The Impact of Agricultural Machinery Socialization Services on the Scale of Land Operation: Evidence from Rural China

Abstract

The aims of this study were to analyze in depth the impact of agricultural machinery socialization services on the scale of land operation and to examine whether agricultural machinery socialization services can effectively alleviate the constraints faced by the scale of land operation. Accordingly, a systematic theoretical analysis and derivation of a farmer production model were first conducted to show that the adoption of agricultural machinery socialization services by farmers alleviated the financial, technical, and labor constraints they faced and facilitated the expansion of the farming scale. Subsequently, the OLS model, three-stage least squares, seemingly uncorrelated Biprobit association model, and Probit model were used based on survey data of maize farmers in 13 Chinese provinces in 2019. The endogeneity of simultaneous decision-making between farmers’ adoption of farm machinery socialization services and operation scale was assessed. Then, we empirically tested the effects of agricultural machinery socialization services on the operation scale under different part-time situations and terrain conditions. The empirical study shows that there is a significant contribution of farmers’ adoption of agricultural machinery socialization services to the scale of land operation. The adoption of agricultural machinery socialization services by purely agricultural household groups has a more significant effect on increasing the scale of land management. Moreover, the adoption of agricultural machinery socialization services significantly increases the scale of land management only among farmers on flat land as compared to farmers on sloping land. The research results help to dissect the relationship between agricultural machinery socialization services and operation scale and provide insights for developing agricultural machinery socialization service systems and expanding operation scale.

1. Introduction

China’s national situation is that of a large country with few farmers; therefore, expanding the scale of agricultural land operation and realizing agricultural modernization are important elements of China’s agricultural reform. According to its latest national agricultural census, the total area of cultivated land in China is 135 million hectares, and the number of agricultural workers is 314 million, while the average area of cultivated land per worker is only 0.43 hectares. Throughout the world’s pioneering countries in agricultural modernization, the importance of large-scale land operations has been repeatedly demonstrated [1]. However, expanding the scale of land operation is often limited by factors such as higher agricultural business risks [2], higher labor costs [3,4], and lack of capital credit [5], which inhibit farmers from engaging in large-scale land operation. The biggest institutional obstacle to the realization of agricultural modernization in China is the smallholder economy [6]; thus, transforming the smallholder economy and expanding the scale of land operation is the inevitable requirement for promoting agricultural modernization and developing modern productive forces.

In recent years, agricultural machinery socialization services have developed rapidly in China’s agricultural industry, which has led to the improvement of agricultural production capacity and boosted the agricultural economy, while helping to weaken the constraints of agricultural production [7]. Liu and Yang (2016) obtained similar findings based on the national rice industry farm household survey data that show that agricultural socialization services have a positive impact on the land operation scale [8]. It is worth noting that the rapid development of services and deepening in all aspects of agricultural production have changed the external conditions of farmers’ production and management decisions [9], affecting the way farmers carry out agricultural production and the choice of factor inputs. What kind of changes, therefore, have agricultural machinery socialized services brought about on land operation scale? Can services relax the constraints of land operation scale and thus help to promote it?

Existing studies have provided a certain research basis for the adoption of farm machinery socialization services by farmers to influence land operation scale, but there is still a lack of systematic theoretical analysis in examining the role of farm machinery socialization services on land operation scale. At the same time, the research methodology often neglects the endogeneity of the interaction between the adoption decision of farm machinery socialization services and the decision of land operation scale at the farm household level. Therefore, this study complements the existing research in the following two respects: first, the theoretical analysis reveals the mechanism by which the service affects the land operation scale of farmers. Second, the OLS model, three-stage least squares, seemingly unrelated Biprobit associative model, and Probit model are used to fully consider the endogeneity problem that may arise from the simultaneous decision-making between services and land operation scale and to more accurately measure the effect of services on land operation scale.

2. Theoretical Analysis

At present, farmers need to expand the scale of land operation through land transfer behavior; then, the logic of the behavior of farmers transferring to the land determines whether the scale of land operation can be realized. This section clarifies the impact of agricultural machinery socialization services on the land operation scale from the perspective of farmers’ behavior of transferring to land. Agricultural machinery socialization services are essentially a reconfiguration of factors, easing the constraints of agricultural labor, technology, and capital, and thus relaxing the constraints on the scale of land operation, as analyzed below:

(1) Alleviating agricultural labor constraints of farm households.

The comparative advantage of non-farm employment in terms of income has prompted many rural laborers to move to non-farm industries, and the age of laborers engaged in agriculture is generally higher, and the insufficient quantity and weaker quality of agricultural labor of farm households has a certain impact on food production and land operation scale. With a more adequate supply of services, farm households can replace agricultural labor by purchasing services. Services can solve the real dilemma of “no one to cultivate the land” for farmers and alleviate the possible constraints of labor factors. When the price of labor rises, services provide farmers with agricultural machinery to replace labor, which breaks the constraints of farmers’ resource endowment, relaxes the restrictions on farmers’ land transfer, and encourages farmers to increase the scale of land operation through land transfer.

(2) Alleviating farmers’ technical constraints.

The high cost of food production, uncontrollable risks, and low returns from grain cultivation make farmers less motivated to engage in food production and less willing to operate on a large scale on land; the introduction of agricultural technologies can improve agricultural returns, but because public agricultural technologies are not yet commonly promoted, most farmers have difficulty in obtaining effective technical support, thus limiting the expansion of land operation scale. Services can act as a transmitter of intellectual capital and human capital, importing high-value-added technology and capital into the grain production process. Grain production in China is dominated by small-scale farmers’ family operations, and the adoption of agricultural technologies by small-scale farmers is mostly passive and unmotivated. In the form of services to farmers, leading-edge technologies such as deep tillage machines and plant protection drones are introduced in food production. It also constructs a modern organization system and applies efficient management methods to improve the technological content and output of food production, effectively attenuating the possible technical limitations of food production and improving the profitability of food production, thus expanding the scale of farmers’ land operations.

(3) Alleviating farmers’ financial constraints.

The financial factor is part of the limitations for expanding the operation scale of farmers’ land. To expand the scale of land operation for food production, it is difficult to rely on labor alone, and agricultural machinery needs to be introduced; for maize production from planting to harvesting through multiple stages, the biological growth characteristics of the crop lead to the strong asset specificity of the machinery applied in different stages. If farmers do not adopt agricultural machinery services but purchase the required agricultural machinery on their own, they are bound to bear high capital expenditure, which restricts farmers from expanding the scale of land operation. The service scale benefits realized by services reduce the unit cost of services and save the capital required for farmers to purchase their agricultural machinery, thus prompting rational farmers to adopt services. These services help farmers to make trade-offs between their production factors and external production factors, achieve a rational allocation of production factors, realize standardized production, and improve the profitability of food production, which in turn helps farmers to expand the scale of land operation.

3. Materials and Methods

3.1. Data

The data used in this study were obtained from a rural survey of maize growers conducted by China Agricultural University in 2019. Data were collected using a combination of both hierarchical and random sampling methods. A total of 13 major maize-producing provinces in China, namely, Hebei, Henan, Anhui, Shandong, Jiangsu, Sichuan, Hunan, Heilongjiang, Gansu, Jilin, Hubei, Liaoning, and Inner Mongolia Autonomous Regions, were selected based on large yields and many participating farmers. This survey team consisted of faculty members from China Agricultural University and their undergraduate, master’s, and doctoral students, and the questionnaire was administered using one-on-one interviews with farmers. After eliminating invalid questionnaires such as missing variables, missing key information, and logical errors, a total of 1048 valid maize grower questionnaires were collected.

