Front. Sustain. Food Syst. Frontiers in Sustainable Food Systems Front. Sustain. Food Syst. 2571-581X Frontiers Media S.A. 10.3389/fsufs.2025.1735427 Original Research How effective is the quality constraint mechanism of agricultural socialized services?—Evidence from 1,138 farmers in Shandong, China Yu Lianghong 1 Writing – original draft Zheng Shan 2 Writing – review & editing Zhang Ying 3 * Conceptualization Writing – review & editing Resources School of Advanced Agricultural Sciences, Peking University, Beijing, China International Business School, Qingdao Huanghai University, Qingdao, China School of Management, the Institute of Marine Development, Ocean University of China, Qingdao, China Correspondence: Ying Zhang, yzhang@ouc.edu.cn 08 01 2026 2025 9 1735427 30 10 2025 30 11 2025 30 11 2025 Copyright © 2026 Yu, Zheng and Zhang. 2026 Yu, Zheng and Zhang https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Constraint mechanisms of agricultural socialized services have been widely established, but how effective is the quality constraint mechanism remains unclear. However, the existing literature has not provided an answer to this question. Therefore, this study theoretically analyzes the effect of quality constraint mechanism on agricultural socialized service and propose the hypothesis, then applies MESR model and 1,138 farmers survey data, empirically examines the impact of constraint mechanism. The study finds that: (1) The constraint mechanism has a positive impact on the quality of agricultural socialized services. Based on contract theory, the intermediary mechanisms focus on two aspects: implicit contracts of trust within familiar societies and explicit contracts of risk-sharing. (2) Under the consideration of counterfactual hypothesis, if farmers with low constraint do not construct constraint mechanisms, the quality of agricultural socialized services will decline by 22%. If farmers with high constraint do not construct constraint mechanisms, the quality of agricultural socialized services will decline by 33%. (3) The results of heterogeneity analysis show that the effect of quality constraint mechanisms on different farmers is in the order of part-time farmers, small-scale pure farmers, and large professional farmers. The influence of quality constraint mechanism on different crops is in the order of vegetable, corn, and wheat. Accordingly, the government should guide farmers in signing service contracts, optimize the service standards system. This study can provide both theoretical and practical guidelines for the construction of constraint mechanisms of agricultural socialized services.

 agricultural socialization agriculture constraint mechanism endogenous switching regression model service quality The author(s) declared that financial support was received for this work and/or its publication. This study was supported by National Natural Science Foundation of China (72503011) and China Postdoctoral Science Foundation (2024M76012). section-at-acceptance Agricultural and Food Economics Introduction

In a principal–agent relationship, the agent may have its own interests and motives to take advantage of the information asymmetry, harming the interests of the principal, thus giving rise to the principal–agent problem (Forges et al., 2024). Agricultural socialized service is a typical principal–agent relationship between the farmers and service organizations. To save operation time and reduce fuel consumption and machinery depreciation, service organizations will take opportunistic behaviors (Xiao and Fang, 2025), such as replacing deep plowing with shallow plowing, sowing seeds at uneven densities, or allowing excessive leakage during harvest. These practices result in poor quality of agricultural socialized services and seriously harms the interests of farmers (Sui and Yu, 2025). In practice, constraint mechanisms are constantly being established to prevent the problem of low quality of agricultural socialized services. The contract is an important constraint tool in the principal–agent relationship. Its clear and explicit terms can regulate the behavior of both parties and reduce potential disputes. If there are loopholes, unclear terms or lack of completeness in the principal–agent contract, it may lead to disputes between the two parties in the implementation process (Luo and Su, 2023). Especially in the case of a complex principal–agent relationship that involves multiple interests and multi-task execution, the importance of a complete contract is more prominent. Currently, the agricultural socialized service contract clearly agrees on the normative code of conduct and work standards that the service organization should abide by Xu et al. (2022). It establishes a penalty mechanism for breach of contract, which punishes the service organization's misconduct, to ensure the quality of agricultural socialized services. So, how effective is the quality constraint mechanism of agricultural socialized services remains unclear. Has it effectively improved the quality of agricultural socialized services? This is the key scientific question that this study tries to answer.

According to the classical principal–agent theory, the principal–agent problem can be solved by means of incentives mechanisms (Dirzka and Acciaro, 2021; Monios and Fedi, 2023), supervision mechanisms (Yu et al., 2023), and constraints mechanisms (Liang Y. et al., 2025; Luo et al., 2025). However, in the practice of agricultural socialized services, it is difficult to implement incentives, because farmers are often unwilling to give extra agricultural surplus to motivate service organizations. Moreover, the supervision mechanism is also difficult to implement, because farmer's supervision capacity is weak (Sui and Yu, 2025). Farmers are more willing to choose to constrain the agricultural socialized service behaviors of service organizations. But, there are few studies on the constraint mechanism of agricultural socialized services. So, the existing studies do not provide a reasonable answer to the scientific question of the effect of quality constraint mechanism on agricultural socialized services.

Based on the above practice issues and theoretical gaps, this study theoretically elaborates the impact of constraint mechanisms on the quality of agricultural socialized services and proposes a theoretical hypothesis, applies the multinomial endogenous switching regression (MESR) model, empirically tests the impact of constraint mechanisms on the quality of agricultural socialized services based on the survey data of 1,138 farmers, then carries out the analysis of heterogeneity and robust test. The innovations of this paper are as follows.

First, research on agricultural socialized services typically relies on theories of division of labor, principal–agent relationships, property rights, and transaction costs. In contrast, this study employs contract theory to analyze the effectiveness of constraint mechanisms in agricultural socialized services and their mediating variables. Second, existing studies focus on the cost-saving and income-generating effects of agricultural socialized services, but few studies pay attention to the quality issue of agricultural socialized services. This study focuses on how to improve the service quality and conducts theoretical and empirical analyses on the effectiveness of the constraint mechanism of agricultural socialized services to realize the innovation of research content. Third, academics commonly use propensity score matching (PSM) and difference in difference (DID) methods to solve the endogeneity, but the PSM method cannot correct the selectivity bias of unobservable, and the DID method cannot be used for cross-sectional data. In this study, we adopt the MESR model, which eliminates the effect of endogeneity by switching endogenous variables into exogenous ones, and also enables parameter estimation in counterfactual situations, making the estimation results more accurate and robust.

 Literature review Research on the quality of agricultural socialized services

The literature on agricultural socialized services is relatively abundant. Scholars have studied the antecedents of agricultural socialized services (Li et al., 2023), the development status (Zhang et al., 2022; Wu et al., 2023), and the economic consequences (Qing et al., 2023), which include the impacts of agricultural socialized services on crop yields, and the transfer of agricultural labor (Yang and Li, 2023), farm household income (Baiyegunhi et al., 2019; Sang et al., 2023), and the impact of agricultural green technology adoption (Lin et al., 2022; Shi et al., 2024). Research indicates that agricultural socialized services exert significantly different impacts on various types of farming households (Mi et al., 2020; Yang and Li, 2023) and different land management scales (Yang and Zhang, 2023). Similarly, the effectiveness of agricultural socialized service constraints is influenced by multiple factors. The same constraint measures may yield varying outcomes across different crops and different service link, necessitating a categorized analysis.

