Resident Empowerment and National Park Governance: A Case Study of Three-River-Source National Park, China

Abstract

The underlying tension between national park development and local community interests presents a significant challenge for contemporary ecological governance. Resident empowerment (RE) is increasingly recognized as a crucial pathway to mitigate this tension and achieve effective national park governance (NPG). However, the intrinsic mechanisms through which RE influences NPG have not been thoroughly explored in existing research. Drawing on the practice of government–resident interaction in China’s national parks, this paper investigates how the decentralization of power can balance the dual goals of environmental protection and social development. Using Three-River-Source National Park as a case study, we employ an ordered Logit regression model to examine the impact of RE on NPG. The study finds that RE is significantly and positively associated with NPG. Its influence is primarily mediated through three mechanisms: an identity effect (enhancing community belonging), an income effect (improving livelihood capabilities), and an environmental effect (strengthening participation in and perception of ecological conservation). Based on this empirical analysis, we recommend policies that further expand residents’ decision-making and management rights and broaden participation channels, thereby promoting the sustainable development and social equity of NPG.

1. Introduction

A central challenge in contemporary environmental governance is the inherent tension between state-led conservation imperatives and the socioeconomic well-being of local communities residing within or near protected areas. Globally, the establishment of national parks and other conservation zones, often implemented through a top-down approach, can restrict residents’ access to traditional livelihoods [1] and exclude them from [2] decision-making processes [3,4]. This exclusion not only raises issues of social equity but can also provoke local resistance [5], ultimately undermining the long-term effectiveness and sustainability of conservation policies [6]. Consequently, finding a governance model that equitably balances ecological protection with community development has become a critical theoretical and practical problem for achieving harmonious human–nature coexistence [7].

In response to this challenge, resident empowerment (RE) has emerged as a key paradigm, widely advocated as a means to mitigate park–community conflicts and enhance governance effectiveness [8,9,10]. However, despite its conceptual appeal, much of the existing research has focused on the normative importance of participation, community management, and social equity [1,11,12,13,14,15,16,17,18,19], without delving deeply into the causal mechanisms at play. A significant gap persists in the empirical literature regarding how empowerment translates into improved governance outcomes; the process is often treated as a “black box”. This lack of mechanistic understanding is particularly pressing as nations worldwide, including China, accelerate the expansion of their national park systems under ambitious ecological agendas. To ensure these large-scale policies are both sustainable and just, there is an urgent need to move beyond advocacy and build an evidence-based understanding of the pathways through which empowerment functions.

The rapid development of China’s national park system provides a particularly salient context for investigating this issue. Characterized by a strong, state-led administrative tradition, China’s approach to conservation has historically been top-down [3,4,20], creating a classic scenario of potential conflict between state objectives and local interests [12]. Yet, within this framework, China is also pioneering unique, government-led empowerment initiatives designed to actively involve residents in conservation efforts. The Three-River-Source National Park serves as a highly representative case study. As China’s largest national park, it has systematically implemented a flagship “one household, one ranger” policy since 2016, which directly integrates thousands of local pastoralist households into the park’s management and monitoring structure. This large-scale, state-driven empowerment program offers a unique natural experiment to empirically scrutinize the effectiveness and underlying mechanisms of a government-led empowerment model.

Therefore, this study aims to fill the identified research gap by “unpacking the black box” of RE within the context of China’s national park governance. Adopting a government–resident interaction perspective, we utilize survey data from 815 households collected in 2023 within the Three-River-Source National Park. By employing an ordered Logit regression model, we first quantitatively assess the overall impact of RE (measured by participation in the ecological ranger policy) on the effectiveness of national park governance (measured by resident satisfaction). We then further investigate the mediating pathways of this relationship, proposing and testing for an identity effect, an income effect, and an environmental effect. Through this empirical analysis, this paper seeks to provide robust evidence on how state-led empowerment can bridge the gap between conservation and development, offering valuable insights for policy optimization in China and other nations facing similar governance challenges.

2. Theoretical Framework and Hypotheses

2.1. The Paradigm Shift Toward Participatory Governance in Protected Areas

Governance is fundamentally a continuous process of negotiating interests among stakeholders [21]. Historically, state-led, top-down approaches to establishing and managing protected areas have demonstrated significant limitations [22]. In such models, local residents are often excluded from the early stages of policy design [23], leading to deficits in both procedural and distributive justice [24].

