Introduction

As China transitions into a phase of high-quality development, green agriculture has become a key driver in China’s efforts to achieve sustainable development, particularly in balancing agricultural productivity with ecological conservation, a topic that has garnered significant global attention1,2,3. Despite substantial government efforts to promote green agricultural technologies, such as conservation tillage, challenges persist. These include limited extension resources, a shortage of extension personnel, and a mismatch between the services provided and the actual needs of farmers4,5. Additionally, environmental constraints, such as natural disasters and climate change, have hindered the widespread adoption of conservation tillage practices by farmers6,7. In 2019 (due to the fact that the latest official statistics of cultivated land area are 2019 data, the proportion of 2019 is still calculated), conservation tillage accounted for less than 6.4% of cultivated land in China8. This low adoption rate is particularly concerning given the well-documented benefits of conservation tillage, which include improved soil health, reduced erosion, and lower greenhouse gas emissions9,10.

One of the key challenges hindering the widespread adoption of conservation tillage technology is the difficulty in fostering effective collaboration among governments, agribusinesses, and rural households (GAR collaboration)11,12,13. GAR collaboration refers to a collaborative effort among governments, agribusinesses, and rural households aimed at improving agricultural productivity for farmers, maintaining ecological balance to meet governmental environmental targets, and promoting agribusiness profitability, ultimately achieving mutual benefits for all stakeholders14. The dual benefits of conservation tillage technology, which encompass both higher crop yields and improved farmland ecological environments, reflect both private economic returns and public ecological goods15. This duality can lead to a “collective action dilemma”, creating a challenge in coordinating parites to share both the costs and benefits of adopting conservation technologies, leading to difficulties for agribusinesses to justify investments without guaranteed returns and for rural households to prioritize long-term ecological benefits over short-term economic gains16,17,18. GAR collaboration helps stakeholders align their short-term economic interests with long-term ecological goals by fostering trust and sharing common goals, such as increasing crop yields, improving soil health, and raising income through joint training programs, policy support, and market-based incentives, thereby they are more likely to work collaboratively toward achieving these objectives, addressing the collective action dilemma.

Existing literature has explored the role of collaboration in agricultural technology adoption, particularly in bilateral collaborations such as public-private partnerships (PPP) and agribusiness-rural household partnerships, including models like “leading enterprises + cooperatives + farmers” and “family farms + leading enterprises”19,20,21. However, most studies have not fully examined the triadic collaboration required for comprehensive green technology adoption22. With the rapid development of the fourth industrial revolution, e-commerce and platform economy have flourished, providing new opportunities for the transformation and upgrading of the agricultural industry. Under this background, governments, agribusinesses, and rural households become more integrated, GAR collaboration becomes crucial in promoting green agriculture. However, in-depth discussions on the mechanism of green agricultural technology adoption from the GAR collaboration perspective remain sparse.

Taking conservation tillage technology as a specific case, this paper aims to illustrate the impact of GAR collaboration on green agricultural development, namely, economic performance, social performance and ecological performance. We introduce a multifaceted framework to explore the mechanisms by which environmental characteristics and collaboration capacity influence the effectiveness of conservation tillage technology adoption. Among them, the environmental characteristics refer to the combination of various natural and artificial factors affecting agricultural production and development. These include natural factors such as climate, soil, terrain, water sources, and the ecological environment, as well as artificial factors such as policies and regulations, market demand, and technological progress23. Our paper mainly focuses on the natural factors related to environmental characteristics, namely soil erosion degree, land scale, and village location, which significantly impact conservation tillage technology adoption due to their uncertainty, complexity, and interdependence24. Collaboration capacity is defined as the ability of governments, agribusinesses, and rural households to work together effectively to achieve shared goals, such as improving agricultural productivity, enhancing environmental sustainability, and boosting economic returns, which is operationalized through various dimensions such as constructive engagement, shared vision, and joint execution and regulation25. Specifically, we examines how collaboration among all parties can enhance the adoption effectiveness. Our study employs a cross-site study design to evaluate both the direct and indirect effects of environmental characteristics and collaboration capacity on the effectiveness of conservation tillage technology adoption. Furthermore, we investigate the mechanism of defection risks and transaction costs. This research contributes to the theoretical foundations of collaborative approaches within the context of green agriculture development and offers practical solutions to enhance the effectiveness of conservation tillage technology adoption.

We makes three key contributions to the existing literature on agricultural technology adoption and collaborative governance: First, we introduce an integrative theoretical framework that combines environmental characteristics and collaboration capacity to examine their joint effects on the effectiveness of conservation tillage adoption. This framework not only advances the theoretical understanding of these relationships but also offers practical insights for policymakers. Second, it provides empirical evidence from rural China on the role of GAR collaboration in promoting the adoption of conservation tillage technology, filling a gap in the literature where few studies have focused on GAR collaboration in China. Third, the study reveals the heterogeneous effects of collaboration capacity, showing that GAR collaboration varies significantly by village size and economic conditions, which underscores the importance of local context in collaborative efforts.