However, there are insufficient research on the quality of agricultural socialized services. The quality of these services reflects farmers' subjective assessment of service organization (Sui and Yu, 2025). Service providers often fail to deliver satisfactory services to clients. In agricultural socialized services, service organizations may exploit information asymmetry to engage in opportunistic behavior, thereby creating service quality risks (Xiao and Fang, 2025). Agricultural risks cause serious hazards (Ding and Xu, 2023; Bai et al., 2024; Chinnannan et al., 2024); therefore, preventing and controlling the risk is particularly important (Hui and Yong, 2023; Mastenbroek et al., 2024). So, constraining service organizations' conduct is particularly crucial. Research on service quality in other industries provides insights for this study. Examples include taxi service quality and customer satisfaction (Kester et al., 2024), and e-commerce customer service quality (Zhang et al., 2024). However, these studies did not validate the effectiveness of quality constraint mechanisms. It should be noted that while agricultural socialized service constraint mechanisms can enhance service quality, the trade-off between constraint costs and benefits remains an important research topic.

 Research on the constraint mechanisms

The principal–agent problem can be solved by means of incentives, supervision, and constraints mechanisms. Existing research is also carried out from these three aspects. First, the principal–agent problem is solved through incentives mechanisms (Schosser, 2019; Dirzka and Acciaro, 2021; Monios and Fedi, 2023). For example, public transport authorities in the UK use incentives to get bus operators to provide more and better public transport services (Mctigue et al., 2020). Superiors use incentives mechanisms to make subordinate units disclose their minimum operating costs (An et al., 2023). Stimulate agents to increase their efforts through reasonable incentive contracts to increase the proportion of high-quality primary products in the supply chain (Liu S. et al., 2024). A reasonable benefit distribution mechanism can help to solve the multiple principal–agent relationship in China's collective construction land market (Yan et al., 2021). Incentivize employees with weak currencies to achieve work goals unrelated to wages (Corgnet et al., 2018).

Second, solving principal–agent problems through supervision mechanisms (Viaggi et al., 2009; Yu et al., 2023). For example, the central government can address the inaction of local governments in environmental governance through environmental protection inspections (Dong et al., 2024) and rent-seeking behavior of local polluting firms (Lin and Xie, 2023). Alternatively, incentives and supervision mechanisms can be used at the same time. Studies have shown that incentives and supervision mechanisms developed by the central government can increase the motivation of local governments in responding to rainstorm warnings, and that the two mechanisms can be substituted for each other (Ma et al., 2024).

Third, the principal–agent problem is solved through the constraint mechanism (Zhang and Li, 2025; Hao et al., 2025). Beyond the agricultural sector, scholars have conducted extensive research on constraint mechanisms, providing valuable insights and inspiration for this study. For example, constraint mechanism in taxi and ride-hailing services (Laeeque and Ali, 2025; Jaydarifard et al., 2025), digital platform economy (Zhang and Liu, 2025), service outsourcing industry (Liang W. et al., 2025), collaborative project delivery (Nwajei et al., 2022), bottom-up public complaints (Dong et al., 2024), and enterprises' rent-seeking behavior (Lin and Xie, 2023). In the agricultural sector, constraint mechanisms can take the form of formal constraints such as institutional regulations or contracts. For instance, beef farming contracts constrain farmers' environmental behaviors (Liang Y. et al., 2025), while corporate low-carbon disclosure systems impose constraints (Zhu et al., 2025). They can also take the form of informal constraints such as social norms and education or training. For example, the role of social norms in constraining farmers' low-carbon behaviors (Feng et al., 2025) and the use of ecological protection education to constrain college students' environmental behaviors (Hou et al., 2025; Luo et al., 2025). Research indicates that agricultural cooperatives should be regulated to prevent them from becoming tools that serve only a select few members (Liu G. et al., 2024). So, the principal can constrain the agent by constructing an optimal contract (Banerjee and Chakraborty, 2023), or constrain the agent by other technical means (Nwajei et al., 2022). Furthermore, the effectiveness of incentives and constraints may differ, raising the question of which approach is more effective—a topic warranting further research. Li and Zhao (2025) found that both incentives and constraints mechanism can motivate farmers to protect farmland, but incentives proved more effective than constraints. Therefore, determining whether incentives or constraints mechanism are more effective in agricultural socialized services requires additional research.

In the practice of agricultural socialized services, it is difficult to implement incentives and supervision mechanisms such as the share of excess production. Farmers are often unwilling to give extra agricultural surplus to motivate service organizations to refrain from opportunistic behavior, and their own supervision capacity is weak (Sui and Yu, 2025). Farmers are more willing to choose to regulate the agricultural socialized service behaviors of service organizations through service standards, service evaluation and other constraint mechanisms, so as to prevent the problem of low quality of agricultural socialized services. However, there are few studies on the constraint mechanism of agricultural socialized services.

 Research gaps

The existing literature on agricultural socialized services and their regulatory mechanisms provides a foundation for this study, but there remain several areas for improvement:

First, risk and return go hand in hand, and most existing studies focus on the cost-saving and income-generating effects of agricultural socialized services, but few studies pay attention to the quality issue of agricultural socialized services.

Second, research on agricultural socialized services typically relies on theories of division of labor, principal–agent relationships, property rights, and transaction costs. However, research analyzing the quality of agricultural socialized services based on contract theory remains scarce.

Third, academics commonly use methods such as PSM and DID to solve the problems of selection bias and endogeneity, but the PSM method cannot correct the selectivity bias of unobservable, and the DID method cannot be used for cross-sectional data.

 Theoretical analysis

In the principal–agent relationship, the inconsistency of interests between the principal and the agent is the root cause of the principal–agent problem (Forges et al., 2024). The inconsistency of interests refers to the dissimilarity or mismatch of the interest objectives of both farmers and service organizations. In agricultural socialized services, farmers pursue agricultural output maximization, while service organizations pursue service profit maximization. When the interests of the principal and agent are inconsistent, the principal–agent problem will inevitably arise (Bogle and Van Kooten, 2013; Vining and Richards, 2016). Generally, principal–agent problems can be solved by constructing constraint mechanisms. Based on contract theory, this study examines the impact of constraint mechanisms on the quality of agricultural socialized services and the intermediary mechanisms involved, focusing on two aspects: implicit contracts within familiar societies and explicit contracts of risk-sharing.

Implicit contracts constraint of trust in rural familiar societies

The implicit constraint mechanism of China's rural familiarity society enhances the quality of agricultural socialized services. In a familiarity society based on personal connections, farmers tend to have stronger trust in service organizations. The lack of trust will reduce the quality of agricultural socialized services (Xiao and Fang, 2025). Currently, the main constraint mechanism of agricultural socialized service comes from the rural familiarity society, service organizations will not reduce the quality of agricultural socialized service for the sake of face (Qiu et al., 2023), and it is difficult for service organizations to carry out the business of agricultural socialized service in the familiarity network if they vocalize the dispute with the farmers. Thus, to a certain extent, service organizations can be restrained from agricultural socialized service behavior and improve the quality of agricultural socialized service. In the process of building a national unified market for agricultural socialized services (Ma et al., 2025), it is necessary to transform from relationship governance to contractual governance (Banerjee and Chakraborty, 2023). Currently, local government agricultural management departments and industry associations are releasing model contracts, formulating service standards for agricultural socialized services, creating a directory of agricultural socialized service organizations, and constructing a credit system for agricultural socialized services, etc., to form a new contractual service constraint mechanism, which will indirectly improve the quality of agricultural socialized services.

