Sustainability and Risks of Rural Household Livelihoods in Ethnic Tourist Villages: Evidence from China

Abstract

This study constructed a quantitative index system evaluating the level of livelihood sustainability of rural households in ethnic tourist villages and identified the main livelihood risks these households encountered. This was achieved by modifying the sustainable livelihood framework proposed by the Department for International Development (DFID) and testing it on three typical ethnic tourist villages in Guangxi, China (Jinkeng, Ping’an Zhuang, and Chengyang Eight). The results showed that the overall livelihood sustainability index was relatively poor. Livelihood assets and transforming structures and processes were the main factors that restricted the sustainability of household livelihoods. Social capital and financial capital values were lowest out of six types of livelihood capital examined. The lack of community participation of households adversely impacted the score of transforming structures and processes. Households encountered various livelihood risks. Generally, the livelihood issues that concerned most households were market, social, and education risks. However, when the farmers were enquired about a single risk that worried them the most, market, financial, and employment risks ranked the highest. Through quantifying the sustainability and risk of rural household livelihoods, guidelines and subsidies could be allocated to promote rural revitalization.

1. Introduction

Since 2020, China’s rural development has entered a new period of consolidating the achievements of poverty alleviation in the preceding years, effectively connecting poverty alleviation with rural revitalization [1]. To promote the effective connection between poverty alleviation and rural revitalization, the livelihoods of all rural residents must be improved [2]. Therefore, it is essential to enhance the sustainability of rural household livelihoods, which is key to rural revitalization [3].

Tourism in China’s ethnic minority areas has thrived in recent years. In 2020, China’s eight provinces and autonomous regions (Inner Mongolia Autonomous Region, Ningxia Hui Autonomous Region, Xinjiang Uygur Autonomous Region, Tibet Autonomous Region, Guangxi Zhuang Autonomous Region Qinghai, Yunnan, Guizhou, Yunnan. The data come from the official websites of the provincial statistical bureaus), mainly inhabited by ethnic minorities, received 2.194 billion tourists, generating a total tourism revenue of CNY 2378.978 billion. Many ethnic regions regard the tourism industry as a leading industry that allows farmers to overcome poverty. These tourism development activities have significantly impacted the lives of local residents, promoting changes and evolution in their livelihoods [4]. However, the effect of tourism development on the livelihood of households in ethnic areas is uncertain [5]. For example, tourism development might create diversified employment opportunities, drive the diversification of livelihoods, reduce livelihood risks, and increase the living standard of local farmers [3,6,7,8]. However, it might also widen the gap between the rich and poor [9] and reduce social capital [10], which would restrict the sustainable development of household livelihoods. Therefore, it is necessary to evaluate the sustainable livelihoods of rural households in typical ethnic tourism villages to implement appropriate actions, considering the key position of ethnic minority areas on the poverty alleviation programs of China, uncertainty on the impact of tourism development on the livelihoods of farmers, and the significance of the livelihood sustainability of farmers to rural revitalization.

A review of the literature found that studies exploring tourism and the sustainability of household livelihoods remain limited [11]. Existing studies on the sustainable livelihoods of farmers in tourist destinations have only systematically analyzed changes to farmer livelihoods after the development of rural tourism [6,12,13] or evaluated the livelihood capital of farmers in rural tourism destinations [9,14,15]. In comparison, quantitative assessments of the sustainability levels of farmer livelihoods in rural tourism destinations are lacking. Furthermore, specific livelihood risks faced by rural households in tourist destinations are not known.

Herein, this study constructed a quantitative index system evaluating the level of livelihood sustainability of rural households in ethnic tourism villages by modifying the sustainable livelihood analysis framework proposed by the DFID in the United Kingdom and integrating the livelihood characteristics of ethnic tourism villages. This study also identified the main livelihood risks faced by rural households in ethnic tourism villages. Three villages (Jinkeng Village, Ping’an Zhuang Village, and Chengyang Eight Village) were selected as case studies. The livelihood sustainability level of local households was based on household survey data. Moreover, differences in livelihood sustainability level and livelihood risk among different types of households and households in different villages were compared. This study seeks to answer the following research questions: (1) How to construct a reasonable quantitative evaluation index system of households’ livelihood sustainability in tourism destinations? (2) What is the specific level of livelihood sustainability of households in ethnic tourism villages in China? (3) What are the livelihood risks faced by households in ethnic tourism villages?

This study has the following contributions: First, this study innovatively constructed a quantitative evaluation index system on the sustainability of rural household livelihoods in ethnic tourism villages based on the DFID sustainable livelihood framework. Second, limitations of the DFID sustainable livelihood framework in analyzing household livelihoods in rural tourism areas were overcome by modifying the DFID framework, enriching the theory of sustainable livelihood. Third, this study has identified the livelihood risks encountered by households in ethnic tourism villages. This compensates for the lack of attention on the livelihood risks in tourist areas. Finally, while enabling ethnic villages is key for alleviating poverty in China, the literature evaluating the livelihood sustainability of households in ethnic tourism villages is relatively lacking. This study focuses on evaluating the livelihood sustainability of households in ethnic tourism villages, enriching relevant research. The results of this study also provide baseline information on the factors impacting the sustainability of farmers in ethnic tourism villages, which could be used to construct effective guidelines and incentives to facilitate rural revitalization.

