From Forest Land Easements to Broader Conservation Agreements: An Analysis of Pathways to Community Support in China’s National Park Pilot

Abstract

Conservation easements (CEs) represent a complex policy instrument designed to mediate the feedback loops within coupled human and natural systems in protected areas. However, their efficacy is often constrained by a lack of systemic understanding of the localized drivers of community support. Building upon the successful implementation of Forest Land Easements (FLEs) within China’s Qianjiangyuan National Park Pilot, this study investigates the potential to expand this policy model to other land types. This study investigates the multilevel factors influencing residents’ willingness to adopt three types of CEs, including forest land (FLE), agricultural land (ALE) and homestead land (HLE) easements in China’s Qianjiangyuan National Park Pilot, the country’s primary CE reform site. We conceptualize a hierarchical support model wherein community participation (CP) and human well-being (HW) interact with support for park management (SM), forming a subsystem that drives decisions within the broader land-use. Utilizing structural equation modelling (SEM) and stepwise regression analysis on survey data from 336 households, we tested this model. The results reveal that SM acts as a critical direct mediator and positive driver of CE acceptance, while CP and HW exert significant indirect effects through SM, demonstrating a key feedback pathway. Regression analyses further elucidate that support for different CE types is driven by distinct configurations of factors, highlighting the heterogeneous nature of subsystems. Notably, livelihood benefits and prior participation experiences emerged as consistent, cross-cutting systemic leverages. It demonstrates that leveraging the implementation experience and community support gained from existing forest land easements is crucial. This study concludes that effective CE design must move beyond one-size-fits-all approaches. It necessitates differentiated, adaptive policies that are coherently aligned with local livelihood subsystems and strategically strengthen participatory feedback mechanisms initiated by successful FLEs. Our findings provide an evidence-based framework for designing resilient, socially sustainable conservation policies in complex protected area systems, grounded in proven practice.

1. Introduction

Conservation easements (CEs) are a prominent environmental policy tool for advancing natural resource conservation [1,2,3] while facilitating adaptive integration of sustainable management practices within global protected areas [4,5]. In the context of global biodiversity conservation frameworks such as the Convention on Biological Diversity (CBD) and the post-2020 Global Biodiversity Framework adopted at COP15, which emphasize the integration of effective area-based conservation measures with social equity and participatory governance, CEs represent a policy instrument that aligns with the goals of protecting 30% of terrestrial and marine areas by 2030 while respecting local livelihoods and rights [6,7]. In particular, forest land easements (FLEs), a specific form of CE applied to forested landscapes, have gained recognition as critical instruments for maintaining forest cover, protecting watersheds, and sequestering carbon, while allowing certain sustainable uses by rights-holders. The concept of CE first emerged in national park management in the 1930s and 1940s, when the United States (U.S.) National Park Service purchased CEs to protect views along a highway corridor [6]. However, the term “conservation easement” did not exist until the 1950s [7]. Building on this legacy, contemporary CEs, including FLEs, are broadly defined as voluntary, legally binding agreements that permanently limit land use to protect its conservation value. Currently, CEs are defined as voluntary, legal agreements that permanently limit land use to protect its conservation value [8]. Evidence shows that CEs can be an effective tool for protecting land at a lower cost than land trusts and public agencies, while allowing the landowner to retain many private rights [9]. CEs are permanent restrictions on land use associated with a property deed: they may prohibit subdivision and development; they may limit the extent of logging or ranching; they may prohibit mining; or they may limit the total building footprint on a property. These restrictions preserve critical habitats and safeguard biodiversity while allowing private landowners to own, use, sell, and bequeath the land subject to the easement stipulations [10]. Therefore, CEs have become a prominent tool for protecting biodiversity in the U.S. [11]. Nevertheless, within the broader scope of transnational conservation policy learning and adaptation, current policymakers often prioritise the structural design of CEs, while empirical analyses of resident support mechanisms and differentiated governance approaches remain underdeveloped. This gap not only limits the effective transfer and adaptation of CE models across different national and regional contexts, particularly in countries with complex land tenure systems and high social–ecological diversity, but may also lead to unintended negative social outcomes after implementation, thereby undermining their alignment with CBD’s dual objectives of conservation efficacy and social inclusivity. Therefore, it is essential to further clarify the influencing factors and response pathways of community support for CEs policies.

In recent years, CEs have been applied to forest conservation in protected areas (PAs) in China [12]. In line with China’s commitment to the COP15 biodiversity targets and its national ecological civilization strategy, CEs are increasingly regarded as an innovative mechanism to reconcile ecological protection with community development in protected areas. National Park pilots in China, such as Qianjiangyuan National Park (QNP) and Wuyishan National Park, have revolutionized the establishment of CEs to maintain the ecological functions rather than the economic functions of collective lands, with various compensations to landowners and users [13,14]. Thus far, QNP has implemented CEs for the longest period and has achieved periodic success in resolving conflicts between local development and conservation. As a relatively new conservation policy in China’s national park management, CEs are implemented without precedent [15], and it is unknown whether they have been a successful practical tool for balancing local development and park conservation. Understanding the success of adapting CEs to the Chinese context in resolving this dilemma is significant not only for efficient park decision-making but also for the broad application of CEs in areas with complex land tenure and conservation demands.

