Skip to content
Frontiers
Search

Original Research

Frontiers in Sustainable Food Systems ·

Volume 10 - 2026 | https://doi.org/10.3389/fsufs.2026.1880279

The impact of agricultural socialized services on grain farmers’ agricultural eco‑economic efficiency

  • Wensu Tang1,2
  • Xiaohui Li3*corresponding author
  1. 1Institute of Agricultural and Rural Development, Jiangxi Academy of Social Sciences, Nanchang, Jiangxi, China
  2. 2College of Economics and Management, Huazhong Agricultural University, Wuhan, Hubei, China
  3. 3Economics Management College, Weifang University of Science and Technology, Weifang, Shandong, China
Article metrics0 viewsView details
PDFCiteMetrics

Abstract

Developing agricultural socialized services (ASS) while promoting green and low-carbon grain production is an important component of agricultural modernization in China. Although previous studies have examined the effects of ASS on agricultural productivity, farm income, and individual input use, it remains unclear whether ASS improve household-level agricultural eco-economic efficiency (AEE), which jointly accounts for economic output, resource use, and environmental costs. Evidence is also limited regarding the factor-allocation channels and production conditions underlying this relationship. Using survey data from 5,992 grain-producing households across 14 provinces in China, this study examines household-level AEE and examines its relationship with ASS adoption using OLS and an instrumental-variable two-stage least squares specification. The results indicate that: (1) ASS adoption is positively associated with grain farmers’ AEE; (2) the estimated relationship varies across farm sizes and crop types, with a more pronounced positive estimate among farmers operating at least 2 ha, but a negative estimate among wheat-growing households; and (3) ASS adoption is associated with a lower share of fertilizer and pesticide expenditure in total variable production costs, whereas no statistically significant labor-reallocation channel is identified. These findings suggest that ASS policies should improve service quality, reduce smallholders’ barriers to service access, develop crop- and production-stage-specific service packages, and promote more precise management of agricultural inputs.

Keywords

1 Introduction

Sustainable agricultural transformation has become a global priority for addressing climate change, ecological degradation, and food security challenges (Pretty, 2008; IPCC, 2023). As agriculture is both an important source of greenhouse gas emissions and a crucial provider of ecological services, improving agricultural eco-economic efficiency (AEE) is essential for balancing agricultural productivity, environmental sustainability, and farmers’ income growth (Tilman et al., 2002; Liu et al., 2020). In developing countries dominated by smallholder farming, the excessive use of chemical fertilizers, pesticides, and fossil-energy-based machinery has generated substantial environmental pressures, including agricultural non-point source pollution, soil degradation, and carbon emissions (Pingali, 2012; Jiang et al., 2024). Meanwhile, fragmented land management, rural labor outmigration, aging agricultural populations, and limited access to green technologies constrain farmers’ ability to adopt sustainable production practices (Peng et al., 2025). Therefore, promoting agricultural green transformation while maintaining agricultural productivity has become a critical challenge. Against this background, agricultural socialized services (ASS), characterized by specialization, market orientation, and professional agricultural operations, have emerged as an important institutional arrangement for facilitating agricultural green development and promoting the integration of smallholders into modern agriculture (Chen et al., 2022).

Existing empirical studies have examined the economic and environmental consequences of ASS from several perspectives. At the household level, agricultural services have been found to improve production efficiency by introducing specialized machinery, professional management, and technology-embedded operations (Deng et al., 2020; Chen et al., 2022). Other studies show that ASS may promote green agricultural development, reduce chemical input dependence, and improve agricultural land-use ecological efficiency or green total factor productivity (Cheng et al., 2022; Li et al., 2024; Wang et al., 2024; Yao et al., 2024). However, the estimated effects are not necessarily uniform. The benefits of machinery and outsourcing services may depend on farm size, land fragmentation, crop characteristics, and the particular production stage being outsourced (Qiu et al., 2021; Qing et al., 2023; Zhang et al., 2023; Zou and Mishra, 2024).

Despite these advances, three questions remain insufficiently addressed. First, much of the existing literature focuses on agricultural productivity, green total factor productivity, land-use ecological efficiency, or individual chemical inputs rather than household-level agricultural eco-economic efficiency that simultaneously incorporates economic output, resource use, and environmental costs. Second, because farmers do not adopt ASS randomly, estimates may be affected by self-selection and omitted household or local characteristics. Third, relatively little is known about whether the relationship between ASS and AEE operates through material-input and household-labor reallocation or whether it varies across farm sizes, crop types, and individual service stages.

To address these gaps, this study uses survey data from 5,992 grain-producing households to estimate household-level AEE and examine its relationship with ASS adoption. In addition to OLS estimation, an instrumental-variable specification is employed to address the potential endogeneity of service adoption. The analysis distinguishes between the fertilizer–pesticide expenditure-share channel and the non-agricultural labor-share channel and further examines heterogeneity across farm sizes, crop types, and service stages.

The results indicate a positive overall relationship between ASS adoption and AEE. The positive estimate is more pronounced among farms operating at least 2 ha, whereas the estimate for smallholders is statistically insignificant. Crop-specific analyses reveal a negative estimate for wheat production, with the service-stage results suggesting that harvesting services are the principal source of this pattern. The mechanism analysis shows that ASS adoption is associated with a lower share of fertilizer and pesticide expenditure in total variable production costs, while no statistically significant non-agricultural labor-reallocation channel is identified.

This study contributes to the literature in three respects. First, it provides household-level evidence based on an AEE measure that incorporates both desirable and undesirable outputs. Second, it distinguishes between material-input expenditure and household-labor reallocation channels. Third, it demonstrates that the estimated environmental consequences of ASS differ across farms, crops, and production stages. Nevertheless, the IV estimates depend on an exclusion restriction that cannot be conclusively verified in an exactly identified model, while the restricted geographic coverage and short survey period limit the generalizability of the findings. The results are therefore interpreted as IV-based evidence rather than definitive causal effects.

2 Theoretical analysis and research hypotheses

AEE refers to the ability of agricultural production systems to maximize economic output while minimizing resource consumption and environmental pollution under given technological conditions (Farrell, 1957; Tone and Tsutsui, 2010). Improving agricultural ecological economic efficiency therefore requires not only higher agricultural productivity but also lower ecological and environmental costs.