 Explicit contracts constraint of risk-sharing in agricultural socialized services

The explicit constraint mechanism of risk-sharing enhances the quality of agricultural socialized services. In addition to implicit constraints, quality risk-sharing is an explicit constraint mechanism for farmers and service organizations. On the one hand, the quality risk of agricultural socialized service is only borne by a single subject (Sui and Yu, 2025), and it is very easy for farmers or service organizations to unilaterally withdraw from the problem. The scientific and reasonable risk-sharing mechanism realizes the diversification of quality risks, significantly reduces the “worries” of the cooperation between service organizations and farmers, strengthens the confidence of each subject to steadily and continuously carry out the value co-creation of agricultural socialized services (Wang and Huan, 2023). On the other hand, the perfect risk-sharing mechanism realizes the reasonable sharing of quality risks among different subjects, maintains the stability of the gain-sharing order, and consolidates the interest relationship between service organizations and farmers. In practice, if the quality risk is more shared by farmers and less shared by service organizations, it will reduce service quality (Sui and Yu, 2025; Xiao and Fang, 2025). Accordingly, this study proposes the following hypothesis.

H: The constraint mechanism positively affects the quality of agricultural socialized services.

 Empirical research design Model construction

The MESR model is applicable when individuals are categorized into two states, each with continuous outcome variables. Additionally, the model addresses endogeneity issues, where individuals' choice behaviors and outcome variables exhibit bidirectional influence. In this study, the decision to establish a constraint mechanism is divided into two states, and the constraint mechanism and service quality are mutually causal. In decision-making contexts, farmers may face multiple alternative decisions and the decisions may affect each other, when the use of traditional single-item regression models may arrive endogenous. Therefore, the MESR model is suitable for this research. For this reason, this study adopts the MESR model proposed by Deb and Trivedi (2006) to analyze, the core idea of the model is to introduce potential choice equations and observed choice equations. The specific steps of the MESR model are as follows.

First, consider the farmers' decision-making equation. Because the Probit model is applicable to ordered categorical variables, and the explanatory variables in this study are measured using a five-point Likert scale, this model is used for the analysis (see Equation 1). In the formula, probit denotes the probability of formulating service constraint mechanism, constraintn denotes the explanatory variables, and Zn is the control variables. δ′ and m are the parameters to be estimated. μn is the random error.

probit=δ′Zn+m×constraitn+μn

Second, consider the service quality determining equation. The equation for determining the quality of agricultural socialized services with a constraint mechanism (experimental group) is shown in Equation 2. The equation for determining the quality of agricultural socialized services without the constraint mechanism (control group) is shown in Equation 3, where qualityna and qualitynb denote the quality levels of agricultural socialized service in the experimental group and the control group, respectively. constraintna and constraintnb denote the influence variables of service quality constraint. εna and εnb are the random error terms.

qualityna=βa′constraitna+εna qualitynb=βb′constraitnb+εnb

Third, the treatment effect assessment equation. To objectively assess the impact of the constraint mechanism on the quality of agricultural socialized services, based on the counterfactual assumption, this study calculates the expected value of agricultural socialized service quality and compares the expected value of service quality with the actual value to measure the average treatment effect of agricultural socialized service quality. The expected value of agricultural socialized service quality in the case of a constraint mechanism is shown in Equation 4, and the expected value of agricultural socialized service quality in the case of an unconstrained mechanism is shown in Equation 5. Similarly, the service quality expectations faced by farmers with constraints in place in the absence of constraints are shown in Equation 6, and those faced by farmers without constraints in the presence of constraints are shown in Equation 7.

E[qualityna|Dn=1]=βaconstraitna+φμaλna E[qualitynb|Dn=0]=βbconstraitnb+φμbλnb E[qualityna|Dn=0]=βaconstraitna+φμaλna E[qualitynb|Dn=1]=βbconstraitnb+φμbλnb

Accordingly, the mean of average treatment effect of the experimental group ATTn and the mean of average treatment effect of the control group AUTn can be calculated. Among them, the average treatment effect of agricultural socialized services quality with constraint mechanisms is the difference between Equation 7 and Equation 4, and the calculation results of ATTn are shown in Equation 8. Similarly, the mean treatment effect of agricultural socialized services quality without constraints is the difference between Equation 6 and Equation 5, and the calculation results of AUTn are shown in Equation 9.

ATTn=(βa′-βb′)constraitna+(φμa-φμb)λna ATUn=(βb′-βa′)constraitnb+(φμb-φμa)λnb Selection of variables

(1) Core explanatory variable. This study selects the constraint mechanism as the core explanatory variable. The constraint mechanism is the effectiveness degree of quality constraints on agricultural socialized services. Due to the lack of direct objective metrics for the constraint mechanism, the core explanatory variable is characterized by a five-point Likert scale, where 1 = very low, 2 = low, 3 = average, 4 = high, and 5 = very high.

(2) Explained variables. Based on the theoretical analysis, the explained variable selected in this study is agricultural socialized service quality. The quality of agricultural socialized services refers to the level of service provided by service organizations and the satisfaction of farmers when farmers entrust the management of agricultural production operations to these organizations. Service quality encompasses pre-production, mid-production, and post-production service quality. Pre-production service quality includes the quality of agricultural input supply, agricultural machinery supply, and farmland preparation. Mid-production service quality includes the quality of sowing, pest and disease control, and fruit harvesting. Post-production service quality includes the quality of product drying, storage, and processing. Farmers evaluate their satisfaction with the service quality provided by these organizations. In addition, to further explore the influence of the constraint mechanism on the service quality of different production links, the pre-production service quality, the mid-production service quality and the post-production service quality of farmers are selected as the explained variables. The quality of agricultural socialized services is based on farmers' subjective evaluations, lacking direct objective measurement indicators. So, the quality of each type of service is characterized by a five-point Likert scale, where 1 = very bad, 2 = bad, 3 = average, 4 = good, and 5 = very good.

(3) Mediating and moderating variables. Based on theoretical analysis and drawing on relevant research, this study selects trust and contract as mediating variables and farm scale as a moderating variable. The trust variable indicates whether farmers trust the service organization. If the service organization is local to the village, trust = 1; if it is from outside the village, trust = 0. The contract variable indicates whether farmers have signed contracts with service organizations, specifying service standards and risk-sharing provisions. If farmers sign a contract with service organizations and share risks, then contract = 1. If no risk-sharing contract is signed, then contract = 0. Farm scale variable is the total area of cropland contracted by farmers. The mean-centered transformation was applied to the farm size variable to reduce multicollinearity and enhance the interpretability of regression coefficients.

(4) Control variables. At the individual level, drawing on related studies (Zhou et al., 2025), this study selects gender, age, education, and political profile as control variables. Among them, the variables gender, age, and education indicate the farmer household head's gender, age, and education level. Political profile indicates whether the household head is a member of the Communist Party of China (CPC) or not, and if so, it takes the value of 1; otherwise, it takes the value of 0. At the household level, according to related research (Zhang, 2024), this study selects health, position in the village, number of people working in agriculture, number of people working in the labor force, and participation in cooperatives as the control variables. Among them, health is the health status of farm household members. The number of farmers and workers indicates the number of laborers in the household who are working at home and working outside, respectively. Cooperative participation is a dummy variable for whether farmers participate in agricultural cooperatives. At the farmland level, drawing on the literatures (Yang and Zhang, 2023; Liao et al., 2025), soil fertility, irrigation convenience, plot dispersion are selected as the control variables. Among them, soil fertility and irrigation convenience are characterized by a five-point Likert scale, and plot dispersion is the number of plots owned by farmers. At the area level, plain, suburban, and distance are selected as control variables in this study. Among them, plain and suburban are the topographic features and economic characteristics of the area where the farm household is located, which are defined as dummy variables. Township distance is the distance of the farm household location from the township government office, which is characterized by a five-point Likert scale.