3.2. Empirical Model Setting

By constructing a theoretical framework, it was found that there is a positive effect of farmers’ adoption of services on the land operation scale, but in fact, there is a simultaneous decision problem between farmers’ adoption of services and land operation scale, and the two affect each other. The resulting model was set up as follows:

$${Land}_i={\alpha }_0+{\delta }_1{Service}_i+{\alpha }_1{Rent}_i+{\alpha }_2{Social}_i+{\sum }_{j=0}^3{\alpha }_{Mj}{M}_{ji}+{\sum }_{j=0}^6{\alpha }_{Nj}{N}_{ji}+{\mu }_i$$

(1)

$${Service}_i={\beta }_0+{\delta }_2{Land}_i+{\beta }_1{Rate}_i+{\sum }_{j=0}^3{\beta }_{Mj}{M}_{ji}+{\sum }_{j=0}^6{\beta }_{Nj}{N}_{ji}+{\nu }_i$$

(2)

The decision equation for the scale of farmers’ land operation is Equation (1), and the decision equation for the purchase of services is Equation (2). The subscript i in the equation shows the i-th farmer and the subscript j shows the j-th variable.${Land}_i$ shows the land operation scale decision variable; ${Service}_i$ indicates the farm machinery socialization service decision variable; ${Rent}_i$ represents land rent; ${Social}_i$ denotes social capital variables; ${Rate}_i$ represents the production commercialization rate of the farm household; $M_{ji}$ represents the individual characteristics of farm households; $N_{ji}$ represents the household characteristics of the farm household; ${\delta }_1$ and ${\delta }_2$ are the main estimated coefficients; ${\mu }_i$ and ${\nu }_i$ are the random error terms.

The land operation scale decision equation mainly measures the effect of services on maize sowing area and farmers’ land transfer decisions and is generally estimated by applying OLS models and three-stage least squares methods. When the explanatory variables are binary categorical variables, a single equation Probit model is usually used for parameter estimation; however, because of the interaction between farmers’ adoption of services and land operation scale decisions, this study then applies a seemingly unrelated bivariate model (Biprobit model, seemingly unrelated bivariate probit). The three-stage least squares (3SLS) model was further applied to test its robustness. The 3SLS model can be used as a reference for the seemingly unrelated bivariate probit model by taking into account the possible correlation between the perturbation terms of differential equations.

There are two endogenous variables in the joint cubic equation of this study, namely the farmer’s adoption of service decision variable and the land operation scale decision variable, and the exogenous variables excluded in the service adoption decision equation are land rent and social capital, and the exogenous variable excluded in the land operation scale decision equation is the production commercialization rate of the farmer. Then, the rank condition and the order condition of the joint cubic equation set are satisfied, and the conditions for conducting estimation are reached and the joint cubic system of equations can be effectively identified.

3.3. Research Hypotheses

The research focus of this paper is to examine the impact of agricultural socialization services on farmers’ land operation scale from the level of farmers. The supply of agricultural machinery socialization services is not equal to the actual inputs of farmers, and only when farmers purchase agricultural machinery socialization services can they really play a role in production and management, and alleviate the resource endowment constraints of farmers in expanding the scale of land operation. Therefore, agricultural machinery socialization services and land can be regarded as two factors that farmers invest in for agricultural production, and the impact of services on the land operation scale can be transformed into farmers’ input choice decisions on the two factors. To this end, this study utilizes a separable farm household production model to analyze how the scale of farm household land operation will change under the conditions that the market for services continues to improve and farmers are free to purchase services.

Based on Bardhan and Udry’s (1999) and Deininger and Jin’s (2008) farm household model [10,11], this study constructs a farm household production model to explain the effect of services on the scale of land operation. The model is simplified to the existence of a well-developed factor market where farmers are rational economic agents seeking to maximize the profitability of agricultural production. It is also presumed that farmers are engaged in exclusively productive agriculture, there is no non-farm employment, no distinction is made between specific types of crops, and only the land factor and the service factor (agricultural machinery socialization services) are in the production function of farmers.

Suppose $\overline{\mathrm{A}}$ is the farmer’s land resource endowment and in the land transfer market, the farmer can transfer the land, and the land rent is r. Then, the actual scale of land operation cultivated by the farmer is A. In the service market, the price of farm machinery socialized service purchased by the farmer is m, and the quantity purchased is S, so the cost of farm machinery socialized service purchased by the farmer is mS. In addition, the unit price of the food product is p. Solve for the farm household to purchase services and the scale of land operation for the optimal solutions S* and A*; in order to make the objective function exist as a great value, it is set f(A, S) to be a strictly concave agricultural production function in which

$$f_S>0, f_A>0, f_{SS}<0, f_{AA}<0,$$

Obtained: $D=\left|\begin{array}{cc}{f}_{SS}& f_{SA}\\ f_{AS}& f_{AA}\end{array}\right|=f_{SS}{f}_{AA}-f_{SA}^2>0$.

The farmer’s objective function can be expressed as

$$\underset{S,A}{\max }\pi =pf(A,S)-mS-(A-\overline{A})r$$

(3)

The first-order conditions are

$$\left\{\begin{array}{c}p{f}_S=m\\ p{f}_A=r\end{array}\right.$$

(4)

Farmers’ decision to invest in services and land in pursuit of the profit maximization goal should satisfy Equation (4) since f_A and f_S are functions of the land factor input A and the service factor input S, respectively. In turn, the full differential containing A, S, m, r, and p is calculated based on Equation (4), which leads to

$$\left\{\begin{array}{c}p{f}_{SS}dS+p{f}_{SA}dA+f_{S}dp-dm=0\\ p{f}_{AS}dS+p{f}_{AA}dA+f_{A}dp-dr=0\end{array}\right.$$

(5)

Solving for the above equation yields

$$dA=\frac{1}{p(f_{AS}{f}_A-f_{AA}{f}_S)}\left[\right(p{f}_{AS}{f}_S-p{f}_{SS}{f}_A)dS+f_{A}dm-f_{S}dr]$$

(6)

This study was conducted to analyze the effect of services on the scale of land operation, so that dm = dr = 0, grain prices p = 0, and it can be derived that

$$\frac{dA}{dS}=\frac{f_{AS}{f}_S-f_{SS}{f}_A}{f_{AS}{f}_A-f_{AA}{f}_S}$$

(7)

The above equation can show that when $f_{AS}>0$, then $\frac{dA}{dS}>0$, the more inputs of socialized services of agricultural machinery, the larger the scale of land operation; when $f_{AS}<0$, then $\frac{dA}{dS}<0$, and the more inputs of services, the smaller the scale of land operation.

If technological progress is not considered, when the input land factor is in relative equilibrium with other factors, the increase in service inputs will increase the comparative scalability of land, thus increasing the marginal output of land. In order to obtain the new equilibrium, the land factor input should also be increased accordingly, which makes both land input and farm machinery socialized service input show the same increase and decrease. Therefore, $f_{AS}>0$ that $\frac{dA}{dS}>0$.

Farmers need to examine the external constraints when facing the decision of whether to expand the scale of land operation. At present, the continuous improvement of the level of services in China has effectively supplied services; through the increased input of services, farmers break the original resource endowment constraints on the scale of land operation and reasonably allocate production factors to realize moderate scale agricultural operation.

In summary, it can be inferred from the theoretical analysis that the increase in service inputs will have a positive effect on the scale of farmers’ land operations.

3.4. Variable Selection

1. Endogenous variables. There are two endogenous variables in this study: one is the decision variable of services; the other is the decision variable of land operation scale. The land operation scale decision variable can be measured by using the maize sown area indicator and the decision indicator of whether the farmer transferred to the land (transferred = 1, not transferred = 0).

2. Explanatory variables of the adoption decision equation of services. The commercialization rate of farmers’ production reflects the degree of marketization and farmers with a higher degree of marketization are more likely to adopt services to improve agricultural returns, while farmers with a lower degree of marketization are relatively more likely to be self-sufficient. Therefore, the exogenous variable that identifies the adoption decision equation of services in this study can be selected as the commercialization rate of farmers.