ASS can significantly improve agricultural ecological economic efficiency through specialization, economies of scale, and technological diffusion. Under traditional smallholder farming systems, farmers are often constrained by fragmented landholdings, insufficient technical knowledge, and limited access to modern agricultural technologies, resulting in excessive use of fertilizers, pesticides, irrigation water, and labor inputs (Pingali, 2012; Chen et al., 2022). By contrast, ASS providers generally possess stronger technical capabilities and professional management experience, enabling them to provide standardized production services and advanced green technologies to farmers (Deng et al., 2020; Cheng et al., 2022). Through mechanized operations and precision agricultural technologies, service organizations can improve resource-use efficiency and reduce unnecessary input redundancy.

Furthermore, ASS facilitate the adoption of environmentally friendly agricultural technologies such as conservation tillage, integrated pest management, water-saving irrigation, and scientific fertilization practices (Knowler and Bradshaw, 2007; Chen et al., 2022). These technologies can effectively reduce agricultural carbon emissions, agricultural non-point source pollution, and soil degradation while maintaining stable agricultural output. ASS also reduce farmers’ transaction costs and uncertainty associated with green technology adoption, thereby increasing farmers’ willingness to engage in sustainable agricultural production (Rogers, 2003; Deng et al., 2020). Consequently, ASS may simultaneously improve economic performance and ecological performance, ultimately enhancing agricultural ecological economic efficiency.

Accordingly, this paper proposes the following hypothesis:

H1: ASS significantly improve grain farmers’ AEE.

Production material allocation efficiency is an important determinant of agricultural ecological economic efficiency. Due to information asymmetry and limited technical capacity, smallholders often face serious distortions in the allocation of agricultural production materials such as fertilizers, pesticides, irrigation water, seeds, and machinery services (Coelli et al., 2005; Li et al., 2024). Such inefficient allocation not only increases production costs but also generates significant environmental externalities.

ASS can improve production material allocation efficiency through technological guidance, specialized operations, and scale management. On the one hand, service organizations can adopt precision agricultural technologies according to soil quality, crop characteristics, and local climatic conditions, thereby reducing the excessive application of fertilizers and pesticides (Zhang et al., 2023). On the other hand, agricultural socialized service providers can achieve economies of scale in machinery operations and agricultural input procurement, helping farmers optimize resource utilization and reduce unit production costs (Diao et al., 2014; Qing et al., 2023). Moreover, ASS facilitate farmers’ access to modern green technologies and standardized production processes, thereby improving input-output matching efficiency.

From an ecological perspective, optimized allocation of agricultural production materials can significantly reduce carbon emissions, chemical residues, and environmental pollution caused by excessive agricultural inputs. From an economic perspective, efficient allocation of production materials contributes to higher agricultural productivity and profitability. Therefore, ASS may improve agricultural ecological economic efficiency by enhancing production material allocation efficiency.

Based on the above analysis, this paper proposes the following hypothesis:

H2: ASS improve AEE by optimizing farmers’ production material allocation efficiency.

Labor allocation efficiency is another important mechanism influencing agricultural ecological economic efficiency. With rapid urbanization and industrialization, rural labor migration has become increasingly common, resulting in labor shortages and aging agricultural populations in many developing countries (Takeshima et al., 2013; Mi et al., 2020). Under labor constraints, farmers may reduce agricultural management intensity or excessively rely on chemical inputs to substitute for labor, thereby negatively affecting agricultural ecological economic efficiency.

ASS can improve labor allocation efficiency by substituting machinery and specialized services for household labor inputs. Mechanized plowing, planting, pest control, and harvesting services can significantly reduce farmers’ dependence on manual labor and alleviate seasonal labor shortages (Diao et al., 2014; Qing et al., 2023). This labor substitution effect enables farm households to allocate labor more efficiently between agricultural and non-agricultural employment activities, thereby improving household income and resource allocation efficiency.

In addition, ASS promote standardized agricultural production and reduce inefficiencies caused by aging labor forces and limited farming skills. Through professional division of labor and specialized management, service organizations can ensure more accurate timing and higher operational quality during agricultural production processes, thereby improving agricultural output and reducing ecological losses (Cheng et al., 2022; Deng et al., 2020). Consequently, the labor-saving and efficiency-enhancing effects generated by ASS may jointly improve agricultural ecological economic efficiency.

Therefore, this paper proposes the following hypothesis:

H3: ASS improve AEE by enhancing farmers’ labor allocation efficiency.

3 Methodology

3.1 Data source

The data used in this study were obtained from a one-time cross-sectional household survey conducted by the research team between October 2022 and February 2023. The survey, entitled the “Survey on Farm Households’ Participation in Agricultural Production Trusteeship,” covered 14 provinces and municipalities, including Heilongjiang, Shandong, Anhui, Shaanxi, Hubei, and Fujian, and spanned seven major geographical regions of China: North, East, Northeast, Central, South, Southwest, and Northwest China. The sample provinces and municipalities were selected by jointly considering regional economic development, agricultural production conditions, and the development of agricultural socialized services. This design provided broad geographic coverage and captured variation in farmers’ participation in ASS across different agricultural production environments. To improve sample representativeness, the survey employed a stratified multistage sampling procedure. Within each sample province or municipality, five counties were first selected using systematic sampling. Two townships were then randomly selected from each sample county, followed by the random selection of five sample villages from each township. The questionnaire collected information on farmers’ participation in agricultural production trusteeship, land transfers, agricultural production and management characteristics, individual characteristics, and household characteristics.

Data cleaning was conducted according to predefined criteria. First, households that did not engage in grain production were excluded because they fell outside the target population of the study. Second, duplicate records and observations containing abnormal information or irreconcilable logical inconsistencies were removed to ensure the accuracy and internal consistency of the survey data. Third, observations with missing information on core DEA inputs or outputs or ASS adoption were excluded because these variables were indispensable for estimating agricultural eco-economic efficiency and constructing the key explanatory variable, respectively. Imputing these key variables could directly affect the construction of the dependent or key explanatory variable and would require strong assumptions that are difficult to verify; therefore, no imputation was performed for these variables. Finally, observations with sown areas exceeding five standard deviations from the sample mean were treated as potential outliers and excluded to reduce the influence of extreme observations on the DEA efficiency estimates. After these data-cleaning procedures, the final analytical sample comprised 5,992 valid observations of grain-producing households. After these data-cleaning procedures,the final analytical sample comprised 5,992 valid grain-producing households. Of these, 3,239 households (54.06%) adopted ASS in at least one production stage, while 2,753 households (45.94%) did not adopt ASS.