Protected areas (PAs) are quintessential coupled human and natural systems, characterized by dense networks of reciprocal feedback between ecological processes and human agency [16]. Understanding community support for conservation is essential for effective protected area governance. Conservation policies such as Conservation Easements (CEs) act not as simple incentive mechanisms, but as complex interventions that reconfigure the structural relationships among local communities, land management institutions, and ecosystem services [12,17,18]. Previous research has often framed such support through a transactional lens, such as social exchange theory, where support is contingent upon a favorable balance of perceived benefits over costs [19]. While valuable, this perspective may not fully capture the hierarchical and mediated relationships through which support for specific policies develops. Community members do not evaluate policies in isolation; their acceptance of interventions like CEs is likely shaped by their prior experience with conservation programs, and their perceptions of how their well-being and participation are valued [20].

To capture this complexity, we propose a Hierarchical Support Model (Figure 1) that situates the willingness to adopt CEs from a structured set of relationships. In this model, ultimate support for CEs is directly driven by trust in the managing institution (SM), which, in turn, is shaped by more foundational system components: the material and perceived state of human well-being (HW) and the processes of community participation (CP). This structure reflects the hypothesis that effective conservation policy requires more than delivering benefits; it requires building legitimate governance through participatory processes that enhance well-being and, in turn, foster institutional trust. We test these systemic relationships using Structural Equation Modeling (SEM), a method suited for disentangling the direct, indirect, and mediating pathways among these latent variables. Our specific hypotheses, derived from this framework, are as follows (see Figure 1):

H1. 

Human well-being (HW) positively influences support for national park management (SM).

H2. 

Human well-being (HW) positively influences support for conservation easements (SC).

H3. 

Support for national park management (SM) positively influences support for conservation easements (SC).

H4. 

Community participation (CP) positively influences human well-being (HW).

H5. 

Community participation (CP) positively influences support for national park management (SM).

H6. 

Community participation (CP) positively influences support for FLEs and other conservation easements (SC).

Figure 1. The hypotheses of the proposed Hierarchical Support Model for CEs. SC refers to support for CEs; HW refers to human well-being; SM refers to support for national park management; CP refers to community participation.

This study applies the model to the context of China’s Qianjiangyuan National Park (QNP), the country’s primary pilot site for CE reforms. QNP presents a critical case study as it features varying degrees of implementation across different easement types: Forest Land Easements (FLEs) are widely established, while Agricultural Land Easements (ALEs) and Homestead Land Easements (HLEs) are in earlier piloting stages. Investigating support across these three types allows us to examine how the proposed hierarchical mechanisms land-use contexts. The findings are poised to provide evidence-based guidance for designing differentiated, adaptive CE policies coherently aligned with local socio-ecological contexts, thereby enhancing the resilience and social sustainability of protected area governance.

2. Methods

2.1. Study Area

QNP is located in Kaihua County, Zhejiang Province in southeastern China (29°23′25.55″ N, 118°12′39.72″ E), covering an area of 252 km² (Figure 2). It is one of the pilot national parks established in 2016, containing 4 towns, 21 administrative villages and 72 natural villages, with more than 9744 community residents within the national park. The park conserves central subtropical evergreen broad-leaved forests and provides important habitats for many rare and endangered wildlife species, such as Manis pentadactyla, Muntiacus crinifrons, and Syrmaticus elliott (IUCN, 2024) [21]. It is also located in a region famous for its tourism destinations, bordered by Mount Huang to the north, Qiandao Lake to the east, and Mount Sanqing to the west. As the only pilot national park established in the densely populated area in southeastern China, the spectacular natural scenery, ecological values, and developed regional economy make QNP a great example of engaging in cross-regional cooperative protection and achieving a balance between conservation and the economy.

Figure 2. The location of QNP. This Chinese map (top left) is supervised and certified by the Ministry of Natural Resources of the People’s Republic of China, with certification code: GS(2019)1673.

In the QNP, state-owned land accounts for only 20.4%, while the rest proportion of land is collective land, which is owned collectively by farmers’ collectives. To better conserve forest resources and conduct ecological restoration, QNP has emerged as the first protected area in China to formally implement FLEs via reform [22]. Under this framework, without the need for land expropriation, servient landowners will transfer a portion of their land rights to the National Park Management Agency (NPMA) through economic incentives, thereby facilitating the unified management of natural resources by the national park. In 2018, FLEs were implemented on 18,342.5 hm² of collective forest land, and 3757 farming households involved had signed easement contracts, with a fixed term until 31 December 2024. NPMA negotiated with local people at two different levels to involve them in entrusting their contracted forest to the community committee through a contract signed between the NPMA and the community committee. Furthermore, as of 2024, nearly 80.04 hm² of farmland within QNP has been incorporated into the scope of ALEs. Under the premise of prohibiting the use of pesticides and fertilizers, prohibiting straw burning, not harming wild animals entering the cultivated land, and allowing protected animals such as crested ibises to forage, landowners can receive an easement subsidy of 800 yuan per mu of farmland. Moreover, the NPMA plans to promote HLEs in the future [23].

2.2. Data Collection

A structured household survey, incorporating pre-survey testing, was conducted between June and August 2023 across all 51 natural villages of QNP. Households were randomly sampled from all local villages within the park. A preliminary test was conducted with local residents (10), village leaders (5), and QNP administrative staff (5) to identify key factors influencing support for conservation efforts (CEs) and to assess primary concerns. The final questionnaire was designed based on the preliminary pre-survey and previous related studies [24,25,26,27], which verified the validity and reliability of the scales, and was appropriately adjusted according to the current situation of QNP and the findings from the pre-survey. The survey yielded 336 complete, clear, and valid responses, resulting in an efficiency response rate of 82%. The random sampling within each of the 51 natural villages was designed to maximize representativeness, and the diversity of villages included (different locations, sizes, and livelihood patterns) helps mitigate but does not eliminate clustering concerns, which is a limitation of the methods in this study. The final questionnaire comprised three sections. The first section collected the respondents’ demographic information, including gender, age, education level, job, and household income. The second section addressed residents’ satisfaction and participation willingness on conservation police of CEs. CEs were included ALEs, FLEs and HLEs. The third section was designed based on the six hypotheses investigating the factors that influence local support for CEs (Appendix A). The factors were selected based on pretest and previous research [24,25,28,29] and included Human Well-being (HW), support for national park management (SM) and community participation (CP) (Table 1).

