Afforestation Through Sand Control: Farmer Participation Under China’s New Round of Grain-for-Green Compensation Policy

Abstract

Within the context of global desertification trends in arid regions, advancing afforestation and sand stabilization efforts are not only vital for human survival but are also key considerations in addressing the challenges of climate change and achieving sustainable development. This study, set against the backdrop of China’s new round of Grain-for-Green compensation policies implemented in 2014, investigates farmers’ behavior in planting economically valuable forests and grasslands driven by compensation incentives. Grounded in the principles of behavioral economics and assuming farmers as rational “economic agents”, this study focuses on farmers residing on the northern and southern slopes of the Tianshan Mountains in Xinjiang. Employing the fuzzy-set qualitative comparative analysis (fsQCA) approach, it examines the intricate causal mechanisms that shape farmers’ involvement or lack thereof in economic forest and grassland activities. These mechanisms are analyzed through the lenses of resource endowment, psychological perception, and external environmental factors. The results indicate that perceived benefits and neighbor imitation serve as essential conditions for non-participation in planting economic forests and grasslands. Three configurational pathways account for participation: farmers motivated by perceived benefits, those guided by the combined influence of “psychological perception and external environment”, and individuals driven by ecological aspirations alongside neighbor imitation. Additionally, four configurational pathways explain non-participation, with two types of farmers identified: those facing a dual deficiency of psychological perception and external environment, and non-high income traditional farmers dependent on agricultural irrigation water.

1. Introduction

Achieving sustainable development has been an important and complex issue over the past decades. Although there is no clear theoretical framework for sustainable development at present [1], people have realized that sustainable development cannot be separated from the inter-coordination among the environment, society, and economy. Especially the sustainability of ecology, which is the foundation of overall development capacity [2]. The deterioration of the ecological environment seriously threatens human survival and limits human development, especially in the context of the global trend of desertification in drylands, where the impact of human activities to land degradation exceeds 70 percent, which affects the livelihoods of more than one billion people [3]. The socioeconomic development of these places is severely constrained by ecological conditions; local farmers, whose livelihoods depend on agricultural activities, are made increasingly vulnerable by desertification. Combating desertification and promoting sustainable development in these areas is therefore crucial.

In order to combat desertification and realize sustainable development at the same time, many countries have introduced some ecological conservation policies, such as Israel’s desert agriculture, and have also made considerable achievements through the planting of drought-tolerant trees for windbreaks and sand fixation, and to improve the climatic environment around farmland, some countries have combined this with the planting of salt-tolerant plants and bioremediation technology to manage saline soil, to a certain extent, curbing land degradation and desertification expansion, and part of the region’s ecological environment has been significantly improved [4,5]. Ecological conditions have been significantly improved, laying the foundation for sustainable agricultural development. The Central Asian countries geographically adjacent to Xinjiang have also adopted analogous ecological restoration policies, which have improved the ecological environment and promoted the sustainable development of local agriculture through a variety of measures, such as planting trees, adjusting the agricultural structure, and promoting sustainable agricultural practices [6,7].

Analogously, since the 1950s, China has been implementing agroforestry systems centered on shelterbelts. To address regional wind–sand hazards, the Chinese government established 4000 km of shelterbelts in areas with serious wind and sand, which effectively safeguarded the agricultural production at that time; in the following decades, the protective forest system has gradually transformed into an integrated agroforestry system, and has been gradually expanded from the wind-erosion-prone areas in the north to northern and southern regions of China to create a large number of farmland forest networks. These measures have played a very important role in improving the ecological environment. However, for a long time, forestry construction has remained a public welfare sector with limited economic benefits. For instance, during the initial phase of the Grain-for-Green Program (GGP), large-scale afforestation initiatives were primarily driven by government subsidies and planning, with farmers playing the role of non-active stakeholders. While this approach advanced forestry development through ecological compensation mechanisms, it also required significant fiscal expenditure from the government, inevitably exacerbating the government’s fiscal strain [8]. This sustained financial “blood transfusion” has been a significant challenge for the government, highlighting the need to achieve self-sufficiency (“blood generation”) through the GGP, which could not only reduce government burdens but also generate economic benefits, contributing more effectively to the achievement of sustainable development at the local level [9].

Thus, during the second round of the GGP, launched in 2014, the government supplemented subsidies by increasing farmers’ income through diversified practices, such as cultivating economic forest fruits, engaging in understory livestock farming, and growing forage crops [10,11]. In these ways, farmers can sell not only fruits and other related agricultural products but also forage crops and Chinese herbs grown in the forest alongside the eggs laid by birds raised in the forest, such as free-range poultry eggs and duck eggs, which are also popular in the market. This approach ensured that even after subsidies ended, farmers would not resume grain monoculture or abandon farmland.

However, another special situation exists in the ecologically fragile arid region of Xinjiang, where our research is concerned, where farmers are caught in a sustainable development dilemma due to ecological and livelihood constraints (e.g., water scarcity and limited job opportunities). For example, in order to alleviate the pressure of water scarcity, the Xinjiang government has vigorously implemented a “land reduction and water-saving” policy, allocating substantial resources for the development of water-efficient irrigation infrastructure. While these facilities have lowered irrigation costs, they have also reduced farmers’ willingness to maintain agricultural activities. Many farmers continue to grow grains, cotton, and vegetables, which increases agricultural water consumption and exacerbates the pressure on already scarce water resources [12,13,14]. Additionally, although the government has invested significantly in public water-saving infrastructure, farmers still face installation costs for using these facilities. For farmers who have already abandoned their land, resuming cultivation would incur prohibitively high costs. The abandonment of arable land thus increases land degradation and accelerates desertification processes [15].

So, if the GGP fails to gain widespread participation from farmers in ecologically fragile arid regions and water-saving infrastructure continues to expand without addressing these challenges, the surging agricultural water demand, coupled with farmland abandonment, will exacerbate land degradation and desertification. Over time, this situation could create a vicious feedback loop of ecological degradation and poverty for farmers in these regions [16,17]. Moreover, to minimize losses, farmers who have already participated in the GGP might choose to revert their land to previous uses or abandon it after the subsidies end, jeopardizing earlier achievements in afforestation and increasing the challenges in sustaining environmental management efforts.

To prevent farmers in ecologically fragile arid regions from being trapped in a vicious feedback loop of ecological degradation and poverty, it is crucial to fully leverage the dual ecological and economic functions of the Grain-for-Green Program (GGP). The success of the GGP in delivering both ecological and economic benefits ultimately depends on the sustained engagement of farmers. Thus, encouraging farmers to actively and consistently engage in the cultivation of economically valuable forests and grassland remains a pivotal yet unresolved challenge.

This study offers notable marginal contributions by combining a complex systems perspective with behavioral economics theories to examine farmers’ participation in afforestation efforts. The theoretical framework, centered on “resource endowment, psychological perception, and external environment”, provides a comprehensive lens to examine the interplay of multiple factors influencing farmers’ decision-making. To elucidate the interdependencies among these tripartite factors, we first performed a configurational analysis of “participating farmers”, which was then followed by a complementary analysis of “non-participating farmers”. We found that different configurational pathways were associated with farmers’ participation and non-participation. Furthermore, psychological perception factors, resource endowment, and external environmental factors interact in diverse combinations, leading to two divergent outcomes, as elaborated in the subsequent analysis.