We conducted a multicollinearity test on the regression variables and found no collinearity among them, meaning the variables are independent. The definitions of the variables are shown in Table 1.

Variables and their measurements.

Type Variable Indicators and their calculation methods/units
Explained variable Quality Quality level of agricultural socialized services, Likert scale
Explanatory variable Constraint Effectiveness of constraints on socialized services, Likert scale
Mediating variable Trust Whether farmers trust the service organization
Contract Whether farmers have signed contracts with service organizations
Moderating variable Scale Area of cultivated land contracted by farmers' families (acres)
Control variable Sex Sex of head of household, Male = 0, Female = 1
Age The age of farmer household head
Education Years of education completed by the head of household (year)
Party Whether the household head is a member of the CPC, No = 0, Yes = 1
Cadres Members of the family serving in the village, None = 0, Yes = 1
Farmers Total number of family members working in agriculture (persons)
Workers Total number of family members working (persons)
Health Health status of family members, five-point Likert scale
Participation Farmers' participation in agricultural cooperatives, no = 0, yes = 1
Fertility Fertility of farmers' contracted cropland, five-point Likert scale
Irrigation Ease of access to farm irrigation, five-point Likert scale
Plot Number of plots of farmland contracted by farmers (plots)
Plain Whether the administrative village is located in a plain, No = 0, Yes = 1
Suburban Whether the administrative village is located in suburban, No = 0, Yes = 1
Distance Distance of farm households from township seat, five-point Likert scale
Data sources

(1) Research area. Shandong is a large agricultural province in China, and agricultural socialized services have achieved remarkable results, creating a “Qilu Model” of agricultural socialized services that is highly influential and exemplary in the whole country. Therefore, this study selects Shandong Province of China as the sample area, which is representative. To ensure that the sample can achieve a relatively uniform distribution within the overall geographical scope of Shandong Province, so as to more accurately and comprehensively reflect the actual situation of the province's agricultural socialized services, the research team selected 12 cities as study area from the east, middle and west, as well as the north and south of Shandong Province. After rigorous and meticulous screening, 20 counties with considerable business volume of agricultural socialized services were identified as sample survey areas. It contains Jimo, Jiaozhou, Pindu, Tengzhou, Wulian, Wenshang, Guangrao, Lizhen, Feixian, Linshu, Leiling, Qihe, Boxing, Huimin, Donge, Shanxian, Jiyang, Zhangqiu, Gaomi, Qingzhou.

(2) Sample selection and questionnaire. The stratified random sampling method was used in the specific sample selection process. First, in each selected sample county, three sample townships were reasonably selected on the basis of their geographical size and agricultural development characteristics. Subsequently, for each sample township, three representative administrative villages were further selected. Finally, in each sample village, the principle of random sampling is strictly adhered to ensure the randomness and objectivity of the sample. On this basis, the research team carried out household surveys and distributed questionnaires to the sample farmers. The authors conducted a comprehensive and in-depth tracking survey of the sample farm households in two waves between 2022 and 2023. Taking fully into account the complex reality of the continuous fragmentation of farm households in the current rural society, it is very likely that some of the sampled households did not participate in agricultural socialized services in the actual survey process. To ensure the validity and relevance of the survey data, we carried out a meticulous screening of the sampled households, and decisively excluded those who did not carry out socialized services. In the first round of farmer survey conducted in 2022, we obtained 582 valid questionnaires, and 556 valid questionnaires were obtained in the year 2023. When the valid questionnaires obtained in these 2 years were summarized and counted, the total number of valid questionnaires from farmers reached 1,138.

(3) Reliability and validity tests of questionnaire. Based on the questionnaire survey data, this study conducted reliability and validity tests on the variables measured using the Likert scale. Reliability analysis was employed to assess the dependability of the sample responses. Cronbach's alpha represents the most common and widely used measurement method. Results from SPSS software indicates that the Cronbach's alpha coefficients for service quality is 0.845, confirming the reliability of the scale. Validity analysis primarily encompasses content validity and construct validity. First, regarding content validity, this study validated the relevance and completeness of the questionnaire content through expert review. Findings indicate that all indicators comprehensively cover the various dimensions of service quality and constraint. Second, construct validity was assessed through confirmatory factor analysis to verify whether the indicators effectively reflect the latent structure of service quality, ensuring the theoretical soundness of the scale. The KMO is 0.888, Bartlett's test significance level is 5%. Construct reliability (CR) value in this study is 0.879, and average variance extracted (AVE) value is 0.622. Results are shown in Table 2. These results indicate that the Likert scale employed in this study possesses high construct validity.

Discriminant validity analysis results.

Variable Question item Factor loadings T CR AVE
Quality Quality of pre-production 0.769 18.362 0.879 0.622
Quality of mid-production 0.840 20.882
Quality of post-production 0.755 17.958
Regression results and analysis Baseline regression results and analysis

To form a comparison between the service constraint mechanism in the actual situation and the counterfactual hypothetical situation, and effectively circumvent the impact of sample selection bias on parameter estimation, this study analyzes the impact of constraint mechanism by using the MESR model, which is capable of overcoming the problem of endogeneity and providing a more accurate parameter estimation. In response to the above research design, this study groups the samples based on the level of service constraints, the constraint of agricultural socialized service is less than 2.5 for the low constraint group, and the constraint of agricultural socialized service is greater than or equal to 2.5 for the high constraint group. Using STATA15 software for regression analysis, with service constraint as the core explanatory variable and agricultural socialized service quality as the explained variable, the model estimation results are shown in the table below.

(1) Results and analysis of farmer decision-making model. As can be seen from Table 3, first the Wald test results are all significant at the 1% level, indicating that the Probit regression model is valid. The goodness-of-fit test statistic R2 also indicates that the overall fit is good. In summary, it can be concluded that the model fitting results are credible. Second, the effects of household head's gender, age, education level, whether party member or not, whether village cadre or not, whether plain or not, whether suburb or not, distance from the township, number of family farmers, plot dispersion, number of family laborers, health status, cooperative participation, soil fertility, irrigation accessibility variables on the decision-making of constraints are in line with the expectations.

Estimation results of decision equation for agricultural socialized services.