3.2 Econometric model specification

3.2.1 Measurement model of AEE

Compared with traditional parametric efficiency estimation methods, Data Envelopment Analysis (DEA) is more suitable for evaluating agricultural production systems characterized by multiple inputs and outputs, as it can simultaneously handle multiple production factors and assign endogenous optimal weights to each indicator. However, previous studies have shown that the efficiency scores generated by conventional DEA models may suffer from serial correlation and endogeneity problems when directly used in second-stage regression analysis, leading to biased statistical inference (Simar and Wilson, 2007; Badunenko and Tauchmann, 2019). To address these issues, this study adopts the two-stage semi-parametric DEA approach proposed by Simar and Wilson (2007), which improves the consistency and reliability of efficiency estimation and subsequent regression analysis.

In the first stage, AEE is estimated using an input-oriented DEA model. Given the focus on resource conservation and environmental performance, the model incorporates both desirable and undesirable outputs. The basic specification is as follows Equations 1–5:

min AEE = 1 m ∑ i = 1 m ( x ¯ x ik ) 1 r 1 + r 2 ( ∑ s = 1 r 1 y ¯ d y sk d + ∑ s = 1 r 1 y ¯ u y qk u ) (1)

s . t . x ¯ ≥ ∑ j = 1 , ≠ k n x ij λ j ; ∑ j = 1 , ≠ k n y qj d λ j ≤ y ¯ d ≤ ∑ j = 1 , j ≠ k n y sj d λ j (2)

x ¯ ≥ x k ; y ¯ d ≤ y k d ; y ¯ u ≥ y k u (3)

λ j ≥ 0 ; i = 1 , 2 , … , m ; j = 1 , 2 , … , n , j ≠ 0 (4)

s = 1 , 2 , … , r 1 ; q = 1 , 2 , … , r 2 (5)

Among these,AEE represents agricultural eco-economic efficiency; n denotes the number of decision-making units (DMUs); each DMU consists of m inputs, r 1 desirable outputs, and r 2 undesirable outputs, x , y d and y u represent the corresponding elements in the input, desirable output, and undesirable output matrices, respectively.

The second stage involves conducting a regression analysis of the AEE values obtained in the first stage against environmental variables. Environmental variables refer to objective conditions that influence AEE and are beyond subjective control. Drawing on existing research (Yang et al., 2018), and considering the conceptual characteristics of AEE, the actual survey context, and data availability, this study selects external factors that genuinely impact farmers' agricultural production but are uncontrollable or unalterable in the short term. The primary environmental variables chosen are the level of information availability and agricultural technology training.

3.2.2 Model specification for the impact of ASS on AEE

This paper focuses on investigating the impact of agricultural social services on the ecological-economic efficiency of agriculture, employing a 2SLS model to mitigate potential endogeneity issues. The specific model is as follows Equations 6–8:

Phase 1 (Instrumental Variable Regression):

AS S i = α 0 + α 1 I V i + α 2 Z i + ε i (6)

ASS ̂ i = α 0 ̂ + α 1 ̂ I V i + α 2 ̂ Z i (7)

Phase 2 (Efficiency Impact Regression):

AE E i = β 0 + β 1 ASS ̂ i + β 2 X i + μ i (8)

In this specification, AS S i represents agricultural socialized services, IV is the instrumental variable, Z denotes the set of control variables affecting ASS, and X refers to the control variables influencing efficiency.

3.3 Variable definitions

3.3.1 Dependent variable

Following previous studies (Rüdenauer et al., 2005; Levidow et al., 2014), AEE in this study refers to the input–output efficiency achieved by farm households during agricultural production, namely obtaining higher agricultural output with lower resource inputs and reduced environmental pollution.

Considering the characteristics of agricultural production and the requirements of green agricultural development, this study constructs an AEE evaluation system consisting of four input indicators, two output indicators, and two environmental indicators (see Table 1). Among the environmental indicators, agricultural carbon emissions are calculated based on energy consumption and agricultural input use during grain production. Referring to the method proposed by Min and Li (2013), the carbon emissions associated with grain production are estimated using weighting coefficients derived from the ratio of grain-sown area to total crop-sown area. Table 2 reports the carbon emission coefficients corresponding to different carbon sources in grain production.

Table 1. Measurement indicators of AEE.

DimensionVariableMeasurement
InputLabor inputTotal agricultural labor days
Land inputCultivated land area (mu)
Agricultural machinery inputExpenditure on agricultural machinery services
Agricultural material inputExpenditure on fertilizers, pesticides, seeds, and agricultural films
Desirable outputAgricultural outputTotal agricultural production value
Undesirable outputAgricultural carbon emissionsCarbon emissions generated during agricultural production
Environmental variableAgricultural technical trainingFrequency of participation in agricultural training
Information accessibilityDistance from village committee to county center

Table 2. Major carbon sources and their carbon emission coefficients for grain production.

CarbonSourceEmission coefficientReference
Pesticide4.9341 kg C·kg−1Oak Ridge National Laboratory
Chemical fertilizer0.8956 kg C·kg−1
Agricultural film5.18 kg C·kg−1The Institute of Resource, Ecosystem and Environment of Agriculture
Machinery0.18 kg C·hm−1Duan et al. (2011)

For efficiency estimation, this study adopts the two-stage semi-parametric DEA approach proposed by Badunenko and Tauchmann (2019) and Simar and Wilson (2007). This method allows for a comprehensive assessment of agricultural production efficiency while incorporating environmental constraints.

3.3.2 Explanatory variable

The explanatory variable in this study is whether grain-producing households purchase ASS. Based on field survey data, the analysis focuses on the utilization of ASS during grain production processes. Drawing on transaction cost theory (Williamson, 1985) and the theory of agricultural division of labor (Hayami and Ruttan, 1985), this study constructs a two-dimensional measurement framework for ASS.

Specifically, ASS include services related to land preparation, sowing, pest and disease control, and harvesting during grain production. If a household purchases socialized services in any production stage, the variable is assigned a value of 1, indicating that the household adopts ASS.

Furthermore, to better capture the intensity of service utilization, this study measures the number of production stages in which ASS are adopted. The variable ranges from 0 to 4, where 0 indicates no use of ASS, while values from 1 to 4 represent the adoption of ASS in one to four production stages, respectively.

3.3.3 Control variables

To reduce potential estimation bias, a series of control variables related to household demographics, family endowments, and production conditions are included in the regression models. Individual characteristics include the gender, age, educational attainment, health status, and migrant work experience of the household head. Household characteristics include the number of agricultural laborers, total household income, and household living standards. Production and operational characteristics include land tenure confirmation, farmland quality, accessibility of mechanized operations, irrigation conditions, ownership of agricultural machinery, and participation in agricultural organizations.