Beginning in the 1970s, a global paradigm shift occurred, moving toward the recognition of participatory and inclusive approaches as essential for effective conservation [25]. It became increasingly apparent that the success of protected areas is intrinsically linked to the ecological environment and the well-being of local stakeholders [26]. Conservationists and policymakers acknowledged that without the active participation and support of local communities and indigenous peoples, the long-term viability of protected areas is constrained [27]. Research consistently indicates that minimizing harm to the interests of local populations is a critical factor for successful conservation outcomes [28], and ensuring residents can exercise guiding and supervisory roles is crucial for resolving stakeholder conflicts [29]. The adoption of more participatory models, for instance, in Poland’s national parks [30]. Participatory governance in Poland’s national parks typically involves the establishment of ‘Scientific Councils’ or ‘Park Councils,’ which include representatives from local governments, non-governmental organizations, scientists, and local community members. Their functions are primarily advisory, providing recommendations on issues such as park management planning, tourism development, and resource utilization. Although these models aim to enhance participation, in practice, they tend to focus more on formal ‘consultation’ rather than deep ‘co-management’ or decentralization of authority. Nevertheless, even such formal communication and consultation mechanisms offer stakeholders a platform to voice concerns, effectively mitigating the acute conflicts that previously arose from purely top-down decision-making, and has proven effective in breaking such deadlocks [30,31,32,33,34].

2.2. Resident Empowerment as a Core Governance Mechanism

Within this participatory framework, resident empowerment (RE) has emerged as a central concept. Empowerment signifies not only an individual’s power to make decisions concerning their own life but also the expansion of democratic participation and access to resources within the community [35,36]. In the context of park governance, RE refers to the process by which governments devolve partial decision-making power and resource management authority to communities and residents [37]. This process, initiated through government–resident interaction and realized through resident participation [37], effectively connects top-down policy execution with bottom-up implementation [38]. It fosters a governance structure characterized by a “tripartite voice” of the government, the community, and its residents [39].

The importance of resident empowerment (RE) lies in its ability to create a more effective and equitable governance landscape. First, it enhances the transparency of decision-making by granting residents rights to voice, participation, and supervision. This allows them to engage directly in the design (e.g., through opinion solicitation), execution (e.g., through roles as ecological rangers), and monitoring of governance rules. Second, RE lowers the threshold for meaningful negotiation and feedback. By leveraging existing community structures and social networks, it creates accessible platforms for residents to express their concerns and provide frontline feedback, thereby enhancing the adaptability and local appropriateness of policy solutions.

2.3. The Mediating Effects of Empowerment on National Park Governance

While the benefits of RE are widely acknowledged, the precise mechanisms through which it improves national park governance (NPG) require deeper empirical investigation. Drawing from established frameworks such as the Theory of Planned Behavior, this study posits that RE influences governance effectiveness through three primary mediating pathways: a recognition effect, an income effect, and an environmental effect. These pathways represent the changes in residents’ attitudes, livelihood conditions, and environmental behaviors that result from empowerment. The conceptual model is illustrated in Figure 1.

RE can open up channels for the upward flow of information and grassroots needs, thereby strengthening residents’ policy understanding and support [8]. Within the “familiar communities” of national parks, direct communication between residents and the government can reduce information asymmetry, mitigate conflicts, and provide policymakers with valuable frontline feedback. This participatory process fosters a sense of ownership and self-identification with the policy, leading to greater approval and support among residents and their social networks [40]. This enhanced policy recognition functions as a positive attitude that promotes effective governance.

RE can create a stronger alignment between residents’ needs and economic opportunities. By devolving governance functions and resources to the community level, empowerment can help residents maximize resource utilization and improve their household income. Furthermore, RE can provide access to new information platforms and employment opportunities, such as ecological ranger positions, which diversifies income sources. This improvement in livelihood security [41] and overall well-being enhances residents’ quality of life and is expected to positively influence their perception of NPG [42].