Compared to existing studies that often focus on single factors (e.g., [18,19,20]), this research examines the causal complexity of interactions between resource endowment, psychological perception, and external environmental factors. For instance, resource endowment elements, such as household income, land concentration, and water scarcity, significantly impact decision-making, as supported by previous findings on farmers’ economic constraints [21]. The inclusion of psychological perception factors—such as ecological aspirations and perceived benefits—addresses the often-overlooked subjective dimensions, aligning with Duan et al. (2021) [22] and Thiam et al. (2024) [23] who emphasize the importance of intrinsic motivation in driving farmers’ actions.

This study also sheds light on the critical role of external environmental factors, such as neighbor imitation and socialized services, in motivating participation among farmers with low intrinsic motivation. These findings align with research by Arhin et al. (2024) [24] and Zhang et al. (2021) [25], which highlights the demonstration effects of community-led initiatives and the value of technical and market support in overcoming barriers to afforestation. By uncovering the interactions between these diverse factors, this study not only enhances the theoretical understanding of farmer behavior but also highlights practical pathways for policy intervention.

From a methodological standpoint, this study highlights the effectiveness of the fuzzy-set qualitative comparative analysis (fsQCA) approach in uncovering multiple configurational pathways that result in both participation and non-participation outcomes. Utilizing fsQCA, this study reveals the antecedent conditions and configurational pathways driving farmers’ participation in planting economically valuable forests and grasslands in the ecologically fragile and arid regions of Xinjiang. This innovative approach offers an alternative to conventional quantitative or qualitative analyses, overcoming limitations such as the inability to account for synergistic effects [26,27]. For instance, this study finds that participation often results from specific combinations of high intrinsic motivation, favorable resource endowment, and supportive external environments—insights that are crucial for tailoring policy measures to different farmer profiles.

Policy wise, this research enhances the optimization of the Grain-for-Green Program (GGP) by identifying practical strategies, such as improving water-saving infrastructure and promoting community-driven services. It underscores that targeted interventions addressing specific combinations of influencing factors can substantially boost participation rates, thereby advancing the dual objectives of ecological restoration and economic sustainability. Moreover, the findings hold broader relevance for afforestation policies in arid regions globally, contributing to international efforts to combat desertification and promote sustainable land management.

By synthesizing these theoretical, methodological, and practical contributions, this study lays a foundation for future research to delve deeper into the dynamic interplay of factors influencing afforestation participation. It addresses gaps in the existing literature while advancing global ecological restoration objectives. So, this study specifically selected farmers from the ecologically fragile and arid regions of Xinjiang as research cases. The realization of local sand control and afforestation depends on the broad participation of farmers. We carried out a survey of 487 farmers across seven counties on the northern and southern slopes of the Tianshan Mountains, focusing on their “participation in the planting of economically valuable forests and grassland”.

The remainder of this paper is organized as follows: Section 2 provides a study review; Section 3 introduces the policy and theoretical analysis; Section 4 introduces the fsQCA methodology; Section 5 discusses the results; Section 6 provides conclusions; Section 7 discusses the policy and theoretical implications; and the final section presents research limitations and future directions.

2. The Study Review

Afforestation is important for ecological sustainability. The existing research has demonstrated that afforestation plays a multi-faceted role in improving the ecological environment [28,29,30], and this point has been widely recognized by the academic community.

However, in the course of actual practice, afforestation activities face more complicated situations. On the one hand, when environmental governance policies are implemented, although ecological benefits are generated, they also pose challenges to farmers’ livelihoods. Meanwhile, some areas have paid attention to the livelihood needs of farmers, increased farmers’ economic income, and improved farmers’ livelihoods, but inappropriate policies have put a relatively large pressure on the local ecology [31], exceeding the carrying capacity of the regional ecology.

This situation is even more acute in the case of afforestation in ecologically fragile arid areas, where the ecological environment is threatened by both desertification and poverty.

Therefore, finding a balance between ecological benefits and economic benefits has become the key to promoting afforestation policies in ecologically fragile and arid areas. None of this would be possible without the active involvement of farmers. Encouraging their enthusiastic engagement is also crucial for ensuring the success of policies and advancing sustainable development [8,9]. This perspective has been acknowledged by scholars as well.

The research indicates that multiple determinants shape farmers’ involvement in forestation initiatives. Notably, economic motivation plays a pivotal role in determining whether farmers adopt afforestation practices [27]. Furthermore, when financial incentives from the government [32,33] or market profit potential [21,34] drive farmers’ afforestation efforts, they are inclined to participate actively, having considered the trade-offs. Consequently, sufficient economic incentives are crucial for sustaining farmers’ commitment to afforestation initiatives that align with governmental policies [35]. In contrast, a deficiency in appropriate financial support, rewards, or market profitability may result in farmers’ willingness to participate, yet their actual afforestation actions and the extent of their implementation are likely to be constrained [36].

However, is it realistic to expect that farmers will participate in afforestation efforts solely based on government financial incentives or potential market profits? This expectation may not always be justified. In their review of the UK’s land use policies over the past century, Westaway et al. (2023) [37] highlighted that relying solely on grants may not sufficiently motivate farmers to engage in tree planting. The effectiveness of financial incentives in promoting afforestation depends on the pre-existing interests and values of farmers or landowners [26,38,39].

Put another way, the recognition of afforestation by farmers is a crucial factor motivating their participation. Additionally, Hopkins et al. (2017) [40] showed that farmers with higher educational attainment and those engaged in environmental conservation are more likely to increase forest planting. A positive attitude and favorable perceptions of afforestation policies significantly influence farmers’ decisions to engage in afforestation [22,41,42]. So, farmers who receive technical training and support are more likely to view afforestation policies positively [43].

Additionally, the availability of land ultimately dictates the level of farmers’ involvement in afforestation [35,36], including factors such as land fragmentation, the scale of land transfer, and the duration of land leases [19].

In addition to land availability, various studies have identified additional factors that affect farmers’ participation in afforestation. These include gender, household size [44], household income [18,45], collective action among farmers [46], mutual cooperation among farmers [20], and the influence of communities and neighbors [25,47]. Also, including publicity and technical training, cultural attitudes [27] drive farmers’ participation in afforestation.

In summary, the current study includes economic factors that directly influence farmers’ participation in afforestation, which they are attracted to it either through short-term benefits or long-term gains; in addition, farmers’ land resource endowment and their socio-institutional conditions (e.g., community support) play the role of indirect influences, which determine the incentive’s feasibility and acceptance; and critically, deeper psychological influences, such as intrinsic values and ecological perceptions, etc., affect farmers’ perceptions of agroforestry practices and resource allocation in the long term. However, the existing research largely concentrates on the impact of individual factors, yet the influence of any single factor is often modulated by its interaction with other elements. This approach often overlooks the complex interplay between various factors and fails to capture commonalities across different cases. Moreover, the subjective factors influencing farmers’ participation are frequently underexplored.

