R2	0.121	0.178	-
Sample size	756	756	756

Note: **, and *** indicate significance at the 5%, and 1% statistical levels, respectively, with standard errors in parentheses.

, the adoption of agricultural machinery socialization services still has a significant negative impact on the relative poverty of rural households, and their adoption can play a role in poverty reduction under different samples. Meanwhile, the estimated results of other control variables are basically consistent with the above. Therefore, the abovementioned conclusions are robust.

4.3. Endogeneity Test

There may be endogeneity problems caused by mutual causality in the benchmark regression model, which makes the above conclusions biased. On the one hand, the adoption of agricultural machinery socialization services has an impact on the relative poverty of farmers by reducing risk; on the other hand, the relative poverty of farmers may also affect the adoption of agricultural machinery socialization services, and farmers with better economic conditions and stronger abilities may be more inclined to adopt them. For this reason, the Extended Regression Model (ERM) was used to perform another regression to avoid possible endogeneity problems. The ERM model is an international cutting-edge method utilized to deal with endogeneity. Compared with the traditional instrumental variable method, which is only applicable to situations where endogenous explanatory variables are continuous, the ERM model is applicable to situations where endogenous explanatory variables are continuous and discrete [47]. At the same time, the harvest service is selected as the proxy index of the agricultural machinery socialization service, and the average harvest service adoption rate of other farmers in the village is used as the instrumental variable for regression by referring to Ma et al. [48]. The regression results show that the adoption of agricultural machinery socialization services significantly inhibits and is conducive to alleviating the relative poverty of farmers (

Table 6. ERM regression results of impact of agricultural machinery socialization service adoption on relative poverty of farmers (N = 2750).

Variable Name	(1)	(2)
	Relative Poverty of Rural Households	Relative Poverty of Rural Households
Agricultural machinery socialization service adoption	−0.607 ***	−0.594 ***
	(0.200)	(0.213)
Whether party members or not		−0.363 ***
		(0.103)
Whether village cadres or not		−0.361 ***
		(0.128)
Size of household labor force		−0.060 ***
		(0.017)
Contracted land area		−0.007
		(0.004)
Deposit balance		−0.159 ***
		(0.030)
Village terrain		0.090
		(0.078)
Village traffic conditions		−0.002
		(0.027)
Level of economic development of the village		−0.027
		(0.036)
Distance from town center		0.016 **
		(0.007)
Distance from the county seat		0.004 **
		(0.002)
Sample size	2750	2750

Note: **, and *** indicate significance at the 5%, and 1% statistical levels, respectively, with standard errors in parentheses.

).

4.4. Heterogeneity Analysis

Based on the above theoretical analysis, we investigated the heterogeneity of the adoption of agricultural machinery socialization services on the relative poverty of farmers under the different characteristics of household labor force quantity and land management scale, and the results are shown in

Table 7. Heterogeneity test results of the impact of agricultural machinery socialization service adoption on relative poverty of farmers (N = 2750).

Variable Name	Number of Household Labor		Scale of Land Operation
	Below or Equal to the Mean	Above the Mean	Below or Equal to the Mean	Above the Mean
Agricultural machinery socialization service adoption	−0.703 ***	−1.060 ***	−0.728 ***	−1.313 ***
	(0.172)	(0.180)	(0.138)	(0.273)
Control variables	Inducted	Inducted	Inducted	Inducted
Constant	−0.663	1.849 ***	0.951 *	−1.388 ***
	(0.675)	(0.566)	(0.506)	(1.129)
Region	Under control	Under control	Under control	Under control
Pseudo R2	0.026	0.053	0.039	0.052
Sample size	1118	1632	1727	1023
Empirical p-value	0.080 *		0.020 **

Note: *, **, and *** indicate significance at the 10%, 5%, and 1% statistical levels, respectively, with standard errors in parentheses. The “empirical p-value” was used to test the significance of coefficient differences between groups and was obtained by using bootstrapping 1000 times.