RE enhances residents’ capacity for environmental improvement. As park stakeholders, residents’ pro-environmental behaviors directly impact ecological protection [43]. Original communities and local governments share intertwined land-use goals deeply rooted in local ecology [44]. RE, through role/capacity/rights enhancement [21], fosters synergistic environmental governance. RE grants decision-making power, expands participation channels, and enables self-regulated resource management rules, thereby enhancing conservation efficiency [28,45]. Empowerment cultivates a sense of belonging, reinforcing their obligation to safeguard natural assets for future generations. Research confirms the human–environment symbiosis: closer personal connections to nature amplify the willingness to commit to environmental stewardship [46], inherently motivating responsible behaviors [47] and ecological improvement. RE activates residents’ intrinsic governance capacity—voluntary engagement proves more effective for sustainable environmental management.

2.4. Hypotheses

Based on the theoretical framework outlined above, this paper proposes the following hypotheses:

Hypothesis 1 (H1). 

Resident empowerment significantly enhances national park governance.

Hypothesis 2 (H2). 

Resident empowerment indirectly and positively affects NPG through a recognition effect (i.e., by enhancing residents’ policy understanding and support).

Hypothesis 3 (H3). 

Resident empowerment indirectly and positively affects NPG through an income effect (i.e., by improving residents’ household income).

Hypothesis 4 (H4). 

Resident empowerment indirectly and positively affects NPG through an environmental effect (i.e., by enhancing local ecological conditions).

3. Research Design and Methodology

3.1. Case Selection and Data Collection

This study selects the Three-River-Source National Park (TRSNP) as its case study. As the largest of China’s first national park pilot projects, it provides a highly representative setting for examining contemporary conservation governance. Within the park’s boundaries, traditional pastoralist knowledge (such as rotational grazing) synergizes with government environmental policies, creating a practical context of “community-based governance” and power devolution to residents. A key innovation in this park is the “one household, one ranger” ecological stewardship program, which represents a concrete implementation of government-led resident empowerment. This program mobilizes residents for active participation in NPG through a comprehensive framework of localized stewardship, hierarchical oversight, and grid-based management. The program, initiated in 2016, had already appointed 17,211 ecological rangers across 53 administrative villages by 2018, demonstrating its extensive reach and significance.

Specifically, this program achieves co-management through the following mechanisms: (1) job binding: at least one member per household is employed as an ecological ranger; (2) performance-based assessment: the ranger’s contract renewal is linked to conservation performance, such as a documented reduction in poaching incidents monitored by camera traps; and (3) social incentives: this includes public recognition through “Outstanding Ranger” awards and enhanced insurance coverage for patrol staff. We explain that this dual-track system of ‘economic security + social recognition’ has significantly reduced conservation-related conflicts.

The data for this study were collected through field surveys conducted in the TRSNP area from July to August 2023. This timing ensures the data’s relevance, aligning with the 2020–2025 revision cycle of the “Overall Plan for Establishing a National Park System”. A mixed sampling strategy, combining simple random and stratified random sampling, was employed to select representative farming and pastoral households at the county, township, and village levels. The surveys gathered detailed information on household livelihoods, demographic characteristics, policy awareness, satisfaction with park policies, and perceptions of the ecological environment. A total of 850 questionnaires were distributed, yielding 815 valid responses, which corresponds to a high response rate of 95.88%. The high level of participation was due to two main factors: (1) administrative coordination: the questionnaires were distributed through the ranger network, whose daily interactions with the community have built a foundation of trust; (2) vested interest: respondents understood that the survey results could inform future adjustments to policies.

3.2. Variable Measurement

The effectiveness of national park governance (NPG), the dependent variable in this study, is operationalized through resident satisfaction (1–5). As a key subjective evaluation, resident satisfaction directly reflects the overall impact of policies on their lives and serves as a robust proxy for the efficacy of governance [48]. The core explanatory variable, resident empowerment (RE), is measured as a binary variable indicating whether a household participates in the “one-ranger-one-household” policy (0/1). This operationalization creates a clear treatment group (participating households) and a control group (non-participating households) for causal analysis.

While acknowledging that resident empowerment (RE) is a multifaceted concept encompassing psychological, social, and political dimensions, this study operationalizes it through participation in the ‘one household, one ranger’ policy. This approach, while being a simplification, offers a distinct advantage for empirical analysis by providing a clear, identifiable treatment group (participating households) and a control group (non-participating households). This binary measure allows for a robust quasi-experimental examination of the policy’s impact, which is the central aim of this paper. Similarly, the effectiveness of national park governance (NPG) is proxied by resident satisfaction. As a key stakeholder group directly affected by park policies, residents’ subjective evaluation serves as a crucial and direct indicator of governance performance from a community perspective. Although this does not capture objective ecological metrics, it directly reflects the perceived success of the policy in balancing conservation with community well-being, which is central to the issue of park–people conflicts.