To address this gap, this paper adopts a complex systems perspective and a configurational approach to explore how these influencing factors interact and combine to drive farmers’ participation in afforestation. Employing a configurational approach based on fuzzy-set theory (fsQCA), we aim to offer a comprehensive, multi-factorial explanation of farmers’ afforestation practices. This approach helps uncover the causal complexity of multiple factors influencing farmers’ participation behavior. Our findings not only help optimize and expand the outcomes of current Grain-for-Green policies but also offer valuable insights for similar afforestation initiatives in other arid regions worldwide. This, in turn, supports the long-term goal of sand control and afforestation in these ecologically vulnerable areas.

3. Policy and Theoretical Analysis

Viewed through a complex lens, farmers’ participation in the planting of economically valuable forests and grassland stems from the comprehensive dynamic interplay of multi-level determinants, characterized by intricate interrelationships among these influencing elements. Often, multiple factors form configurations that jointly affect farmers’ behavior. By integrating relevant theories from behavioral economics, planned behavior, and social exchange, this study proposes that farmers’ participation in forest and grassland management is the result of a multidimensional interaction between “resource endowment, psychological perception, and external influence”.

Specifically, farmers initially make rational decisions based on their available resources. Their attitudes toward relevant policies and ecological conservation act as mediating factors influencing their participation in planting. External factors, such as the demonstration effect of neighboring farmers and technical guidance on forest and grassland management (including planting techniques and sales), encourage farmers to break away from traditional production methods. These three dimensions—resource endowment, psychological perception, and external influence—work synergistically to drive farmers’ participation in the planting of economically valuable forests and grassland.

From a configurational perspective, identifying the pathways through which different types of farmers engage in cultivating economically valuable forests and grassland enables a needs-tailored analysis of heterogeneous agricultural communities. This, in turn, facilitates their participation and enhances the policy’s effectiveness.

3.1. Resource Endowment: Resource Endowment Serves as the Foundation for Farmers’ Decision-Making

Resource endowment is the sum of the various resources that a country, region, or individual possesses for economic development, at the core of which are the factors of production that can be utilized. Referring to relevant studies, we choose to use household total income, land concentration, water-saving irrigation coverage, and water scarcity level as the measurement indicators of farmers’ resource endowment [18,48].

As the north and south slopes of the Tianshan Mountains in this research are the oasis agricultural agglomeration areas in Xinjiang, these regions represent the highest population density and the greatest pressure on water use. The initial investments and costs associated with participating in the planting of economically valuable forests and grassland in this region—such as capital, farmland, water-saving infrastructure, and irrigation water resources—are key factors influencing farmers’ decisions. These resources represent farmers’ core assets and form the foundation for their initial cost–benefit analysis in agricultural decision-making. Under the guidance and support of the new round of the Grain-for-Green policy, and based on the “rational economic agent” assumption, farmers are expected to make rational assessments of their resource endowments and select the optimal choice that aligns with their goal of maximizing profits.

Firstly, the planting of economically valuable forests and grassland requires farmers to have a strong ability to bear risks. The higher the family’s total income, the stronger the farmers’ risk-bearing capacity. Furthermore, in water-scarce regions, participating in the cultivation of economic forests requires the installation of water-saving facilities, which also requires more funding.

Secondly, when farmland is excessively fragmented and lacks sufficient concentration, it impedes the large-scale cultivation of staple crops and raises production costs across activities such as plowing, planting, and harvesting. The unified installation of water-efficient irrigation systems would incur prohibitive costs. And the adoption of drip irrigation technology is primarily propelled by government-led infrastructure initiatives; land fragmentation often hinders farmers from fully capitalizing on these public investments.

Finally, in ecologically fragile arid regions, the cost of agricultural water usage plays a crucial role in influencing agricultural management. In areas with low water scarcity, the pressure to exploit groundwater is lower, and water prices are lower, making it more suitable for growing water-intensive economic crops. However, in regions with severe water scarcity, farmers must invest in water-saving irrigation systems to participate, which increases investment costs. Consequently, based on their current preferences, low-income farmers may be reluctant to change their operational strategies.

3.2. Psychological Perception: Farmers’ Psychological Foundation as an Intermediary

Psychological perception is the psychological assessment and judgment of policies or behaviors formed by farmers in the decision-making process, based on subjective cognition, emotions, and values, and is an intrinsic driver of farmers’ behavioral choices. Drawing on our previous research and related studies, we have chosen to use ecological aspiration and perceived benefits as indicators for measuring the farmers’ psychological perception [41,49].

From the perspective of rational smallholders, farmers tend to exhibit conservative, risk-averse characteristics, coupled with a natural attachment to their farmland, which may hinder them from overcoming their cautious attitudes. However, farmers are not entirely rational actors; perception forms the foundation of their decision-making. Their behavioral decisions are deeply rooted in psychological perception, which acts as the foundational psychological mechanism influencing and shaping their choices.

The extent of farmers’ ecological aspirations and their perceived benefits from policies related to the planting of economically valuable forests and grassland directly shape their psychological perception. This perception, in turn, can influence or modify their initial rational assessments, which will further affect or change the results of farmers’ initial rational weighing.

Farmers’ willingness to engage in the latest phase of the GGP (Grain-for-Green Program) hinges on their subjective cost–benefit calculus, with anticipated returns emerging as the primary driver of enrollment decisions. From the perspective of farmers’ mental accounting, if they understand the policy’s content—such as the ability to choose between planting economic forests or ecological forests, intercropping, understory farming, or combining forestry and grassland systems—their expectations of economic returns improve. This allows them to recognize that economic benefits can still be attained even after the subsidy period concludes, thereby fostering long-term participation.

However, if farmers have insufficient perceived benefits—for example, they face multiple transitional challenges after the subsidy ends, including reduced land output, difficulties in agricultural and livestock operations, finding alternative income sources, and labor resource allocation—the difficulties encountered could intensify livelihood pressures, potentially diminishing their prospects for sustainable advantages.

Thus, the higher farmers’ perceived benefits from the compensation policy for the planting of economically valuable forests and grassland, the more positively they evaluate its implementation. This mitigates the uncertainty risk, making them more inclined to adopt forest and grassland management. Additionally, the higher farmers’ ecological aspiration levels, the stronger their motivation to protect the environment, which fosters a behavioral inclination toward forests and grassland management.

From a development economics perspective, when farmers possess high levels of ecological aspiration and perceived benefits, they are more likely to actively and voluntarily participate in forest and grassland development under the dual economic and ecological incentives of the new round of the Grain-for-Green Program.

3.3. External Environment: External Environment as an Inducement for Farmers to Participate in Economic Forest and Grassland Management

External environment refers to the guidance and influence of external factors, such as policy, technology, the market, and society, on the participation behavior of farmers’ in planting economic forests and grasses, the core of which is to reduce the risk of farmers’ decision-making and promote behavioral change through institutional support, resource supply, and the demonstration effect. Drawing on related research, we choose neighbor imitation and socialized services as measures of farmers’ external environment [50,51].

The planting of economically valuable forests and grassland in Xinjiang include both ecological timber forests and economic crops. Ecological forests primarily consist of native species, such as tamarisk and sacsaoul, while economic forests and grasslands include crops like fragrant pears, grapes, jujube, walnuts, and forage plants, such as alfalfa for livestock. These measures are of great significance for promoting local farmland protection, water conservancy construction, and the value of local ecosystem services. This study focuses on the Grain-for-Green model involving economically valuable forests and grasslands.