3. Explanatory variables of the land operation scale decision equation. The land operation scale decision equation mainly measures the impact of services on maize sowing areas and farmers’ land transfer decisions. (Land operation scale decision variables: first, the maize sown area of the farmer was taken as the natural logarithm; second, whether the maize grower transferred to the land was used for robustness testing: transferred = 1, indicating that the farmer transferred to the land, i.e., expanded the operation scale. Not transferred = 0, indicating that the farmer did not transfer into the land, i.e., did not expand the scale of operation.) Land rent, as a way to reflect land equity in the economy, is an influencing factor of land operation scale behavior, and this study selects the logarithmic form of the average transfer price of farmers’ land (RMB/mu) in 2018 as the land rent variable. In addition, because Chinese rural society is basically an acquaintance society, long-term, stable, and friendly social relationships among farmers can help secure support from friends and neighbors, thus realizing the economic reciprocal benefit claims [12]. Social capital formed through social relationship networks plays a more important role in the production and life of farm households. Social capital expresses the ability of farmers to obtain resources through interpersonal relationships in social networks and is the sum of social resources that farmers possess. It has been demonstrated that social capital plays a significant role in farmers’ expansion of land operation scale [13], and this study selects the logarithmic form of the number of relatives and friendly acquaintances who communicate with each other in various ways (a meeting, call, WeChat, etc.) during the Spring Festival as the social capital variable.

4. Control variables. This study introduces two types of control variables, agricultural operation of farming households and farming household characteristics. The control variables of farming operation of farming households were selected from household demographic characteristics (number of farming household members, proportion of farming labor force), land resources (land quality), household assets (farm machinery assets), and farming dependence (proportion of non-farm income), which consisted of five variables. Farm household characteristics included a total of three variables: the household head’s age, education level, and health status. A set of regional dummy variables was introduced in order to control for the differences between different regions in terms of climatic factors, hydrological conditions, geography, and the current status of agricultural production, which are difficult to observe, so as to avoid the influence of regional factors on the adoption of services and the scale of land operation. The above relevant variables are defined and statistically described as shown in

Table 1. Definition of variables and descriptive statistics.

Type	Variable and Code	Variable Definition	Min	Max	Mean	Standard Deviation
Endogenous variables	Decision variables of services	Whether farmers adopt services: Adopted = 1, not adopted = 0	0	1	0.569	0.495
	Number of adopted items of services	If the number of adopted items is 0, the farmers did not adopt services, and if the number of adopted items is 5, the farmers adopted 5 services.	0	5	1.473	1.601
	The degree of adoption of services	The sum of socialized services of farm machinery for tillage, seeding, plant protection, irrigation and drainage, and harvesting of maize production by farmers accounted for the proportion of maize sown area	0	5	1.151	1.629
	Decision variables of land operation scale	Farmers’ maize sowing area (mu)	0.2 1	120	12.289	16.825
	Behavioral variables of land operation scale	Whether farmers are transferred to land: Yes = 1, no = 0	0	1	0.481	0.499
Explanatory variables	Land rent 2	The average transfer price of farmers’ land in 2018 (RMB/mu)	0	1200	292.862	177.151
	Social capital	Number of relatives communicating with each other during the Spring Festival + Number of acquaintances communicating with each other during the Spring Festival	0	780	49.901	53.118
	Commercialization rate	Maize sales (kg)/total maize production (kg)	0	1	0.799	0.350
Control variables	Age of household head	Age of head of household (years)	20	79	51.838	10.272
	Education level of household head	Illiterate = 1, elementary school = 2, middle school = 3, high school = 4, college = 5, college or above = 6	1	6	2.776	0.929
	Health status of household head	Good = 1, fair = 2, poor = 3, incapable of work = 4	1	4	1.416	0.628
	Number of family members	Number of farm household members (persons)	1	12	3.938	1.503
	Share of the agricultural labor force	Number of household agricultural laborers/Number of household members	0.125	1	0.631	0.242
	Average agricultural machinery assets per mu	The current value of agricultural machinery assets per acre for farmers	0	5000	285.226	649.815
	The proportion of non-farm income	(Income from work and family non-farm business RMB + property income RMB + income from five guarantees of minimum living income RMB + income from disaster relief RMB + other income RMB)/(Income from agricultural business RMB + income from work and family non-farm business RMB + property income RMB + income from five guarantees of minimum living income RMB + income from disaster relief RMB + other income RMB)	0	1	0.574	0.373
	Land quality	Barren land = 1, low-quality land = 2, medium-quality land = 3, high-quality land = 4, very fertile land = 5	1	5	2.962	0.855
Region dummy variable	Eastern region	Is it located in the eastern region? Yes = 1; No = 0	0	1	0.395	0.489
	Central region	Is it located in the central region? Yes = 1; No = 0	0	1	0.223	0.417
	Western region	Is it located in the western region? Yes = 1; No = 0	0	1	0.222	0.415
	Northeast region	Is it located in the northeast region? Yes = 1; No = 0	0	1	0.160	0.366

Note(s): Data were obtained from rural field research; because cross-sectional data are prone to heteroskedasticity, to ensure the validity of the model, natural logarithms were taken for land rent, social capital, age of household head, number of household members, maize sown area, and average farm machinery assets per mu to attenuate the possible heteroskedasticity problem. ¹ As the rural survey conducted by China Agricultural University in 2019 on maize growers in 13 provinces nationwide, i.e., including small-scale farmers with a maize sowing area of 0.2 mu and large-scale farmers with a maize sowing area of 120 mu, resulted in a large standard deviation in the maize sowing area (mu) of farmers. ² Due to a large number of missing values of land rent, the missing land rent was generated gradually according to the village, township, county—urban area, and province—autonomous region directly under the central government to which the researched farmers belonged, and the average value of regional land rent was selected based on village, township, county—urban area, and the province—autonomous region directly under the central government in turn to define the variable.

.

4. Results

4.1. Overall Estimation

Firstly, farmers’ land operation scale behavior was expressed as the logarithm of maize sown area, and the OLS model and 3SLS model were constructed for overall regression separately, with the same settings of control variables. The specific estimated results are shown in

Table 2. OLS model estimated results of services for the scale of land operation.

Variable	OLS (1)		OLS (2)		OLS (3)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.164 *** (0.047)	— —	— —	— —	— —	— —
Number of adopted items of services	— —	— —	0.133 ** (0.065)	— —	— —	— —
Degree of adoption of services	— —	— —	— —	— —	0.125 *** (0.041)	— —
Decision variables of land operation scale (logarithm)	— —	0.056 *** (0.019)	— —	0.051 *** (0.015)	— —	0.059 *** (0.017)
Land rent (logarithm)	0.092 *** (0.014)	— —	0.080 *** (0.015)	— —	0.076 *** (0.015)	— —
Social capital (logarithm)	0.010 (0.027)	— —	0.008 (0.030)	— —	0.007 (0.031)	— —
Commercialization rate	— —	0.095 ** (0.042)	— —	0.155 *** (0.058)	— —	0.087 ** (0.035)
Age of household head (logarithm)	−0.233 ** (0.111)	0.036 (0.068)	−0.196 * (0.117)	0.110 (0.266)	−0.191 (0.120)	0.085 (0.068)
Education level of household head	−0.051 ** (0.025)	0.044 *** (0.016)	−0.068 ** (0.027)	0.047 ** (0.021)	−0.064 ** (0.027)	0.053 *** (0.018)
Health status of household head	−0.099 *** (0.035)	0.014 (0.021)	−0.123 *** (0.037)	0.025 (0.043)	−0.127 *** (0.037)	0.021 (0.026)
Number of family members (logarithm)	0.058 *** (0.008)	0.051 (0.053)	0.064 *** (0.009)	0.049 (0.032)	0.066 *** (0.009)	0.052 (0.039)
Share of the agricultural labor force	0.048 (0.108)	0.035 (0.066)	0.049 (0.118)	0.060 (0.054)	0.052 (0.119)	0.049 (0.057)
Average agricultural machinery assets per mu (logarithm)	0.046 *** (0.006)	−0.009 ** (0.004)	0.045 *** (0.006)	−0.021 ** (0.009)	0.045 *** (0.006)	−0.014 ** (0.006)
Proportion of non-farm income	−0.035 (0.072)	0.034 (0.044)	−0.078 (0.096)	0.017 (0.015)	−0.097 (0.078)	0.029 (0.041)
Land quality	0.027 (0.025)	0.012 (0.016)	0.035 (0.028)	0.061 (0.054)	0.041 (0.028)	0.014 * (0.008)
Region dummy variable	Controlled	Controlled	Controlled	Controlled	Controlled	Controlled
Constant	2.304 *** (0.522)	−0.438 *** (0.097)	1.495 *** (0.550)	−1.019 *** (0.311)	1.556 *** (0.563)	−0.315 (1.177)
F Statistic	61.19	38.51	52.53	26.80	52.59	58.18
Prob > F	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
R2	0.425	0.230	0.417	0.190	0.411	0.295
Sample size	1048