3.3.4 Instrumental variable

Potential endogeneity may arise from omitted variables and self-selection bias in the adoption of ASS. To address this issue, this study employs an instrumental variable (IV) approach and estimates the model using two-stage least squares (2SLS), following the methods adopted by Angrist and Pischke (2009), Wooldridge (2010) and Stock and Watson (2020).

The instrumental variable selected in this study is the level of support provided by village committees for ASS. The rationale is that village committee support can effectively reflect local institutional support and policy encouragement for the development of ASS. This variable directly affects farmers’ willingness to adopt ASS and the accessibility of service resources, thereby influencing the scale and level of ASS adoption. At the same time, village committee support is unlikely to directly affect household AEE except through its influence on ASS adoption, thus satisfying the exogeneity requirement for a valid instrumental variable. Accordingly, this study employs the IV-2SLS approach to identify the causal effect of ASS on AEE.

Following the validation approach of Jiang et al. (2026), we further test the relevance and exogeneity of the instrument. When AEE is regressed on ASS adoption, village committee support, and all control variables, the coefficient on village committee support is 0.0008 (p = 0.362), indicating no significant direct effect on efficiency and thus supporting the exclusion restriction. Conversely, in the first‑stage regression where ASS adoption is the dependent variable, village committee support yields a coefficient of 0.123 (p < 0.001) and the F‑statistic for its exclusion is 519.04 (p < 0.001), confirming a strong correlation with ASS adoption and effectively ruling out weak instrument concerns.

To further examine the exclusion restriction of the instrumental variable, we tested whether village committee support directly affects farmers’ adoption of green technologies. The results show that village committee support has no significant effect on green technology adoption after controlling for household and production characteristics. This suggests that village committee support is unlikely to influence agricultural ecological economic efficiency through the green technology adoption channel, providing additional evidence supporting the validity of the instrumental variable. Due to space constraints, the detailed regression results are not reported but are available upon request.

3.3.5 Mechanism variable

To further explore the internal mechanisms through which ASS affect AEE, this study constructs two mechanism variables.

  • (1) Proportion of production material input

The proportion of production material input is measured as the ratio of expenditures on pollution-intensive agricultural inputs, such as fertilizers and pesticides, to total agricultural production costs. A higher ratio indicates stronger dependence on chemical inputs and a potentially greater environmental burden.

  • (2) Labor reallocation

Labor reallocation is measured as the proportion of household labor engaged in non-agricultural activities to total household labor. A higher value indicates a greater degree of labor transfer from agricultural to non-agricultural sectors and a more optimized household labor allocation structure. Agricultural socialized services can substitute for agricultural labor input, thereby releasing household labor resources for non-farm employment and improving the efficiency of labor allocation.

Descriptive statistics for all variables are reported in Table 3.

Table 3. Variable definitions and descriptive statistics.

VariablesDefinitionMeanStd. Dev.
AEEDEA-based AEE score (0–1)0.7880.075
ASSWhether the household purchases agricultural socialized services: 0 = no; 1 = yes0.5410.498
Gender0 = female; 1 = male0.8440.363
AgeYears48.3639.643
Education levelYears of schooling9.9931.931
Health status1 = very unhealthy; 2 = unhealthy; 3 = average; 4 = relatively healthy; 5 = very healthy4.0900.942
Migrant work experienceWhether the household head has migrant work experience: 0 = no; 1 = yes0.5130.499
Number of agricultural laborersNumber of household agricultural laborers1.9270.886
Household total incomeHousehold total income (logarithm, 10,000 yuan)2.3530.951
Household living standard1 = low; 2 = lower-middle; 3 = middle; 4 = upper-middle; 5 = high2.7230.868
Land tenure confirmationWhether land tenure has been officially certified: 0 = no; 1 = yes0.8930.309
Farmland quality1 = very poor; 2 = poor; 3 = average;4 = good; 5 = very good3.4200.829
Mechanization suitabilityWhether field roads are suitable for machinery operation: 0 = no; 1 = yes0.8940.464
Irrigation conditionWhether irrigation is convenient: 0 = no; 1 = yes0.6690.471
Machinery ownership0 = no; 1 = yes0.6850.464
Cooperative participationWhether the household has joined a cooperative: 0 = no; 1 = yes0.5790.494
Village committee support1 = very unsupported; 2 = relatively unsupported; 3 = neutral; 4 = relatively supportive; 5 = very supportive3.5021.143
Proportion of production material inputRatio of chemical fertilizer and pesticide expenditures to total agricultural production costs (%)0.4940.209
Labor reallocationShare of non-agricultural labor in total household labor (%)0.2610.325

4 Results and discussion

4.1 Estimating the impact of ASS on AEE

This section estimates the causal effect of ASS on agricultural eco‑economic efficiency. To address potential endogeneity arising from self‑selection and omitted variables, we employ an instrumental variable two‑stage least squares (IV‑2SLS) estimator, using village committee support for ASS as the instrument. Table 4 reports the OLS and 2SLS results.

Table 4. Regression results of ASS on AEE.

VariableAEE
OLS2SLS
ASS0.014*** (0.019)0.020*** (0.007)
Control variablesControlledControlled
Wald chi2380.060***
Weak instrument test (F-statistic)521.220***
N5,9925,992

*** denote significance at the 1% level. Standard errors are in parentheses.

The OLS estimate shows that ASS is positively associated with eco‑economic efficiency (coefficient = 0.014, p < 0.01). The 2SLS result confirms that, after correcting for endogeneity, ASS continues to exert a positive and statistically significant effect (coefficient = 0.020, p < 0.01). Notably, the 2SLS coefficient is larger than the OLS estimate, indicating a downward bias in the OLS. This pattern is consistent with a negative selection mechanism: farmers with lower unobserved efficiency are more likely to rely on ASS, thereby suppressing the OLS estimate toward zero.

We conduct several diagnostic tests to assess the validity of the instrument. First, the first‑stage F‑statistic for the excluded instrument is 521.22, which far exceeds the rule‑of‑thumb threshold of 10, strongly rejecting the weak instrument hypothesis. Second, we perform a falsification test by including the instrument directly in the outcome equation alongside ASS and all controls; the coefficient on village committee support is 0.0008 (p = 0.362), indicating no significant direct effect on eco‑economic efficiency and supporting the exclusion restriction. The Wald chi‑squared statistic of the second stage is 380.06 (p < 0.01), confirming that the model has satisfactory explanatory power after instrumentation.