2. Theoretical Basis and Literature Review

2.1. Sustainable Livelihood Analysis Framework and Its Modification

The concept of sustainable livelihood originated from the deepening understanding of poverty of Sen [16] and Chambers and Conway [17]. As a paradigm shift in international development thinking, sustainable livelihood analysis methods have become widely used in rural development, poverty alleviation, and livelihood selection [18]. Moreover, multiple sustainable livelihood analysis frameworks have been developed [10]. Among these frameworks, the sustainable livelihood analysis framework proposed by the DFID in the UK is the most widely recognized framework that is used in livelihood research [19]. Thus, this framework was selected to examine the sustainability of the livelihoods of rural households in ethnic tourism villages in this study.

According to the DFID framework (Figure 1), the livelihood systems of farmers have five components: vulnerability context, livelihood assets, transforming structures and processes, livelihood strategies, and livelihood outcomes [19]. Of these, livelihood assets are at the core of sustainable livelihoods, including human, natural, physical, financial, and social capital [20,21]. The vulnerability context is the external environment that significantly impacts the sustainability of the livelihoods of rural households [22]. Transforming structures and processes are fundamental in shaping the capital and outcomes of livelihoods. Livelihood strategy is important for rural households to make a living and realize sustainable livelihoods [19]. Livelihood outcome is the achievements and objectives that the livelihood strategies of rural households target [14].

The DFID framework provides a standardized tool and systematic research concept for studying the livelihoods of rural households. However, limitations exist, especially in the context of rural tourism. First, it does not integrate the positive role and value of traditional cultural elements [23,24,25]. The DFID framework perceives culture as a rigid structure that hinders farmers from obtaining other livelihood development opportunities [23]. However, traditional or vernacular culture is an indispensable asset for rural communities to attain sustainable development, especially in ethnic areas that develop tourism [6,23,24]. Ethic cultures and traditional skills are crucial elements in increasing the tourism attractiveness of ethnic regions and enhancing the sustainability of local household livelihoods. Therefore, the value and role of culture should be fully reflected in the sustainable livelihood framework [3,25].

Furthermore, the existing framework does not sufficiently account for the community participation of rural households [26]. The involvement of local residents in decision-making processes and political governance could effectively protect their own interests [22,26,27]. For rural households in rural tourism areas, participating in local tourism development affairs and local governance could substantially impact their livelihood outcomes [22,26]. However, sustainable livelihood approaches focus more on policy and institutional levels, with no systematic focus on community development and community participation at the grassroots level [26].

Therefore, when designing an evaluation index system on the sustainability of rural household livelihoods in ethnic tourism villages here, the DFID sustainable livelihood framework was modified to overcome the identified flaws. Specifically, this study added “cultural capital” to “livelihood capital,” and added “rural household community participation level” to “transforming structures and processes” to improve the reliability of the evaluation results.

2.2. Literature Review

2.2.1. Study on the Sustainability of Farmers’ Livelihood in Tourism Destinations

Rural Tourism has long been considered to be an effective catalyst of rural development and revitalization [6]. Vaishar and Šťastná [28] defined rural tourism as tourism activities carried out in the countryside. An introduction of tourism to a rural community inevitably affects local livelihoods [3,4]. Support for tourism stems from the perceived benefits of tourism development, particularly increased employment opportunities and income growth [6,13]. However, the impacts of tourism development on the livelihoods of local farmers may not always be as positive as expected [6,29]. Therefore, the sustainability of rural households’ livelihood in rural tourism destinations have received growing attention recently [4,30]. When analyzing the livelihood sustainability of rural households in rural tourism destinations, scholars mainly carry out their research from two aspects: first, scholars systematically and qualitatively analyze the specific impacts of tourism development on farmers’ livelihoods. Lasso and Dahles [5] found that tourism development in the Komodo village of Indonesia led the local fishing community to give up their fishing and become fully dependent on selling souvenirs. Although the transformation of traditional livelihoods offers more returns for local people in the short term, it is too early to believe that tourism provides sustainable livelihood strategies for local community because of the potential threats to this new livelihood. Su et al. [6] and Su et al. [12] adopted a sustainable livelihood approach to analyze the impact of tea tourism on the livelihood sustainability of farmers in Hetu Town. The results showed that tourism improved the overall livelihood sustainability by enhancing farmers’ livelihood options and livelihood diversity. However, tourism also increased the within-community income gap, which may generate social risks in the long run. Pasanchay and Schott [13] examined both livelihood benefits and costs of operating a homestay based on the DFID sustainable livelihood framework. The findings showed that homestays contributed to reduce poverty and achieve a sustainable community. However, they also undermined the community’s crucial social and natural capitals.