Note: *, **, *** indicate that the variable coefficient estimates are significant at the 10%, 5%, and 1% statistical levels, respectively. The values in parentheses are standard errors. Same below.

.

Table 2 presents the empirical results of the OLS regression for the full sample of farmers, reporting the estimated coefficients of the variables in the agricultural machinery socialization service equation and the land operation scale equation. The empirical results of the land operation scale equation show that the service decision variable has a positive effect on the land operation scale and passes the 1% significance test with an estimated coefficient of 0.164, indicating that the service has a facilitating effect on the expansion of the land operation scale.

To conduct a robustness test for the expansion of land management scale by services, the variable of whether to adopt services or not is replaced, to more accurately measure the behavior of farmers in adopting services. First, the core explanatory variable is replaced with the number of service adoption items, where the number of adoption items is 0 to indicate that farmers do not adopt services, and the number of adoption items is 5 to indicate that farmers adopt 5 services. The variable for the number of service adoption items passes the significance test at the 5% statistical level with an estimated coefficient of 0.133, which confirms that services have a facilitating effect on expanding the scale of land management. The core explanatory variable was then replaced with the degree of service adoption, i.e., the sum of acres of tillage, planting, plant protection, irrigation and drainage, and harvesting services adopted by farmers as a proportion of the sown acreage of corn. The service adoption variable passes the significance test at the 1% statistical level with an estimated coefficient of 0.125, again verifying that services contribute to the expansion of the land operation scale.

However, due to the interaction between farmers’ decisions to adopt services and land operation scale decisions, parameter estimation by a single equation OLS model may be biased; therefore, the focus of this study is on the regression results of the three-stage least squares model (

Table 3. Estimated results of the 3SLS model of services on the scale of land operation.

Variable	3SLS (1)		3SLS (2)		3SLS (3)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.103 *** (0.026)	— —	— —	— —	— —	— —
Number of adopted items of services	— —	— —	0.096 *** (0.023)	— —	— —	— —
The degree of adoption of services	— —	— —	— —	— —	0.109 *** (0.028)	— —
Decision variables of land operation scale (logarithm)	— —	0.073 ** (0.029)	— —	0.050 ** (0.024)	— —	0.061 ** (0.030)
Land rent (logarithm)	0.039 *** (0.009)	— —	0.015 *** (0.004)	— —	0.041 *** (0.009)	— —
Social capital (logarithm)	0.015 (0.031)	— —	0.009 (0.006)	— —	0.025 (0.027)	— —
Commercialization rate	— —	0.079 *** (0.028)	— —	0.066 *** (0.021)	— —	0.075 ** (0.031)
Age of household head (logarithm)	−0.168 (0.152)	0.058 (0.076)	−0.215 (0.172)	0.037 (0.033)	−0127 (0.158)	0.053 (0.047)
Education level of household head	−0.016 ** (0.008)	0.028 * (0.016)	−0.015 *** (0.006)	0.022 ** (0.010)	−0.015 ** (0.007)	0.024 * (0.013)
Health status of household head	−0.097 (0.108)	0.011 (0.025)	−0.081 (0.074)	0.021 (0.016)	0.077 (0.069)	0.031 (0.027)
Number of family members (logarithm)	0.057 ** (0.029)	0.072 (0.092)	0.032 ** (0.015)	0.039 (0.051)	0.033 (0.028)	0.048 (0.038)
Share of the agricultural labor force	0.016 (0.034)	0.017 (0.018)	0.017 (0.028)	0.020 (0.024)	0.027 (0.030)	0.022 (0.025)
Average agricultural machinery assets per mu (logarithm)	0.078 *** (0.020)	−0.011 ** (0.005)	0.048 *** (0.012)	−0.072 ** (0.035)	0.064 ** (0.026)	−0.028 ** (0.013)
Proportion of non-farm income	−0.084 (0.076)	0.075 (0.120)	−0.097 (0.139)	0.034 (0.029)	−0.092 (0.010)	0.025 (0.021)
Land quality	0.049 (0.082)	0.013 (0.016)	0.047 (0.061)	0.015 (0.020)	0.054 (0.047)	0.015 (0.011)
Region dummy Variable	Controlled	Controlled	Controlled	Controlled	Controlled	Controlled
Constant	4.146 *** (1.537)	−0.731 ** (0.350)	2.393 ** (1.116)	−1.936 *** (0.717)	2.815 *** (1.048)	−1.042 *** (0.299)
Sample size	1048

).

The analysis of farmers’ adoption of agricultural machinery socialization services affecting their land operation scale is used to determine the mechanism of the impact of agricultural machinery socialization services on promoting land operation scale. The results of the empirical analysis shown in Table 3 can be used to determine whether agricultural machinery socialization services and land operation scale are directly related. From the land operation scale equation, it can be obtained that the service decision variable, the number of service adoption items, and the degree of service adoption all have a positive influence effect on the land operation scale of farmers, and the estimated results all pass the significance test at the 1% statistical level. It can be obtained from the agricultural machinery socialization service equation that the land operation scale decision variable has a significant positive effect on service, indicating that there is a bidirectional feedback effect between the two kinds of decision-making of service and land operation scale, which is in line with the findings of Yang et al. (2019) [7]. It is worth mentioning that, based on Yang et al.’s study, the measurement of service in this study is no longer limited to whether or not to adopt the service consisting of 0 and 1 variables, but includes adding the number of service adoption items and the degree of service adoption to more accurately measure the impact of agricultural machinery socialization services on land operation scale.

This study focuses on the impact of services on land operation scale in order to analyze the mechanism of the role of services in expanding land operation scale. Based on 3SLS (1), 3SLS (2), and 3SLS (3), it can be found that services measured in all three forms show a positive correlation with the land operation scale, and all of them pass the significance test at 1% statistical level, which verifies the proposed research hypothesis. Adoption of services by farmers can allocate labor, technology, and capital, effectively alleviating the constraints of factors on land operation scale, thus producing an upgrading effect on land operation scale.