As a robustness check, we re‑estimate the model using limited information maximum likelihood (LIML), which is more robust to weak instruments, and the results remain virtually unchanged. The positive and significant relationship between ASS and eco‑economic efficiency is robust across alternative specifications, including reduced sets of control variables and different standard error adjustments.

We also note that the sign and significance of the control variables are largely consistent with expectations. For instance, agricultural labor force size and land quality positively contribute to efficiency, while off‑farm work experience and machinery accessibility show negative associations, potentially reflecting a reallocation of household resources away from meticulous farm management. These patterns are stable across OLS and 2SLS specifications, reinforcing the plausibility of the empirical model.

Previous studies have documented that ASS promotes smallholder agricultural production, particularly through the substitution of modern service inputs for traditional labor and through the technology‑embedding effect arising from advanced equipment and management practices (Ma et al., 2018; Zhang et al., 2017). The evidence consistently suggests that, by providing specialized, large‑scale, and standardized services, ASS enables smallholder farmers to use resources more efficiently and reduce waste in the production process, thereby improving agricultural eco‑economic efficiency and facilitating the green transformation of agricultural production. Our findings align with and extend this literature by showing that the effect remains robust even after accounting for endogeneity, and that the true causal impact may have been underestimated in naïve OLS models.

4.2 Robustness checks

To verify the robustness of the benchmark findings, several robustness checks are conducted by replacing the core explanatory variables, adopting alternative estimation methods, and excluding part of the sample. Specifically, this study further measures ASS using service expenditures and participation forms, re-estimates the models using the Tobit and IV-Reg2 methods, and excludes rice-growing households from the sample to reduce potential estimation bias caused by crop-specific production characteristics. The robustness test results are reported in Supplementary Tables S1–S3. Overall, the estimated coefficients of ASS remain positive and statistically significant across different model specifications, indicating that the benchmark regression results are robust and reliable.

4.3 Heterogeneity analysis

4.3.1 Heterogeneity by farm size

The effect of ASS on agricultural eco‑economic efficiency may differ between smallholder and large‑scale farmers due to their distinct production and management characteristics. Following the World Bank and Xu et al. (2018), we classify farms with a cultivated land area of <2 ha (30 mu) as smallholders and those with 2 ha (30 mu) or more as large‑scale farms. We then re‑estimate the IV‑2SLS model separately for the two groups.

Table 5 reports the heterogeneous effects of ASS across different farm scales. The estimated coefficient of ASS remains positive and statistically significant for large-scale farms, indicating that participation in ASS increases AEE by approximately 0.019. By contrast, although the estimated effect for smallholders is also positive, it does not reach statistical significance. These findings imply that large-scale farmers benefit more from ASS in terms of AEE improvement. One plausible explanation is that large‑scale farmers, who command more land and capital, are better positioned to adopt and integrate specialized services into their operations (Zou and Mishra, 2024). They can achieve substantial efficiency gains through the mechanization and standardized management that ASS provides (Yang and Li, 2022). In contrast, smallholders, with their limited land area, may struggle to realize the benefits of scale embedded in these services. The cost of service contracting might even offset potential efficiency improvements, leading to an insignificant net effect on eco‑economic performance (Qiu et al., 2021). Moreover, transaction costs and coordination challenges in service provision for fragmented small plots can further dampen the impact (Qian et al., 2022). These results underscore the importance of tailoring ASS promotion strategies to the needs and constraints of different farm size groups.

Table 5. The impact of ASS on AEE by farm scale groups.

VariableAEE
Large‑scale farmsSmallholders
ASS0.019*** (0.009)0.016 (0.010)
Controls variablesControlledControlled
Wald chi296.560***205.280***
Weak instrument test (F-statistic)33.250***54.860***
N2,8603,132

*** denote significance at the 1% level. Standard errors are in parentheses.

4.3.2 Heterogeneity by crop type

Differences in production practices across crops may generate heterogeneous effects of ASS on eco‑economic efficiency. We compute crop‑specific efficiency scores and re‑estimate the IV‑2SLS model for rice, wheat, maize, and soybean separately. Table 6 presents the results.

Table 6. The impact of ASS on AEE by crop type groups and service stage.

VariableAEE
Panel A. Crop‑specific effects of ASS on AEE
RiceWheatMaizeSoybean
VariableAEE
Panel B. Effects of ASS by service stage for wheat farmers
Land preparationSowingPest controlHarvesting
ASS0.015 (0.010)−0.074*** (0.017)−0.003 (0.008)0.011 (0.012)
Controls variablesControlledControlledControlledControlled
Wald chi268.400***139.150***89.000***30.680***
Weak instrument test (F-statistic)14.250***19.820***35.250***12.440***
N1,3501,7162,386540
ASS0.009(0.005)0.020***(0.005)0.004(0.005)−0.023*(0.006)
N1,096865876982

*** denote significance at the 1% level. Standard errors are in parentheses.

The findings reveal that ASS exerts a significantly negative effect on AEE among wheat farmers, reducing the efficiency score by approximately 0.074. By contrast, no statistically significant effects are observed for rice, maize, or soybean production.

To better understand this negative effect, we further decompose ASS into four major production-stage categories for wheat farmers, including land preparation, sowing, pest control, and harvesting services. The results indicate that the negative effect is primarily driven by harvesting services. Specifically, participation in harvesting services is associated with a significant reduction in AEE, whereas pest control services do not exhibit a significant effect, and land preparation and sowing services are positively associated with efficiency.

This heterogeneous pattern may reflect a mismatch between the agronomic requirements of wheat cultivation and the service configurations currently available in the market. The literature suggests that farmers’ demand for agricultural socialized services varies considerably across production stages. Mechanized services in land preparation and harvesting tend to exhibit relatively high adoption rates, whereas outsourcing of fertilizer and pesticide application remains less prevalent due to greater difficulties in monitoring service quality and input application practices(Wang and Huang, 2025; Qing et al., 2023). In wheat production, critical growth‑stage management practices such as precise fertilization and pest control are heavily reliant on timely and knowledge‑intensive interventions. If available ASS packages cannot effectively cover these sensitive tasks, the resulting gap may compromise eco‑economic outcomes. Moreover, the high degree of standardization in wheat harvesting services may drive farmers toward excessive mechanization rather than ecologically sound practices, partially explaining the negative coefficient (Ma et al., 2018).

A plausible explanation is that wheat production in China is already highly mechanized, particularly during the harvesting stage. Harvesting services are therefore characterized by a high degree of standardization and scale-oriented operation. Although such services can reduce labor requirements and improve operational efficiency, they may not adequately account for the ecological management needs of individual plots. Consequently, the ecological costs associated with highly standardized harvesting operations may partially offset their productivity gains, resulting in lower AEE.