The three mediating pathways—the recognition, income, and environmental effects—are measured using composite indicators constructed from specific survey questions designed to capture residents’ perceptions of policy consultation, income changes, and environmental improvement. To isolate the effect of RE and account for confounding factors, the models also include a series of control variables, given the significant influence of respondents’ education level, income level, gender, and other factors on environmental perception [49,50], we elected to incorporate both household-level characteristics (e.g., annual per capita income, household size, labor force proportion, and social networks) and individual attributes of the respondent (e.g., gender, education level, and employment status) into the control variables. To address potential issues arising from different units and dimensions, some data were normalized before the analysis. A detailed description of all key variables is provided in

Table 1. Detailed description of key variables.

Variable		Indicators	Weight
Explanatory Variable	RE	Is your household a “one post, one household” policy family? (Yes = 1, No = 0)	1/1
Dependent Variable	NPG	How satisfied are you with the ecological projects in the national park? (1–5, with 1 being very dissatisfied and 5 being very satisfied)	1/1
Mediating Variables	Environmental Effect	Do you agree that the ecological environment has been improving in recent years? (1–5, with 1 being strongly disagree and 5 being strongly agree)	1/1
	Income Effect	How has your household’s income changed? (Significantly decreased = 1, Slightly decreased = 2, Hardly any change = 3, Slightly increased = 4, Significantly increased = 5)	1/3
		What is your household’s income and expenditure situation? (Expenses greater than income = 1, Basically balanced = 2, Expenses less than income = 3)	1/3
		How has your household’s source of income changed? (Lost = 1, Decreased = 2, Unchanged = 3, Increased = 4)	1/3
	Recognition Effect	How well do you understand ecological policies? (1–5, with 1 being not at all familiar and 5 being very familiar)	1/3
		Do you think ecological projects should continue to be implemented? (Yes = 1, No = 0)	1/3
		Do you feel that the implementation process of ecological protection has fully respected the wishes of the herdsmen? (Not at all = 0, Consulted = 1, Consulted and adopted = 2)	1/3

, their descriptive statistics are presented in

Table 2. Descriptive statistics.

Variable		Obs	Mean	Std. Dev.	Min	Max
Explanatory Variable	RE	815	0.582	0.494	0	1
Dependent Variable	NPG	815	3.088	1.154	1	5
Mediating Variables	Environmental Effect	815	4.080	1.163	1	5
	Income Effect	578	0.458	0.227	0	1
	Recognition Effect	694	0.661	0.186	0	1
Control Variables	Annual per capita income (in 10,000)	815	9.248	21.833	0	316.110
	Education Level	815	1.676	1.204	1	5
	Household Head	815	0.839	0.368	0	1
	Gender	815	0.701	0.458	0	1
	Total household population	804	3.632	1.819	1	13
	Labor force participation rate	797	0.529	0.305	0	1
	Fiscal subsidies (in 10,000)	815	1.813	7.648	0	155.980
	Fixed household income (in 10,000)	814	1.875	6.258	0	89
	Migrant work experience	815	0.223	0.417	0	1
	Social network extent	815	1.515	0.804	1	4

, and the frequency distribution for resident satisfaction in

Table 3. The frequency distribution for resident satisfaction.

Variable/Indicators	1	2	3	4	5
How satisfied are you with the ecological projects in the national park? (1–5)	14.23%	13.01%	29.20%	36.81%	6.75%
How satisfied are you with the ecological compensation policy? (1–5)	25.89%	25.89%	60.12%	1.84%	0.98%

.