Variable Low constraint group High constraint group
Coefficient Standard error Coefficient Standard error
Sex 0.069 1.201 0.109 0.189
Age 0.216 0.231 0.291*** 0.078
Education 0.335*** 0.000 0.317 0.278
Party 0.361*** 0.037 0.311*** 0.011
Cadres 0.252*** 0.019 0.272*** 0.008
Farmers 0.105*** 0.009 0.169*** 0.000
Workers −0.053 0.183 −0.112 0.184
Health 0.011 0.074 0.154 0.135
Participation 0.267*** 0.000 0.116 0.127
Fertility 0.326*** 0.064 0.397*** 0.007
Irrigation 0.146*** 0.003 −0.315 0.284
Plot 0.199*** 0.004 0.196*** 0.000
Plain 0.017*** 0.005 0.241 0.312
Suburban 0.046*** 0.001 0.096 0.084
Distance 0.023 0.035 0.316*** 0.011
Constant term 1.416*** 0.085 4.153*** 0.074
Wald X2 235.873*** 125.382***
Log Likelihood 3734.763 3972.642
R2 0.478 0.396
ρ 0.082** 0.043**
Test for difference in coefficients between groups 0.013***

**, *** indicate significant at the 5%, and 1% levels, respectively.

In addition, after regressing the samples in two groups, direct comparison of the size of the coefficients will produce bias, so this study also conducted a test for the difference in coefficients between low constraint and high constraint groups. The criterion for determining the significance of the difference in coefficients between two groups is whether the interaction term is significant. Commonly used tests are Chow test, Suest test, and Fisher's Permutation test. According to the estimated results of Chow test of the interaction term model, the P-value of the test of difference in coefficients between groups is calculated to be 0.013. The results of the Suest test and the Fisher's Permutation test are basically similar, indicating that there is a significant difference between the low constraint group and the high constraint group.

(2) Results and analysis of treatment effect estimation model. It can be seen from Table 4 that, on the one hand, the constraint mechanism in the group of low constraint has a positive treatment effect on the quality of agricultural socialized service with a significance level of 1%. The average treatment effect is 0.300, this indicates that for farmers with low levels of constraints, establishing constraint mechanisms can improve the quality of agricultural socialized services by 30%. So, the constraint mechanism has a positive effect on the quality of agricultural socialized services. The current actual service quality is 1.663. To achieve the desired effect, the constraint must be increased by at least 2 units. On the other hand, in the high constraint group, the constraint mechanism has a positive treatment effect on the quality of agricultural socialized services with a significance level of 1%. The average treatment effect is 1.093, this indicates that for farmers with high levels of constraints, establishing constraint mechanisms can improve the quality of agricultural socialized services by 109.3%. So, the constraint mechanism has a positive effect on the quality of agricultural socialized services. The current actual service quality is 4.378, which has achieved the desired effect.

Treatment effects of constraint mechanism of agricultural socialized service.

Groups A B C = A-B Quality decline rate (%) 95% confidence interval DW test
Actual quality of services Quality under counterfactuals ATT ATU
Low-constraint group 1.663 1.363 0.300*** – 22 [1.023, 2.571] 1.898
High-constraint group 4.378 3.285 – 1.093*** 33 [3.196, 4.998] 2.015

*** indicate significant at the 1% levels.

A cross-sectional comparison within the group revealed that the constraint mechanism was more effective for farmers in the high-constraint group than for those in the low-constraint group. This may be attributed to the stronger sense of contractual commitment among farmers in the high-constraint group. Therefore, the constraint mechanism should first be promoted among farmers with a higher contractual spirit. In summary, the constraint mechanism has a positive effect on the quality of agricultural socialized services, and hypothesis H is proved.

In addition, considering the counterfactual assumption, if the farmers in the low constraint do not take agricultural socialized service constraints, the estimation results show that the quality of agricultural socialized service will decrease from 1.663 to 1.363, with a decrease of 22%. It means that if farmers with low constraint do not construct constraint mechanisms, the quality of agricultural socialized services will decrease 22%. If farmers in the high constraint do not take agricultural socialized service constraints, the estimation results show that the quality of agricultural socialized services will decrease from 4.378 to 3.285, with a decrease of 33%. It means that if farmers with high constraint do not construct constraint mechanisms, the quality of agricultural socialized services will decrease 33%.

 Analysis of mediating and moderating effects

(1) Analysis of mediating effects. Based on theoretical analysis, this section examines two pathways through which the constraint mechanism influences the quality of agricultural socialized services. Using trust level and risk-sharing contracts as mediating variables, regression analysis was conducted via a mediation effect model. The results are presented in Models 1–6 of Table 5. First, regarding the trust mediation variable, the effect of constraints on trust is 0.102, with a significance level of 5%. Trust's effect on service quality is 0.139, with a significance level of 5%. The mediation effect of trust is confirmed, indicating that the constraint mechanism indeed influences the quality of agricultural socialized services through the trust pathway. This is particularly evident in cross-regional agricultural socialized services, where farmers generally trust local service organizations but exhibit lower trust in service providers outside their village or township. Consequently, non-local agricultural machinery operators typically need to find local intermediaries to connect with farmers when providing such services. Second, regarding the risk-sharing contract intermediary variable, the constraint mechanism's effect on risk-sharing contracts was 0.088, with a significance level of 10%. The impact of risk-sharing contracts on service quality is 0.114, with a significance level of 5%. The mediation effect test for risk-sharing contracts passes, indicating that constraint mechanisms do indeed influence the quality of agricultural socialized services through the risk-sharing pathway. This occurs because signing risk-sharing contracts binds the interests of service organizations to farmers, creating incentive compatibility. Consequently, service organizations will not engage in opportunistic behavior for their own benefit.

(2) Analysis of moderating effects. This study incorporates the interaction term between constraints and farm scale as a moderating variable into the model, yielding the moderation effect results shown in Model 7 of Table 5. We can know that the coefficient for the constraint variable is 0.153, significant at the 10% level. The coefficient for the interaction term is 0.095, significant at the 5% level. This indicates that farm scale exerts a positive moderating effect. Specifically, larger farm scale enhances the effectiveness of constraint mechanisms and elevate the quality of agricultural socialized services. Conversely, excessively small farms limit the efficacy of such mechanisms. This may stem from two factors: First, larger farmers typically possess stronger contractual awareness, greater social capital, and more robust claim-making capabilities, thereby exerting more effective constraints on service organizations. Second, service providers, seeking sustainable operations, are reluctant to compromise service quality for major clients.

Results of mediating and moderating effects.

Variable Mediating variable 1 Mediating variable 2 Moderating variable
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
Quality Trust Quality Quality Contract Quality Quality
Constraint 0.191*** 0.102** 0.191*** 0.088* 0.153*
Trust 0.137**
Contract 0.114**
Constraint × Scale 0.095**
Control Yes Yes Yes Yes Yes Yes Yes
Constant 4.027 4.053** 4.059*** 2.989** 3.763*** 4.733** 4.554
R2 0.265 0.301 0.298 0.334 0.411 0.389 0.277

**, *** indicate significant at the 5% and 1% levels, respectively.

 Heterogeneity analysis and robust test Heterogeneity analysis

(1) Heterogeneity of farmer types. Based on the above criteria for the division of farm household types, this study divides the sample farmers into large professional farmers, part-time farmers, and small-scale pure farmers. Then we estimate the three sub-samples separately, and obtains the results of the heterogeneity analysis distinguishing the farmer types, as shown in Table 6, Models 1–3. As can be seen from the table, the constraint mechanism has a significant positive impact on the service quality of small-scale pure farmers, the service quality of part-time farmers, and the service quality of large professional farmers. This impact is, in order of magnitude, part-time farmers, small-scale pure farmers, and large professional farmers. This may be because most part-time farmers spend the majority of their time working away from home, and lack the time to supervise service organizations. Constraint mechanisms can address the shortcomings of on-site supervision for part-time farmers.

(2) Heterogeneity of different crops. The main crops of agricultural socialized services in the research area are wheat, corn and vegetables. For this reason, this study divides agricultural socialized service crops into three categories: wheat, corn, and vegetables. Then, estimates the three sub-samples separately, obtaining the results of the heterogeneity analysis distinguishing crop types as shown in Table 6, Models 4–6. As can be seen from the table, the regression coefficient of constraint variable is all positive. It means that the constraint mechanism positively affects the agricultural socialized service quality of wheat, corn, and vegetable. The order of influence is:, vegetables > corn > wheat. This may be due to the high economic value of vegetables and the significant economic losses resulting from poor service quality. Constraint mechanisms can substantially reduce such losses.