5 Mechanism analysis: how ASS improve AEE

The baseline results establish that ASS exert a positive causal effect on AEE. Following the analytical framework proposed by Jiang (2022), this study further examines whether material input reallocation and labor reallocation serve as mediating channels linking ASS and AEE. The results of the mediation analysis (Table 7) reveal that the positive effect of ASS on AEE operates primarily through a material input reallocation channel. Conducted through a causal steps approach within an IV-2SLS framework, the mediation analysis reveals that the positive effect of ASS on AEE operates primarily through a material input reallocation channel.

Table 7. Mediating effects of ASS on AEE.

VariableEfficiency(total effect)Fert-pest ratio (ASS → Med)Labor ratio (ASS → Med)Efficiency (direct + mediators)
ASS0.020*** (0.007)−12.313** (5.658)0.045 (0.028)0.020*** (0.007)
Fert-pest ratio−0.000** (0.000)
Labor reallocation−0.002 (0.004)
Control variablesControlledControlledControlledControlled
First-stage F (IV)110.490***110.330***108.370***93.410***
N5,992

* and denote significance at the 1% and 5% level. Standard errors are in parentheses. The coefficient on the fertilizer–pesticide ratio in the efficiency equation is −0.00004 (standard error = 0.00002) and is statistically significant at the 5% level.

The first-stage results for the mediators (Columns 2 and 3) show that ASS adoption significantly reduces the ratio of chemical fertilizer and pesticide expenditure to total variable input costs by approximately 12.3 percentage points. This finding suggests that outsourcing production tasks to specialized service providers facilitates a more cost-effective and environmentally sustainable allocation of material inputs.

In contrast, ASS does not exert a statistically significant effect on the labor reallocation, implying that the anticipated transfer of household labor from agricultural to non-agricultural sectors has not yet emerged, at least within the scope captured by this proxy variable.

When both mediators are incorporated into the efficiency equation (Column 4), the fertilizer and pesticide expenditure ratio displays a negative and statistically significant association with eco-economic efficiency, whereas the coefficient on the labor reallocation variable remains insignificant. Meanwhile, the direct effect of ASS on AEE remains positive and highly significant, and its magnitude changes only marginally relative to the total effect. This pattern indicates that the mediating role of material input reallocation exists but remains limited in magnitude, while the labor reallocation channel fails to generate a statistically identifiable mediating effect.

Nevertheless, the statistically significant indirect effect operating through the fertilizer and pesticide expenditure ratio confirms that material input reallocation serves as an important mediating mechanism linking ASS and AEE. Specifically, the adoption of outsourced agricultural services reduces the intensity of chemical input use per unit of output, thereby contributing to improvements in AEE. This finding is consistent with previous studies suggesting that specialized agricultural services help mitigate the excessive use of agrochemicals through more standardized and technology-intensive production practices (Wang et al., 2024; Zhang et al., 2023).

By contrast, the labor reallocation channel, proxied by the share of non-agricultural labor in total household labor, does not exhibit a statistically significant mediating effect. One possible explanation is that ASS primarily substitute for labor-intensive production activities rather than directly inducing an immediate reallocation of household labor from agricultural to non-agricultural employment. In addition, labor transfer decisions are typically influenced by local labor market conditions, non-farm employment opportunities, and household livelihood strategies, and therefore tend to occur gradually over time (Caunedo and Kala, 2021). As a result, the labor reallocation effect may not be fully captured in cross-sectional data and does not generate a statistically significant mediating effect in the present analysis.

6 Discussion

The positive association between ASS adoption and AEE is consistent with previous evidence that agricultural services improve production efficiency and environmental performance through specialization, technology-embedded operations, and access to modern management practices (Chen et al., 2022; Cheng et al., 2022). It also accords with studies reporting gains in agricultural land-use ecological efficiency and green total factor productivity following service adoption (Li et al., 2024; Yao et al., 2024). Because AEE incorporates economic output, resource use, and environmental costs, the present results indicate that the efficiency gains documented in earlier work may be accompanied by improvements in the balance between production and environmental pressure. Nevertheless, the heterogeneous estimates show that this relationship is not uniform across production contexts.

The positive farm-size estimate is concentrated among farms operating at least 2 ha, whereas the estimate for smallholders is not statistically significant. This pattern is consistent with studies emphasizing the scale economies associated with machinery and specialized services (Qian et al., 2022; Zou and Mishra, 2024). Larger farms can spread coordination and contracting costs over a greater cultivated area and may be better able to integrate standardized operations into farm management. For smallholders, fragmented plots and the fixed costs of coordinating service provision can reduce the efficiency gains from outsourcing, in line with the smallholder mechanization-service paradox discussed by Qiu et al. (2021). The farm-size results therefore suggest that the performance of ASS depends partly on whether farm organization and plot conditions are compatible with standardized service delivery.

The crop-specific results also differ from a uniform positive relationship. In particular, the negative estimate for wheat contrasts with evidence that outsourced fertilizer and pesticide application can improve the technical efficiency of wheat production (Zhang et al., 2023). The two findings concern different service compositions and outcomes: the present analysis measures adoption across production stages and evaluates eco-economic efficiency, whereas Zhang et al. examine particular input-application services and conventional technical efficiency. The stage-specific estimates are consistent with this distinction. Sowing services are positively associated with wheat AEE, while the negative estimate is concentrated in harvesting services. Mechanized harvesting can reduce labor requirements and accelerate operations, but highly standardized service delivery may yield smaller environmental gains when it does not accommodate plot-level ecological and agronomic conditions.

The material-input result is broadly consistent with studies showing that specialized services can reduce farmers' dependence on chemical inputs through standardized and technology-intensive operations (Wang et al., 2024; Zhang et al., 2023). However, the observed decline concerns the expenditure share of fertilizer and pesticides rather than their physical application rates, so it should be interpreted as a change in the cost composition of variable inputs. The absence of a statistically significant labor-reallocation channel differs from studies emphasizing the labor-releasing effects of mechanization and service outsourcing (Caunedo and Kala, 2021; Zou and Mishra, 2024). This result does not preclude labor saving within agricultural production; rather, it indicates that such saving is not reflected in a higher share of household members engaged in non-agricultural work in the survey data. Entry into non-farm employment also depends on local employment opportunities, skills, and household livelihood strategies and may occur with a time lag.