Table 1. Variables and questions of the SEM model.

Latent Variable	Observed Variable	Question	Measurement
Support for Conservation Easements	SC_1	Would your family be willing to support for participating in the FLEs?	1–5 Likert scales (Very unwilling–Very willing)
	SC_2	Would your family be willing to support for participating in the ALEs?
	SC_3	Would your family be willing to support for participating in the HLEs?
Human Well-being	HW_1	Do you think participating in CEs has increased your income?	1–5 Likert scales (disagree–agree)
	HW_2	Do you think participating in CEs has improved your quality of life?
	HW_3	Do you think participating in CEs has improved environmental safety and health?
	HW_4	Do your efforts in National Park conservation match the benefits you have received?
Support for national park management	SM_1	Are you willing to support National Park conservation?	1–5 Likert scales (Very unwilling–Very willing)
	SM_2	Are you willing to comply with National Park management regulations?
	SM_3	Are you willing to monitor and report violations of conservation regulations?
	SM_4	Are you willing to engage in ecological conservation advocacy?
	SM_5	Are you willing to provide suggestions for the development of the National Park?
	SM_6	Are you willing to participate in conservation management activities of the National Park?
Community Participation	CP_1	To what extent does the National Park management listen to and respect community opinions and demands?	1–5 Likert scales (disagree–agree)
	CP_2	Do you acknowledge a cooperative relationship with the National Park management?
	CP_3	How well do you understand the planning, construction, and management of the National Park?

Social Exchange Theory provides a robust framework for understanding multi-party collaboration, trust-building, and pro-environmental behavior by positing that individuals engage in interactions based on an evaluation of anticipated costs and benefits [26]. Informed by this perspective, we conceptualized support for park management as a multi-faceted construct. It was measured by gauging respondents stated willingness across a spectrum of engagement levels, ranging from basic support and regulatory compliance to more active roles such as monitoring, advocacy, providing suggestions and participating in conservation management activities.

Furthermore, based on the community participation model [25], we assessed the degree to which residents felt empowered to move beyond passive compliance toward active involvement. This involved evaluating their perceptions of whether park management listened to and respected community opinions, whether a cooperative relationship was acknowledged, and their understanding of park planning and management processes.

For measuring Human Wellbeing, our approach is guided by the multi-dimensional framework outlined in the Millennium Ecosystem Assessment [27], which emphasizes fundamental material needs, health, security, social relationships, and freedom of choice and action. While the Human Development Index integrated factors such as income, education, and life expectancy to assess Human Wellbeing, this study specifically focused on evaluating residents’ perceptions of Human Wellbeing improvements caused by CEs four dimensions: economic income, quality of life, environmental safety and health, and the alignment between conservation efforts and benefits was adopted in national parks (as shown in Table 1). All the designed questions were simple, clear, general, and non-sensitive. A measurement approach based on 5-point Likert scales was used to measure each question.

2.3. Data Analysis

Data processing and statistical modelling in this study were conducted using specialized software, where the descriptive statistics and regression analysis were performed in Microsoft Excel and SPSS version 26.0, while structural equation modelling was implemented in AMOS version 26.0 (IBM, New York, NY, USA). With a sample size of 336 households, the survey results carry a margin of error of 5.35% at a 95% confidence level. This sample size was adequate for performing the SEM analysis [27,30].

Reliability indicates the consistency and stability of measurement outcomes, while validity assesses the internal structural soundness of the questionnaire. The survey data demonstrated satisfactory reliability and validity, where the reliability was 0.896, and the validity was 0.843. All latent variables exceeded the threshold value of 0.70. We employed Cronbach’s alpha (α) and the Guttman split-half coefficient to evaluate reliability, and the results showed that all constructs showed Cronbach’s α scores above 0.70 and Guttman coefficients above 0.50, confirming adequate internal consistency across measurement items. Cronbach’s α values above 0.7 suggest good reliability, those between 0.35 and 0.7 indicate acceptable reliability, and values below 0.35 reflect poor reliability. A split-half coefficient exceeding 0.5 is generally considered acceptable [31]. SEM parameters were estimated using maximum likelihood estimation, allowing simultaneous testing of all hypothesized directional relationships. This approach is particularly suitable for examining latent variables, such as perceptions and attitudes, that are difficult to measure directly, thereby minimizing subjective measurement errors [32]. In our research, SEM was applied to analyze how various factors influence local support for conservation policies, including conservation easements. The method evaluates latent constructs through observed variables while accounting for measurement validity and reliability [33].