Nevertheless, in the process of policy promotion, farmers may adhere to traditional agricultural practices, lack professional knowledge and market information about the planting of economically valuable forests and grassland, and have concerns about the sales channels for such products. These factors, along with a risk-averse decision-making inertia, make farmers hesitant to alter their established production methods, thereby impeding the adoption of economically valuable forest and grassland planting.

Therefore, only robust external influencing conditions can effectively motivate farmers to overcome their dependence on traditional crop cultivation and redirect their production preferences toward the planting of economically valuable forests and grassland.

Firstly, professional technical guidance and training can help farmers to enhance planting efficiency and increasing survival rates. Additionally, through socialized services, farmers can access market opportunities for selling forest and grassland products. Secondly, early adopters of the planting of economically valuable forests and grassland serve as role models, with their practices and economic benefits influencing those who have not yet participated. This demonstration effect, combined with collective action among farmers, further encourages greater participation. In summary, the technical and market services provided by socialized support help alleviate farmers’ concerns, while the demonstration effect of neighbor imitation and collective action stimulate farmers’ enthusiasm for participation, ultimately attracting more farmers to engage in the planting of economically valuable forests and grassland.

In summary, this study integrates eight influencing factors across three dimensions—resource endowment, psychological perception, and external influence—into a unified analytical framework to examine farmers’ participation in the planting of economically valuable forests and grassland. It aims to reveal the underlying mechanisms driving farmers’ participation from multiple perspectives and explore how the multidimensional interaction of these factors jointly influences the pathways that lead to farmers’ engagement in planting activities (Figure 1).

4. fsQCA Methods

4.1. fsQCA Methods

The fuzzy-set qualitative comparative analysis (fsQCA) method has significant advantages over other methods, such as conventional regression models, in terms of dealing with complex causality, combining qualitative and quantitative analysis, its applicability to small and medium sample size studies, and data versatility.

For example, traditional logistic regression will attempt to identify the independent effects of each factor, but the problem is that this method ignores how these factors combine to work together. In contrast, fsQCA focuses on the combination of conditions rather than the net effect of a single variable. And logistic regression requires a large sample size to ensure statistical stability, otherwise it is susceptible to the influence of multicollinearity, leading to unreliable results. In contrast, fsQCA performs robustly in small- and medium-sized samples because it does not rely on statistical significance but reveals causality through the logic of conditional combinations. Therefore, our study adopts fsQCA as our research methodology, which, as an analytical method that combines qualitative and quantitative features, is more suitable for this study compared to both traditional quantitative research methods and qualitative analysis methods.

So, we employ the fsQCA methodology to examine the diverse and intricate pathways that either impede or facilitate farmers’ engagement in the planting of economically valuable forests and grassland. Such diverse and intricate pathways imply that different combinations of conditional variables, i.e., groupings, can produce the same result.

In fsQCA, necessity and adequacy analyses constitute essential analytical phases, and their concepts are central [47].

The necessary condition denotes an essential element or set of elements that must invariably exist to enable the occurrence of the targeted result. Without this condition, the outcome cannot be achieved, making it essential for the result to materialize. In other words, the non-existence of a necessary condition constitutes a “disabling” factor for the outcome. A sufficient condition, on the other hand, refers to a specific factor or configuration of factors whose presence guarantees the manifestation of the anticipated result. Stated differently, if the sufficient condition is met, the outcome will inevitably follow. In this case, the presence of a sufficient condition serves as an “enabling” factor for the outcome.

To summarize, this research employs the fsQCA 3.0 software for qualitative comparative analysis, seeking to investigate and resolve the following key inquiries:

• Are there necessary conditions that hinder or facilitate farmers’ participation in the planting of economically valuable forests and grassland?
• What configurational pathways hinder or facilitate farmers’ participation in the planting of economically valuable forests and grassland?

4.2. Variable Measurement and Calibration

4.2.1. Outcome Variable

Drawing upon the conceptual foundation previously established, the outcome variable of this investigation is therefore defined as “Participation” and “Non-participation”.

4.2.2. Condition Variables

(1)

Resource Endowment Level: Farmers’ capital endowment is represented by total household income. Land endowment is reflected in three factors: land concentration, water-saving irrigation coverage, and water scarcity level. The specific measurement methods are shown in

Table 1. Variables and their measurements.

Variable Type	Variable	Sub-Variable	Measurement Criteria
Condition Variable	Resource Endowment Level	Household Total Income	Total household income in 2020 (includes income from crop farming, livestock, off-farm work, property income, etc.) (in CNY 10,000 logged)
		Land Concentration	Total farmland area/number of land plots (unit: mu/plot)
		Water-Saving Irrigation Coverage	Water-saving irrigated area/total household farmland area
		Water Scarcity Level	Water scarcity severity is quantified using a 7-grade measurement system, with higher scores reflecting more pronounced scarcity conditions.
	Psychological Perception Level	Ecological Aspiration	Satisfaction with local climate and environment, investment in Grain-for-Green, surrounding farmers’ participation, and current ecological protection measures
		Perceived Benefits	Does participation in the planting of economically valuable forests and grassland increase total household income? 1 = Yes; 0 = No
	External Environment Level	Neighbor Imitation	Do surrounding farmers invest significant land and effort in the planting of economically valuable forests and grassland? 1 = Yes; 0 = No
		Socialized Services	Are technical support services provided by the village for the planting of economically valuable forests and grassland? (includes planting techniques and sales) 1 = Yes; 0 = No
Outcome Variable	Participation Behavior	Participation	Has the farmer participated in the planting of economically valuable forests and grassland? 1 = Yes; 0 = No

.

(2)

Psychological Perception Level: Farmers’ psychological perception is represented by their ecological aspiration and perceived benefits. Ecological aspiration is calculated by adding up the scores. Perceived benefits assess the impact of economic forests and grassland on household benefits from 2015 to 2020, encompassing both ecological- and income-related effects. Adverse effects or the absence of influence are assigned a value of 0, while positive impact is coded as 1.

(3)

External Environment Level: this study considers the availability of socialized services and the degree of neighbor imitation to measure the impact of the external environment.

4.2.3. Variable Calibration

To guarantee calibration precision, appropriate calibration techniques are chosen for distinct variables, taking into account the unique attributes of conditional variable datasets and the patterns observed in descriptive statistical analysis, in alignment with the calibration framework proposed by Basurto and Speer (2012) [52]. According to Ragin’s (2008) [53] research, we set three anchors for continuous variables and discrete variables, including the following: full membership, crossover point, and full non-membership. These three anchors assist in understanding the extent to which variables within a configuration path belong to the cases corresponding to that path. For instance, when variable X is entirely non-member in a specific configuration path, all instances of variable X in the cases associated with this configuration path are also completely non-member. Here, we adopt different standards for data calibration, which are tailored to the specific statistical features of the dataset to ensure accuracy and reliability [47]. The calibration method for the different variables is as follows: We determine the crossover points, complete membership, and complete non-membership for “total family income” through mean and standard deviation as 1.984706, 3.46487, and 0.504542, respectively. Since the data of “irrigation coverage rate” are in the form of percentage, with 387 farmers having an irrigation coverage rate of 100% and 70 farmers having an irrigation coverage rate of 0%, we set 0.5 (50%) as the crossover point, 1 (100%) as complete membership, and 0 (0%) as complete non-membership, which is consistent with the practical context observed in the field. Due to the excessive dispersion of the data of “land concentration”, if the mean and standard deviation are used to determine the degree of complete membership, it will result in an excessively high degree of complete membership, thereby excluding most cases from complete membership; so compared to the mean, the method of using quantiles is more secure [54], so we use 75%, 50%, and 25% as complete membership, crossover point, and complete non-membership, respectively. Similarly, since “ecological desire” and “water shortage degree” are discrete variables, we choose 3, 2, and 1 as the thresholds of complete membership, crossover point, and complete non-membership for “ecological desire” based on the frequency distribution of the variable, and choose 6, 4, and 2 as the thresholds of complete membership, crossover point, and complete non-membership for “water shortage degree” (

Table 2. Variable calibration anchor points and descriptive statistics.