3.3. Model Specification

Given that the dependent variable (resident satisfaction) is measured on an ordered categorical scale, this study employs an ordered Logit regression model for the empirical analysis. The analytical strategy proceeds in two stages. First, to test the overall impact of RE on NPG (Hypothesis 1), we specify the following baseline econometric model:

$$NPG={\alpha }_0+{\alpha }_{1}RE+{\alpha }_{2}Z+\mu$$

(1)

Second, to test the mediating roles of the recognition, income, and environmental effects (Hypotheses 2, 3, and 4), we construct a mediation model based on the causal steps approach, specified as follows:

$$\mathrm{M}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{R}\mathrm{E}+{\mathsf{\beta }}_2\mathrm{Z}+\mathsf{\epsilon }$$

(2)

$$\mathrm{N}\mathrm{P}\mathrm{G}={\mathsf{\gamma }}_0+{\mathsf{\gamma }}_1\mathrm{R}\mathrm{E}+{\mathsf{\gamma }}_2\mathrm{M}+{\mathsf{\gamma }}_3\mathrm{Z}+\mathsf{\delta }$$

(3)

In these equations, NPG represents national park governance, RE is resident empowerment, M denotes the mediating variables (the three effects), and Z is a vector of control variables. The Greek letters (α, β, γ) are the coefficients to be estimated, while μ, ϵ, and δ are the random disturbance terms.

4. Results

4.1. Baseline Regression Results

Table 4. Overall regression results.

Variable	(1) NPG	(2) NPG
RE	0.544 *** (4.22)	0.549 *** (4.17)
Annual per capita income (in 10,000)		−0.0065 (−1.02)
Education Level		1.145 * (2.52)
Whether the household head		0.0899 (0.46)
Gender		0.0958 (0.65)
Total household population		0.0409 (1.03)
Labor force participation rate		0.0541 (0.23)
Fiscal subsidies (in 10,000)		0.0026 (0.17)
Fixed household income (in 10,000)		−0.0028 (−0.15)
Migrant work experience		0.0471 (0.30)
Social network extent		0.0847 (0. 86)
Constant/cut4	2.973 *** (18.18)	3.682 *** (9.95)
N	815	797

T statistics in parentheses, * p < 0.1, *** p < 0.01.

presents the results of the ordered Logit regression model examining the impact of resident empowerment (RE, is measured as a binary variable “participation/non-participation in the ‘one household, one ranger’ policy”) on national park governance (NPG, is proxied by a subjective resident satisfaction scale “1–5”). The model in Column (1), which excludes control variables, shows that RE has a positive and statistically significant effect on NPG at the 1% level. After introducing a comprehensive set of household and individual control variables in Column (2), the coefficient for RE remains positive and highly significant. These results provide strong initial evidence that resident empowerment (RE) is significantly and positively associated with the effectiveness of national park governance (NPG), thereby supporting Hypothesis 1. Moreover, the significant effect of respondents’ education level (β = 1.145, p < 0.1) may reflect a positive correlation between environmental awareness and participation attitudes, which is generally consistent with the findings of Jasmina et al. [51] and Pokharel et al. [52].

4.2. Robustness Test

To ensure the validity of our baseline findings, we conducted a series of robustness tests, the results of which are summarized in

Table 5. Robustness test.

Variable	(1) Replace	(2) Tail Reduction	(3) Tobit	(4) PSM
RE	0.655 *** (4.92)	0.560 *** (4.19)	0.305 *** (3.73)	0.561 *** (3.67)
Control Variables	YES	YES	YES	YES
Constant/cut4	0.803 * (2.36)		2.484 *** (11.95)	3.702 *** (9.01)
N	797	797	797	661

T statistics in parentheses, * p < 0.1, *** p < 0.01.

. First, we replaced the dependent variable (satisfaction with ecological projects) with an alternative measure (satisfaction with ecological compensation policy), and the positive and significant effect of RE persisted (Column 1). Second, to mitigate the potential influence of outliers, we winsorized the core variables at the 10% level; the result remained consistent (Column 2). Third, we re-specified the model using a Tobit regression, which again yielded results consistent in sign and significance with the original Ologit model (Column 3).

Finally, to address potential endogeneity arising from self-selection bias (i.e., non-random participation in the ranger policy), we employed Propensity Score Matching (PSM). As shown by the balance test in

Table 6. Balance test.