The exogenous variables in the land operation scale equation are land rent and social capital. Land rent has a positive effect on the land operation scale and passes the 1% significance test. When farmers expand the scale of land operation, they should consider both land rent and the income after land transfer; when the income brought by deducting the cost of land transfer is positive, farmers will tend to expand the scale of land operation. When land with better infrastructure and higher quality is available, farmers are willing to expand the scale of land operation even if the land rent is more expensive. On the contrary, if the land is of poor quality, farmers are not willing to expand the scale of land operation even if the land rent is very cheap. Social capital has no significant effect on the scale of farmers’ land operations, which is different from the expected effect. This may be mainly because the development level of China’s land transfer market is improving at this stage, the informal land transactions are gradually decreasing, and the transaction cost of the land transfer process is reduced, and social capital is no longer needed to screen the suitable transaction objects. The exogenous variable of the adoption decision equation of services is the commercialization rate of agricultural production, and the commercialization rate of agricultural production has a significant positive effect on all three forms of service adoption variables. Farmers with higher commercialization rates of agricultural production tend to adopt services to improve quality and efficiency.

There is a significant positive effect of agricultural machinery assets per mu on the scale of land operation in the control variables of the agricultural operation category. The more agricultural machinery assets a farmer has, the more advantageous his resource endowment is, and the farmer will tend to expand the scale of land operation. It is worth noting that the negative effect of average agricultural machinery assets per acre on the adoption of services is significant, and choosing farm machinery socialization services and purchasing farm machinery are two ways for farmers to mechanize maize production, and they are substitutes for each other. Farmers with sufficient farm machinery endowment tend to use their own farm machinery rather than purchasing farm machinery socialization services. The number of farm household members has a significant positive effect on the scale of land operation, and the larger the number of farm household members, the richer the farm household labor factor, and the greater the possibility of engaging in large-scale land operation.

Among the control variables of farm household characteristics, the education level of the household head has a significant negative effect on the scale of the farm business, and a significant positive effect on the adoption of services, indicating that farmers with higher education level are more likely to be employed in non-farm industries with higher labor compensation. The growth of the agricultural economy cannot be achieved without the investment of human capital, and the higher education level of the household head may improve the application level of learning and mastering new technologies, and the farmers will tend to adopt the services.

4.2. Robustness Analysis

The replacement land operation scale decision variable was used as a land operation scale behavior variable, i.e., whether maize growers transferred to land or not, with different estimation methods to perform robustness tests.

Table 4. Model estimated results of services on the scale of land operation (replacement indicators).

Variable	Probit (1)		Biprobit (2)		3SLS (3)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.335 *** (0.094)	— —	1.705 *** (0.062)	— —	0.182 *** (0.061)	— —
Behavioral variables of land operation scale	— —	0.346 *** (0.093)	— —	1.243 *** (0.012)	— —	0.175 *** (0.046)
Land rent (logarithm)	0.133 *** (0.028)	— —	0.074 *** (0.007)	— —	0.033 *** (0.007)	— —
Social capital (logarithm)	−0.144 (0.312)	— —	0.024 (0.021)	— —	−0.040 * (0.022)	— —
Commercialization rate	— —	0.265 *** (0.102)	— —	0.148 *** (0.041)	— —	0.107 *** (0.024)
Age of household head (logarithm)	0.157 *** (0.046)	0.137 (0.216)	0.215 (0.194)	0.207 (0.319)	0.001 (0.006)	0.005 (0.007)
Education level of household head	−0.152 *** (0.050)	0.065 ** (0.032)	−0.036 ** (0.015)	0.029 *** (0.011)	−0.072 *** (0.027)	0.037 *** (0.006)
Health status of household head	0.011 (0.007)	0.020 (0.068)	0.092 (0.060)	0.072 (0.098)	0.039 (0.032)	0.017 (0.029)
Number of family members (logarithm)	0.023 ** (0.011)	0.030 (0.028)	0.054 (0.153)	−0.059 (0.132)	0.016 * (0.009)	0.012 (0.033)
Share of the agricultural labor force	0.035 (0.117)	0.097 (0.220)	0.331 (0.287)	0.037 * (0.022)	0.021 (0.015)	0.022 (0.017)
Maize sown area (logarithm)	−0.020 (0.041)	0.012 ** (0.005)	−0.024 *** (0.005)	0.028 *** (0.007)	−0.023 *** (0.004)	0.037 ** (0.015)
Average agricultural machinery assets per mu (logarithm)	0.025 ** (0.011)	−0.031 *** (0.011)	0.030 *** (0.009)	−0.029 *** (0.008)	0.032 *** (0.007)	−0.008 *** (0.003)
Proportion of non-farm income	−0.13 (0.014)	0.010 (0.013)	−0.161 (0.115)	0.094 (0.083)	0.026 (0.061)	0.008 (0.049)
Land quality	−0.084 (0.062)	0.043 *** (0.005)	−0.074 (0.056)	0.021 (0.042)	0.008 (0.024)	0.011 (0.017)
Region dummy variable	Controlled	Controlled	Controlled	Controlled	Controlled	Controlled
Constant	−0.353 (1.029)	−3.397 *** (0.986)	−0.608 (0.865)	−2.312 *** (0.866)	−0.158 (0.207)	0.581 *** (0.147)
Log-likelihood value	−636.84	−646.84	−964.73
Wald χ2	823.21
prob > χ2	0.0000
Sample size	1048

(1) shows through the land operation scale equation of the single Probit model that the adoption of service decision by farmers has a positive effect on farmers’ land operation scale behavior and is significant at the 1% statistical level. Table 4 (2) shows that the Wald test of the seemingly uncorrelated association model can be applied to analyze the argument; compared with the single equation Probit estimation, both estimated results show that farmers’ decision to adopt services has a significant positive effect on farmers’ land operation scale behavior, and farmers’ land operation scale behavior also has a significant positive effect on their adoption behavior of services. This indicates that there is a two-way feedback effect between the adoption of services and the scale of land operation. The results of the three-stage least squares (3SLS) in Table 4 (3) are approximately the same as those of the seemingly uncorrelated association model, which proves that the regression results of the seemingly uncorrelated association model are robust and reliable.

The estimated results in Table 4 show that the robustness estimation was carried out using whether maize farmers transferred their land or not as a land operation scale behavior variable. The results from the land operation scale equation show that the decision of farmers to adopt services has a significant positive effect on the land transfer behavior of farmers by applying Probit, Biprobit, and 3SLS models and all pass the test of significance at the 1% statistical level. Similarly, the decision of farmers to transfer land has a significant positive effect on their adoption of services. The estimated results of the land operation scale equation show that the estimated coefficients of the decision-making variables of farmers’ adoption of services are significantly positive, indicating that the purchase of services by farmers has a significant positive impact on their decision to transfer land, which verifies the hypothesis of the previous researchers. The mechanism is that the purchase of services by farmers can effectively alleviate the labor, technology, and financial constraints faced by land operation scale.

Compared with the empirical results in Table 3, after replacing the land operation scale indicators and estimation methods, the core explanatory variable of agricultural machinery socialization services maintains the original significance and sign direction, and its coefficient estimate changes less, which indicates that the adoption of agricultural machinery socialization services by farmers has a facilitating effect on the scale of land operation, and the results have strong robustness.

4.3. Heterogeneity Analysis

Analyzed in terms of the multiple constraints on farm households’ access to services, the part-time employment situation of farm households has a certain degree of influence on both the adoption behavior of services [14] and land operation scale behavior. The essence of services is to replace internal factors with external factors, and farmers, as rational economic people, will tend to allocate labor within the household to industries with higher rates of return in order to obtain the maximum return, and then there may be a situation where the number of agricultural labor in the household is insufficient or where it is difficult to be competent in agricultural production, at which time they will tend to adopt services.

As China’s urbanization, industrialization, and modernization of agriculture and rural areas continue to accelerate, the urbanization process of part-time farm households is not completely based on the current stage of the socioeconomic situation and institutional system, and the coexistence of farming and working patterns arise from the household and individual farm household levels [15]. This study defines farm households with non-agricultural income not exceeding 10% of total income as purely agricultural households and farm households with non-agricultural income exceeding 10% of total income as part-time farm households. The effects of services on the scale of land operation were analyzed separately based on pure agricultural households and part-time farmers (

Table 5. Model estimated results of effect of the adoption of services by pure agricultural households on the scale of land operation.