The SEM system is represented by the following matrix equations:

$$\eta =\beta \eta +\varphi \xi +\lambda$$

(1)

where $\eta$ denotes endogenous latent variables, $\xi$ represents exogenous latent variables, and $\varphi$ and $\lambda$ correspond to coefficient matrices and error vectors linking these variables.

$$X={\wedge }_X\xi +\delta$$

(2)

$$Y={\wedge }_Y\eta +\epsilon$$

(3)

Here, $X$ refers to observed variables associated with exogenous latent variables $\xi$, $Y$ denotes observed variables related to endogenous latent variables $\eta$, and ${\wedge }_X$ and ${\wedge }_Y$ are matrices of correlation coefficients.

Model fitting was conducted in AMOS V26.0. Following standard SEM procedures, we first examined potential estimation violations before evaluating overall model fit [34]. A key assumption is the normality of data distribution. Typically, absolute skewness values greater than 3 and kurtosis values exceeding 10 (or a stricter threshold of 8) suggest non-normality. Testing with SPSS V26.0 confirmed that all variables had skewness absolute values below 1 and kurtosis absolute values under 2, meeting the normality assumption and justifying the use of maximum likelihood estimation in AMOS. Subsequently, stepwise regression analysis was performed in SPSS V26.0 to further examine influencing factors. This method iteratively selects independent variables that best explain variation in the dependent variable, making it suitable for identifying key predictors of local support for different conservation easement types (agricultural, forest, and homestead land). After comparing alternative regression specifications based on data characteristics and model fit, the stepwise approach was adopted for its effectiveness in clarifying distinct drivers behind support for each easement category.

This study methodologically draws upon and continues the Structural Equation Modeling (SEM) and stepwise regression analysis framework applied in Hu et al. (2022) [19] to study community perceptions of tourism development in the US Grand Canyon National Park, as this framework has proven effective in disentangling multi-level variable relationships within complex social–ecological systems. Simultaneously, we have highlighted the key adaptations and extensions of this study: (a) Different Research Context and Core Variables: Shifting the analytical context from perceptions of tourism development to support for conservation easement policies; core constructs were adapted from tourism-related concepts to Community Participation (CP), Human Well-being (HW), Support for Park Management (SM), and Support for Conservation Easements (SC), aligning with the context of conservation policy research. (b) Deepening the Theoretical Framework: Explicitly introducing the Hierarchical Support Model to provide a more systematic interpretation of feedback loops within conservation policy–community interactions, moving beyond the primarily Social Exchange Theory perspective used in the prior study. (c) Contrasting and Extending the Case Study: Extending the analysis from a single, mature Forest Land Easement case to include three easement types (FLEs, ALEs, HLEs) at varying implementation stages (widely established, piloted, planned). This allows for comparative analysis across land types, testing the explanatory power of the methodology across different subsystems.

3. Results

3.1. Demographic Analysis

Table 2 presents the demographic profile of the surveyed respondents. The sample was predominantly male (73.2%), with nearly all respondents (92.6%) being ancestral residents of the area. In terms of age distribution, the two largest groups were respondents aged 40–59 (43.9%) and those over 60 years old (43.6%), together accounting for the majority of the sample. Geographically, more than half of the respondents (56.5%) resided inside the boundaries of the national park. Educational attainment was generally limited, with 44.7% of respondents having completed only primary school or less. Additionally, a small proportion (5.2%) reported having lived in the community for at least three decades.

Table 2. The demographic profile of the respondents.

Variable	Variable Meaning	Sample Size	%	Variable	Variable Meaning	Sample Size	%
Gender	Male	221	73.2	Education level	Primary school or below	134	44.7
	Female	81	26.8		Junior high school	119	39.7
Age	18–24	2	0.6		Senior high school	32	10.7
	25–39	33	10.5		Associate’s degree	11	3.7
	40–59	138	43.9		Bachelor’s degree	4	1.3
	60 or older	137	43.6	Duration of residency	Ancestral residence	286	92.6
	Other	4	1.3		Parents’ migration	16	5.2
Residence location	Inside	178	56.5		Contemporary migration	2	0.6
	Outside	137	43.5		Immigration	5	1.6

3.2. Structural Equation Model

The initial hypothetical model was tested and required further revision. Based on the initial model test results, only HW, CP, and SM were found to have statistically significant impacts on SC. Therefore, model correction was carried out by using HW, CP, and SM for model construction. The absolute fit, relative fit, and simple fit indices were used to evaluate the overall suitability of the model. All the fit indices met the corresponding standards, indicating that the overall fit of the model was very good.

The hypotheses of the theoretical models were verified (Figure 3 and Table 3). Among the six hypotheses proposed, SM was found to have a positive impact on support for CEs, and CP was found to have a positive impact on HW and SM. The negative impact of CP on support for CEs was not found to be significant, nor were the negative impacts of HW on support for CEs and national park management. Finally, SM was confirmed to be an important factor in support for CEs and CP. However, the findings did not confirm that residents’ HW and CP are driving factors for support for CEs. Residents’ willingness to participate in CEs policies is significantly positively affected by their satisfaction with the national park management. The higher the residents’ satisfaction, the more willing they are to support CEs policies. Furthermore, the residents’ level of community participation has a positive impact on support for national park management. Therefore, the degree of community participation has an indirect positive impact on residents’ willingness to support CEs policies.

Figure 3. The structural model of Support for CEs.

Table 3. The results of hypothesis verification.

Hypothesis		Parameter Estimate	Critical Ratio Value	Significance	Results
H1	HW positively impacts SC	0.044	0.501	0.616	Failed
H2	HW positively impacts SM	0.010	0.158	0.875	Failed
H3	SM positively impacts SC	0.819	7.477	***	Validated
H4	CP positively impacts HW	0.526	11.949	***	Validated
H5	CP positively impacts SM	0.217	4.453	***	Validated
H6	CP positively impacts SC	0.081	1.177	0.239	Failed

Note. SC refers to support for CEs; HW refers to human well-being; SM refers to support for national park management; CP refers to community participation. ***: p < 0.001.