The article is structured as follows: The second section outlines the theoretical framework and hypotheses. The third section elaborates on the study design, including data, variable, and methods. The fourth section presents the empirical analysis and results. The fifth section investigates the mechanisms of transaction costs and defection risks between GAR collaboration and the effectiveness of conservation tillage technology adoption. The sixth section concludes the study and provides policy implications.

Theoretical framework

Our study is grounded in the Integrative Framework for Collaborative Governance (IFCG) by Emerson, Nabatchi, and Balog26, which emphasizes the impact of the system’s context on collaborative outcomes. We use this framework to explore the roles of environmental characteristics and collaboration capacity in influencing the effectiveness of conservation tillage technology adoption (as shown in Fig. 1). Environmental characteristics, including soil quality, land scale, and geographical location, affect the feasibility of conservation tillage and influence GAR collaboratio27. Based on this refined framework, the hypotheses are developed to explore how environmental characteristics directly affect adoption effectiveness and indirectly affect through collaboration capacity.

Fig. 1
figure 1

Theoretical analysis framework.

Full size image

Theoretical framework and hypotheses

Theoretical hypotheses

Impact of collaboration capacity on the effectiveness of conservation tillage technology adoption

As a dynamic mechanism of GAR collaboration, collaboration capacity reflects the operational strength of GAR collaboration25. High collaboration capacity enhances the ability of stakeholders to align their actions and share resources, thereby reducing uncertainty and improving the overall effectiveness of conservation tillage technology adoption28.

Constructive engagement refers primarily to the time and resource investment made by stakeholders to support collaboration29. This includes both tangible resources, such as financial and labor contributions, and intangible resources, such as time and commitment to meetings, training, and information sharing. Sufficient resource investment and sustained involvement from all parties are critical for ensuring the success of GAR collaboration and optimizing resource allocation30. Constructive engagement makes parties build trust by participating in regular meetings and training, enhance communication by sharing real-time information, increase cohesion by investing financial and labor, all of which enhance the effectiveness of conservation tillage technology adoption. This study characterizes constructive engagement across four dimensions: financial contribution, labor contribution, meeting organization, and training participation. The following hypothesis was formulated:

H1a: Constructive engagement significantly promotes the effectiveness of conservation tillage technology adoption.

Shared vision plays a key role in ensuring that the goals and objectives of governments, agribusinesses, and rural households are aligned, reflecting the relational dimension of GAR collaboration in conservation tillage technology adoption, encompassing trust, reciprocity, recognition, and commitment24,31. When stakeholders share common goals and a clear understanding of what will happen when they are achieved, such as increased food production, improved soil fertility, and improved ecosystems, they are more likely to work collaboratively toward achieving these objectives. This shared understanding helps mitigate conflicts, enhances communication, and fosters commitment among all parties32. By focusing on a unified objective, parties can collaborate their actions more effectively, leading to better resource allocation and improved decision-making. As a result, the adoption process becomes more efficient, reducing delays caused by conflicting interests among parties and ensuring clearer roles and responsibilities, which enhances the successful implementation of conservation tillage technology33.

Given the difficulty of directly measuring the shared vision of different parties at the village level, we use indicators such as village committee elections, responsiveness to the one-case-one-discussion (public deliberation) system, and commitment to public affairs to reflect the degree of shared vision among different subjects in the collaboration process34,35.

H1b: A shared vision significantly promotes the effectiveness of conservation tillage technology adoption.

Joint execution and regulation reflect the functional dimension of GAR collaboration in conservation tillage technology adoption, encompassing procedural and institutional arrangements, leadership, rules, and regulations25,36. Despite differences in responsibilities, interests and resource conditions37, when parties adhere to common rules and procedures, they can collectively manage resources, address any emerging conflicts, and ensure that all parties adhere to agreed-upon rules12. Further, it can reduces opportunistic behavior and ensured parties focused on achieving common goals, enhancing the stability of collaboration. This in turn facilitates the more efficient and timely implementation of conservation tillage technology, ensuring that resources are allocated effectively and that each party fulfills its responsibilities, leading to higher technology uptake rates38,39. Our study uses indicators such as supervision measures, punishment measures, personnel allocation, and regulation implementation to reflect joint execution and regulation in conservation tillage technology adoption.

H1c: Joint execution and regulation significantly promote the effectiveness of conservation tillage technology adoption.

Impact of environmental characteristics on GAR collaboration capacity and conservation tillage technology adoption

Environmental characteristics are fundamental in affecting the effectiveness of conservation tillage technology adoption. These characteristics include natural factors soil erosion, economic factors like land scale, and geographical factors such as village location.