Variable	Unmatched (U)/ Matche (M)	Mean		Standard Deviation (%)	T-Test
		Treatment Group	Control Group		t	p > |t|
Annual per capita income (in 10,000)	U	9.575	8.999	2.6	0.360	0.716
	M	8.440	8.011	1.9	0.460	0.643
Education Level	U	1.771	1.554	18.1	2.500	0.013
	M	1.772	1.691	6.7	0.960	0.335
Whether the household head	U	0.838	0.841	−0.9	−0.130	0.900
	M	0.837	0.837	0.0	−0.000	1.000
Gender	U	0.685	0.722	−8.1	−1.120	0.263
	M	0.683	0.643	8.6	1.260	0.210
Total household population	U	3.689	3.584	5.8	0.800	0.421
	M	3.689	3.752	−3.5	−0.510	0.607
Labor force participation rate	U	0.534	0.522	3.6	0.510	0.612
	M	0.536	0.510	8.5	1.280	0.200
Fiscal subsidies (in 10,000)	U	1.346	2.032	−11.8	−1.750	0.080
	M	1.283	1.184	1.7	0.640	0.522
Fixed household income (in 10,000)	U	1.970	1.775	3.1	0.430	0.667
	M	1.777	1.288	7.7	2.060	0.040
Migrant work experience	U	0.229	0.210	4.7	0.650	0.516
	M	0.228	0.246	−4.2	−0.620	0.536
Social network extent	U	1.516	1.518	−0.2	−0.030	0.976
	M	1.520	1.474	5.7	0.870	0.382

and the kernel density plots in Figure 2, the matching process successfully balanced the covariates between the treatment and control groups, ensuring their comparability. The regression analysis on the matched sample (Table 4, Column 4) provides strong evidence to support the hypothesis that the positive impact of RE on NPG remains statistically significant at the 1% level. Collectively, these tests demonstrate that our primary finding is robust across different model specifications and analytical approaches.

4.3. Mechanism Test

Having established a robust direct effect, we next investigate the mediating pathways through which RE influences NPG, corresponding to Hypotheses 2, 3, and 4. We test for the recognition, income, and environmental effects using a mediation analysis framework, with results from the bootstrap tests presented in

Table 7. Bootstrap Test.

Variable	(1) Recognition Effect	(2) Income Effect	(3) Environmental Effect
Indirect Effect	0.112 *** (3.79)	0.220 ** (2.88)	0.0335 * (1.97)
Direct Effect	0.215 * (2.39)	0.106 (1.46)	0.272 *** (3.36)
Total Effect	0.327 *** (3.54)	0.325 ** (3.12)	0.305 *** (3.74)
N	680	565	797

T statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

.

The test for the recognition effect is detailed in

Table 8. Recognition effect test.

Variable	(1) Satisfaction	(2) Recognition Effect	(3) Satisfaction
RE	0.549 *** (4.17)	0.0624 *** (4.39)	0.419 ** (2.86)
Recognition effect			2.958 *** (7.32)
Control Variables	YES	YES	YES
Constant/cut4	3.682 *** (9.95)	0.623 *** (56.23)	5.861 *** (11.86)
N	797	694	680

T statistics in parentheses, ** p < 0.05, *** p < 0.01.

. The results indicate that RE is significantly linked to an increase in residents’ policy recognition (Column 2). When both RE and policy recognition are included in the model predicting NPG (Column 3), the coefficient of RE remains significant but is reduced in magnitude. The bootstrap test (Table 7, Column 1) confirms a significant indirect effect, as well as a significant direct effect. This pattern indicates a partial mediation, confirming that RE is associated with improved NPG, a relationship that is partially mediated by increasing residents’ understanding of and support for conservation policies. Thus, Hypothesis 2 is supported.

Next, we examine the income effect, with results presented in

Table 9. Income effect test.

Variable	(1) Satisfaction	(2) Income Effect	(3) Satisfaction
RE	0.549 *** (4.17)	0.0579 ** (3.01)	0.310 (1.81)
Income Effect			8.744 *** (16.90)
Control Variables	YES	YES	YES
Constant/cut4	3.682 *** (9.95)	0.422 *** (27.80)	8.003 *** (14.41)
N	797	578	565

T statistics in parentheses, ** p < 0.05, *** p < 0.01.

. The analysis shows that RE has a significant positive impact on residents’ household income (Column 2). In the full mediation model (Column 3), the income effect is highly significant, while the direct effect of RE on NPG becomes statistically insignificant. This finding is corroborated by the bootstrap test (Table 7, Column 2), which reveals a significant indirect effect but an insignificant direct effect. This pattern signifies a dominant pathway, providing strong evidence that RE enhances NPG primarily by improving the economic well-being of residents. Therefore, Hypothesis 3 is supported.