3.3. Exploration of Factors Influencing Different Types of CEs

To further examine whether the support for different types of CEs was influenced by different factors, we gradually incorporated the measurement variables derived from SEM and demographic variables into the stepwise regression model. The results of the stepwise regression model for the three types of CEs were found to be significant (p < 0.05) (Table 4). All models demonstrated that the higher the communities’ willingness to engage in conservation, the more willing they were to support CEs. Although the three different types of CEs were found to be influenced by different factors, they were commonly influenced by livelihood improvements and benefits. For the FLEs policy, which has been implemented on a larger scale in QNP, respondents’ willingness to support this policy is significantly correlated with livelihood and spatial accessibility factors such as their education level and residential location and previous environmental experience. Residents with higher education levels and living in areas with better road accessibility had a higher willingness to support FLEs. It is worth noting that the community’s previous experiences of participating in ecological compensation, CEs, and other protection agreements significantly influenced the support for FLEs.

Table 4. The stepwise regression models of the results of influencing factors on respondents’ willingness to support CEs.

Model		Unstd. Coef		Std. Coef	t	Sig	Collinearity Statistics
		B	SE	Beta			Tolerance	VIF
HLEs	Constant	1.806	0.326		5.534	0.000
	WC	0.312 ***	0.076	0.246	4.084	0.000	0.693	1.442
	WR	0.184 *	0.079	0.143	2.312	0.021	0.657	1.522
	CB	−0.171 **	0.054	−0.196	−3.149	0.002	0.650	1.537
	IL	0.136 *	0.055	0.163	2.454	0.015	0.574	1.743
	SI	0.209 *	0.086	0.126	2.424	0.016	0.938	1.066
		SS	df	MS	F	Sig
		41.751	5	8.350	15.095	0.000
ALEs	Constant	−0.128	0.567		−0.225	0.822
	WC	0.252 *	0.120	0.193	2.107	0.037	0.564	1.773
	RU	0.271 ***	0.071	0.278	3.833	0.000	0.899	1.112
	LI	−0.370 **	0.137	−0.195	−2.701	0.008	0.910	1.098
	WR	0.269 *	0.123	0.185	2.189	0.030	0.659	1.517
	EA	0.237 *	0.115	0.170	2.058	0.042	0.691	1.447
		SS	df	MS	F	Sig
	Regression	48.341	5	9.668	18.112	0.000
FLEs	Constant	2.488	0.298		8.353	0.000
	WC	0.528 ***	0.065	0.448	8.170	0.000	0.903	1.108
	EPP	−0.138 **	0.042	−0.192	−3.270	0.001	0.792	1.263
	LI	−0.374 ***	0.078	−0.261	−4.809	0.000	0.920	1.087
	EL	−0.095 **	0.029	−0.182	−3.278	0.001	0.883	1.132
	DR	−0.051 **	0.016	−0.181	−3.090	0.002	0.796	1.256
	IE	0.128 **	0.045	0.159	2.860	0.005	0.875	1.142
		SS	df	MS	F	Sig
	Regression	52.033	6	8.672	25.156	0.000

Note. Unstd. Coef: Unstandardized Coefficient; B: The unstandardized regression coefficient; SE (Std. Error): The standard error of the coefficient B; Std. Coef (Standardized Coefficient/Beta): The standardized regression coefficient.; VIF: Variance Inflation Factor; The independent variables included in the regression model were willingness to conservation (WC); willingness to comply with park management regulations (WR); agreeing that the cost matches the benefits of conservation (CB); agreeing that the establishment of a national park has improved the quality of life (IL); economic forests as a source of income (SI); risk of unemployment (RU); willingness to comply with park management regulations (WR); willingness to engage in conservation advocacy (EA); participation in ecological compensation, CEs, and other protection agreements (EPP); living inside the national park (LI); the respondent’s education level (EL), the distance from county roads (DR); agreeing that establishment of a national park improves environmental safety and health (IE). *** p < 0.001, ** p < 0.01, * p < 0.05.

For the ALEs policy, which has only been piloted on a small scale, and the planned HLEs policy, residents place greater emphasis on factors related to their own income, well-being, benefits, and potential risks. They were significantly influenced by factors such as agreeing that the cost matched the benefits of conservation, having economic forests as a source of income, and the risk of unemployment. Residents living within QNP have a lower willingness to support ALEs than those living outside of QNP.

4. Discussion

4.1. Key Findings and Theoretical Implications

Our results revealed a clear hierarchical structure in how community support for conservation policies develops, but not as we initially predicted. The findings required a nuanced interpretation that distinguishes between direct and indirect effects. The results showed that support for national park management was the direct driver of residents’ support for CEs policies, and community participation was an indirect driver. And human well-being directly impacts residents’ support for the CEs policy. Despite the absence of direct effects, the model reveals a robust and theoretically significant indirect pathway. Community participation significantly enhances both human well-being (H4 supported) and support for park management (H5 supported). Support for park management (SM), in turn, is the primary and powerful direct driver of support for conservation easements (H3 supported, β = 0.819, p < 0.001). This confirms that SM acts as a full mediator in the model.