Firstly, natural disasters, particularly soil erosion, have a substantial impact on agricultural production6. The adoption of conservation tillage technology has been proven to effectively reduce soil erosion, helping to mitigate operational risks and economic losses associated with natural disasters40. For instance, straw-returning technology, when combined with no-tillage or deep-loosening practices, improves crop resilience to natural disasters41. Secondly, conservation tillage technology exhibits economies of scale due to farming mechanization42. As land scales increases, the adoption of conservation tillage technology becomes more efficient, reducing the costs associated with labor, negotiation, supervision, and equipment use, making the conservation tillage technology more cost-effective for larger farms. Thirdly, the location of a village plays a crucial role in resource utilization and value realization43. Villages located closer to economic and administrative centers benefit from improved access to agricultural markets, lower transportation costs, and better access to extension services that provide technical support and resources for implementing conservation tillage practices. This proximity not only makes resource allocation more efficient but also increases the market value of agricultural products, thus enhancing the effectiveness of conservation tillage adoption.

Environmental characteristics also have an indirect influence on the effectiveness of conservation tillage adoption by shaping GAR collaboration. Specifically, soil erosion increase the need for collaboration. When soil erosion increase are high, collaborative action is required to mitigate these risks. This need enhances GAR collaboration, indirectly promoting the effectiveness of conservation tillage technology adoption by improving risk management and resource allocation. Furthermore, a village’s land scale also impacts effectiveness of conservation tillage technology adoption indirectly through collaboration capacity44. For large-scale farms and cooperatives, a larger land scale helps reduce transaction costs in collaboration by minimizing the need for extensive collaboration and oversight. This efficiency in managing resources and operations also promotes improvements in conservation farming techniques, further enhancing the effectiveness of conservation tillage technology adoption45,46. For smallholders, however, as land scales expand, the costs associated with monitoring and managing collaboration efforts increase. These higher transaction costs can strain their limited resources, reducing their ability to effectively collaborate with other stakeholders and potentially weakening their effectiveness of conservation tillage technology adoption. Thus, land scale indirectly affects adoption effectiveness through collaboration capacity, depending on the size and management complexity of collaborating entities. Fianlly, villages located closer to economic and administrative centers have better market access, reduced transportation costs, and more efficient communication. When parties can interact more easily due to geographical proximity, they are able to collaboration more effectively, indirectly boosting the effectiveness of conservation tillage technology adoption.

H2: Environmental characteristics have both a direct impact and an indirect impact on the effectiveness of conservation tillage technology adoption through collaboration capacity.

Data and methods

Study area

We conducted a survey in Ningxia, Shaanxi, and Gansu Province (as shown in Fig. 2), which are located in the arid and semi-arid regions of the Eurasian continent. The natural environment in these regions is influenced by the westerly climate, plateau monsoon climate, and East Asian monsoon climate, resulting in predominantly dry conditions with low levels of rainfall. The average annual temperature ranges from 4 to 14 °C, and annual precipitation is around 200–800 mm. The combination of a dry and rainless climate, rugged terrain, loose soil, and human intervention has led to the Loess Plateau being one of the areas most severely affected by soil erosion.

Taking into account the natural conditions, crop varieties, cultivation patterns, and farming systems in the northwest region, the government has implemented a series of measures to increase the adoption of conservation tillage. These measures include farmland fertility protection subsidies, agricultural machinery purchase subsidies, and conservation farming integrated innovation and demonstration projects.

Data collection

The research methodology followed the guidelines and regulations of the Humanities and Social Sciences Research Committee of Yangzhou University, which complies with the Declaration of Helsinki. The research proposal received approval from this committee. Informed consent was obtained from all subjects prior to their participation in this study. Each participant was thoroughly informed about the purpose of the research, and their anonymity was guaranteed.

Data were collected using a multi-stage stratified random sampling method across three provinces. The selection of these regions was primarily due to their location within the Loess Plateau gully area in northwest China. This area has historically struggled with soil erosion due to infertile soils and fragile ecological environments, making conservation tillage technology adoption essential for promoting agricultural green transformation and sustainable development.

Shaanxi’s Yulin, Gansu’s Qingyang, and Ningxia’s Guyuan were selected as research areas because they were demonstrated by sites established by the Ministry of Agriculture and the Provincial Agricultural Machinery Bureau in the Loess Plateau gully region, and the successful implementation of conservation tillage practices. Subsequently, 2–3 counties were selected from each city using a stratified sampling method, considering factors such as project layout, economic conditions, and geographical location. Furthermore, 3–5 townships were chosen in each county based on representativeness. Building upon this, 4–6 villages were sampled by simple random sampling.

To ensure data accuracy, we conducted individual or group interviews with various subjects such as government personnel engaged in conservation tillage technology promotion, village cadres, cooperative leaders, enterprise executives, large-scale landholders, and family farm owners. A total of 173 villages provided valid information through these interviews. The survey questionnaires collected data on basic village conditions, soil erosion, conservation tillage technology training, collaboration in implementation, and the effect of conservation tillage technology adoption.

Fig. 2
figure 2

Study area.