Variables	Descriptive Statistics				Calibration
	Mean	S.D.	Maxi	Mini	Full Membership	Crossover Point	Full Non-Membership
Household Total Income (logged)	1.985	1.485	5.736	−6.908	3.465	1.985	0.505
Land Concentration	43.2152	84.535	760	0.8	40	15	6.775
Water-Saving Irrigation Coverage	0.791	0.389	1	0	1	0.5	0
Water Scarcity Level	3.834	1.975	7	1	6	4	2
Ecological Aspiration	1.5895	1.06	3	0	3	2	1
Perceived Benefits	0.17	0.372	1	0	1	-	0
Neighbor Imitation	0.20	0.401	1	0	1	-	0
Socialized Services	0.30	0.459	1	0	1	-	0
Participation	0.339	0.474	1	0	1	-	0

).

5. Results and Analysis

5.1. Descriptive Statistical Analysis

Firstly, prior to conducting the configurational analysis of farmer cases, descriptive analyses were carried out for the full sample, as well as for participants and non-participants separately, followed by a t-test to evaluate the differences between the two groups. The statistical characteristics of the primary variables are summarized in

Table 3. Descriptive statistics of variables and t-test.

Variable Names		Mean (Standard Deviation)			Mean Difference (2) − (3) = (4)
		Full Sample (1)	Participants (2)	Non-Participants (3)
Resource Endowment Level	Household Total Income	18.065 (30.98)	42.509 (47.241)	7.803 (8.577)	34.707 *** (38.664)
	Land Concentration	43.212 (84.535)	80.361 (125.887)	27.616 (51.997)	52.745 *** (73.89)
	Water-Saving Irrigation Coverage	0.821 (0.363)	0.86 (0.329)	0.805 (0.376)	0.056 (−0.047)
	Water Scarcity Level	3.834 (1.975)	5.326 (1.873)	3.207 (1.656)	2.119 *** (0.217)
Psychological Perception Level	Ecological Aspiration	1.589 (1.06)	2.424 (0.705)	1.239 (0.986)	1.185 *** (−0.281)
	Perceived Benefits	0.132 (0.338)	0.431 (0.497)	0.006 (0.076)	0.425 *** (0.421)
External Environment Level	Neighbor Imitation	0.16 (0.367)	0.424 (0.496)	0.05 (0.217)	0.374 *** (0.279)
	Socialized Services	0.265 (0.442)	0.563 (0.498)	0.14 (0.347)	0.423 *** (0.151)
Participation Behavior	Participation	0.296 (0.457)	1 (0.000)	0 (0.000)	1 (0.000)

Note: (1) *** indicate significance levels of 1%. (2) The last column presents the t-test results for the mean differences between the planting participants and non-participants. (3) The household total income values in this table are not log transformed.

.

From Table 1, it can be observed that the average income level of participating farmers is CNY 425,090, significantly higher than the CNY 78,380 average for non-participating farmers, with a difference of CNY 347,070 at the 1% significance level. Regarding household water-saving irrigation coverage, participating farmers exhibit higher levels compared to non-participants; however, the difference does not reach statistical significance.

Farmers who participate in planting economically valuable forests and grasslands report significantly higher perceived benefits from the new round of Grain-for-Green policies compared to non-participants. Additionally, the participating group faces notably higher levels of land concentration and water scarcity than the non-participating group. This may be attributed to the greater challenges associated with managing such farmland, which incentivizes farmers to join the Grain-for-Green Program.

The level of socialized services accessible to participating farmers is also significantly higher, surpassing that of non-participants with a statistical significance at the 1% level. Such findings imply that participating farmers have better access to technical support and other resources. Moreover, neighbor imitation and ecological aspirations are markedly stronger among participants, underscoring their critical role in motivating farmers to engage in planting economically valuable forests and grasslands.

However, the t-test analysis above merely highlights differences in selected variables between the two groups, offering only a preliminary understanding of how these factors might influence participation decisions. While the t-test results suggest potential relationships, they do not provide definitive insights into the impact of these factors on farmers’ participation. To explore this further, this study avoids relying on traditional regression analysis, which has been the standard approach in previous research. Revealing the underlying reasons for farmers’ participation requires moving beyond an “atomic perspective” that analyzes the impact of an individual element in isolation. Given the sophisticated nature of actual environments, the true environment in which farmers operate cannot be fully interpreted through conventional regression analysis.

Therefore, we implement the Qualitative Comparative Analysis (QCA) methodology, taking a holistic perspective. By treating these variables as antecedent conditions with potential influence, we use “configurational effects” to examine how these variables interact and work together to shape farmers’ participation in the planting of economically valuable forests and grassland.

5.2. Necessity Analysis

As a fundamental step in our analytical process, to ascertain if the antecedent condition variables are essential for the outcome, we initially evaluated whether each condition variable represents a requisite factor. The so-called necessary condition means that when a certain result occurs, it always exists, and then it is a necessary condition for achieving the specified outcome, and conversely, when the condition does not exist, the result will not be produced. In the necessity analysis of QCA, the consistency level serves as a critical indicator to evaluate whether a condition variable is an essential prerequisite for the achieving outcome. A condition is deemed essential when its consistency measure surpasses the threshold of 0.9 [52]. This indicates that in scenarios where the consistency measure surpasses 0.9, implying its presence in at least 90% of cases, the occurrence of the outcome invariably requires the presence of this condition. Consequently, the necessary condition encompasses all instances of the observed result, effectively forming its superset.

Drawing upon the empirical data presented in

Table 4. Necessity analysis of antecedent conditions.

Condition Variables	Participation		Non-Participation
	Consistency	Coverage	Consistency	Coverage
Household Total Income	0.821	0.496	0.351	0.504
~Household Total Income	0.179	0.104	0.649	0.896
Land Concentration	0.662	0.414	0.395	0.586
~Land Concentration	0.338	0.19	0.605	0.81
Water-Saving Irrigation Coverage	0.86	0.311	0.804	0.689
~Water-Saving Irrigation Coverage	0.14	0.231	0.196	0.769
Water Scarcity Level	0.753	0.475	0.351	0.525
~Water Scarcity Level	0.247	0.138	0.649	0.862
Ecological Aspiration	0.696	0.52	0.27	0.48
~Ecological Aspiration	0.304	0.149	0.73	0.851
Perceived Benefits	0.431	0.969	0.006	0.031
~Perceived Benefits	0.569	0.194	0.994	0.806
Neighbor Imitation	0.424	0.782	0.05	0.218
~Neighbor Imitation	0.576	0.203	0.95	0.797
Socialized Services	0.563	0.628	0.14	0.372
~Socialized Services	0.438	0.176	0.86	0.824

Note: the symbol (~) represents the antecedent condition as a non-high-level state; for example, ~ gross household income represents the level of non-high gross household income. And all (~) symbols in this article have the same meaning.