Second, analyzing the sustainability of farmers’ livelihood by quantitatively evaluating the level of farmers’ livelihood assets. Shi and Li [31] measured the sustainable livelihood of farmers based on livelihood assets and systematically discussed the synergistic relationship between the multi-functional development of rural tourism and the sustainable livelihood of farmers. Ma et al. [8] evaluated the level of livelihood assets of farmers in Jiuzhaigou Nature Reserve and analyzed the specific impact of livelihood assets on farmers’ income. The results showed that tourism improved the level of livelihood assets of local farmers, and livelihood assets had a significant positive impact on farmers’ income. Su et al. [32] measured the livelihood capital of rural households in the beautiful south of Southern China and analyzed the specific strategies of different types of households responding to seasonal fluctuations in tourism. These studies have laid a solid foundation for the sustainable livelihood theory. However, the studies with the former research perspective cannot intuitively reflect the sustainability level of tourism livelihoods. The literature with the latter research perspective evaluated the livelihood sustainability only based on the level of livelihood capital, ignoring the other aspects of sustainable livelihood framework, which have some limitations. A few scholars have tried to build a more intuitive quantitative evaluation index system for farmers’ livelihood sustainability in tourist destinations. Li et al. [3] constructed a livelihood sustainability index for assessing the livelihood sustainability of rural households in tourism destinations. This index includes livelihood capital, livelihood strategy, and the interlinkage between livelihood and environment. Liu [33] designed an evaluation index system for the livelihood sustainability of rural households in rural tourism destinations based on the DFID sustainable livelihood analysis framework and the pressure-state-response model. This evaluation index system includes three dimensions of vulnerability context, livelihood capital, and livelihood strategies, and eight specific indicators. The above literature analysis shows that scholars have not paid enough attention to the intuitive quantitative evaluation of the livelihood sustainability level of farmers in tourist destinations currently. Although some scholars, such as Li et al. [3], Liu [33], etc., have made valuable attempts in constructing the quantitative evaluation index system of livelihood sustainability of rural households in tourist destinations, these evaluation index systems can be further revised and improved.

2.2.2. Research on Tourism in Ethnic Minority Areas

Ethnic tourism is a form of tourism in which the “quaint” customs of indigenous or exotic peoples as tourist attractions [34]. Previous studies have proven that the development of ethnic tourism will bring some positive effects to minority areas. For instance, Samarathunga [35] pointed out that ethnic tourism has a huge potential in contributing to the socio-economic development of local area. Kuhn et al. [36] found that ethnic tourism has effectively promoted the formation of cultural food knowledge. Sinclair-Maragh and Bernard Simpson [37] verified that the development of ethnic tourism can effectively promote the ethnic identity of residents. Wasudawan [38] certified that ethnic tourism can effectively alleviate poverty. These positive impacts demonstrate the value of ethnic tourism development. However, ethnic tourism also causes some negative impacts or risks, which is not conducive to the sustainable development of ethnic tourism. These negative impacts include disempowered local communities because of ethnic tourism and enabling power inequality [39]. Ethnic culture has been commercialized and loses its authenticity [40]. Elites with considerable power are even more likely to benefit from tourism activity [41]. Furthermore, ethnic tourism has destroyed the local culture and traditions [42]. However, less attention has been paid to the sustainable livelihood of farmers in ethnic tourism destinations. Although scholars such as Ramaano [43] and Shang and Wang [44] have analyzed the livelihood capital of farmers in ethnic tourism areas and the impact of tourism development on farmers’ livelihood, few studies quantitatively measure the sustainability level of farmers in ethnic tourism destinations or pay attention to the livelihood risk faced by farmers in ethnic tourism destinations. Therefore, research on sustainable livelihood of farmers in ethnic tourism areas needs to be enriched.

3. Study Area and Data Sources

3.1. Study Area

This study selected Jinkeng Village, Ping’an Zhuang Village, and Chengyang Eight Village in Guangxi Province, China as the study areas (Figure 2 and Figure 3). Chengyang Eight village is a typical traditional Dong village located on the banks of Linxi River in Linxi Town, Liuzhou City. This village is composed of eight Dong nationality stockades. The village belongs to national AAAA-level tourist scenic spots, “national characteristic tourism towns”, “villages with Chinese minority characteristics”, and the first batch of “Chinese landscape villages” [45]. Although this village includes eight stockades, only four core stockades (Ma’an, Ping, Yan, and Pingtan) are currently undergoing tourism development. These four stockades encompass 959 rural households with 3940 inhabitants (The data were provided by Chengyang Eight Village’ committee). This study focused on these four stockades involved in tourism development.

The villages of Jinkeng and Ping’an Zhuang belong to the Longji Terraced Fields Scenic Area. These two villages are located in Longji Town, Longsheng Autonomous County of Guilin City. Guilin is a famous tourism city in China. In 2021, the city received 122.4 million tourists, a year-by-year increase of 19.5%. The total tourism consumption reached CNY 150.3 billion, a year-by-year increase of 21.8% (The data were taken from the official website of Guilin Municipal Bureau of Culture, Radio, Television, and Tourism (http://wglj.guilin.gov.cn/xxgk/fdzdggnr/sjfb/202204/t20220412_2252657.html, accessed on 23 April 2022)). Jinkeng Village is inhabited by the Hongyao nationality. The village encompasses 292 rural households with more than 1380 inhabitants (The data were provided by Jinkeng Village’ committee). All of the villagers of Ping’an Zhuang Village belong to Zhuang nationality, with 219 households and 826 inhabitants (The data were provided by Ping’an Zhuang Village’ committee). Both villages form the core of the Longji Terraced Fields Scenic Area (a national AAAA-level tourist scenic spot). Ping’an Zhuang Village belongs to the first group of “villages with Chinese minority characteristics”, and Jinkeng village is rated as a “Chinese classic village landscape”. The development of tourism in these three ethnic villages is of a certain scale that significantly impacts the livelihoods of local households. Thus, all three villages are representative case studies.

3.2. Data Sources

To obtain the primary data, rural households were surveyed using structured questionnaires and semi-structured interviews. A pre-survey was conducted in the three villages from 1 to 9 June 2021. The questionnaire was then revised based on problems encountered in the interviews with local village committee members and households, as well as in the questionnaire distribution process. Experts in the field of tourism livelihood were also invited to help improve the questionnaire.