Full size image

Method

Structural equation modeling is a linear statistical modeling technique that analyzes the intrinsic structural mechanism of related phenomena by statistical hypothesis testing. Structural equation models are capable of explaining complex relationships between multiple dependent and independent variables, including their direct and indirect relationships, through latent variables. In addition, structural equation modeling can account for the possible effects of missing variables on model estimates. Moreover, by analyzing the correlation coefficients between variables and deriving the real causal relationship according to theory and experience, the endogenous problem of the fixed model can be weakened to a certain extent. Therefore, our study used AMOS statistical software to perform structural equation analysis to test (a) the impact of environmental characteristics on collaboration capacity, (b) the impact of collaboration capacity on adoption effectiveness, and (c) the impact of environmental characteristics on adoption effectiveness (as shown in Fig. 1).

In order to verify the robustness of the path analysis, we also used the Ordered Logistic model and the OLS model for robustness testing:

$$\:{\text{E}\text{f}\text{f}\text{e}\text{c}\text{t}\text{i}\text{v}\text{e}\text{n}\text{e}\text{s}\text{s}}_{\text{i}\text{k}}={{\upalpha\:}}_{0}+{{{\upalpha\:}}_{1}\text{C}\text{o}\text{l}\text{l}\text{a}\text{b}\text{o}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{v}}{+{{\upalpha\:}}_{2}\text{C}\text{i}\text{r}\text{c}\text{u}\text{m}\text{s}\text{t}\text{a}\text{n}\text{c}\text{e}}_{\text{i}\text{w}}+{{\upepsilon\:}}_{1}$$
(1)
$$\:{\text{C}\text{o}\text{l}\text{l}\text{a}\text{b}\text{o}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{v}}={{\upbeta\:}}_{0}+{{{\upbeta\:}}_{1}\text{C}\text{i}\text{r}\text{c}\text{u}\text{m}\text{s}\text{t}\text{a}\text{n}\text{c}\text{e}}_{\text{i}\text{w}}+{{\upepsilon\:}}_{2}$$
(2)

where \(\:{\text{E}\text{f}\text{f}\text{e}\text{c}\text{t}\text{i}\text{v}\text{e}\text{n}\text{e}\text{s}\text{s}}_{\text{i}\text{k}}\) represents the effectiveness of conservation tillage technology adoption of \(\:{\text{v}\text{i}\text{l}\text{l}\text{a}\text{g}\text{e}}_{\text{i}}\), k = 1, 2, 3 respectively corresponds to economic, ecological and social performance. \(\:{\text{C}\text{o}\text{l}\text{l}\text{a}\text{b}\text{o}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{v}}\) indicates collaboration capacity of all parties of \(\:{\text{v}\text{i}\text{l}\text{l}\text{a}\text{g}\text{e}}_{\text{i}}\), v = 1, 2, 3 respectively corresponds constructive engagement, shared vision, and joint execution and regulation.\(\:\:{\text{C}\text{i}\text{r}\text{c}\text{u}\text{m}\text{s}\text{t}\text{a}\text{n}\text{c}\text{e}}_{\text{i}\text{w}}\) denotes environmental characteristics of \(\:{\text{v}\text{i}\text{l}\text{l}\text{a}\text{g}\text{e}}_{\text{i}}\), w = 1, 2, 3 respectively represents the soil erosion degree, land scale, and village location. \(\:{{\upalpha\:}}_{0},\:{{\upalpha\:}}_{1},\:{{\upalpha\:}}_{2},\:{{\upbeta\:}}_{0},\:\:{{\upbeta\:}}_{1}\) are a series of coefficients to be estimated, \(\:{{\upepsilon\:}}_{1},\:{{\upepsilon\:}}_{2}\) are the random error.

On this basis, we further examine the mechanism of GAR collaboration on the effectiveness of conservation tillage technology adoption from the aspects of transaction costs and defection risks.

$$\:{\text{C}\text{o}\text{s}\text{t}}_{\text{i}}={{\upgamma\:}}_{0}+{{{\upgamma\:}}_{1}\text{C}\text{o}\text{l}\text{l}\text{a}\text{b}\text{o}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{v}}{+{{\upgamma\:}}_{2}\text{C}\text{i}\text{r}\text{c}\text{u}\text{m}\text{s}\text{t}\text{a}\text{n}\text{c}\text{e}}_{\text{i}\text{w}}+{{\upepsilon\:}}_{3}$$
(3)
$$\:{\text{R}\text{i}\text{s}\text{k}}_{\text{i}}={{\upvartheta\:}}_{0}+{{{\upvartheta\:}}_{1}\text{C}\text{o}\text{l}\text{l}\text{a}\text{b}\text{o}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}}_{\text{i}\text{v}}{+{{\upvartheta\:}}_{2}\text{C}\text{i}\text{r}\text{c}\text{u}\text{m}\text{s}\text{t}\text{a}\text{n}\text{c}\text{e}}_{\text{i}\text{w}}+{{\upepsilon\:}}_{4}$$
(4)

\(\:{\text{C}\text{o}\text{s}\text{t}}_{\text{i}}\) and \(\:{\text{R}\text{i}\text{s}\text{k}}_{\text{i}}\) denote transaction cost and defection risk, respectively. \(\:{{\upgamma\:}}_{0},\:{{\upgamma\:}}_{1},\:{{\upgamma\:}}_{2},\:{{\upvartheta\:}}_{0},\:{{\upvartheta\:}}_{1},\:{{\upvartheta\:}}_{2}\) are the estimated coefficients, and \(\:{{\upepsilon\:}}_{3},\:{{\upepsilon\:}}_{4}\) are the random error.