Variable	OLS (1)		3SLS (2)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.293 *** (0.096)	— —	0.199 ** (0.057)	— —
Decision variables of land operation scale	— —	0.068 *** (0.023)	— —	0.099 ** (0.045)
Land rent (logarithm)	0.107 *** (0.029)	— —	0.076 *** (0.018)	— —
Social capital (logarithm)	−0.085 (0.058)	— —	−0.091 (0.112)	— —
Commercialization rate	— —	0.094 *** (0.027)	— —	0.107 *** (0.024)
Age of household head (logarithm)	−0.187 *** (0.058)	0.126 (0.158)	−0.075 ** (0.036)	0.105 (0.129)
Education level of household head	−0.103 *** (0.029)	0.096 ** (0.043)	−0.055 *** (0.010)	0.028 *** (0.007)
Health status of household head	0.145 ** (0.070)	0.076 (0.084)	0.061 (0.089)	0.047 (0.062)
Number of family members (logarithm)	0.160 ** (0.079)	0.062 (0.054)	0.083 *** (0.029)	0.032 (0.033)
Share of the agricultural labor force	0.033 (0.025)	0.066 (0.046)	0.026 (0.033)	0.041 (0.037)
Average agricultural machinery assets per mu (logarithm)	0.048 *** (0.011)	−0.029 ** (0.013)	0.027 *** (0.006)	−0.023 ** (0.011)
Proportion of non-farm income	−0.089 (0.067)	0.017 (0.022)	0.029 (0.025)	0.027 (0.019)
Land quality	−0.014 (0.030)	0.009 (0.016)	0.016 (0.022)	0.011 (0.017)
Region dummy variable	Controlled	Controlled	Controlled	Controlled
Constant	3.300 ** (1.323)	−1.684 *** (0.591)	5.179 ** (2.172)	−2.858 *** (1.102)
F Statistic	13.00	7.07
Prob > F	0.000	0.000
R2	0.347	0.203
Sample size	304

and

Table 6. Model estimated results of the adoption of services by part-time farmers on the scale of land operation.

Variable	OLS (1)		3SLS (2)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.105 ** (0.050)	— —	0.077 ** (0.034)	— —
Decision variables of land operation scale	— —	0.037 *** (0.011)	— —	0.058 *** (0.022)
Land rent (logarithm)	0.082 *** (0.012)	— —	0.051 *** (0.016)	— —
Social capital (logarithm)	−0.015 (0.030)	— —	−0.036 (0.074)	— —
Commercialization rate	— —	0.034 ** (0.015)	— —	0.090 ** (0.039)
Age of household head (logarithm)	0.126 (0.114)	0.021 (0.017)	0.030 (0.042)	0.023 (0.019)
Education level of household head	−0.034 ** (0.015)	0.032 * (0.017)	−0.026 *** (0.009)	0.024 *** (0.006)
Health status of household head	0.042 (0.036)	0.034 (0.031)	0.131 (0.107)	0.059 (0.045)
Number of family members (logarithm)	0.026 (0.017)	0.040 (0.036)	0.018 (0.027)	0.019 (0.013)
Share of the agricultural labor force	0.015 (0.011)	0.021 (0.017)	0.044 (0.050)	0.032 (0.027)
Average agricultural machinery assets per mu (logarithm)	0.031 *** (0.008)	−0.008 ** (0.004)	0.036 *** (0.013)	−0.012 ** (0.006)
Proportion of non-farm income	−0.028 (0.054)	0.012 (0.009)	0.127 (0.161)	0.084 (0.155)
Land quality	0.023 (0.028)	0.008 (0.018)	0.016 (0.014)	0.012 (0.011)
Region dummy variable	Controlled	Controlled	Controlled	Controlled
Constant	2.775 *** (0.525)	−1.354 *** (0.357)	3.142 *** (1.108)	−1.255 *** (0.449)
F Statistic	52.31	37.44
Prob > F	0.000	0.000
R2	0.479	0.268
Sample size	744

).

As the degree of non-farm transfer of rural labor continues to deepen, the development of agricultural machinery socialization services further weakens the physical strength and time constraints of the labor force, helping to alleviate the shortage of labor engaged in food production. Analyzing the situation of part-time farmers (Table 6 is based on the estimated results of pure agricultural households, while

Table 7. Model estimated results of the adoption of services by flat-land farmers on the scale of land operation.

Variable	OLS (1)		3SLS (2)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.391 *** (0.073)	— —	0.222 *** (0.061)	— —
Decision variables of land operation scale	— —	0.066 *** (0.020)	— —	0.158 ** (0.069)
Land rent (logarithm)	0.182 *** (0.064)	— —	0.051 *** (0.014)	— —
Social capital (logarithm)	−0.161 (0.171)	— —	−0.006 (0.005)	— —
Commercialization rate	— —	0.109 *** (0.034)	— —	0.104 *** (0.035)
Age of household head (logarithm)	0.169 *** (0.057)	0.047 (0.063)	0.001 (0.006)	0.018 (0.034)
Education level of household head	−0.154 ** (0.070)	0.050 ** (0.023)	−0.072 *** (0.027)	0.031 *** (0.008)
Health status of household head	−0.020 (0.017)	0.022 (0.018)	0.039 (0.032)	0.017 (0.025)
Number of family members (logarithm)	0.118 *** (0.041)	0.058 (0.053)	0.047 ** (0.023)	0.012 (0.019)
Share of the agricultural labor force	0.024 (0.025)	0.032 (0.027)	0.021 (0.015)	0.015 (0.011)
Average agricultural machinery assets per mu (logarithm)	0.041 ** (0.017)	−0.013 *** (0.004)	0.032 *** (0.007)	−0.039 *** (0.014)
Proportion of non-farm income	−0.037 (0.047)	0.013 (0.024)	0.026 (0.041)	0.016 (0.013)
Land quality	0.013 (0.015)	0.021 (0.018)	0.008 (0.014)	0.017 (0.019)
Region dummy variable	Controlled	Controlled	Controlled	Controlled
Constant	1.956 *** (0.584)	−0.949 *** (0.347)	1.963 *** (0.429)	−1.006 ** (0.419)
F Statistic	42.46	22.37
Prob > F	0.000	0.000
R2	0.377	0.185
Sample size	879

is based on the estimated results of part-time farmers), the adoption of farm machinery socialization services by both pure agricultural households and part-time farmers will have a significant positive effect on land operation scale, and the degree of the effect of pure agricultural households is higher than that of part-time farmers.

Since it is inappropriate to use the service decision variable as an exogenous variable in the land operation scale equation, and estimation using the OLS model may result in biased estimated results, this study will moreover examine the estimation of the equation using the 3SLS. The estimated results based on the 3SLS show that the adoption of services by purely agricultural households has a positive effect on the land operation scale with an estimated coefficient of 0.199 and passes the test of significance at the 5% statistical level.

The estimated results based on 3SLS (Table 6) show that the adoption of services by part-time farmers has a positive effect on the land operation scale with an estimated coefficient of 0.077 and passes the test of significance at a 5% statistical level. This indicates that the adoption of services by both purely agricultural and part-time farmers has a positive effect on land operation size for different part-time situation farmers. The adoption of services by part-time farmers can positively influence land operation scale compared to that of purely agricultural households. This promotes the organic connection between Chinese farmers and modern agricultural development during the period of agricultural structural transformation and has more far-reaching policy implications.

One study found [16] that part-time farmers will use their labor income to purchase agricultural machinery socialization services, which can alleviate the labor constraints caused by part-time employment and does not significantly affect the cultivation of food crops. Farmers’ part-time employment will make the labor force transfer to non-agricultural industries, and the adoption of agricultural machinery socialization services compensates for the “quality” and “quantity” of the labor force. The adoption of agricultural machinery socialization services provides support for the use of agricultural machinery by part-time farmers, encouraging them to scale up their land operations and facilitating the integration of small farmers into modern agricultural development.