The formal survey was conducted in Chengyang Eight Village from 11–15 June 2021, and in Jinkeng and Ping’an Zhuang villages from 15–20 August 2021. The survey was carried out on a household basis. The average survey time for each household was 35–40 min. Respondents were mainly household heads or the adult labor force of households. Because the villages were ethnic, local people with higher education levels were hired as language translators to prevent language communication barriers. Following the way of Elahi et al. [46,47], the respondents could refuse to answer any question if it was sensitive to them and made them feel uncomfortable. A total of 500 questionnaires were distributed, and 466 valid questionnaires were returned, with an effective rate of 92.3%. Overall, there were 230, 124, and 112 valid questionnaires from Chengyang Eight Village, Jinkeng Village, and Ping’an Zhuang village, respectively. The basic situation of the surveyed farmers is shown in

Table 1. Situation of the investigated households.

Subject	Variable	Frequency	%	Subject	Variable	Frequency	%
Gender	Male	252	54.1		5–6	215	46.1
	Female	214	46.9		7–8	55	11.8
Education level	Primary school and below	143	30.7		9 and above	25	5.4
	Junior high school or technical secondary school	151	32.4	Age	18–35 years old	89	19.1
	High school or junior college	117	25.1		36–45 years old	233	50.0
	University and above	55	11.9		46–55 years old	114	24.5
Family population	1–2	5	1.1		56–65 years old	19	4.1
	3–4	166	35.6		66 years old or older	11	2.4

Data sources: the data were obtained through questionnaires.

.

The questionnaire was separated into three parts. First, basic information on rural households was obtained, including demographic characteristics, population structure, and income level. Second, the state of livelihoods in rural households was recorded, including household livelihood capital, livelihood strategies, livelihood outcomes, policies and institutions, and the vulnerability context. Third, the livelihood risks faced by local households were recorded. Information about livelihood risks was obtained by asking farmers: “What livelihood issues are you currently worried about? Of these livelihood issues, which one are you most worried about?” Based on the results of the pre-survey, information on market risks, financial risks, social risks, employment risks, natural risks, education risks, pension risks, health risks, policy risks, and marriage risks were incorporated into the questionnaire.

4. Materials and Methods

4.1. Rural Household Categories

In addition to the comparative analysis of the sustainable livelihoods of rural households in different villages, this study further classified households according to their livelihood strategies. Drawing on existing research [32,48,49] and considering the actual situation in the study villages, rural households were classified into five types based on the prominent livelihood activities of households and the ratio of income from these livelihood activities to total household income.

The five types of rural households were pure tourism households (income from tourism accounted for 100% total household income), tourism-oriented households (income from tourism accounted for >50% and <100% total household income), balanced households (neither tourism income nor income from other livelihood activities accounted for >50% total household income), laboring-oriented households (income from engaging in industrial or engineering work accounted for >50% total household income), and agricultural-oriented households (households with >50% total income from agricultural activities).

Pure tourism households selected catering, accommodation, supermarkets for tourism, and national handicraft stores as their livelihood activities. Income from tourism was the only source of household income. This type of rural household accounted for 9.2% of surveyed households (n = 43 households). There were 188 tourism-oriented households, accounting for 40.3% of surveyed households. Tourism-oriented households were mainly engaged in tourism activities, but also engaged in industrial or engineering work temporarily, or planted a small number of crops during the off-season. There were 104 balanced households, accounting for 22.3% of surveyed households. The livelihood combination of this type of household was mostly “tourism operation + year-round migrant work/rental income/employment of enterprises and institutions + agricultural planting” with diversified income sources and relatively balanced income distribution. The livelihoods of laboring-oriented households (21.2% of surveyed households, n = 99 households) were mainly local work or migrant work, combining small-scale agricultural planting. There were 32 agricultural-oriented households (6.9% of surveyed households). This type of household was mainly engaged in crop farming or animal husbandry, combining short-term migrant work.

4.2. Establishment of the Evaluation Index System of Livelihood Sustainability

Drawing on existing research on sustainable tourism livelihoods and considering the livelihood characteristics of rural households in ethnic tourism villages, an evaluation index system of livelihood sustainability was constructed by modifying the DFID sustainable livelihood framework. The index system included five first-level indicators; namely, vulnerability context, livelihood assets, livelihood strategies, transforming structures and processes, and livelihood outcomes. Under these five first-level indicators, 16 second-level indicators and 38 third-level indicators were selected (

Table 2. Evaluation index system of the livelihood sustainability of households in ethnic tourism villages.