These findings necessitate a re-conceptualization of the roles of HW and CP. They are not primary drivers of support. Rather, they function as foundational system components that build the institutional trust necessary for policy acceptance. Community participation enhances human wellbeing and, independently, strengthens trust in the managing institution. This trust then becomes the proximal cause of support for specific policies. This pattern of influence underscores that community support for conservation policies is not merely a transactional calculation of costs and benefits, but an emergent outcome of interrelated socio-ecological subsystems. These findings are consistent with social exchange theory, which posits that individual engagement is influenced by perceived returns from an interaction [26]. There is a similar finding from a study of Gunung Mulu National Park, which suggested that human wellbeing could influence locals’ support for the planning and policies implemented in the national parks or WHS areas [35]. These results are consistent with the social exchange theory”, which assumes that the level of individual engagement in the exchange process depends on the costs and benefits of the exchange [26,36,37]. The theory is built on the assumption that the exchange between two parties, whether individuals, industries, or groups, depends entirely on the relationship between them. Social exchange theory is not limited to industry or business, it is also applicable to society, communities, and individuals [29].

However, this study extends the social exchange theory, which suggests that individuals’ engagement depends on perceived returns from interactions. While social exchange theory suggests that support depends on perceived benefits, our findings reveal that this relationship is mediated. Residents do not simply exchange their support for direct benefits. Rather, benefits and participatory experiences shape their trust in the governing institution, and it is this institutional trust that ultimately enables acceptance of specific policy interventions. The distinction has important implications. It suggests that building institutional legitimacy may be as important as delivering material benefits, and that participatory processes serve a dual function of delivering tangible outcomes and building relational capital. The strong positive effect of SM on SC underscores the centrality of institutional trust. When residents trust the park management and perceive it as legitimate, responsive, and worthy of cooperation, they are substantially more willing to accept specific conservation policies. This finding aligns with research on institutional trust in environmental governance [38,39].

This integration shifts the analytical focus from a static assessment of net benefits to a dynamic understanding of how long-term feedback loops, among participatory processes, perceived well-being, and institutional trust, co-produce policy legitimacy and compliance. In the context of national park establishment in China, potential conflicts and exchanges of interest may exist between park officials and local community residents regarding ecological conservation and community development [19]. These conflicts and benefits exchanges could affect the residents’ support for relevant conservation policies enacted by national parks, such CEs, thereby influencing the future coordination of the relationship between PAs and community development, as well as community planning.

Importantly, this study builds upon the implementation of Forest Land Easements (FLEs) in Qianjiangyuan National Park. It provides a crucial institutional reference, demonstrating that a policy allowing residents to receive economic subsidies without altering land tenure can tangibly demonstrate that conservation benefits outweigh costs, thereby fostering initial support. The results of the stepwise regression model for supporting FLEs policy further identified that support was higher among residents with better road access and higher education levels, linking support to livelihood advantages and capacity. This indicates that when residents enjoy the benefits of convenient transportation in their daily lives, they are more inclined to support ecological conservation. Residents with higher levels of education tend to have greater support for conservation policies. It can be observed by survey field that the overall livelihood conditions of residents with higher levels of education were better, as they often engage in higher-paying professions and run small businesses at home. Thus, a good household livelihood and convenient community living conditions are key factors in enhancing residents’ support for CEs policies.

Theoretically, this study made a progression from the FLEs’ demonstration of basic exchange logic, illustrating a methodological and conceptual extension of prior work [19]. While prior research effectively applied social exchange theory to tourism perceptions, the current study demonstrates how embedding this logic can uncover the deeper, multi-level pathways (e.g., CP → SM → SC) through which support for complex, long-term conservation policies is generated and sustained. While the model specifies participation as antecedent to institutional trust based on the temporal sequence of policy implementation in QNP, the relationship between CP and SM is likely reciprocal over longer time horizons. Trust built through initial participation may enable deeper subsequent engagement, creating a virtuous cycle. Longitudinal data would be needed to fully disentangle the causal direction of this relationship. In summary, the key theoretical contribution of this study is not the discovery that “wellbeing and participation drive support”, but the revelation of how they matter: indirectly, through the critical mediating mechanism of institutional trust. This nuanced understanding provides a more solid foundation for policy design than simplistic assumptions about direct benefit–support linkages.

4.2. The Role of Past Experience and Participation

Our study further identified community participation as a significant determinant influencing residents’ support for CEs policies within national park communities. This finding extends the framework of social exchange theory, which traditionally emphasizes economic exchanges confined to formal and contractual benefits, by incorporating community participation as a form of social resource exchange Blau (2017) [40]. The study demonstrated a positive correlation between the degree of community participation and the level of support for CEs policies among respondents. This relationship can be explained by the fact that when residents are actively involved in the decision-making processes that affect their environment, they tend to develop a stronger sense of stewardship and investment in the outcomes of these policies. Such involvement can also lead to the development of more equitable and effective conservation strategies that align with the needs and values of the community [41].

The results indicated that the level of support for different types of CEs is influenced by different factors. However, a common feature is that support for conservation policies often stems from the benefits or assistance that residents have received from previous policies. These findings of this study are similar to those in the field of psychology, suggesting that people’s past experiences influence their decision-making and preferences for things, which is due to the specific cognition they have developed toward certain matters as a result of their past experiences [42,43]. Drawing from the implementation experience of FLEs, this study finds that residents’ prior involvement in ecological compensation, easements, and other conservation agreements significantly affects their support for FLEs. One possible explanation for this result is that FLE reform is the most widely implemented conservation policy in QNP, and almost all residents involved have participated in the reform and received corresponding compensation. However, it is clear that the surveyed residents, based on their previous participation experience, are not satisfied with FLEs. Consequently, it was found that the higher the degree of previous participation, the lower the willingness to participate in the future. This study aligns with the research by Peng et al. [44], which indicates that past environmental experiences can influence future environmental decision-making.