Heterogeneity regression results for farmers and service quality types.

Variable Heterogeneity of farmer types Heterogeneity of crop types
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Small-scale pure farmers Part-time farmers Professional farmers Wheat Corn Vegetable
Constraint 0.083** 0.132** 0.039** 0.096*** 0.115 0.185***
Constant 3.707*** 3.743*** 3.733*** 3.759*** 3.713*** 3.781***
Control Yes Yes Yes Yes Yes Yes
R2 0.306 0.535 0.288 0.537 0.534 0.560

*, **, *** indicate significant at the 10%, 5%, and 1% levels, respectively.

 Robustness tests

To enhance the reliability of the above findings, this study conducts robustness tests from variable replacement test and the division of samples test.

(1) Robustness test of variable substitution. This study employs four new indicators to replace the explanatory variables for robustness testing. First, the data with zero values in the constraint mechanism are left-merged to 0, and it means the constraint mechanism is not constructed. The data with a number of adoptions greater than zero are right-merged to 1, which means the constraint mechanism is constructed. According to data processing above, the binary categorical independent variable data are formed. Second, the completeness of service contract terms can be used to represent constraint variable. Professional legal personnel evaluate the completeness of each contract, and higher contract integrity indicates the stronger constraints. Third, breach records can be used to represent constraint variable. The more breaches occur, the weaker the constraints become. Fourth, use third-party assessment scores from village collective to represent constraint variable. The village collective leader shall evaluate the performance of the service organization.

On this basis, regression analysis is carried out using the OLS model to compare whether there is a difference in the results. The regression results are shown in Table 7. In Model 1, it can be seen that the coefficients and significance of the model regression results are basically consistent with the benchmark regression model, which verifies the robustness of the model estimation results. In Model 2, contract completeness can enhance the quality of agricultural socialized services. In Model 3, breach records undermine the quality of agricultural socialized services. In other words, the fewer the breach records, the higher the service quality. In Model 4, Evaluation scores exert a positive influence on the quality of agricultural socialized services. It means the higher the third-party assessment scores, the better the service quality. The results from the above tests all indicate that the empirical results of the benchmark regression are robust.

(2) Robustness test of sample division. In this study, we take the sample division according to the area and links of agricultural socialized service, and conduct the robustness test through the sample replacement method. This study groups the samples according to the area of agricultural socialized service and the link numbers of agricultural socialized service. According to the average land contract size of Chinese farmers and the average land operation size of farmers in the sample area, this study takes 10 mu as the boundary and divides the sample into two sub-samples: “the area of agricultural socialized services is less than or equal to 10 mu” and “the area is greater than 10 mu.” The estimation results of the sample divided by the area of agricultural socialized services are shown in Table 8.

Robustness test results for independent variable substitution.

Variable Model 1 Model 2 Model 3 Model 4
Quality Quality Quality Quality
Constraint 0.191***
Contract completeness 0.142**
Breach records −0.175**
Assessment scores 0.189*
Constant 3.775*** 2.531** 1.972** 3.006***
Control Yes Yes Yes Yes
R2 0.293 0.302 0.376 0.298

*, **, *** indicate significant at the 10%, 5%, and 1% levels, respectively.

Robustness test results for the delimited sample.

Variable Model 1 Model 2
Service area ≤ 10 mu Service area > 10 mu
Coefficient Standard error Coefficient Standard error
Constraint 0.186*** 0.013 0.177*** 0.009
Constant 4.759*** 1.149 4.223*** 0.208
Control Yes Yes
R2 0.307 0.396
Difference between groups 0.058***

*** indicate significant at the 1% levels.

According to the regression results of Model 1 and Model 2, the significance level of the constraints variable and the coefficients size are generally consistent with the baseline regression. In addition, this study divides the sample into groups with and without village positions, and then conducts regressions separately to compare whether there are significant changes in the results. The results show that the estimation results in the model with and without positions are more consistent with those of the benchmark model, which verifies the robustness of the benchmark regression model estimation results. Furthermore, this study also adopts the truncated-tailed approach to sample censoring, eliminating outliers, and then conducting robustness tests, which are not shown due to space limitations.

 Conclusions and policy implications Conclusions

In the practical context of the continuous construction of agricultural socialized service constraints mechanism, according to contract theory, this study analyzes the implementation effect of agricultural socialized service quality constraints mechanism from the theoretical and empirical levels. Using data from 1,138 farmers in China, we apply the MESR model to empirically test the impact of constraints mechanism on agricultural socialized service quality. The main conclusions are as follows:

(1) The constraint mechanism has a positive impact on agricultural socialized service quality. Based on contract theory, the intermediary mechanisms focused on two aspects: implicit contracts within familiar societies and explicit contracts of risk-sharing. This result still holds after a series of robustness tests.

(2) Under the consideration of the counterfactual assumption, if farmers with low constraint do not construct constraint mechanisms, the quality of agricultural socialized services will decrease from 1.663 to 1.363, with a decrease of 22%. If farmers with high constraint do not construct constraint mechanisms, the quality of agricultural socialized services will decrease from 4.378 to 3.285, with a decrease of 33%.

(3) The results of heterogeneity analysis show that the effect of constraint mechanisms on the service quality of different farmer types is in the order of part-time farmers, small-scale pure farmers, and large professional farmers. The influence of the constraint mechanism on the quality of different crops is in the order of vegetable, corn, and wheat.

 Policy implications

(1) We should require the signing of agricultural socialized service contracts, specifying the decision-relevant elements, including scope and quality KPIs, inspection frequency and method, penalty ranges and triggers, timelines for claims/ADR, third-party oversight, public disclosure, and complaint/blacklist processes. Then we can pair them with a simple roadmap that clarifies roles for local government, cooperatives, providers, and third parties, indicative timelines and resource needs, and a small set of monitoring indicators such as coverage, compliance, re-inspection failure, and complaint rates.

(2) The standards of agricultural socialized services should be optimized. When formulating service standards, the contents of services, operational procedures and technical requirements should be specified and defined. For example, clear numerical requirements should be given for key indicators such as planting density, so that they can be implemented by service organizations. Service standards should be regularly evaluated, and the standard system should be updated in a timely manner in accordance with market demand and technological progress. Monitor the effectiveness of the implementation of the service standards and listen to feedback from farmers and service organizations on a regular basis, so as to continuously improve and enhance the quality and applicability of the service standards.

(3) For part-time farmers, enhance the transparency of service information and the rapid dispute resolution mechanism. In addition, the government should establish a credit evaluation system of agricultural socialized service for different types of farmer. Establish a credit database system and agricultural socialized service platform to record and manage trustworthy operation and dishonest behavior. The rating results will be made public and transparent in the village committee office to urge service organizations to consciously maintain a good credit record. For high-value crops such as vegetables, we can promote third-party quality inspection. For mid-production and single link, we should strengthen service quality constraints. The government should establish a directory of agricultural socialized services and set entry thresholds. Service organizations with poor quality will be subject to disciplinary measures to restrict their activities in the agricultural socialized service market.

 Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

 Ethics statement

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent from the [patients/ participants OR patients/participants legal guardian/next of kin] was not required to participate in this study in accordance with the national legislation and the institutional requirements.

 Author contributions

LY: Writing – original draft. SZ: Writing – review & editing. YZ: Conceptualization, Writing – review & editing, Resources.

 Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References An Q. Tao X. Chen X. (2023). Nested frontier-based best practice regulation under asymmetric information in a principal–agent framework. Eur. J. Oper. Res. 306, 269–285. doi: 10.1016/j.ejor.2022.07.035 Bai X. Pan K. Shoaib N. Sun X. Wu X. Zhang L. (2024). Status of phthalate esters pollution in facility agriculture across China: spatial distribution, risk assessment, and remediation measures. Sci. Total Environ. 908:168416. doi: 10.1016/j.scitotenv.2023.16841637944601 Baiyegunhi L. J. S. Majokweni Z. P. Ferrer S. R. D. (2019). Impact of outsourced agricultural extension program on smallholder farmers' net farm income in Msinga, KwaZulu-Natal, South Africa. Technol. Soc. 57, 1–7. doi: 10.1016/j.techsoc.2018.11.003 Banerjee S. Chakraborty S. (2023). Optimal incentive contracts with a spiteful principal: single agent. Math. Soc. Sci. 122, 29–41. doi: 10.1016/j.mathsocsci.2023.01.005 Bogle T. Van Kooten G. C. (2013). Options for maintaining forest productivity after natural disturbance: a principal–agent approach. Forest Policy Econ. 26, 138–144. doi: 10.1016/j.forpol.2012.09.005 Chinnannan K. Somagattu P. Yammanuru H. K. Reddy U. Nimmakayala P. (2024). Health risk assessment of heavy metals in soil and vegetables from major agricultural sites of Ohio and West Virginia. Biocatal. Agric. Biotechnol.57:103108. doi: 10.1016/j.bcab.2024.103108 Corgnet B. Gómez-Miñambres J. Hernán-González R. (2018). Goal setting in the principal–agent model: weak incentives for strong performance. Games Econ. Behav. 109, 311–326. doi: 10.1016/j.geb.2017.12.017 Deb P. Trivedi P. K. (2006). Maximum simulated likelihood estimation of a negative binomial regression model with multinomial endogenous treatment. Stata. J. 6, 246–255. doi: 10.1177/1536867X0600600206 Ding Y. Xu J. (2023). Global vulnerability of agricultural commodities to climate risk: evidence from satellite data. Econ. Anal. Policy 80, 669–687. doi: 10.1016/j.eap.2023.09.013 Dirzka C. Acciaro M. (2021). Principal-agent problems in decarbonizing container shipping: a panel data analysis. Transport. Res. D Transport Environ. 98:102948. doi: 10.1016/j.trd.2021.102948 Dong J. Liu M. Zhang W. Jiang H. Wang J. Zhang B. . (2024). The effects of top-down political inspection on bottom-up public complaints in China. J. Clean. Prod. 476:143658. doi: 10.1016/j.jclepro.2024.143658 Feng Y. Feng Y. Liu Z. (2025). Normative influences on carbon offset behavior: insights from organic farming practices. Sustainability 17:1638. doi: 10.3390/su17136152 Forges F. Koessler F. Salamanca A. (2024). Interacting mechanisms: a perspective on generalized principal–agent problems. J. Math. Econ. 114:103023. doi: 10.1016/j.jmateco.2024.103023 Hao H. Yao E. Pan L. Chen R. Wang Y. Xiao H. (2025). Exploring heterogeneous drivers and barriers in MaaS bundle subscriptions based on the willingness to shift to MaaS in one-trip scenarios. Transport. Res. A Policy Prac. 199:104525. doi: 10.1016/j.tra.2025.104525 Hou W. Yang Q. Cao Y. Luo L. Ding J. Wang Q. . (2025). Impact of ecological education on university students' environmentally sustainable behavior—evidence from China. Sustainability 17:6051. doi: 10.3390/su17136051 Hui M. Yong F. (2023). Risk preferences and the low-carbon agricultural technology adoption: evidence from rice production in China. J. Integr. Agric. 22, 2577–2590. doi: 10.1016/j.jia.2023.07.002 Jaydarifard S. Yigitcanlar T. Alexander P. (2025). Risk factors and safety strategies for mitigating violations, harassment and assault in taxi and ride-hailing services. Transport Rev. 45, 776–800. doi: 10.1080/01441647.2025.2512251 Kester S. Ong A. German J. D. Dangaran P. C. Jethro B. Paz J. . (2024). Service quality and customer satisfaction analysis among motorcycle taxi transportation in the Philippines through SERVQUAL dimensions and social exchange theory. Case Stud. Transport Policy 15:101139. doi: 10.1016/j.cstp.2023.101139 Laeeque S. Ali M. (2025). When passengers misbehave: exploring the pathways from customer incivility to service sabotage among ride-hailing drivers. J. Services Market. 39, 426–441. doi: 10.1108/JSM-10-2024-0551 Li Y. Huan M. Jiao X. Chi L. Ma J. (2023). The impact of labor migration on chemical fertilizer use of wheat smallholders in China- mediation analysis of socialized service. J. Clean. Prod. 394:136366. doi: 10.1016/j.jclepro.2023.136366 Li Z. Zhao X. (2025). Incentives and constraints: an optimization mechanism for improving the performance of farmland protection entities. PLoS ONE 20:e0320750. doi: 10.1371/journal.pone.032075040455864 Liang W. Yu W. Yao X. (2025). Does the development of the service outsourcing industry enhance green product exports? Quasi-natural experimental evidence from China. Energy Econ. 149:108727. doi: 10.1016/j.eneco.2025.108727 Liang Y. Zhang P. Bi W. Qi Q. Xie F. (2025). Contract farming and pro-environmental behavior: insights from beef cattle farmers. Front. Sustain. Food Syst. 9:1590726. doi: 10.3389/fsufs.2025.1590726 Liao L. Guo J. Peng Y. Liu Y. Ling Y. Tang Y. (2025). Agricultural socialized services and grain yield per unit area: empirical evidence from Jiangxi Province, China. Front. Sustain. Food Syst. 9:1611236. doi: 10.3389/fsufs.2025.1611236 Lin B. Xie J. (2023). Superior administration's environmental inspections and local polluters' rent seeking: a perspective of multilevel principal–agent relationships. Econ. Anal. Policy 80, 805–819. doi: 10.1016/j.eap.2023.09.023 Lin Y. Hu R. Zhang C. Chen K. (2022). The role of public agricultural extension services in driving fertilizer use in rice production in China. Ecol. Econ. 200:107513. doi: 10.1016/j.ecolecon.2022.107513 Liu G. Chen C. Fu X. Liu Y. Khan N. Luo L. (2024). Has the agricultural cooperatives served each member fairly? A new perspective based on utilization level of member services. PLoS ONE 19:e0294439. doi: 10.1371/journal.pone.029443938295086 Liu S. Wang L. Shi X. Ouyang S. Yang L. (2024). Incentive contract design considering quotas production: a principal-agent perspective. Heliyon 10:e24137. doi: 10.1016/j.heliyon.2024.e2413738293419 Luo L. Yang Q. Qiao D. Cao Y. Ding J. Ma H. . (2025). How does environmental education affect college students' waste sorting behavior: a heterogeneity analysis based on educational background. J. Environ. Manage. 