, a comprehensive analysis of consistency measurements for individual conditions in the resource endowment, psychological perception, and external environment dimensions with respect to farmers’ participation in the planting of economically valuable forests and grassland are generally below 0.9. This indicates that none of these single conditions is necessary for achieving this outcome.

At the same time, in the necessity analysis for non-participation in the planting of economically valuable forests and grassland, the consistency levels for ~perceived benefits (0.994) and ~neighbor imitation (0.95) exceed 0.9. This indicates that in the configurational pathways for non-participation, ~perceived benefits and ~neighbor imitation are necessary conditions.

5.3. Adequacy Analysis

Through the adequacy analysis of condition configurations, we can investigate the various permutations of antecedent conditions that adequately result in favorable results. The adequacy analysis of condition configurations is primarily performed by constructing and analyzing a truth table, followed by an evaluation of the resulting configurations.

To begin with, in constructing the truth table, we followed the guidance of QCA method experts and accounted for the large number of cases in our study. Specifically, we established separate frequency thresholds for the 144 participating farmers and the 343 non-participating farmers. This approach aims to eliminate the influence of contradictory configurations and avoid the uncertainty of extreme cases. Both the raw consistency and the PRI (Proportional Reduction in Inconsistency) values were set at 0.8. When the consistency and PRI values of a condition configuration with the outcome are equal to or greater than 0.8, the configuration is considered a sufficient condition for the outcome [47,55].

Secondly, in analyzing the truth table, each condition in the QCA can exist in two states: present or absent. Thus, 2ⁿ possible configurations can be generated for n antecedent conditions. However, not all logically possible configurations are supported by empirical cases, leading to the presence of logical remainders in the truth table analysis. To address this, a standardized analysis of the truth table was implemented, generating three distinct categories of solutions: complex, intermediate, and parsimonious.

The complex method is characterized by the exclusion of all residual components. In comparison, the intermediate solution is more selective, incorporating only those logically residual components that align with theoretical predictions and empirical evidence. Meanwhile, the parsimonious solution takes a comprehensive approach by including all residual components, thereby simplifying the overall framework. Among these, the intermediate solution is considered superior as it neither eliminates necessary conditions nor deviates from theoretical and empirical evidence. The intermediate solution is reported in this paper, and the parsimonious solution is used to separate primary and secondary conditions. Core conditions are those that overlap in both the intermediate and parsimonious solutions, whereas peripheral conditions are only present in the moderate approach [47,55].

Finally, grounded in the research situation, a thorough examination will be carried out for each path. When naming the configurations, this paper considers the coherence of all configurations and the uniqueness of each configuration [56].

5.3.1. Pursuing Profits Is Important, but It Is Not the Only Solution

In practice, for the outcome of participation, as indicated in

Table 5. Configurational pathways for participation in planting economically valuable forests and grassland.

	Participation
	①	②	③
Household Total Income
Land Concentration
Water-Saving Irrigation Coverage
Water Scarcity Level
Ecological Aspiration
Perceived Benefits
Neighbor Imitation
Socialized Services
Consistency	0.99	1	0.958
Raw Coverage	0.149	0.101	0.054
Unique Coverage	0.079	0.031	0.054
Number of Explainable Cases	31	5	17
Overall Solution Consistency	0.984
Overall Solution Coverage	0.233

Note: In the standardized analysis of the truth table for participation, because the qualitative inclusion algorithm cannot deal with redundant qualitative terms, users need to select the qualitative entailment terms (The prime implicants algorithm is a minimalization method of Boolean functions, which can produce prime implicants, that is, the logically simplest expressions, which logically include all original expressions and cannot be further simplified. However, when multiple equivalent original expressions are generated through the prime implicants algorithm, at least one of them can be considered redundant, and this situation is called redundant prime implicants. When redundant prime implicants are generated through the prime implicants algorithm, it must be selected by researchers based on theoretical and practical experience [53]). Based on the relevant theories and field investigations, this paper selects “ecological desire ∗ neighborhood imitation” as the qualitative implication terms, respectively, where “∗” indicates “and”, and all “∗” in this paper have the same meaning. indicates the presence of a core condition. indicates the presence of a peripheral condition. indicates the absence of a core condition. indicates the absence of a peripheral condition. A blank indicates that the condition is irrelevant (can be present or absent).

, the overall consistency of the configurational pathways leading to non-participation was 0.984. The overall solution coverage reached 23.3%. The consistency levels for the following three pathways were 0.99, 1, and 0.958, respectively—all exceeding the consistency threshold of 0.8. Therefore, these pathways can be regarded as sufficient conditions for farmers’ participation in the planting of economically valuable forests and grassland, and they exhibit equivalency.

We categorize the farmers in these three pathways into the following three types:

(1)

Farmers Driven by Benefit Perception

In Pathway 1 (household income ∗ land concentration ∗ water-saving irrigation coverage ∗ water scarcity ∗ ecological aspiration ∗ perceived benefits ∗ socialized services), farmers following this pathway have a solid operational foundation in terms of income, land concentration, and water-saving infrastructure, while also grappling with the challenge of limited irrigation water resources. Driven by the core condition—high perceived benefits—they participate in the planting of economically valuable forests and grassland. Additionally, high ecological aspirations and robust socialized services significantly influence their decisions.

(2)

Farmers Jointly Driven by Psychological Perception and Neighbor Imitation

In Pathway 2 (household income ∗ water-saving irrigation coverage ∗ water scarcity ∗ ecological aspiration ∗ perceived benefits ∗ neighbor imitation ∗ socialized services), these farmers have a solid operational foundation in terms of income and water-saving infrastructure. Although they are also facing the pressure of scarce irrigation water resources, regardless of whether their arable land is scattered or concentrated, farmers will actively participate in planting economic forests and grasslands, driven by three core conditions: ecological aspiration, high sense of benefit, and high neighborhood emulation.

(3)

Farmers Jointly Driven by Ecological Aspiration and Neighbor Imitation

In Pathway 3 (household income ∗ land concentration ∗ water-saving irrigation coverage ∗ water scarcity ∗ ecological aspiration ∗ ~ perceived benefits ∗ neighbor imitation ∗ ~ socialized services), farmers have a solid operational foundation, particularly in terms of income, land concentration, and water-saving infrastructure. Despite lacking perceived benefits and socialized services, the two core conditions—high ecological aspiration and high neighbor imitation—drive their participation in the planting of economically valuable forests and grassland.

In summary, across the three pathways, it is clear that farmers commonly operate in water-scarce environments while maintaining a solid operational foundation. In such conditions, factors related to psychological perception and the external environment interact to encourage farmers’ participation in planting economically valuable forests and grasslands:

Firstly, “Farmers Driven by Benefit Perception” are the most susceptible to policy incentives. With adequate support from socialized services, such as technical assistance and training, these farmers show a greater inclination to participate actively in planting economically valuable forests and grasslands, primarily driven by profit-oriented motives. Therefore, this configurational pathway has the highest raw coverage, indicating that profit-oriented farmers constitute the largest proportion among all participating farmers.