Although the results of this study do not directly show that past participation has affected residents’ support for ALEs, the location of the residents’ dwellings indirectly reflects this consequence [45]. This speculation stems from the ALE in QNP being piloted only within the national park, excluding residents outside its boundaries from participating in the reform. Consequently, while park residents have been influenced by the ALE policy after joining, their satisfaction remains low, leading to weaker support than those living outside the park. The relationship between satisfaction and support is often reciprocal and influential, particularly in the context of services, policies, or social dynamics. From the perspective of social exchange theory, when community residents feel that conservation policies and the issues themselves are beneficial to their livelihoods and well-being, they may develop a psychology of reciprocity [46], thereby increasing their satisfaction with and support for these issues. The implementation journey of FLEs demonstrates that establishing this positive cycle of “benefits–satisfaction–support” is key to promoting the sustainability of conservation policies.

4.3. Inheritance and Adaptation of the Methodological Framework

The methodological framework of this study draws upon and extends the effective approach employed in our prior analysis of community perceptions in protected areas [19]. Specifically, we have adopted the analytical paradigm combining Structural Equation Modeling (SEM) and stepwise regression analysis to decipher complex relationships among multi-level variables. This paradigm has proven effective in previous research for revealing the systemic mechanisms underlying the formation of community attitudes.

The primary adaptations and distinctions from the prior study are as follows. Firstly, the core variables and research context are different. This study shifts the analytical focus entirely from “perceptions of tourism development” to “support for conservation easement policies”. Accordingly, we constructed a measurement model centered on community participation, human well-being, and management support to fit the specific context of policy evaluation within China’s national parks. Secondly, this study is based on a deeper theoretical framework. While retaining the perspective of social exchange theory, this paper explicitly constructs a Hierarchical Support Model. This aims to more systematically capture the feedback loops between socio-ecological subsystems within conservation policies. Last but not least, compared to the prior analysis of a single, this study expands the research subjects to three types of easements at different implementation stages (FLEs widely implemented, ALEs piloted locally, HLEs in the planning phase), which represent a comparative case design. This design allows us to test the explanatory power of the same analytical framework across different land types and livelihood subsystems, thereby revealing the heterogeneous impacts of the policies.

4.4. Practical Recommendations for CE Design

Based on this study’s findings, the levels of residents’ support for ALEs, FLEs, and HLEs are influenced by different factors. A possible reason is that the current implementation levels of these three forms of CEs vary across QNPs, and their impact differs accordingly. Based on the three different situations and considering the results of this study, it is evident that the residents of QNP are not very satisfied with the FLEs that have been widely implemented and the ALEs that are being piloted in some areas. Residents who have participated in these two conservation policies, based on their prior participation experiences, have a low willingness to participate and low levels of support. ALEs are more closely related to livelihoods; when residents face a higher risk of unemployment, they attempt to participate in ALE reform to obtain certain subsidies that can alleviate difficulties in their livelihoods. Therefore, it is suggested that the department should engage in the current work to improve local satisfaction towards FLEs and ALEs, thereby increasing the willingness of HLEs to participate.

There are also differences in the support levels of residents living inside and outside QNP. Those residing within park boundaries show lower support for the FLEs and ALEs, indicating that the current FLE-based compensation mechanism has not yet achieved a satisfactory balance between conservation and local well-being. For proposed HLEs, families that are more dependent on local resources tend to be more supportive of this policy’s implementation, particularly when they perceive tangible welfare improvements from park management. The more they can feel the improvement in their welfare brought about by the construction and management of the national park, the more they support this policy. This suggests that while CE reforms should be differentiated according to conservation targets and land types, the foundational principle, validated by the FLE case, is to resolve livelihood constraints and enhance community welfare, as support under social exchange theory derives from perceived net benefits.

The results of this study show that in QNP, CEs, as a conservation policy, have preliminarily achieved ecological effects and some community support. Nevertheless, many issues require further discussion and analysis, e.g., why residents living within the park have lower support for ALEs and FLEs, and whether the implementation of HLEs will be negatively affected by problems associated with the other two types of CEs. Building on the FLE implementation precedent, it is necessary to establish differentiated CE schemes by designing CE systems that integrate refined conservation needs with the well-being of residents. This approach should more comprehensively consider community interests, enhance satisfaction with CE participation, and thereby promote the more effective adoption of CEs. Such schemes would enable the unified management of natural resources, achieving both conservation goals and community sustainability.

Furthermore, community participation must be prioritized, as communities are the collective landowners and key stakeholders in implementing CEs policies [45]. The FLE experience highlighted in this study confirms that enhanced community engagement increases willingness to participate in CEs and amplifies the perceived value of conservation. Practices observed in Sanjiangyuan National Park, China, demonstrate that involving local residents in roles such as rangers, through ecological monitoring, environmental education, and conservation labor, can improve community well-being and foster stronger policy support [47]. Thus, we advocate for multi-faceted community participation, from policy awareness to active involvement in conservation management and decision-making. Learning from the participatory pathways established under FLEs, this approach can strengthen policy comprehension and ownership, creating a positive feedback loop wherein participants become both beneficiaries and advocates of conservation.