First-Level Indicators	Second-Level Indicators	Third-Level Indicators	Value Assignment Standard	Reference
Vulnerability context (A1)	Natural environment vulnerability (B1)	Limited by natural environment (C1)	Extent to which household livelihoods are restricted by the natural environment: very high = 0, high = 0.25, general = 0.5, low = 0.75, never = 1	DfID [19], Zhao et al. [18], Shen et al. [22]
		Frequency of disasters and shocks (C2)	Frequency at which households are impacted by various disasters and shocks: very frequent = 0, frequent = 0.25, general = 0.5, sometimes = 0.75, never = 1
	Market environment vulnerability (B2)	Seasonal fluctuation (C3)	Degree of seasonal fluctuation in household livelihoods: very high = 0, high = 0.25, general = 0.5, low = 0.75, never = 1
		Market conditions (C4)	Degree of household livelihood activities facing market shrinkage and market fluctuation: very high = 0, high = 0.25, general = 0.5, low = 0.75, never = 1
Livelihood assets (A2)	Natural capital (B3)	Cultivated land area (C5)	Cultivated land area per capita	Li et al. [50], Su et al. [32]
		Cultivated land quality (C6)	Quality of cultivated land of households: very poor = 0, poor = 0.25, general = 0.5, relatively fertile = 0.75, very fertile = 1
	Human capital (B4)	Total household Labor (C7)	Number of family laborers (labor workers aged 18–65 who are not in school and are healthy)	Li et al. [3], Hua et al. [51], Ding et al. [52], Wang et al. [25]
		Education level (C8)	Education of household Members: uneducated = 0, primary school = 0.25, junior high school or technical secondary school = 0.50, high school or junior college = 0.75, university and above = 1
		Health status (C9)	Health status of household members: very bad = 0, bad = 0.25, general = 0.50, good = 0.75, very healthy = 1
	Physical capital (B5)	Fixed assets of Household (C10)	Number of fixed assets owned by households	Ma et al. [24], Huang et al. [4], Wang et al. [53]
		Housing quality (C11)	Types of housing of households: brick and tile house = 0.25, concrete house = 0.5, brick and wood = 0.75, wooden house = 1
		Length of housing residence (C12)	Resident time of household house: more than 35 years = 0, 26–35 years = 0.25, 16–25 years = 0.50, 6–15 years = 0.75, 5 years and less = 1
		Housing area (C13)	Housing area per capita
	Financial capital (B6)	Annual household Income (C14)	Annual income per capita of households	Wang et al. [53], Li et al. [3]
		Bank loans (C15)	Ease of loan from bank: very difficult = 0, relatively difficult = 0.25, general = 0.5, relatively easy = 0.75, very easy = 1
		Borrowing from relatives and friends (C16)	Ease of borrowing from relatives and friends: very difficult = 0, relatively difficult = 0.25, general = 0.5, relatively easy = 0.75, very easy = 1
	Social capital (B7)	Close relatives and friends (C17)	Number of relatives and friends who can help households when they encounter difficulties	Wang et al. [25], Li et al. [3], Huang et al. [4]
		Neighborhood relationships (C18)	Relationship between households and their neighbors or villagers: very bad = 0, relatively bad = 0.25, general = 0.5, relatively good = 0.75, very good = 1
		Village cadres and management staff (C19)	Relationship with village cadres or local management staff: very bad = 0, relatively bad = 0.25, general = 0.5, relatively good = 0.75, very good = 1
		Leadership potential (C20)	Number of relatives/friends who are village cadres or in government agencies
	Cultural capital (B8)	Cultural cognition (C21)	Level of understanding of national traditional culture: completely unknown = 0, do not understand = 0.25, general = 0.5, understand = 0.75, understand very well = 1	Li et al. [3], Ma et al. [24], Wang et al. [25],
		Willingness of cultural inheritance (C22)	Willingness of family members to inherit national traditional culture: very unwilling = 0, unwilling = 0.25, general = 0.5, willing = 0.75, very willing =1
		Mastery of manual skills (C23)	Level of mastery of family members with ethnic traditional handicrafts: no mastery = 0, low mastery = 0.25, general mastery = 0.5, high mastery = 0.75, complete mastery =1
Livelihood strategies (A3)	Livelihood diversification (B9)	Number of household livelihoods (C24)	Number of rural household livelihood strategies	Su et al. [6], Wu and Pearce [54], Guo and Yang [10]
	Livelihood freedom (B10)	Freedom of household livelihoods mode (C25)	Ability of households to freely choose their livelihoods and mobilize among different livelihood activities: very low = 0, low = 0.25, general= 0.5, high = 0.75, very high = 1
Transforming structures and processes (A4)	Level of community participation (B11)	Participate in industry development decision making (C26)	Whether household participates in the decision making of local industry development: no = 0, yes = 1	Shen et al. [22], Brocklesby and Fisher [26]
		Participation in local governance (C27)	Whether household participates in local governance: no = 0, yes = 1
		Participation in community organizations (C28)	Whether household participates in various cooperative organizations in the community: no = 0, yes = 1
		Power of community organizations (C29)	Role of local community organizations in local governance: very small = 0, small = 0.25, general = 0.5, large = 0.75, very large = 1
	Government governance level (B12)	Satisfaction with government governance (C30)	Degree of satisfaction of households with local government governance: very dissatisfied = 0, dissatisfied = 0.25, general = 0.5, satisfied = 0.75, very satisfied = 1	DfID [19], Guo and Yang [10], Cui et al. [55]
	Policies and institutions (B13)	Industrial development policies (C31)	Number of industrial development policies enjoyed by households
		Support policies and welfare policies (C32)	Number of support policies and welfare policies enjoyed by households
Livelihood outcomes (A5)	Economic sustainability (B14)	Living standard (C33)	Degree of improvement of the overall living standard of households: very low = 0, low = 0.25, general = 0.5, high = 0.75, very high = 1	Su et al. [6], Su et al. [12], Cui et al. [55], Shang et al. [56]
		Employment opportunities (C34)	Degree of increasing in employment opportunities that rural households can obtain: very low = 0, low = 0.25, general = 0.5, high = 0.75, very high = 1
	Natural resource sustainability (B15)	Resource utilization (C35)	The degree of sustainable use of the natural resources owned by households: very low = 0, low = 0.25, generally = 0.5, high = 0.75, very high = 1
		Ecological environment (C36)	Degree of optimization of the ecological environment in the area where households live: very low = 0, low = 0.25, generally = 0.5, high = 0.75, very high = 1
	Socio-cultural sustainability (B16)	Cultural inheritance (C37)	Degree of effective inheritance of local national culture: very low = 0, low = 0.25, general = 0.5, high = 0.75, very high = 1
		Social relationship (C38)	Degree to which the relationship between households has become more harmonious: very low = 0, low = 0.25, general = 0.5, high = 0.75, very high = 1

).