It is worth noting that the adoption of agricultural machinery socialization services by purely agricultural households contributes more to the scale of land operation than that of part-time farmers. This may be due to the information asymmetry between farm machinery socialization service providers and adopters, where service adopters are difficult to directly observe the true level of effort of farm machinery socialization service providers, and the spatially widespread nature of field crops determines that their corresponding agricultural labor also has to be carried out in a larger space, which makes labor supervision in agricultural production extremely difficult [17]. While agricultural machinery socialization services are essentially hired labor, hired labor is different from self-employed labor that requires certain labor supervision. Thus, the possibility of moral hazard in the adoption of agricultural machinery socialization services by part-time farmers is higher than that of purely agricultural households.

By contrast, when farmers have land transfer behaviors and make land management scale decisions, the probability of choosing agricultural machinery socialization services is greatly increased due to the fact that maize production is a seasonal operation, and the suitable operating time for each production link is relatively short, and it is difficult to complete all the links of maize production relying on internal family labor only.

Service is an effective path to realizing capital substitution for labor, and it plays an important role in both rural revitalization and agricultural modernization development. Topographical conditions are an important factor affecting the level of regional agricultural mechanization in China, and the two extreme areas of agricultural mechanization development in China are mountainous and hilly areas and plain areas, and the level of regional agricultural mechanization affects the ease of adoption of services and the degree of labor substitution by farmers. In this study, farmers were grouped into flat-land farmers and slope farmers (including slope and mountainous areas) based on the topography of the actual farmland operated by farmers at the end of the year as flat land or slope mountainous areas; the effects of services on the scale of land operation based on flat-land farmers and slope farmers were analyzed separately as shown in Table 7 and

Table 8. Model estimated results of the adoption of services by sloping farmers on the scale of land operation.

Variable	OLS (1)		3SLS (2)
	Definition of Land Operation Scale: Logarithm of Maize Sown Area
	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation	Land Operation Scale Equation	Agricultural Machinery Socialization Service Equation
Decision variables of services	0.088 (0.099)	— —	0.165 (0.124)	— —
Decision variables of land operation scale	— —	0.035 (0.043)	— —	0.146 (0.185)
Land rent (logarithm)	0.084 *** (0.030)	— —	0.043 *** (0.013)	— —
Social capital (logarithm)	0.114 (0.077)	— —	−0.042 * (0.025)	— —
Commercialization rate	— —	0.075 ** (0.034)	— —	0.095 *** (0.029)
Age of household head (logarithm)	0.029 (0.024)	0.059 (0.048)	0.035 (0.066)	0.084 (0.083)
Education level of household head	−0.024 *** (0.008)	0.039 * (0.021)	−0.011 *** (0.004)	0.009 *** (0.003)
Health status of household head	0.012 (0.008)	0.051 (0.039)	0.073 * (0.043)	0.038 (0.035)
Number of family members (logarithm)	0.029 (0.020)	0.033 * (0.020)	0.020 (0.013)	0.014 (0.011)
Share of the agricultural labor force	0.027 (0.024)	0.079 (0.083)	0.028 (0.021)	0.026 (0.024)
Average agricultural machinery assets per mu (logarithm)	0.019 * (0.010)	−0.005 ** (0.002)	0.046 *** (0.008)	−0.007 * (0.004)
Proportion of non-farm income	−0.082 (0.145)	0.021 (0.017)	0.038 (0.069)	0.019 (0.052)
Land quality	−0.019 (0.058)	0.069 (0.054)	0.012 (0.029)	0.018 (0.022)
Region dummy variable	Controlled	Controlled	Controlled	Controlled
Constant	2.769 *** (0.730)	1.597 *** (0.566)	−0.364 (0.742)	0.735 *** (0.175)
F Statistic	33.33	13.57
Prob > F	0.000	0.000
R2	0.672	0.419
Sample size	169

.

Agricultural machinery socialization services are generally more sensitive to topographical features, and in terms of cultivated land topography, the effect of services on land operation scale is significantly better for farmers on flat land than for farmers on sloping land.

For the plain area suitable for mechanical cultivation, the adoption of agricultural machinery socialization services by farmers can increase the scale of land management; at the same time, the decision of land-scale management is also an important factor limiting the development of agricultural machinery socialization services, and the open and flat terrain in the plain area is suitable for the use of high-power agricultural machinery for cultivation. From the point of view of matching terrain characteristics and farm machinery power, the adoption of high-power farm machinery by flat-land farmers is favorable to the scale of land operation; however, most of China’s current agricultural operations are still dominated by small-scale farmers, and individual farmers do not have the economic conditions and operating conditions to directly purchase high-power farm machinery on their own, and if the small-scale farmers purchase their own inexpensive small-power farm machinery, it will also directly increase the farmer’s. If small-scale farmers buy their own cheap and small-power farm machinery, it will also directly increase their production costs, which may reduce the scale of their land management. The above dilemma may be the reason why farmers in the plains have adopted the socialized agricultural machinery service to realize the operation of high-power agricultural machinery.

Based on the three-stage least squares model of the land operation scale in Table 7 (2), it can be seen that the adoption of agricultural machinery socialization services by flat-land farmers has a positive effect on the land operation scale, with an estimated coefficient of 0.222, and the estimated results pass the 1% significance test. The plain area is suitable for agricultural machinery operation, and the application of services significantly reduces the labor time of the labor force of flat-land farmers engaged in corn production, and the labor force can have time to invest in non-agricultural industries, and there is a part-time transfer of the labor force. The development of services provides sufficient motivation for flat-land farmers to expand the scale of land management and optimizes the conditions for large-scale land management.

Compared with flat land, the infrastructure, such as mechanized roads in the sloping mountainous land, is weaker, and it is more difficult for farm machinery to operate. Moreover, based on the three-stage least squares model of the land operation scale in Table 8 (2), it can be seen that the adoption of farm machinery socialization services by farmers in sloping land does not have a significant effect on the land operation scale. This may be due to the fact that the increase in land slope inhibits the realization of farm machinery socialization services in the maize production process of farmers; sloping farmers increase the difficulty of unified production operations and the use of large-scale machinery, which is not conducive to the scale of operation, thus reducing the probability of farmers adopting farm machinery socialization services. It may also be because the data are cross-sectional and only represent the original terrain state, which does not reflect the modifications people made to the terrain during this part of the study, thus affecting its significance, but this does not negate the constraining effect of sloping land on the adoption of farm machinery by farmers.

5. Discussion

Agricultural machinery socialization services are an important part of the socialized agricultural service system and have important practical significance in promoting the scale of land operation. Agricultural machinery socialization services refer to the machinery operation services in agricultural production provided by agricultural machinery service organizations for farmers. It specifically includes tillage, seeding, plant protection, irrigation and drainage, and harvesting links [18]. The agricultural machinery socialization service in the latter part of the article is replaced by service, which is hereby stated. The service is a concept with Chinese characteristics. Since 1983, when the concept of “social services” was first introduced in Chinese Central Document No. 1, it has been closely associated with the Chinese agricultural reform and policy discourse. The concept of agricultural services is directly used in some international academic literature [19,20]. However, Chinese farmers operate on a small scale with a low level of division of labor, making it difficult to complete the agricultural production aspect of the social division of labor [21]. Therefore, rapidly developing agricultural services can explain the reference to agricultural services in the context of “socialization” in the Marxist-guided political documents of the Chinese Communist Party. Agricultural services are commonly used not only in China but also in other countries. In Indonesia in Asia, when faced with a general increase in real wages in both the agricultural and non-agricultural sectors, a significant number of farmers resorted to agricultural services and rented land to increase the scale of their farming operations, thus inducing agricultural machinery to replace labor [22]. Agricultural services have also become increasingly common in Africa, firstly because wages for performing agricultural labor have been rising, making the opportunity cost of agricultural production higher and negatively affecting agricultural productivity and smallholder welfare [23]; secondly, because they can provide timely and affordable mechanization services to farmers who cannot afford agricultural machinery. Since 2007, Ghana has been providing subsidized agricultural machinery to individual farmers and private enterprises to provide agricultural services to smallholder farmers across Ghana, with its agricultural services focused on land preparation services and tillage services [24]. The Nigerian government invests in the machinery needed for mechanized agricultural operations to reduce the cost of agricultural machinery operations and promote the adoption of agricultural services by smallholder farmers [25].