Variables

Table 1 shows a descriptive statistics analysis of the variables under investigation, including the effectiveness of conservation tillage technology adoption, collaboration capacity, shared vision, joint execution and regulation, as well as environmental characteristics and mechanisms variables.

The effectiveness of conservation tillage technology adoption. Given that green agriculture aims to achieve a comprehensive integration of natural, ecological, and socioeconomic goals, we examine the effectiveness of conservation tillage technology adoption from three perspectives: ecological performance, economic performance, and social performance. Utilizing customer satisfaction theory, our paper evaluates the effectiveness of conservation tillage technology adoption based on different subjects’ perceived quality differences before and after technology adoption. By evaluating these three dimensions, we aim to provide a holistic understanding of the effectiveness of conservation tillage technology adoption in promoting green agriculture.

Collaboration capacity

Collaboration capacity primarily consists of constructive engagement, shared vision, and joint execution and regulation. Constructive engagement is measured in terms of financial contribution, labor contribution, meeting organization, and training participation. Shared vision is measured by voter turnout in village elections, responsiveness to the “one case, one debate” system, and commitment to public affairs. These indicators indicate the level of participation and collaboration in achieving a shared vision for the adoption of conservation tillage technology. The observed variables of the shared vision are all ordered variables from 1 to 5. Joint execution and regulation encompass the implementation of management measures and regulations within the GAR collaboration. This is evaluated based on supervision measures, punishment measures, personnel allocation and regulation implementation. All observed variables are binary variables of 0–1, with 0 being no and 1 being yes (the calculation process of the above variables is shown in Appendix B).

Environmental characteristics

Considering the characteristics of the research subject and the constraints imposed by the availability of data, our study primarily examines soil erosion degree, land scale, and village location.

Soil erosion degree

The study area is located in the arid and semi-arid regions of Northwest China. Due to the lack of appropriate technological measures and institutional arrangements, the degree of soil erosion can partly measure the impact of natural disasters and climatic conditions on agricultural production. Consequently, the degree of soil erosion is quantitatively determined by the percentage of cultivated land area affected by water and soil loss.

Land scale

We focus on technology adoption within the agricultural production system, where land serves as a fundamental resource. Therefore, the land scale is measured as the logarithm of the village’ s cultivated area (mu). This metric provides a standardized measure of land size, facilitating comparisons across different villages.

Village location

Village location has significant implications for realizing the market value of its resources. The closer a village is to a major administrative center and markets, the more favorable it is for achieving the market value of its resources47. Therefore, our study measures village location based on the distance of the village from the township government.

Table 1 Variable definition and descriptive statistics.
Full size table

Results

Path analysis

Pearson moment product correlations between variables were below.80 (see Appendix Table A) and VIF values were < 3, so we considered the degree of multicollinearity acceptable. Since we divide the effectiveness of technology adoption into three dimensions: economic performance, ecological performance, and social performance, we constructed three SEM models to evaluate the impact of environmental characteristics and collaboration capacity on these three dimensions. A total of 173 questionnaires were used to test the fit of the model. The results show that χ2/d.f.=1.978 (less than 3), RMSEA = 0.075 (less than 0.08), GFI = 0.984 (greater than 0.9), and the other fitting indexes (CFI, IFI, etc.) are all greater than the standard value of 0.9, so the construct validity of the three models is good.

First, we analyze the impact path of environmental characteristics on collaboration capacity. Both the degree of soil erosion and land scale had a significant positive impact on the shared vision and constructive engagement, but the impact of land scale on the joint execution and regulation was significantly negative. The village location only had a significant negative impact on the shared vision. The effects of the other paths were not significant. This is about the same as we expected. The larger the land scale, the higher the cost of supervision and management by all parties, thus reducing the joint execution and regulation dimension of collaborative capacity. The more remote the village location, the less easy it is to communicate with the outside world, and the more difficult it is for all parties to effectively form a coherent goal.