4.3. Weighting of Indicators and Livelihood Sustainability Measurement Method

4.3.1. Determination of Indicator Weighting by Combined Weighting Methods

Because the original data had different dimensions and orders of magnitude, the method of extreme difference normalization was used to standardize it. Then, indicators were weighted. Weighting methods can be subjective or objective, with each having advantages and disadvantages [57]. The subjective weighting method mainly depends on the experience of experts, with the results being influenced by their opinions. In comparison, the objective weighting method determines the weighting of indicator quantitatively based on the information provided by the observation values of the primary data to avoid the subjectivity of expert scoring [25]. However, this leads to the objective method having higher data quality requirements [57]. Studies have increasingly combined subjective and objective weighting methods to determine the weighting of indicators [57,58]. This approach reduces the subjective bias of expert judgment and issues of incomplete data or poor data quality [57,58]. The current study combined the subjective weighting method of the analytic hierarchy process (AHP) and objective entropy weighting method to determine the weight of indicators.

(1)

Analytic hierarchy process method to determine the weight of indicators

A four-level hierarchical structure model was first constructed using the evaluation index system. The target layer was the livelihood sustainability level. The criterion layer included five first-level indicators. The sub-criterion layer included 16 second-level indicators. The scheme layer included 38 third-level indicators.

Second, a judgment matrix was constructed for each level in the hierarchy. A nine-scale value method was used to score and judge the relative importance of indicators [59]. Six experts were consulted, including scholars studying ethnic tourism, poverty, and sustainable livelihoods (three people) and staff from the local poverty alleviation office (three people).

Then, the weight of each level in the hierarchy was calculated according to the expert-based judgment matrix. The consistency of the judgment matrix was then verified. Specifically, rationale for the weights was confirmed by calculating the consistency ratio (CR) of the judgment matrix. The weighting of each index was calculated using SPSSAU (

Table 3. Weights of the indicators.

First-Level Indicators	The Weights of AHP Method (%)	The Weights of Entropy Method (%)	The Comprehensive Weights(%)	Second-Level Indicators	The Weights of AHP Method (%)	The Weights of Entropy Method (%)	The Comprehensive Weights (%)	Third-Level Indicators	The Weights of AHP Method (%)	The Weights of Entropy Method (%)	The Comprehensive Weights (%)
A1	10.9	9.3	10.1	B1	66.7	76.1	71.4	C1	33.3	11.4	22.4
								C2	66.7	88.6	77.6
				B2	33.3	23.9	28.6	C3	33.3	45.5	39.4
								C4	66.7	54.5	60.6
A2	45.1	44.1	44.6	B3	6.2	6.8	6.5	C5	66.7	66.7	66.7
								C6	33.3	33.3	33.3
				B4	8.4	6.3	7.4	C7	49.0	53.6	51.3
								C8	19.8	35.7	27.7
								C9	31.2	10.7	21.0
				B5	11.1	12.5	11.8	C10	38.7	43.6	41.2
								C11	14.0	14.5	14.3
								C12	19.8	20.0	19.9
								C13	27.5	21.8	24.6
				B6	33.3	23.6	28.4	C14	49.0	43.3	46.1
								C15	31.2	26.9	29.1
								C16	19.8	29.8	24.8
				B7	24.2	42.0	33.1	C17	41.8	84.9	63.3
								C18	20.5	4.9	12.7
								C19	21.4	7.0	14.2
								C20	16.3	3.2	9.8
				B8	16.8	8.8	12.8	C21	40.0	23.1	31.5
								C22	40.0	38.5	39.2
								C23	20.0	38.4	29.3
A3	10.5	9.4	10.0	B9	66.7	56.5	61.6	C24	100.0	100.0	100.0
				B10	33.3	43.5	38.4	C25	100.0	100.0	100.0
A4	13.2	32.8	23.0	B11	49.0	81.0	65.0	C26	32.9	11.0	21.9
								C27	32.9	11.3	22.1
								C28	14.2	17.7	16.0
								C29	20.0	60.0	40.0
				B12	19.8	2.5	11.1	C30	100.0	100.0	100.0
				B13	31.2	16.5	23.9	C32	66.7	86.6	76.6
								C32	33.3	13.4	23.4
A5	20.3	4.4	12.4	B14	49.0	28.2	38.6	C33	66.7	54.5	60.6
								C34	33.3	45.5	39.4
				B15	19.8	35.9	27.8	C35	66.7	64.3	65.5
								C36	33.3	35.7	34.5
				B16	31.2	35.9	33.6	C37	33.3	57.1	45.2
								C38	66.7	42.9	54.8

Data sources: the data were calculated by the authors.

). All weighted values obtained in this study passed the consistency verification, showing that expert judgement on the indicators had good consistency.