Based on China’s national agricultural situation of small farmers in a large country, the development of the country’s services can help make up for the small and scattered scale of farmers’ family operations, drive the specialized division of labor in maize production, and then realize the transformation of China’s grain production from traditional small-scale farmer operations to modern large-scale operations and improve the degree of intensification and specialization of agricultural-scale operations [26,27]. However, with the expansion of the land operation scale, many countries and regions have experienced a gradual decrease in agricultural productivity. Further, studies have more often attributed the factors influencing the scale of farmers’ land operations to the property rights system, part-time employment [28], human capital [29], differences in farmers’ household resource endowment [30], etc. Only relatively few scholars have considered the impact of services on the scale of land operations.

In fact, the development of services is an inevitable choice to actively promote moderate-scale land operations in China. At the same time, services are also the basic guarantee of land operation scale. Services arise from the combination of the agricultural division of labor and specialization and influence the scale of operation of specialized agricultural production [31]; services discover the “intersection” between farmers and economies of scale, and promote the efficient allocation and utilization of social resources, which also improves agricultural productivity [32,33]. The role of services in the land operation scale is fully affirmed by academics, and Jiang et al. [34] conducted an empirical study based on village-level survey data and concluded that village-level irrigation and drainage services, pest control services, and agricultural machinery tillage services have significant positive effects on land operation scale. In addition, vigorously fostering various types of service organizations can promote the development of land operation scale [35].

Based on Chinese agricultural development practice, it is concluded that although the agricultural moderate-scale operation of land transfer has achieved some success, it is only a sufficient and unnecessary condition for improving agricultural production efficiency, and moderate-scale operation of socialized services of agricultural machinery is the sufficient and necessary condition for improving agricultural production efficiency. The progress of agricultural science and technology enhances the moderate-scale threshold for agricultural operation subjects, and the improvement of socialized services of agricultural machinery provides support and guarantee for the intensive, specialized, and organized in-depth development of moderate-scale agricultural operation. With regard to the relationship between operation scale and socialized services, it was found that the expansion of farmers’ rice cultivation scale significantly promotes the adoption of labor-intensive farm machinery socialized services, probably because farmers with larger rice field sizes face more severe labor supply constraints [36], especially under the classical analysis paradigm based on price and cost. Only the scale of farm machinery socialized services is sufficiently large enough to reduce the average operating cost of machinery per mu, which will motivate many small farmers to actively adopt farm machinery socialization services [37]; farmers’ expansion of land operation scale will motivate them to adopt farm machinery and technology, and there is a positive impact of moderate land operation scale on farm machinery socialization services [38,39]. However, there are also findings supporting an inverted U-shaped relationship between land operation scale and farmers’ decision to adopt socialized services, but the critical constraint between land operation scale and socialized services will increase if we consider the further improvement of socialized service division of labor and the decrease in transaction costs, so that for Chinese small farmers whose land operation scale is generally small, an increase in their land operation scale will help them to adopt socialized services [40].

6. Conclusions

In order to examine the impact of agricultural machinery socialization services on land operation scale, this study, after systematic theoretical analysis and the derivation of the farmer production model, shows that the adoption of agricultural machinery socialization services by farmers alleviates the financial, technological, and labor constraints they face, and is conducive to the expansion of land operation scale by farmers. Subsequently, based on the survey data of maize farmers in 13 provinces in China in 2019, the OLS model, three-stage least squares, seemingly uncorrelated Biprobit linkage model, and Probit model were used to comprehensively analyze the effect of farmers’ adoption of agricultural machinery socialization services on land operation scale, and it was concluded that the adoption of agricultural machinery socialization services by farmers has a significant role in promoting the scale of land operation. The research hypothesis was verified. In addition, the adoption of agricultural machinery socialization services by both purely agricultural households and part-time farmers has a significant positive effect on their land operation scale, but the intensity of the effect is higher for purely agricultural households. Terrain condition is a key factor influencing the operation of agricultural machinery, and the adoption of agricultural machinery socialization services by farmers on flat land has a positive effect on the scale of land operation, while the adoption of agricultural machinery socialization services by farmers on sloping land does not have a significant effect on the scale of land operation.

In the context of China’s accelerated urbanization process, “who is going to farm the land” and agricultural comparative interests and other issues are becoming increasingly prominent; agricultural machinery socialized services are the external market for labor, machinery, as well as other factors of production revitalization and effective use. In essence, the scale of land operation is the result of the allocation of labor factors and land resources by farmers according to their own comparative advantages. The scale of land management and services both influence each other and work together.

At present, the government should further improve the service system, and make up for the shortcomings in the service system, in order to better promote the transformation of the farmers’ economy to modern-scale agriculture. In the agricultural extension department, to make full use of the service, it is necessary to strengthen the service system construction with all kinds of scientific and technological elements in the form of socialized services being added to the agricultural production link at the same time, to encourage farmers to exchange land and other forms of transfer to achieve the moderate scale of the farmers’ operating land and to ensure that it is centralized and continuous. This would reduce the service required for the land obstacles arising from the transaction costs.

In practice, there are still problems such as an unsound service system, low level of science and technology, poor service capacity, narrow service coverage, and mismatch between service supply and demand, which make it difficult to alleviate the constraints faced by farmers in expanding the scale of land operation in terms of capital, technology, and machinery. Therefore, the government should strengthen the construction of the service system to better promote the transformation and development of the traditional agricultural management mode to specialization, scale, and modernization.

From the development history and experience of developed countries, it can be found that land operation scale is an important initiative for modern agricultural development. Although the level of economic development and resource endowment conditions vary from country to country, both the United States and Canada, which are rich in land resources, and Japan and South Korea, where there is a tense relationship between people and land, have continuously introduced policies to promote land operation scale. Since China’s 1978 implementation of the household contract responsibility system of land resources equally divided in each household, and the different fertility of the land cut distribution to make it as fair as possible; however, it has led to increased land fragmentation, and farmers’ land management scale is small. Therefore, unlike the large-scale agriculture in Europe and the United States, and other countries, small farmers in China are still farmers in general and improve the socialized service system of agricultural machinery oriented to small farmers in order to seek an appropriate scale of operation and to promote the organic convergence of small farmers and modern agriculture. This can be achieved through the establishment of professional cooperatives and other means of uniting small farmers who do not participate in land transfers, increasing the accessibility and economic effectiveness of the services available to small farmers and promoting greater acceptance of agricultural machinery socialization services by farmers to make up for the shortcomings of the small and dispersed scale of farmers’ household operations, in order to realize the scaling up and modernization of agriculture.

7. Limitations of the Study

Finally, it should be noted that, due to the limitations of paper length and data availability, this paper fails to examine the impact of socialized agricultural machinery services on the scale of land operation in a disaggregated manner, which is a shortcoming of this paper.

Because this study focuses on the effect of services on the scale of land operation, it examines how to influence the scale of land operation through services. However, the effects of services on other aspects, such as farmers’ welfare and agricultural production performance, are not considered, which will be a direction for future research.