Finally, the results for the environmental effect are presented in

Table 10. Environmental effect test.

Variable	(1) Satisfaction	(2) Environmental Effect	(3) Satisfaction
RE	0.549 *** (4.17)	0.0379 ** (2.85)	0.489 *** (3.69)
Environmental Effect			0.305 *** (5.26)
Control Variables	YES	YES	YES
cut4	3.682 *** (9.95)	0.136 (1.32)	4.891 *** (11.15)
N	797	815	797

T statistics in parentheses, ** p < 0.05, *** p < 0.01.

. RE is found to have a significant positive impact on residents’ perception of ecological improvement (Column 2). When both RE and the environmental effect are included in the model (Column 3), the coefficient for RE remains significant, though its magnitude is slightly reduced. The bootstrap test (Table 8, Column 3) provides strong evidence to support that both the direct and indirect effects are statistically significant. This indicates another case of partial mediation where RE enhances NPG both directly and indirectly by fostering tangible improvements in the local environment. Hypothesis 4 is therefore also supported.

Furthermore,

Table 11. Summary of mechanism test.

Variable	(1) Recognition Effect	(2) Income Effect	(3) Environmental Effect
Result	Partial mediation	Dominant pathway	Partial mediation

provides a detailed presentation of the mediating effect findings. From Table 7 and Table 11, the results indicate that the income effect serves as the dominant mediating pathway, suggesting that economic incentives are a core driver of residents’ support for park management. This contrasts with the partial mediation of recognition and environmental effects, revealing that non-economic factors also play a significant, albeit supplementary, role and require long-term cultivation to translate into governance acceptance.

5. Discussion

This study provides robust empirical evidence that a government-led RE model can significantly enhance the effectiveness of NPG using the Three-River-Source National Park in China as a case study. Our findings move beyond a simple affirmation of participation, delving into the specific mechanisms that drive this positive relationship. The analysis reveals that the income effect acts as a dominant pathway, a particularly crucial finding which suggests that for empowerment to be successful in this context, it must deliver tangible economic benefits that improve local livelihoods. Simultaneously, the partial mediation of the recognition and environmental effects indicates that while economic improvements are paramount, fostering a sense of ownership, policy support, and tangible ecological improvement are also vital components of effective governance. These results contribute a nuanced, quantitative perspective to the broader literature on community-based conservation, demonstrating the efficacy of a state-driven empowerment paradigm.

The theoretical and practical implications of these findings are significant. Theoretically, this study advances our understanding of empowerment by empirically validating a multi-dimensional causal model (encompassing the recognition, income, and environmental pathways) within a non-Western context. It contributes the concept of “government-led empowerment” as a distinct paradigm, contrasting with models that prioritize community autonomy, and shows how state intervention can proactively foster, rather than hinder, community engagement. Practically, the “one household, one ranger” policy serves as a compelling model for other regions seeking to reconcile the often competing goals of conservation and development. It demonstrates that state-led programs, if designed to provide direct and substantial benefits to residents, can effectively transform local communities from passive subjects into active stewards of their environment.

However, the findings of this study should be interpreted in light of several limitations. First, and most importantly, our operationalization of the core constructs, while pragmatic, represents a simplification. ‘Resident Empowerment (RE)’ was measured as a binary variable based on policy participation, which does not capture the quality, degree, or subjective experience of empowerment felt by individuals. Future research could employ multi-item scales to explore these nuances. Similarly, proxying NPG with resident satisfaction, though valuable, provides only a subjective community perspective. It does not account for objective ecological outcomes (e.g., biodiversity metrics) or institutional performance (e.g., conflict resolution efficiency), which are essential for a holistic assessment of governance effectiveness. This paper acknowledges that simplified operations may overlook the substantive depth of empowerment effects, focusing more on formal governance outcomes, potentially leading to an overestimation of intervention effectiveness. Second, the cross-sectional nature of our data captures a single point in time, which limits our ability to draw definitive causal inferences or analyze the dynamic evolution of the relationship between empowerment and governance. Additionally, while the Propensity Score Matching (PSM) method effectively addresses self-selection bias based on observable characteristics, it cannot account for potential endogeneity arising from unobservable factors. For instance, households with a higher intrinsic environmental consciousness or greater pre-existing trust in government might be more likely to both participate in the ranger program and report higher satisfaction. Such unobserved heterogeneity could create a confounding effect. While this study’s robustness checks provide confidence in the findings, future research employing panel data with methods like Difference-in-Differences (DID) or sourcing appropriate instrumental variables (IV) would be invaluable for further strengthening the causal claims. Finally, the generalizability of our conclusions may be constrained by the unique context of the TRSNP, with its specific ecological fragility, the cultural attributes of Tibetan pastoral communities, and the overarching Chinese administrative system. Specifically, the finding that the income effect acts as a dominant pathway may be highly context dependent. This mechanism is likely prominent in settings like the TRSNP, where empowerment is delivered via direct, government-funded salaries to pastoralist communities. In national parks where community benefits are derived from less direct sources, such as tourism revenue sharing or entrepreneurial opportunities, the mediating roles of recognition, social capital, or environmental values might become more pronounced. Therefore, caution is warranted when applying this specific mechanistic finding to different socioeconomic and governance contexts.