Secondly, “Farmers Driven by Psychological Perception and Neighbor Imitation” demonstrate equally high demands for both economic profits and ecological benefits. All factors at the psychological perception and external environment levels play a core driving role for these farmers.

Lastly, “Farmers Jointly Driven by Ecological Aspiration and Neighbor Imitation” have a higher demand for ecological benefits compared to the profit-driven farmers in Pathway 1. Their ecological rationality surpasses their economic rationality. Motivated by the pursuit of ecological benefits and influenced by their peers, they participate despite the challenges. This particular pathway exhibits the lowest raw coverage, suggesting that this group represents the smallest segment among all participating farmers.

5.3.2. It Is Not Only the Lack of Perceived Benefits That Hinders Farmers’ Participation

For the outcome of non-participation, a frequency threshold of six was set. According to the configuration in

Table 6. Configurational pathways for non-participation in planting economically valuable forests and grassland.

Condition Configurations	Non-Participation
	①	②	③	④
Household Total Income
Land Concentration
Water-Saving Irrigation Coverage
Water Scarcity Level
Ecological Aspiration
Perceived Benefits
Neighbor Imitation
Socialized Services
Consistency	0.978	0.964	0.964	0.958
Raw Coverage	0.419	0.455	0.275	0.144
Unique Coverage	0.129	0.164	0.063	0.034
Number of Explainable Cases	118	127	63	72
Overall SolutionConsistency	0.961
Overall Solution Coverage	0.68

Note: indicates the presence of a core condition. indicates the presence of a peripheral condition. indicates the absence of a core condition. indicates the absence of a peripheral condition. Blank indicates that the condition is irrelevant (can be present or absent).

, the overall consistency level of configurational pathways associated with non-participation was found to be exceptionally high, reaching 0.961. The consistency levels for the following four pathways were 0.978, 0.964, 0.964, and 0.958, respectively, all exceeding the consistency threshold of 0.8. Therefore, these pathways can be regarded as sufficient conditions for farmers’ non-participation in the planting of economically valuable forests and grassland, and they exhibit equivalency.

In Table 6, since we found that there are four pathways that can produce the result that farmers do not participate in the planting of economic forests and grasslands—in which paths 1 and 2 can be grouped into one category and paths 3 and 4 can be grouped together, which each constitute a second-order equivalent configuration [57]—we categorize the configurational pathways for non-participation into the following two types:

(1)

Farmers Facing Dual Deficiencies in “Psychological Perception + External Environment”

In Pathway 1 (~ household income ∗ ~ ecological aspiration ∗ ~ perceived benefits ∗ ~ neighbor imitation ∗ ~ socialized services), these farmers generally have non-high incomes, and their non-participation is primarily driven by the three core conditions of non-high ecological aspiration, non-high neighbor imitation, and non-high socialized services, with non-high household income and non-high perceived benefits playing a supporting role in reinforcing their decision not to participate.

In Pathway 2 (water-saving irrigation coverage ∗ ~ ecological aspiration ∗ ~ perceived benefits ∗ ~ neighbor imitation ∗ ~ socialized services), these farmers possess a solid foundation in water-saving infrastructure. Similar to Pathway 1, farmers’ non-participation is primarily driven by the three core conditions of non-high ecological aspiration, non-high neighbor imitation, and non-high socialized services, with non-high perceived benefits playing a supporting role in reinforcing their decision not to participate.

Therefore, Pathways 1 and 2 indicate that in the absence of technical support from socialized services and the influence of neighbor imitation, farmers with insufficient ecological aspiration and non-high perceived benefits are ultimately unable to commit to the planting of economically valuable forests and grassland, as driven by their profit-seeking motives. Field surveys reveal that non-participating farmers predominantly opted to cultivate grain crops, with surplus labor seeking external employment. This behavior poses challenges for achieving the policy’s goal of combating desertification and creating sustainable forests.

(2)

Non-High Income Traditional Farmers Dependent on Land Endowments

In Pathway 3 (~ household income ∗ water-saving irrigation coverage ∗ ~ water scarcity ∗ ~ perceived benefits ∗ ~ neighbor imitation ∗ ~ socialized services), farmers generally have non-high incomes and face limited water resource constraints, despite possessing a strong foundation in water-saving infrastructure. Their non-participation is primarily driven by the two core conditions of non-high household income and non-high water scarcity. Additionally, non-high perceived benefits, non-high neighbor imitation, and non-high socialized services serve as auxiliary factors, reinforcing their decision not to participate.

In Pathway 4 (~ household income ∗ land concentration ∗ water-saving irrigation coverage ∗ ~ water scarcity ∗ ~ ecological aspiration ∗ ~ perceived benefits ∗ ~ neighbor imitation), farmers generally have non-high incomes. Despite facing limited water resource constraints, they possess high land concentration and a strong foundation in water-saving infrastructure. Their non-participation is primarily driven by the two core conditions of non-high household income and non-high water scarcity. Additionally, non-high ecological aspiration, non-high perceived benefits, and non-high neighbor imitation play supporting roles in reinforcing their decision not to participate.

Therefore, Pathways 3 and 4 indicate that, when irrigation water resources in their environment are relatively abundant, their insufficient ecological aspiration and lack of the neighbor demonstration effect and socialized service support lead to a fear of the uncertainties of policy. As a result, their production decisions tend to favor traditional crops, such as cotton and vegetables, which are water-intensive.

Field surveys further reveal that the water-saving irrigation facilities utilized by these farmers were constructed under the unified leadership of the local government. These facilities reduce the water costs. Consequently, due to the endowment effect, farmers with access to these facilities are more inclined to continue allocating their farmland to cotton and vegetable production. This reinforces their status quo bias, further hindering their willingness to participate in the GGP.

5.4. Robustness Test

Robustness tests were carried out in this study to examine the configurational pathways of farmers who participated and those who did not [55]. QCA, as a set-theoretic approach, is regarded as robust when minor adjustments to operational parameters yield results that preserve clear subset relationships and do not significantly affect the substantive conclusions drawn from the research [58,59].

This study aligns with the approach introduced by Ordanini et al. (2014) [60]; we first addressed the relatively small sample size of “participation” by lowering the frequency threshold from four to three and increasing the PRI consistency from 0.8 to 0.85. The resulting configurations were largely consistent, demonstrating stability in the analysis (see Appendix A).

Second, for the larger sample size of “non-participation”, we lowered the frequency threshold, reducing it from six to five. Additionally, the PRI consistency was increased from 0.8 to 0.85. The resulting configurations also remained largely consistent, and the subset relationship was clear (see Appendix B).

These robustness tests demonstrate that the outcomes of our research are robust and reliable.

6. Conclusions

In this investigation, we utilized survey data obtained from 487 farmers across seven counties (cities) in northern and southern Tianshan, Xinjiang. By employing the fuzzy-set qualitative comparative analysis (fsQCA) method, we examined the determinants that drive both participation and non-participation in the planting of economically valuable forests and grasslands. Grounded in behavioral economics theory and the “rational economic agent” hypothesis, and adopting a multidimensional analytical framework of “resource endowment—psychological perception—external environment”, this study reached the following conclusions:

(1)

The involvement of farmers in the cultivation of economically valuable forests and grassland is characterized by multiple complex causal relationships. Both non-participation and participation behaviors demonstrate attributes of equifinality and conjunctural causation, highlighting the multifaceted nature of decision-making in this context.