Next, we analyze economic performance, ecological performance, and social performance separately. For economic performance, path analysis explains 30% of the variation in economic performance (as shown in Fig. 3). The degree of soil erosion negatively affects the economic performance of technology adoption. Land scale significantly positively affects economic performance of technology adoption. The villages location did not show significant differences. This suggests that high erosion rates may lead to a decrease in land productivity and increase the difficulty of technology application. Large-scale land provides better experimental conditions and economic returns for the implementation of conservation tillage technology. In addition, shared vision, constructive engagement and joint execution and regulation all have significant positive effects on economic performance. This shows that the enhancement of collaboration capacity can significantly increase the economic benefits of all parties.

As far as ecological performance is concerned, path analysis explains 20% of the variation in economic performance (as shown in Fig. 4). Land scale has a significant positive impact on ecological performance, while village location has a significant negative impact on ecological performance. This is in line with our expectations. Villages in remote areas may lack sufficient technical support and training, making it difficult for farmers to master new technologies, which in turn affects the improvement of ecological performance. Similarly, shared vision, constructive engagement and joint execution and regulation all have significant positive effects on ecological performance. This means that when parties share common goals and actively participate in the regulation of technology adoption, they are more likely to take concerted action to improve the environment.

For social performance, path analysis explains 79% of the variation in dependent economic performance (as shown in Fig. 5). Soil erosion, land scale and village location have no significant effect on social performance. This is because the promotion of social performance depends more on the comprehensive influence of political and economic factors. In addition, only the shared vision dimension of collaboration capacity has a positive impact on the improvement of social performance, while constructive engagement and joint execution and regulation are not significant. This may be because groups that share common goals are more likely to establish trust mechanisms that enhance closeness in their relationships.

Fig. 3
figure 3

Path of structural equation model.

Full size image
Fig. 4
figure 4

Path of structural equation model 2.

Full size image
Fig. 5
figure 5

Path of structural equation model 2.

Full size image

Heterogeneity analysis

We further discuss the heterogeneous effects of village size on the collaboration capacity to promote conservation tillage technology adoption. Village size is a crucial factor affecting GAR collaboration. With the rapid development of urban-rural integration and the in-depth promotion of the rural revitalization strategy, measures such as village mergers and residential community placements have significantly altered village sizes. These changes impact the institutional arrangements for GAR collaboration, which in turn affects the adoption of conservation tillage technology.

On one hand, GAR collaboration requires a certain group size to achieve “cluster effects.” Pamela (1988) pointed out that the influence of group size on the supply of collective goods depends primarily on the production function of those goods. Collective goods can achieve the same exclusionary effects as private goods by preventing non-payers from consuming public goods through market segmentation and price setting. When the attributes of these goods are closer to public goods, an increase in group size favors their supply.

On the other hand, larger group sizes lead to higher management costs for planning, organizing, directing, communicating, and controlling, which increases the likelihood of “free-riding” behavior. Conversely, smaller group collaboration incurs lower organizational and collaboration costs, resulting in stronger cohesion, shared moral values, and effective mechanisms for penalizing non-compliant rural households.

Since a unified standard for dividing village types by population size has not yet been formed, drawing on the provisions of Beijing’s current local standard “Village Planning Standards”, villages with a permanent population of less than 200 people are defined as small, villages with a population of 200–600 people are defined as medium-sized, and villages with a population of 600–1000 people are defined as large, villages with more than 1001 people are defined as super-large villages. In view of the fact that only 2.3% of the villages in the research sample we used had a population larger than 600 people, villages with a population of more than 200 people were defined as medium-large.

In view of the fact that the distribution and variance of the samples in the two groups were not consistent after grouping, Fisher’ s Permutation test based on Bootstrap was used and the sampling number was set to 1000 times to test the coefficient difference between the groups, and the results are shown in Table 2. According to the estimation coefficients of the core explanatory variables and the empirical P value of Fisher’ s Permutation test, constructive engagement has a stronger effect on conservation tillage technology adoption in small villages than in medium-large villages, and it is mainly reflected in ecological performance. This may be because, small villages tend to have a tighter community structure, and the positive effects of ecological protection measures can be directly observed. Shared vision also had a more significant impact on the environmental performance of conservation tillage technology adoption in small villages. It may be a shared vision is easier to form and spread in small villages. For small villages, joint execution and regulation had a stronger effect on the effectiveness conservation tillage technology adoption, and this difference was further verified by the test of coefficient difference between groups.

Table 2 Heterogeneity analysis results.
Full size table

Robustness test

Replacement of explanatory variables

To ensure the robustness of our findings, we replaced the measures of the effectiveness of conservation tillage technology adoption. We employed factor analysis to combine the three dependent variables—economic performance, social performance, and ecological performance—into a single variable representing the overall effectiveness of conservation tillage technology adoption. As shown in Table 3, the results indicate that constructive engagement, shared vision, and joint execution and regulation all pass the significance test, demonstrating a significant positive effect on the overall effectiveness of conservation tillage technology adoption.

To mitigate the impact of extreme values and ensure data stability, we applied a 1% and 5% winsorization treatment to the sample data48. The results, presented in Table 3, reveal that after winsorization, constructive engagement, shared vision, and joint execution and regulation continue to have significantly positive impacts on the effectiveness of conservation tillage technology adoption. Additionally, land scale is found to significantly influence adoption effectiveness. These results further validate the reliability of our conclusions.