. Referring to the research of Lian Yujun et al. [49], the “empirical p-value” was used to determine whether there is a significant difference between the group regression coefficients. According to the “empirical p-value”, the impact of the adoption of agricultural machinery socialization services on the relative poverty of farmers is significantly different under the characteristics of the size of the household labor force and the scale of land management. Specifically, compared with farmers with a small number of family members involved in labor, those with more members have a higher negative impact on relative poverty. A possible reason for this is that for farmers with a larger labor force, the labor release effect of agricultural machinery socialization services is more obvious, accelerating the differentiation of the family labor force in agriculture and non-agricultural industries, promoting the growth of non-agricultural income among farmers, and thus alleviating relative poverty. Compared with farmers with a small land management scale, those with larger land management scales who adopt agricultural machinery socialization services experience a higher degree of negative impact on relative poverty. One possible reason for this is that for farmers with a large land management scale, the service use cost will decrease significantly and the income increase effect will be more obvious due to the existence of a service scale effect.

4.5. Mechanism Test

The mechanism verification results of the adoption of agricultural machinery socialization services to alleviate the relative poverty of farmers by reducing risk are shown in

Table 8. Test results of the impact of the mechanism of agricultural machinery socialization service adoption on relative poverty of farmers (N = 2750).

Variable Name	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Relative Poverty of Rural Households	Natural Risk	Relative Poverty of Rural Households	Employment Risk	Relative Poverty of Rural Households	Agricultural Machinery Investment Risk	Relative Poverty of Rural Households
Agricultural machinery socialization service adoption	−0.858 ***	−1.020 ***	−0.862 ***	−0.906	−0.864 ***	−0.044	−0.858 ***
	(0.123)	(0.024)	(0.123)	(0.686)	(0.123)	(0.187)	(0.123)
Natural risk impact			0.101 ***
			(0.002)
Employment risk					0.847
					(0.612)
Agricultural machinery investment risk							0.174
							(0.115)
Control variables	Inducted
Constant terms	0.514	−11.860 ***	0.524	−3.160 *	0.522	−5.519 ***	0.552
	(0.429)	(3.992)	(0.429)	(1.697)	(0.429)	(0.837)	(0.430)
Observations	2750	2750	2750	2750	2750	2750	2750
Pseudo R2 or R2	0.039	0.104	0.039	0.068	0.039	0.078	0.039

Note: *, and *** indicate significance at the 10%, and 1% statistical levels, respectively, with standard errors in parentheses.

.

First, the mediation effect of natural risk on the relationship between the adoption of agricultural machinery socialization services and the relative poverty of farmers is tested. According to the regression results in column (1), the adoption of agricultural machinery socialization services has a significant negative impact on the relative poverty of farmers, and it is conducive to alleviating the relative poverty of farmers. At the same time, the results in column (2) show that the adoption of agricultural machinery socialization services has a significant negative impact on the impact of natural risk, which is conducive to reducing the impact of natural risk. In addition, the regression results in column (3) show that after the introduction of the variable of natural risk, the adoption of agricultural machinery socialization services still has a significant negative impact on the relative poverty of farmers, and the impact of natural risk has a significant positive impact on the relative poverty of farmers. It can be seen that natural risk plays a partial mediating role in the relationship between the adoption of agricultural machinery socialization services and the relative poverty of farmers, and the mediating effect accounts for 12% of the total effect.

Second, the mediation effect of the employment risk on the relationship between the adoption of agricultural machinery socialization services and the relative poverty of farmers is examined. According to the results in column (4), the adoption of agricultural machinery socialization services has no significant impact on the employment risk. The results show that after the introduction of the variable of employment risk, the adoption of agricultural machinery socialization services still has a significant negative impact on the relative poverty of farmers, but the impact of employment risk has no significant impact. According to the mediation effect test steps in the previous model construction, the bootstrap method was used. The test results showed that employment risk had no intermediary effect on the relationship between the adoption of agricultural machinery socialization services and the relative poverty of farmers.

Thirdly, the mediating effect of agricultural machinery investment risk on the relationship between the adoption of agricultural machinery socialization services and the relative poverty of farmers is tested. According to the results in column (6), the adoption of agricultural machinery socialization services has no significant impact on agricultural machinery investment risk. At the same time, the regression results in column (7) show that after the variable of agricultural machinery investment risk is introduced, the adoption of agricultural machinery socialization services still has a significant negative impact on the relative poverty of farmers, but the agricultural machinery investment risk has no significant impact. According to the mediation effect test steps in the previous model construction, the bootstrap method was used, and the test results showed that the agricultural machinery investment risk had no intermediary effect on the relationship between the adoption of agricultural machinery socialization services and the relative poverty of farmers.