(2)

The configurational pathways for participating farmers can be categorized into three types: farmers driven by benefit perception; farmers jointly driven by psychological perception and neighbor imitation; and farmers jointly driven by ecological aspiration and neighbor imitation.

Each pathway is closely related to farmers’ strong ecological aspirations in relatively water-scarce environments, and these farmers generally have a solid operational foundation in terms of income, land concentration, and water-saving infrastructure. First, “Farmers Driven by Benefit Perception” are primarily motivated by economic profit. They exhibit high perceived benefits from planting economically valuable forests and grasslands and receive robust support from socialized services. Second, “Farmers Driven by Psychological Perception and Neighbor Imitation” demand both economic profits and ecological benefits equally. They benefit from the dual advantages of psychological perception (high ecological aspiration and high perceived benefits) and the external environment (neighbor imitation and socialized services). Third, “Farmers Jointly Driven by Ecological Aspiration and Neighbor Imitation” have a stronger demand for ecological benefits. These farmers exhibit strong ecological aspirations and are motivated by the demonstration effects of neighbor imitation, which are reinforced through collective action.

(3)

The configurational pathways of non-participating farmers can be classified into two distinct types: those experiencing dual deficiencies in both psychological perception and the external environment, and traditional farmers who heavily depend on agricultural irrigation water.

Each pathway is intricately connected to farmers’ perceived benefits and neighbor imitation, while two of the pathways are also significantly influenced by water scarcity. However, participation behavior is not solely influenced by the perceived benefits in any pathway. Farmers experiencing dual deficiencies in both psychological perception and the external environment lack key intrinsic motivators, such as ecological aspirations and perceived benefits, as well as extrinsic drivers like neighbor imitation and access to socialized services. Their behavior is primarily driven by these two dimensions working together. On the other hand, low-income traditional farmers who rely on agricultural irrigation water are highly dependent on the availability of relatively abundant irrigation resources. Without intrinsic psychological drivers or external environmental support, these farmers are more inclined to cultivate high-water-demanding cash crops, like cotton and vegetables, which offer greater expected returns. Their behavior is predominantly shaped by their low-income situation and the availability of abundant irrigation resources, with the absence of psychological perception and external environmental factors serving only as secondary influences.

7. Policy and Theoretical Implications

7.1. Policy Implications

In summary, the Grain-for-Green Program (GGP), as one of the most pivotal and significant agroforestry policies in China, has been instrumental over the past two decades in mitigating wind and sand erosion while enhancing regional ecological environments.

The high regional overlap between the GGP and poverty makes it necessary to change the Government’s “subsidy-dependent” policy to a “self-sustaining” policy, in order to promote local sustainable development, achieve ecological improvement, and reduce poverty and increase the incomes of farmers. It is only through the promotion of sustainable local development that ecological improvements and the reduction in poverty and increase in income for farmers can be realized.

So, the above analysis of the antecedent conditions and configurational pathways for farmers’ non-participation and participation behaviors provides the following implications for the voluntary and sustained participation of farmers in the cultivation of economically valuable forests and grasslands, as well as similar policies in other arid zones around the globe:

First, enhance farmers’ positive expectations toward planting economically valuable forests and grasslands. Given the significant impact of ecological aspiration and perceived benefits on farmers’ motivation to participate, it is essential to leverage diverse media channels, such as radio, television, and social media, to promote environmental protection awareness. This will enhance farmers’ comprehension of issues such as desertification and climate change, promoting a heightened awareness of environmental protection. When farmers develop intrinsic motivation at the psychological perception level, coupled with sufficient perceived benefits and ecological aspiration, their enthusiasm for independently and sustainably managing forest and grassland resources will be greatly enhanced.

Second, provide tailored socialized services for economically valuable forests and grasslands in arid regions. Given the significant role of neighbor imitation and socialized services, it is crucial to establish the supportive external ecological or social surroundings for farmers to participate in planting. This includes fostering the influence of skilled individuals within villages and continually improving socialized services by offering professional technical guidance, training, and market sales support. Additionally, leveraging the role of local water user associations in promoting water-saving facilities is essential. Collective action among farmers can motivate more individuals to adopt water-saving systems, thus promoting the wider implementation of the economic forest and grassland management model.

Finally, promote water-adaptive planting in arid regions. Local governments, in guiding farmers to participate in the management of economically valuable forests and grasslands, should ensure seamless coordination across various policies. While promoting water-saving agricultural practices, particular attention should be paid to optimizing cropping patterns in areas facing significant water scarcity. Farmers should be incentivized to manage economically valuable forests and grasslands, encouraging diversified planting systems that are compatible with the local water resource availability.

In planning the long-term implementation of the Grain-for-Green Program, it is essential to balance the dual goals of sustained sand control and regional climate improvement with foundational policies like local “land reduction and water saving” initiatives and the national food security strategy. By doing so, the relevant policies can be better integrated and improved.

7.2. Theoretical Implications

In summary, our study makes the following contributions. We move beyond the limitations of previous research that relied solely on theories like social exchange theory to validate the influence of individual factors on farmers’ participation behaviors. Such studies failed to explore the configurational effects of resource endowment, psychological states, and external environments on the mechanisms influencing farmers’ participation behaviors.

Drawing on the complex systems perspective and integrating relevant theories of behavioral economics, planned behavior, and social exchange, we constructed a theoretical framework of “resource endowment–psychological perception–external environment”. This framework analyzes the driving mechanisms behind farmers’ non-participation and participation through factors of the dimensions of resources, psychology, and environment, establishing a more comprehensive and holistic analytical model.

Furthermore, constrained by methodological limitations, traditional research has often relied on correlation-based causal inference. This approach is inadequate for unraveling the more intricate causal relationships present in real-world contexts. As a result, such studies have fallen short of uncovering the nuanced differences in the mechanisms that either impede or facilitate farmers’ participation in the cultivation of economically valuable forests and grassland. In contrast, our research, by analyzing the configurational effects among factors within pathways, identifies multiple pathways that can either hinder or promote participation. This highlights the causal complexity underlying farmers’ behaviors and offers insights from a configurational perspective that can be applied to similar cases in other regions.

8. Research Limitations and Future Directions

First, understanding how to motivate farmers to participate in afforestation activities remains a valuable area for ongoing research. Future research could focus on gathering additional relevant data and applying dynamic QCA methods to further explore the intricate causal relationships between various influencing factors and farmers’ afforestation behaviors across diverse spatial and temporal contexts. Additionally, while the survey data utilized in this study benefit from their structured nature, they fall short in providing an in-depth examination of the observed phenomena. Future research could integrate grounded theory to explore farmers’ situations more comprehensively.

Finally, due to the limitations of the QCA method in accommodating a restricted number of antecedent conditions, this study was constrained to analyzing only the key variables that significantly influence the outcomes. Future research could adopt alternative theories and perspectives to explore other antecedent conditions not covered in this study.