Several aspects are not fully captured in the present analysis. First, the one-time cross-sectional survey does not observe changes in service use, AEE, or household employment over time and therefore cannot assess the persistence of the estimated relationships or delayed adjustments in labor allocation. Second, ASS adoption is defined as the purchase of services in any of four production stages. This measure does not distinguish provider type, service quality, contract terms, frequency, or the intensity of service use within a stage. Third, the fertilizer-pesticide mechanism is measured by its expenditure share rather than physical application rates, while the labor measure records participation shares rather than working time, earnings, or seasonal mobility; changes in input prices and the composition of other costs may therefore affect these measures. Fourth, the AEE indicators rely on household-reported production data and standardized carbon-emission coefficients. Direct measurements of input quantities, soil and water outcomes, and emissions would provide a more complete account of environmental performance. Finally, the sample consists of grain-producing households in 14 Chinese provinces, and the observed relationships may differ for other crops, regions, or service-market and institutional settings. Panel data and more detailed measures of service characteristics and environmental outcomes would allow these issues to be examined more directly.

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.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fsufs.2026.1880279/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Arcelia Toledo-López, Instituto Politécnico Nacional, Mexico

Reviewed by: Cai Baozhong, Hunan University of Science and Engineering, China

Weili Zhu, Beijing Forestry University, China

Statements

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.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research was supported by the project "High-level Talent Research Start-up" Fund Project of Weifang University of Science and Technology (SKRC2025005).

Data availability

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

Ethics approval

Ethical approval was not required for the studies involving humans. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

WT: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Writing – original draft, Writing – review & editing. XL: Funding acquisition, Project administration, Supervision, Validation, Writing – review & editing.