Table 3 Replacement of explanatory variables results.
Full size table

Replacement of statistical method

We reperformed the regression using the ordered logistic model and the OLS model, and the regression results are shown in Table 4. The results show that collaboration capacity has a positive impact on the economic, ecological and social performance of the effectiveness of conservation tillage technology adoption. In terms of economic performance, constructive engagement, shared vision and joint execution and regulation all significantly improved income levels. In terms of ecological performance, shared vision and constructive engagement significantly improved ecological performance, but joint execution and regulation had no significant impact. In terms of social performance, the shared vision significantly promoted the improvement of relationships, while the effects of constructive engagement and joint execution and regulation were not significant. Environmental characteristics have a direct impact on the adoption effectiveness, such as the degree of soil erosion has a negative impact on economic performance, land scale has a positive impact on economic and ecological performance, and village location has a negative impact on ecological performance. Environmental characteristics also indirectly affect the adoption effectiveness through collaboration capacity, such as the degree of soil erosion and land scale promote constructive engagement, but the expansion of land scale may reduce constructive engagement.

Table 4 Replacement of statistical method results.
Full size table

Mechanism of GAR collaboration on the effectiveness of conservation tillage technology adoption

Our paper builds on the studies conducted by Feiock49 and Wu et al.28 to examine how GAR collaboration can optimize factor allocation and improve technological efficiency by reducing transaction costs and mitigating defection risks, ultimately enhancing the effectiveness of conservation tillage technology adoption. Due to the “limited rationality” of participants, disputes over benefit distribution and opportunistic behavior increase the risk of collaboration failure when faced with information asymmetry and complex tasks. Conversely, effective collaboration reduces transaction costs, ensuring that the net benefits outweigh management costs, thereby maximizing overall benefits.

Mechanism of transaction costs

GAR collaboration improves the effectiveness of technology adoption by reducing transaction costs. The difficulty of GAR collaboration is related to the regional scope of governance, the complexity of the problems faced, and the number of participating entities. The larger the scope of governance and the more complex the problems, the higher the transaction costs of negotiation, supervision, and implementation. Williamson pointed out that transaction costs are equivalent to increasing the “friction” in economic activities, and institutional arrangements play a crucial role in reducing these costs50. Furthermore, coordinated actions among different parties can improve mutual understanding and reduce transaction costs, thereby enhancing the efficiency of solving complex problems51.

Transaction costs can be decomposed into the costs of solving information problems and addressing incentive problems. On the one hand, GAR collaboration reduces information asymmetry and enhances cross-administrative governance capacity through information sharing, thus improving the effectiveness of conservation tillage technology adoption28. On the other hand, rural households play a pivotal role in conservation tillage technology adoption. By implementing policy incentives that ensure equitable and premium prices, along with establishing institutional safeguards to mitigate adverse selection, GAR collaboration strengthens rural household participation and enhances their satisfaction, thereby improving the effectiveness of conservation tillage technology adoption.

Mechanism of defection risks

Defection risk refers to the problems that emerge when the decision of one participant in an agreement results in worse conditions for the other participants due to conflicting interests49. GAR collaboration improves the effectiveness of technology adoption by reducing defection risks. Multiple collaboration and communication among participants not only enhance their motivation but also facilitate the decentralization and non-hierarchy of rights. This promotes information flow and sharing within the network, while reducing risk factors such as conflict and contract implementation52.

As pointed out by Putnam, the trust formed during the collaboration process facilitates the accumulation of social capital through collaboration and reciprocity among all parties, thus addressing various challenges faced during collaboration53. Therefore, when facing conflicts caused by poor collaboration, betrayal of collaboration, and unfair distribution of benefits, participants can establish embedded defection mechanisms and binding contractual mechanisms through institutional frameworks and the promotion of GAR collaboration54. This approach reduces the probability of moral hazard and adverse selection while altering the costs and benefits of opportunistic behavior, thereby lowering defection risks.

Mechanism analysis

To validate the proposed mechanism, we used the entropy method to downscale constructive engagement, shared vision, and joint execution and regulation into collaboration capacity, which was then included as an independent variable in the regression model. The dependent variable adopts the violation frequency to reflect the defection risks encountered during conservation tillage technology adoption. This was done to explore whether collaboration capacity could reduce defection risks and consequently improve the effectiveness of technology adoption. Table 5 shows the mechanism analysis results. The estimation results in column (1) show that collaboration capacity significantly reduces violation frequency.

In the subsequent analysis, cross-village governance capacity for soil and water conservation and farmers’ satisfaction are used as dependent variables to reflect transaction costs. On the one hand, the dependent variable uses the cross-village governance capacity for soil and water conservation to reflect cross-regional governance capacity, aiming to assess whether collaboration capacity could reduce transaction costs by improving cross-village governance of soil and water conservation, thereby improving the effectiveness of conservation tillage technology adoption. The estimation results in column (2) show that collaboration capacity significantly improves cross-village governance for soil and water conservation.