These limitations highlight several promising avenues for future research. Longitudinal studies are needed to track the long-term impacts of empowerment policies on residents’ livelihood strategies, social capital, and pro-environmental behaviors over time. Future work could also employ mixed-methods approaches, combining quantitative surveys with in-depth qualitative interviews and ethnographic observation to further “open the black box” of the mediating pathways and to understand the nuanced lived experiences of empowerment. In subsequent research, it would be valuable to further distinguish between formal empowerment (e.g., task delegation) and substantive empowerment (e.g., participation in decision-making), as well as between instrumental satisfaction (economic rewards) and value-based satisfaction (governance recognition). Furthermore, comparative case studies contrasting government-led, co-managed, and community-led empowerment models across different national and cultural contexts would be invaluable for identifying more universal principles of effective conservation governance. Finally, exploring the role of emerging technologies, such as digital platforms for enhancing participatory monitoring or blockchain for increasing the transparency of ecological compensation, could offer new insights into fostering more inclusive and sustainable governance models.

Although the “one household, one ranger” model in TRSNP reflects Chinese characteristics, its underlying mechanism—prioritizing economic incentives while fostering cultural identity—can offer valuable insights for developing countries. Future research should incorporate cross-park panel data and objective ecological indicators (e.g., NDVI index) to enhance generalizability.

6. Conclusions and Recommendations

In the context of growing tensions between ecological protection and community development in China’s national parks, this study investigated the role of resident empowerment (RE). Using 2023 survey data from the Three-River-Source National Park and an ordered Logit model, we demonstrate that a government-led empowerment model is significantly and positively associated with national park governance (NPG). This conclusion is validated by a series of robustness tests. Our mechanism analysis reveals that this positive effect is driven by powerful income, recognition, and environmental effects, underscoring that achieving a win-win outcome for conservation and development hinges on delivering tangible economic benefits while fostering social equity and a sense of shared purpose.

Based on these findings, and acknowledging that current empowerment practices still face challenges related to the depth and quality of participation, we offer several policy recommendations. First, the scope and pathways of empowerment should be expanded by granting residents greater authority in decision-making and management. (1) based on the finding that income is a dominant pathway, we now propose moving beyond a uniform subsidy. We recommend establishing differentiated ecological compensation standards, de-linking the patrol subsidy from household income and re-linking it to performance-based floating rewards (e.g., a community conservation effectiveness fund). (2) To build upon the partial mediation effect of identity, we suggest concrete measures such as awarding an official “Ecological Steward” certification, which provides participants with ceremonial honor and greater rights to participate in decision-making processes, thereby strengthening their sense of ownership. Innovating participatory models, such as establishing formal community co-management committees, can ensure that empowerment translates into deep-level influence rather than superficial consultation. Second, initiatives must be supported by robust capacity-building programs, including specialized skills training (e.g., in conservation techniques or tourism management) and the development of sustainable livelihood support systems, such as targeted ecological compensation (conflict buffer mechanism) and franchise-based business opportunities. Finally, these efforts require strong institutional and financial safeguards, including the refinement of fiscal transfer policies to ensure precise and sustainable funding, as well as the exploration of diversified financing mechanisms, such as carbon credit trading or ecological banking, to alleviate fiscal constraints and secure the long-term success of national park governance (NPG).