(2)

Entropy method to determine indicator weighting

The specific gravity of each indicator was calculated as:

$$f_{ij}=P_{ij}/{\displaystyle \sum _{i=1}^n{P}_{ij}}$$

(1)

The entropy of each indicator was calculated as:

$$e_j=-(1/\ln n){\displaystyle \sum _{i=1}^n{f}_{ij}\ln }{f}_{ij}$$

(2)

The weight of the jth indicator was calculated as:

$$W_j=(1-e_j)/{\displaystyle \sum _{j=1}^m(1-e_j)}$$

(3)

Through these steps, the weighting of each third-level indicator was calculated. The weights of first- and second-level indicators were obtained by summing the weights of third-level indicators. The weights of second- and third-level indicators were then normalized to obtain a final weight (Table 3).

(3)

Determining the combined weighting of indicators

Following Zhang et al. [57], the equal weighted average method was used to obtain comprehensive weightings, which were the average of subjective and objective weights (Table 3).

4.3.2. Measurement of the Sustainability of Rural Household Livelihoods

Based on the standardized scores and corresponding weights of the various measurement indicators, a weighted average model was constructed to assess the sustainability level of farmer livelihoods. First, the standardized score P_ijk of each third-level indicator was multiplied by its weight W_ijk and was merged to form the final evaluation value of each second-level indicator.

$$P_{ij}={\displaystyle \sum _{k=1}^m{W}_{ijk}{P}_{ijk}}$$

(4)

Similarly, based on the evaluation value, P_ij, and corresponding weight, W_ij, of each second-level index, the evaluation value, P_i, of each first-level index was obtained. Based on the evaluation values, P_i, and weight, W_i, of the five first-level indicators of the sustainable livelihood framework, a comprehensive evaluation model of the sustainability of rural household livelihood was established.

$$F_{SL}={\displaystyle \sum _{i=1}^5{P}_i{W}_i}$$

(5)

where F_SL represents the livelihood sustainability index, reflecting the sustainability level of household livelihoods. The livelihood sustainability index and the evaluation values of first- and second-level indicators were between 0 and 1. The value 1 indicates that the livelihood sustainability level of rural households, or the level of a single indicator, was in an excellent state. In contrast, the value 0 indicates that the level is very poor. A value equal to 0.5 is intermediate. The larger the livelihood sustainability index, or the evaluation value of a single first-level indicator and a single second-level indicator, the higher the livelihood sustainability of rural households, or the development level of the single indicator.

5. Results

5.1. Livelihood Sustainability of Rural Households

5.1.1. Livelihood Sustainability of All Rural Households

The livelihood sustainability of rural households was poor (mean: 0.387;

Table 4. Evaluation results of livelihood sustainability level of households in ethnic tourism villages.

First-Level Indicators	Weights (%)	Single Evaluation Value of First-Level Indicator	Second-Level Indicators	Single Evaluation Value of Second-Level Indicator	Weights (%)
Vulnerability context	10.1	0.380	Natural environment vulnerability	0.363	71.4
			Market environment vulnerability	0.422	28.6
Livelihood assets	44.6	0.350	Natural capital	0.386	6.5
			Human capital	0.347	7.4
			Physical capital	0.577	11.8
			Financial capital	0.280	28.4
			Social capital	0.228	33.1
			Cultural capital	0.595	12.8
Livelihood strategies	10.0	0.457	Livelihood diversification	0.500	61.6
			Livelihood freedom	0.388	38.4
Transforming structures and processes	23.0	0.329	Households’ community participation level	0.250	65.0
			Government governance level	0.483	11.1
			Policies and institutions	0.504	23.9
Livelihood outcomes	12.4	0.572	Economic sustainability	0.578	38.6
			Natural resource sustainability	0.626	27.8
			Socio-cultural sustainability	0.521	33.6
Evaluation index of livelihood sustainability in ethnic tourism villages: 0.387

Data sources: the data were calculated by authors.

). Out of the five first-level indicators, the livelihood outcome index was highest (0.572), and was slightly higher than the intermediate level. The household survey showed that most farmers participating in tourism development believed that their living standards had been significantly improved, natural resources had been more effectively utilized, and local national culture had been effectively inherited with the development of the local tourism industry. The livelihood strategy index was ranked second (0.457) after livelihood outcomes. In the overall surveyed area, 85.4% of households were engaged in two or more livelihood activities, with a high percentage of households participating in diverse livelihood types. However, the livelihood freedom of households was low. Most households that mainly engaged in farming and working beyond the village stated that they lacked the skills and funds needed to participate in other livelihood activities. The vulnerability context index value was poor (0.380) because of the relatively high frequency of disasters, such as landslides (three areas), landslides in terraces (Jinkeng Village and Ping’an Zhuang Village), fires (three areas), and floods (Chengyang Eight Village), which had a certain impact on household livelihoods. The seasonality of the tourism industry and impact of coronavirus (COVID-19) caused tourism livelihood households to face strong market vulnerability.

The lowest scores were obtained for livelihood assets and transforming structures and processes, with index values of 0.350 and 0.329, respectively. These two components clearly restrict the livelihood sustainability of local households. Out of the various livelihood capital index values, the highest were obtained for cultural capital and physical capital, followed by natural capital and human capital. Financial capital and social capital were the lowest. Cultural capital and physical capital index values both exceeded 0.5, as most households in the villages were familiar with traditional culture and had mastered the traditional handicrafts of their own nation. Tourism operators were able to apply traditional culture to tourism management. Tourism reception facilities and the daily fixed assets of rural households have also improved, further enhancing the physical capital of local households. The natural capital index was low (0.386). The per capita cultivated land areas of Jinkeng Village, Ping’an Zhuang Village, and Chengyang Eight Village were 0.521, 0.486, and 0.502 mu (1 mu = 1/15 ha. The data were obtained by survey), respectively, indicating that the natural assets of local households were limited. Local governments also expropriated large amounts of land resources to construct tourism infrastructure and service facilities, further aggravating the loss of household natural capital.