5. Discussion

Traditional smallholder farmers are vulnerable to natural risks such as floods and droughts; if their income is damaged, they can fall into relative poverty. With the continuous non-agricultural transfer of rural labor factors, especially the outward transfer of high-quality labor such as young and middle-aged people, agricultural production is faced with a decline in labor quality and shortage constraints such as aging and feminization, and it also induces employment risk, thereby reducing the agricultural operating income of farmers. On the one hand, the busy and idle periods in agricultural production are relatively fixed. When the demand for a large number of employees in the busy season cannot be satisfied in a timely manner, the risk of labor shortages in agricultural production will increase. On the other hand, due to the difficulty in measuring the quality of agricultural labor, the phenomenon of moral hazard and adverse selection may occur in the hired labor force, which intensifies the uncertainty of agricultural production and further reduces the production and operational income of farmers. In addition, small-scale farmers have strong capital rigidity and high management scale constraints, and the purchase of agricultural machinery is prone to the impact of agricultural machinery investment locking risk, resulting in damage to farmers’ income and an increase in the probability of relative poverty.

Risk and vulnerability are the main reasons for farmers descending into or returning to poverty. Improving farmers’ risk-coping ability can effectively alleviate relative poverty [50]. With the outward flow of rural labor force elements and the continuous progress of agricultural technology, agricultural machinery socialization services have gradually become a new risk management tool for farmers before and after the event by using modern agricultural production machinery and equipment to prevent risk. However, the impact of agricultural machinery socialization services on the relative poverty of farmers is heterogeneous, and the poverty reduction effect is affected by the size of the household labor force and the scale of land management. In the process of using agricultural machinery socialization services for poverty reduction, emphasis should be placed on using the services to reduce the external constraints of farmers’ relative poverty, so as to give full play to the effect of such agricultural services for poverty reduction.

The adoption of agricultural machinery socialization services can help reduce the relative poverty of farmers and reduce the degree of relative poverty. Based on the division of labor theory and technology change theory, it can be seen that agricultural machinery socialization services affect the relative poverty of farmers through the mechanism of efficiency improvement, labor substitution, and risk reduction. This paper focused on the risk reduction mechanism of agricultural machinery socialization services, divided the risk into natural risk, employment risk, and agricultural machinery investment risk, and carried out intermediary mechanism testing. According to the test results, the adoption of agricultural machinery socialization services mainly affects the relative poverty of farmers by reducing the impact of natural risks, which further verifies the positive impact of these services on the ability to cope with natural disasters and to alleviate relative poverty, and it puts forward higher requirements for optimizing the development system of agricultural machinery socialization services. Due to the rapid development of urbanization, many farmers have moved towards non-agricultural industries; strengthening mechanization in developing countries plays an important role in guaranteeing food security and improving labor productivity [9]. As agricultural machinery is characterized by high operational scale constraints and asset specialization, the adoption of agricultural machinery socialization services in developing countries dominated by small-scale farmers can effectively avoid agricultural machinery investment risk. At the same time, farmers can reduce agricultural operational losses by reducing natural risks, accelerate the non-agriculture transfer of labor while reducing labor hiring costs through the labor substitution effect, and increase their income, which is conducive to the promotion of poverty reduction among farmers in developing countries.

6. Conclusions

Based on the micro-survey data of 2750 farmers in Henan Province, this paper empirically tested the impact and mechanism of the adoption of agricultural machinery socialization services on the relative poverty of farmers, discussed the heterogeneity of this adoption, and drew the following conclusions.

The adoption of agricultural machinery socialization services has a significant negative impact on the relative poverty of farmers, which alleviates poverty by reducing the impact of natural risks. At the same time, the impact of the adoption of agricultural machinery socialization services on the relative poverty of farmers is heterogeneous under the characteristics of different household labor force size and land management scale.

Firstly, it can be seen that the improvement in the social service supply and demand system of agricultural machinery should be included in the scope of farmers’ relative poverty management, and an agricultural machinery social service system with the goal of reducing poverty among farmers should be formulated to enhance farmers’ risk resistance ability. Secondly, a supporting system of employment skills training for farmers should be constructed to promote the realization of non-agricultural employment for the rural labor force, as well as to promote the adoption of socialized services for agricultural machinery, which is conducive to increasing the agricultural wage income and business income of farmers. Thirdly, the government needs to organize training on agricultural machinery to help farmers to use them properly, which will also help to increase the demand for agriculture machinery. Finally, we should encourage and support the orderly connection between the large-scale management of agricultural land and the large-scale management of services and build a coordination mechanism of “walking on two legs” to better serve the agricultural production and poverty reduction of rural households.