References

  1. Angrist J. D., Pischke J.-S. (2009). Mostly Harmless Econometrics: An Empiricist's Companion. Princeton: Princeton University Press.
  2. Badunenko O., Tauchmann H. (2019). Simar and Wilson two-stage efficiency analysis for Stata. Stand. Genomic Sci. 19, 950 – 988. doi: doi:10.1177/1536867X19893640
  3. Caunedo J., Kala N. (2021). Mechanizing Agriculture: Impacts on Labor and Productivity. NBER Working Paper No. 29061. Cambridge, MA, United States: National Bureau of Economic Research. doi: doi:10.3386/w29061
  4. Cheng Y., Zhang D., Wang X. (2022). Green development effect of agricultural socialized services: an analysis based on farming households' perspective. Resour. Sci. 44, 1848 – 1864. doi: doi:10.18402/resci.2022.09.09
  5. Chen T., Rizwan M., Abbas A. (2022). Exploring the role of agricultural services in production efficiency in Chinese agriculture: a case of the socialized agricultural service system. Land 11: 347. doi: doi:10.3390/land11030347
  6. Coelli T., Rao D. S. P., O'Donnell C. J., Battese G. E. (2005). An Introduction to Efficiency and Productivity Analysis. New York: Springer.
  7. Deng X., Xu D., Zeng M., Qi Y. (2020). Does outsourcing affect agricultural productivity of farmer households? Evidence from China. China Agric. Econ. Rev. 12, 673 – 688. doi: doi:10.1108/CAER-12-2018-0236
  8. Diao X., Cossar F., Houssou N., Kolavalli S. (2014). Mechanization in Ghana: emerging demand, and the search for alternative supply models. Food Policy 48, 168 – 181. doi: doi:10.1016/j.foodpol.2014.05.013
  9. Duan H. P., Zhang Y., Zhao J. B., Bian X. M. (2011). Carbon footprint analysis of farmland ecosystem in China. J. Soil Water Conserv. 25, 203 – 208. doi: doi:10.13870/j.cnki.stbcxb.2011.05.020
  10. Farrell M. J. (1957). The measurement of productive efficiency. J. R. Stat. Soc. Ser. A 120, 253 – 281. doi: doi:10.2307/2343100
  11. Hayami Y., Ruttan V. W. (1985). Agricultural Development: An International Perspective. Rev. and expanded ed. Baltimore: Johns Hopkins University Press.
  12. IPCC (2023). Climate Change 2023: Synthesis Report. Geneva: Intergovernmental Panel on Climate Change.
  13. Jiang T. (2022). Mediating effects and moderating effects in causal inference. China Ind. Econ. 5, 100 – 120. doi: doi:10.19581/j.cnki.ciejournal.2022.05.005
  14. Jiang T., Wang M., Zhang W., Zhu C., Wang F. (2024). A comprehensive analysis of agricultural non-point source pollution in China: current status, risk assessment and management strategies. Sustainability 16: 2515. doi: doi:10.3390/su16062515
  15. Jiang W., Zhu J., Zhang Q., Razzaq A. (2026). Internet use shaped farmer adoption of integrated pest management in rural central China. Front. Sustain. Food Syst. 9: 1708905. doi: doi:10.3389/fsufs.2025.1708905
  16. Knowler D., Bradshaw B. (2007). Farmers' adoption of conservation agriculture: a review and synthesis of recent research. Food Policy 32, 25 – 48. doi: doi:10.1016/j.foodpol.2006.01.003
  17. Levidow L., Lindgaard-Jørgensen P., Nilsson Å., Assimacopulos D. (2014). Eco-efficiency improvements in industrial water-service systems: assessing options with stakeholders. Water Sci. Technol. 69, 2113 – 2121. doi:,. doi:10.2166/wst.2014.131
  18. Li P., He L., Zhang J., Han H., Song Y. (2024). Research on the impact of agricultural socialization services on the ecological efficiency of agricultural land use. Land 13: 853. doi: doi:10.3390/land13060853
  19. Liu Y., Sun D., Wang H., Wang X., Yu G., Zhao X. (2020). An evaluation of China's agricultural green production: 1978–2017. J. Clean. Prod. 243: 118483. doi: doi:10.1016/j.jclepro.2019.118483
  20. Ma W., Renwick A., Grafton Q. (2018). Farm machinery use, off-farm employment and farm performance in China. Aust. J. Agric. Res. Econ. 62, 279 – 298. doi: doi:10.1111/1467-8489.12249
  21. Min R., Li G. C. (2013). Grain's production efficiency and its spatial distribution in China from the “two-oriented (resource and environment)” perspective. Econ. Geogr. 33, 144 – 149. doi: doi:10.15957/j.cnki.jjdl.2013.03.023
  22. 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: doi:10.1016/j.jclepro.2020.120636
  23. Peng Y., Guo J., Liu Y., Liao L., Jiang S., Tang Y. (2025). Will land transfer improve grain farmers’ adoption of agricultural green production technology? Evidence from Jiangxi Province, China. Humanit. Soc. Sci. Commun. 12: 466. doi: doi:10.1057/s41599-025-04767-2
  24. Pingali P. (2012). Green revolution: impacts, limits, and the path ahead. Proc. Natl Acad. Sci. USA 109, 12302 – 12308. doi:,. doi:10.1073/pnas.0912953109
  25. Pretty J. (2008). Agricultural sustainability: concepts, principles and evidence. Philos. Trans. R. Soc. Lond. B Biol. Sci. 363, 447 – 465. doi:,. doi:10.1098/rstb.2007.2163
  26. Qian L., Lu H., Gao Q., Lu H. (2022). Household-owned farm machinery vs. outsourced machinery services: the impact of agricultural mechanization on the land leasing behavior of relatively large-scale farmers in China. Land Use Policy 115: 106008. doi: doi:10.1016/j.landusepol.2022.106008
  27. Qing C., Zhou W., Song J., Deng X., Xu D. (2023). Impact of outsourced machinery services on farmers' production behavior: evidence from Chinese rice farmers. J. Environ. Manag. 327: 116843. doi: doi:10.1016/j.jenvman.2022.116843
  28. Qiu T., Shi X., He Q., Luo B. (2021). The paradox of developing agricultural mechanization services in China: supporting or kicking out smallholder farmers? China Econ. Rev. 69: 101680. doi: doi:10.1016/j.chieco.2021.101680
  29. Rogers E. M. (2003). Diffusion of Innovations. 5th ed. New York: Free Press.
  30. Rüdenauer I., Gensch C. O., Grießhammer R., Bunke D. (2005). Integrated environmental and economic assessment of products and processes. J. Ind. Ecol. 9, 105 – 116. doi: doi:10.1162/108819805775248061
  31. Simar L., Wilson P. W. (2007). Estimation and inference in two-stage, semi-parametric models of production processes. J. Econ. 136, 31 – 64. doi: doi:10.1016/j.jeconom.2005.07.009
  32. Stock J. H., Watson M. W. (2020). Introduction to Econometrics. 4th ed. Boston: Pearson.
  33. Takeshima H., Nin-Pratt A., Diao X. (2013). Mechanization and agricultural technology evolution, agricultural intensification in sub-Saharan Africa: typology of agricultural mechanization in Nigeria. Am. J. Agric. Econ. 95, 1230 – 1236. doi: doi:10.1093/ajae/aat045
  34. Tilman D., Cassman K. G., Matson P. A., Naylor R., Polasky S. (2002). Agricultural sustainability and intensive production practices. Nature 418, 671 – 677. doi:, doi:10.1038/nature01014
  35. Tone K., Tsutsui M. (2010). An epsilon-based measure of efficiency in DEA: a third pole of technical efficiency. Eur. J. Oper. Res. 207, 1554 – 1563. doi: doi:10.1016/j.ejor.2010.07.014
  36. Wang L., Lyu J., Zhang J. (2024). Explicating the role of agricultural socialized services on chemical fertilizer use reduction: evidence from China using a double machine learning model. Agriculture 14: 2148. doi: doi:10.3390/agriculture14122148
  37. Wang X., Huang Y. (2025). Agricultural socialized services in China's smallholder farming systems: a systematic review of service types, benefits, and measurement challenges. Front. Agron. 7: 1638637. doi: doi:10.3389/fagro.2025.1638637
  38. Williamson O. E. (1985). The Economic Institutions of Capitalism. New York: Free Press.
  39. Wooldridge J. M. (2010). Econometric Analysis of Cross Section and Panel Data. 2nd ed. Cambridge: MIT Press.
  40. Xu Z., Zhang J., Lv K. (2018). The scale of operation, term of land ownership and the adoption ofinter-temporal agricultural technology: an example of “straw return tosoil directly”. Chin. Rural. Eco. 3, 61 – 74. doi: doi:10.20077/j.cnki.11-1262/f.2018.03.005
  41. Yang Q., Huang, Zou (2018). Is agricultural social service conducive to the improvement of agricultural production efficiency? An empirical analysis based on three stage DEA model. J. China Agric. Univ. 23, 232 – 244. doi: doi:10.11841/j.issn.1007-4333.2018.11.24
  42. Yang S., Li W. (2022). The impact of socialized agricultural machinery services on land productivity: evidence from China. Agriculture 12: 2072. doi: doi:10.3390/agriculture12122072
  43. Yao W., Zhu Y., Liu S., Zhang Y. (2024). Can agricultural socialized services promote agricultural green total factor productivity? From the perspective of production factor allocation. Sustainability 16: 8425. doi: doi:10.3390/su16198425
  44. Zhang X., Yang J., Thomas R. (2017). Mechanization outsourcing clusters and division of labor in Chinese agriculture. China Econ. Rev. 43, 184 – 195. doi: doi:10.1016/j.chieco.2017.01.012
  45. Zhang Y., Yin Y., Li F., Duan W., Xu K., Yin C. (2023). Can outsourcing improve the technical efficiency of wheat production with fertilization and pesticide application? Evidence from China. J. Clean. Prod. 415: 137952. doi: doi:10.1016/j.jclepro.2023.138587
  46. Zou B., Mishra A. K. (2024). Modernizing smallholder agriculture and achieving food security: an exploration in machinery services and labor reallocation in China. Appl. Econ. Perspect. Policy 46, 1662 – 1691. doi: doi:10.1002/aepp.13433

Citation

Tang, W. & Li, X. (2026). The impact of agricultural socialized services on grain farmers’ agricultural eco‑economic efficiency. Frontiers in Sustainable Food Systems, Volume 10. https://doi.org/
Received
Accepted
Published
Volume
10 - 2026

Copyright

Copyright © 2026 Tang and Li.

License: https://creativecommons.org/licenses/by/4.0/

*Correspondence: Xiaohui Li

Article metrics

Views
0
Downloads
0
Citations
0
Shares
0

Related articles

Original ResearchMaking inequities visible: a call for participatory food system dashboards for SDG monitoringRositsa T. Ilieva, Yvette Ng, Nevin CohenReviewAvocado (Persea americana Mill) seed starch: composition, extraction, and applications—a comprehensive reviewNehal Chavda, Rajdip Khaniya, Margi Gokalani, Ajay SinghBrief Research ReportComparison of food carbon footprints using 20- and 100-year timeframesKatie Osborn, Silas Crocker, Kira Presley, Alexis Shenkiryk, Maggie Spire, Jennifer Molidor, Mansha Kapur, Jennifer A. Jay