On the other hand, the dependent variable of farmers’ satisfaction is used as an indicator to demonstrate how incentive mechanisms implemented through GAR collaboration reduce transaction costs. The estimation results in column (3) show that collaboration capacity significantly improves rural household satisfaction.

Table 5 Mechanism analysis results.
Full size table

Discussion

The data structure of this study is pivotal in understanding the relationships and impacts of various factors on the adoption of conservation tillage technology. Data for this study were collected through structured field surveys across three provinces (Shaanxi, Gansu, and Ningxia), regions selected based on their active participation in conservation tillage practices. We employed a multi-stage stratified random sampling method to ensure the representativeness of the sample. A total of 173 valid samples were collected, consisting of village-level data on technology adoption, environmental characteristics, and collaboration among governments, agribusinesses, and rural households (GAR). Data collection was conducted in partnership with local agricultural authorities and cooperatives, and all data were cross-checked with official statistics where available. However, the reliance on self-reported survey data introduces potential bias, such as recall bias, and the cross-sectional nature of the data limits the ability to capture long-term trends in technology adoption.

While the data collected offer valuable insights into green technology adoption, several limitations must be acknowledged. First, the reliance on self-reported data introduces the potential for bias, including recall bias and over-reporting of benefits. Second, the cross-sectional nature of the data limits our ability to observe changes over time. As this study captures a single point in time, it cannot account for the dynamic processes involved in long-term technology adoption. Third, although the data were cross-checked with official records, regional differences in data availability may affect the comparability of results across different study sites.

Conclusions and policy implications

Conclusions

This article investigates conservation tillage technology adoption in the Ningxia, Shaanxi, and Gansu provinces, using conservation tillage technology as a case study. By constructing a theoretical framework of environmental characteristics, collaboration capacity, and adoption effectiveness, this study explores the pathways, heterogeneity effects, and mechanisms through which GAR collaboration enhances the effectiveness of conservation tillage technology adoption. The main conclusions are as follows:

Firstly, the study highlights the importance of collaboration capacity in enhancing the effectiveness of conservation tillage technology adoption. Constructive engagement, shared vision, and joint execution and regulation each contribute significantly to the economic, ecological, and social performance of conservation tillage technology adoption. Secondly, environmental characteristics not only have a direct impact but also affect the effectiveness of conservation tillage technology adoption indirectly through collaboration capacity. Thirdly, the mechanism through which GAR collaboration enhances technology adoption is primarily through reducing defection risks and transaction costs.

Policy implications

Enhancing GAR collaboration is pivotal for fostering transformation and innovation in green agriculture. Shared vision is pivotal in influencing the adoption of conservation tillage technology. Thus, constructing communication and learning platforms, nurturing trust, and cultivating multi-party learning, communication, and collaborative management networks are imperative. Developing an incentive-compatible mechanism for conservation tillage technology adoption and establishing pathways for GAR collaboration, such as digitization, green branding, diversification, and contextualizing benefits, will bolster resource sharing and value co-creation.

Moreover, the complexity of environmental characteristics significantly impacts collaboration capacity and the effectiveness of conservation tillage technology adoption. A broader environmental context is crucial for devising sustainable development strategies. Boosting investments in modern sciences and technologies, like remote sensing and big data, enhances comprehension of environmental dynamics. Additionally, assessing the value of eco-agricultural products and establishing frameworks for economic, social, and ecological impacts offers institutional and technological backing for GAR-driven green agri-development.

Finally, improving the effectiveness of conservation tillage technology adoption through GAR collaboration entails reducing costs and risks through mechanism designs. Robust structural, organizational, procedural mechanisms, and efficient management are key. Concerted efforts across technological supply, production services, information sharing, regulatory coordination, and execution supervision are needed to maximize the contributions of all parties. A vertically integrated model addresses information asymmetry, while risk-sharing and benefit-sharing alleviate external constraints, fostering green agriculture development.

While our study offers important contributions, there are several limitations that should be acknowledged. First, the analysis relies on cross-sectional data, which limits our ability to capture the dynamic and evolving nature of collaboration and technology adoption over time. Second, the geographical scope of this study is limited to specific regions in China, which may restrict the generalizability of the findings. Finally, potential measurement errors, particularly in self-reported data on collaboration engagement and technology adoption, may affect the accuracy of the results. Future research could explore several key areas. First, longitudinal studies should be conducted to assess the long-term impacts of collaboration on technology adoption. Second, further studies could expand the study to other regions and countries with diverse agricultural practices and collaboration structures could provide more comprehensive insights. Finally, examining how digital tools and emerging technologies can facilitate collaboration among stakeholders may offer valuable insights into accelerating the adoption of green technologies.