389:126064. doi: 10.1016/j.jenvman.2025.12606440466314 Luo S. Su S.-P. (2023). Design of power mutation and exercise mode of the natural resource assets principal-agent mechanism: from the perspective of incomplete contract theory. J. Nat. Resources 38, 2360–2371. doi: 10.31497/zrzyxb.20230911 Ma R. Liu J. An S. (2024). Early warning response to rainstorm: designing a model with incentive and supervision mechanisms based on the principal-agent theory. Int. J. Disaster Risk Reduct. 111:104683. doi: 10.1016/j.ijdrr.2024.104683 Ma X. Hou X. Cui Y. Ma J. (2025). Do socialized agricultural services promote smallholder participation in large markets? Evidence from grain farmers in China. Agribusiness 2025, 1–12. doi: 10.1002/agr.22014 Mastenbroek A. Gumucio T. Nakanwagi J. (2024). Gender, agricultural risk perceptions, and maize seed systems: a case study of drought-tolerant maize varieties in Uganda. Agric. Syst. 217:103912. doi: 10.1016/j.agsy.2024.103912 Mctigue C. Monios J. Rye T. (2020). The principal-agent problem in contracting public transport provision to private operators: a case study of the UK Quality Contract Scheme. Utilities Policy 67:101131. doi: 10.1016/j.jup.2020.101131 Mi Q. Li X. Gao J. (2020). How to improve the welfare of smallholders through agricultural production outsourcing: evidence from cotton farmers in Xinjiang, Northwest China. J. Clean. Prod. 256:120636. doi: 10.1016/j.jclepro.2020.120636 Monios J. Fedi L. (2023). The principal-agent problem in hierarchical policy making: a failure of policy to support the transition to LNG as an alternative shipping fuel. Marine Policy 157:105846. doi: 10.1016/j.marpol.2023.105846 Nwajei U. O. K. Bølviken T. Hellström M. M. (2022). Overcoming the principal-agent problem: the need for alignment of tools and methods in collaborative project delivery. Int. J. Project Manage. 40, 750–762. doi: 10.1016/j.ijproman.2022.08.003 Qing C. Zhou W. Song J. Deng X. Xu D. (2023). Impact of outsourced machinery services on farmers' green production behavior: evidence from Chinese rice farmers. J. Environ. Manage. 327:116843. doi: 10.1016/j.jenvman.2022.11684336459784 Qiu H. Feng M. Chi Y. Luo M. (2023). Agricultural machinery socialization service adoption, risks, and relative poverty of farmers. Agriculture 13:1787. doi: 10.3390/agriculture13091787 Sang X. Luo X. Razzaq A. Huang Y. Erfanian S. (2023). Can agricultural mechanization services narrow the income gap in rural China? Heliyon 9:e13367. doi: 10.1016/j.heliyon.2023.e1336736816315 Schosser J. (2019). Consistency between principal and agent with differing time horizons: computing incentives under risk. Eur. J. Oper. Res. 277, 1113–1123. doi: 10.1016/j.ejor.2019.03.044 Shi R. Shen Y. Du R. Yao L. Zhao M. (2024). The impact of agricultural productive service on agricultural carbon efficiency—from urbanization development heterogeneity. Sci. Total Environ. 906:167604. doi: 10.1016/j.scitotenv.2023.16760437858831 Sui Y. Yu L. (2025). Quality risk identification and fuzzy comprehensive assessment of land trusteeship services in China. Land 14:2027. doi: 10.3390/land14102027 Viaggi D. Bartolini F. Raggi M. (2009). Combining linear programming and principal–agent models: an example from environmental regulation in agriculture. Environ. Model. Softw. 24, 703–710. doi: 10.1016/j.envsoft.2008.10.014 Vining A. R. Richards J. (2016). Indigenous economic development in Canada: confronting principal-agent and principal–principal problems to reduce resource rent dissipation. Resources Policy 49, 358–367. doi: 10.1016/j.resourpol.2016.07.006 Wang Y. Huan M. (2023). The effects of socialized agricultural services on rural collective action in the irrigation commons: evidence from China. Agric. Water Manage. 289:108519. doi: 10.1016/j.agwat.2023.108519 Wu C. Li D. Zhang X. Pan J. Quan L. Yang L. . (2023). China's agricultural machinery operation big data system. Comput. Electron. Agric. 205:107594. doi: 10.1016/j.compag.2022.107594 Xiao J. Fang X. (2025). Risk identification and spatiotemporal evolution in rural land trusteeship. Land 14:1132. doi: 10.3390/land14061132 Xu Y. Lyu J. Xue Y. Liu H. (2022). Intentions of farmers to renew productive agricultural service contracts using the theory of planned behavior: an empirical study in Northeastern China. Agriculture 12:1471. doi: 10.3390/agriculture12091471 Yan L. Hong K. Chen K. Li H. Liao L. (2021). Benefit distribution of collectively-owned operating construction land entering the market in rural China: a multiple principal–agent theory-based analysis. Habitat Int. 109:102328. doi: 10.1016/j.habitatint.2021.102328 Yang S. Li W. (2023). The impact of socialized agricultural machinery services on the labor transfer of maize growers. Agriculture 13:1249. doi: 10.3390/agriculture13061249 Yang S. Zhang F. (2023). The impact of agricultural machinery socialization services on the scale of land operation: evidence from rural China. Agriculture 13:1594. doi: 10.3390/agriculture13081594 Yu L. Zheng S. Gao Q. (2023). Government environmental regulation strategy for new pollutants control in mariculture. Marine Policy 150:105545. doi: 10.1016/j.marpol.2023.105545 Zhang M. Li X. (2025). Understanding the relationship between coopetition and startups' resilience: the role of entrepreneurial ecosystem and dynamic exchange capability. J. Business Indus. Market. 40, 527–542. doi: 10.1108/JBIM-04-2024-0226 Zhang M. Zhang Y. Weng Z. Chen Z. (2022). Design and service effect evaluation of agricultural social service platform based on 5G and cloud computing. Wireless Commun. Mobile Comput. 2022:4949242 doi: 10.1155/2022/4949242 Zhang T. (2024). Agricultural socialized services can facilitate the rapid development of the science and technology backyard. Front. Agric. Sci. Eng. 11, 210–212. doi: 10.15302/J-FASE-2024541 Zhang Y. Liang C. Li X. (2024). Understanding virtual agents' service quality in the context of customer service: a fit-viability perspective. Electron. Commer. Res. Appl. 65:101380. Zhang Y. Liu R. (2025). Digital platform economy, financing constraints, and labor market economic development. Finance Res. Lett. 85:108122. doi: 10.1016/j.frl.2025.108122 Zhou Y. He L. Ke X. Zhang E. Zhu J. Lin A. (2025). Impact of agricultural machinery purchase subsidies on the sustainable and intensive utilization of cultivated land: a perspective on agricultural machinery socialization services. J. Rural Stud. 119:103798. doi: 10.1016/j.jrurstud.2025.103798 Zhu C. Su Y. Fan R. Xu R. Li B. (2025). Research on the optimal incentive and constraint mechanisms for corporate carbon information disclosure considering different market contexts: a network-based evolutionary game analysis. Energy Econ. 142:108207. doi: 10.1016/j.eneco.2025.108207

Edited by: Xin Wang, Longyan University, China

Reviewed by: Lei Luo, Sichuan Agricultural University, China

Ravi Kumar Bommisetti, Akal University, India