In summary, building on the FLE precedent, effective CE expansion requires establishing differentiated schemes that are coherently aligned with local livelihood systems, institutionalize genuine participatory feedback, and ensure tangible co-benefits to reinforce the positive socio-ecological feedback loops essential for resilient and socially sustainable conservation.

4.5. Limitations and Future Research

This study has several limitations that should be acknowledged. First, while our sample size (N = 336) meets established guidelines for SEM [30,48] and post hoc fit indices indicate adequate power to detect meaningful effects, we did not conduct formal a priori power analysis. Future research, particularly studies planning complex multi-group comparisons or testing for small effect sizes, should incorporate a priori power calculations to ensure adequate sample sizes for detecting hypothesized effects. And this sample also limits our capacity to conduct multigroup analysis (MGA) across multiple subgroups simultaneously. MGA represents a more rigorous approach to examining structural heterogeneity, as it allows for a formal statistical comparison of path coefficients across theoretically defined groups while preserving the hierarchical support model’s theoretical framework [49]. Future research with expanded sampling should prioritize MGA to test, for example, whether the structural pathways differ between: (a) residents inside vs. outside the national park; (b) those with vs. without prior FLE participation experience; and (c) groups with high vs. low well-being perceptions. Second, our cross-sectional design captures perceptions at a single time point, limiting causal inference. Although our model is theoretically grounded in the sequence of policy implementation, alternative causal pathways, such as trust leading to participation, are plausible. Longitudinal research is needed to disentangle causal directions and examine how support evolves as residents gain experience with different easement types. Third, although our data were collected across multiple villages, our analysis did not account for potential clustering at the village level. Residents within the same village may share unobserved characteristics due to shared history, social networks, or local institutions, leading to correlated errors and underestimated standard errors if not addressed. Future research with larger samples should employ multilevel modeling or cluster-robust standard errors to account for this potential non-independence of observations. Fourth, our model does not include measures of latent environmental values or general conservation attitudes. The omission of these variables could introduce omitted-variable bias, as environmental values may influence both how residents perceive their wellbeing and participation, and their support for conservation policies. We initially attempted to measure general conservation attitudes, but the items failed to meet acceptable psychometric standards (Cronbach’s α < 0.60; factor loadings < 0.50) and were therefore excluded to maintain measurement model integrity. Future research should employ alternative, context-validated measures of conservation attitudes or consider behavioral proxies for environmental values. Fifth, regarding sample representativeness, while our sample’s demographic profile (predominantly older males) accurately reflects the composition of resident populations in this rural region, where younger adults and working-age females often migrate to urban areas for employment, it may not capture the perspectives of absent household members. These individuals may maintain connections to the community and could influence household decisions about participating in conservation policy. Future research should, where possible, include the perspectives of migrant community members. Sixth, the exploratory regression analyses used to examine CE-type differences are subject to well-documented methodological limitations, including sample dependence and model instability [35,36]. These results should be viewed as hypothesis-generating rather than confirmatory. Future research should employ more robust methods, particularly multi-group SEM with expanded samples, to test whether the structural pathways differ across easement types and population subgroups.

Despite these limitations, the core SEM findings, particularly the mediating role of institutional trust and the moderating effects revealed by multi-group analysis, provide a robust foundation for understanding how community support for conservation policies develops. The transparent acknowledgment of these limitations only strengthens the credibility of these core contributions and provides clear directions for future inquiry.

5. Conclusions

This study analyzed the determinants of community support for diversified conservation easements in a Chinese national park, with the widely implemented Forest Land Easement (FLE) serving as the foundational case and primary entry point. The findings demonstrate that support is not a linear outcome of benefit provision, but a product of interrelated socio-ecological pathways operating within a hierarchical structure. The structural equation model identified support for park management (SM) as the critical predictor and full mediator of CE acceptance. Human well-being and community participation, while not directly influencing CE support, exert important, indirect impacts through SM. This reveals a key pathway through which participatory engagement and improvements in human well-being build institutional trust, which in turn enables the acceptance of specific conversation policies. This feedback dynamic was first established and observed within the context of FLEs, providing a transferable model for extending conservation agreements to other land types.

Regression results further elucidated the heterogeneity of local socio-ecological subsystems. Support for forest, agricultural, and homestead easements was associated with distinct configurations of factors, though livelihood benefits and prior program experiences consistently functioned as cross-cutting drivers. Notably, prior experience with FLEs emerged as a particularly influential factor across easement types, underscoring its role as a reference point for community evaluation of new conservation instruments. The generally moderate levels of support for existing easements, particularly their linkage to past participatory experiences, suggest that historical gaps in process equity and benefit perception may constrain current policy legitimacy. These perceptions have been notably shaped by residents’ direct encounters with FLE implementation, informing their attitudes toward subsequent easement programs. These results carry implications for the design of area-based conservation instruments in line with global frameworks such as the post-2020 Global Biodiversity Framework. They indicate that effective CE implementation requires context-sensitive and adaptive governance approaches that: (1) align easement design with localized livelihood systems and land-use traditions; (2) institutionalize meaningful participatory mechanisms that enhance procedural fairness and agency; and (3) ensure that conservation interventions deliver tangible, equitable co-benefits to reinforce positive socio-ecological feedbacks.

In summary, this research provides a systematic, evidence-based perspective, grounded in the empirical case of forest easements, on how conservation policies can be embedded within complex human–environment interactions. It suggests that achieving socially sustainable and ecologically effective outcomes depends on governance structures that recognize, engage with, and adapt to the dynamic couplings between community well-being, institutional trust, and conservation practice, as exemplified and learned through the implementation pathway of Forest Land Easements.