The human capital index (0.347) was slightly below that of natural capital. Most household members have a low level of education, mostly primary or junior high school, which greatly restricted the development of household livelihoods. The index values of financial capital and social capital were low (0.280 and 0.228, respectively), clearly reducing the sustainability level of household livelihoods. Financial capital was low because the number of tourists visiting ethnic tourism villages sharply decreased due to COVID-19, negatively impacting income. There was also a large income gap between households with different livelihood strategies. In particular, it was harder for households that did not participate in tourism development to obtain loans from banks. Most farmers also stated that it was difficult to borrow money from relatives and friends. Regarding social capital, most households stated that they lacked relatives and friends working in government departments. Furthermore, the relatives and friends of rural households could only provide limited support to households when they encountered difficulties, due to the strong homogeneity of the economic status of rural households with relatives/friends and the weakening of traditional ethics.

In terms of transforming structures and processes, the implementation of a series of poverty alleviation policies from central to local governments, and the distribution of ticket income from local scenic spots, have helped to improve household livelihoods. However, the lack of community participation by farmers significantly limited the sustainability of household livelihoods. Rural households in all three regions stated that they rarely participated in community organization, industrial development decision making, or local governance. Although Jinkeng Village and Ping’an Zhuang Village have established tourism management committees, these community organizations have become a mere formality, rather than fulfilling a key role.

5.1.2. Comparative Analysis of the Sustainability Level of Household Livelihoods across Villages

The livelihood sustainability index was highest in Jinkeng Village (0.441), followed by Ping’an Zhuang Village (0.380) and Chengyang Eight Village (0.361) (

Table 5. Evaluation results of livelihood sustainability level of households in different villages.

First-Level Indicators	Single Evaluation Value of First-Level Indicator			Second-Level Indicators	Single Evaluation Value of Second-Level Indicator
	Jinkeng Village	Ping’an Zhuang Village	Chengyang Eight Village		Jinkeng Village	Ping’an Zhuang Village	Chengyang Eight Village
Vulnerability context	0.430	0.361	0.362	Natural environment vulnerability	0.434	0.357	0.308
				Market environment vulnerability	0.420	0.371	0.497
Livelihood assets	0.392	0.372	0.316	Natural capital	0.364	0.391	0.395
				Human capital	0.368	0.370	0.325
				Physical capital	0.668	0.564	0.534
				Financial capital	0.370	0.366	0.189
				Social capital	0.257	0.250	0.202
				Cultural capital	0.568	0.520	0.646
Livelihood strategies	0.486	0.480	0.429	Livelihood diversification	0.449	0.510	0.460
				Livelihood freedom	0.546	0.432	0.379
Transforming structures and processes	0.443	0.262	0.301	Households’ community participation level	0.271	0.250	0.216
				Government governance level	0.512	0.538	0.441
				Policies and institutions	0.878	0.166	0.466
Livelihood outcomes	0.578	0.560	0.575	Economic sustainability	0.625	0.636	0.525
				Natural resource sustainability	0.596	0.553	0.677
				Socio-cultural sustainability	0.510	0.478	0.547
Evaluation index of livelihood sustainability of households in Jinkeng village: 0.441 Evaluation index of livelihood sustainability of households in Ping’an Zhuang village: 0.380 Evaluation index of livelihood sustainability of households in Chengyang Eight village: 0.361

Data sources: the data were calculated by authors.

). The livelihood outcome index of households in all three villages exceeded 0.5. The livelihood strategy index was highest in Jinkeng (0.486), followed by Ping’an Zhuang (0.480) and Chengyang Eight (0.429). All three villages had high levels of livelihood source diversification. However, the livelihood freedom of households was low in Chengyang Eight, resulting in it having the lowest livelihood strategy index of the three villages.

The livelihood assets index values of the three villages are Jinkeng (0.392) > Ping’an Zhuang (0.372) > Chengyang Eight (0.316). Regarding different types of livelihood capital, cultural and physical capital were above the intermediate level for all three villages. Chengyang Eight village, with relatively low level of tourism development, had a more robust cultural atmosphere and local households had a higher degree of understanding of local culture. Consequently, this village had the highest cultural capital index of the three villages. Jinkeng Village had a high physical capital index with a well-developed tourism industry, including strong infrastructure and tourism service facilities. Social capital restricted the improvement of livelihood assets in the households of all three villages, requiring focus. The financial capital index of Chengyang Eight Village (0.189) was significantly lower than that of the other two villages, due to it having a less developed tourism industry. The proportion of rural households participating in tourism development, and the per capita income of farmers, was higher in the other two villages. It was also easier for households in Ping’an Zhuang and Jinkeng to obtain bank loans.

Households in Ping’an Zhuang and Chengyang Eight village had similar vulnerability context index scores that were significantly lower than those of households in Jinkeng. Chengyang Eight had the lowest natural environment vulnerability score, due to the impact of floods, fires, and landslides. However, Chengyang Eight had the highest score for the market environment vulnerability index. Visitors to this village were mainly from Guangxi, with generally low numbers. One of the main tourism sources in the village is a sketching base in the village, which attracts teachers and students of art institutes, painters, and photography enthusiasts. Consequently, farmers in this village do not have a strong perception of the impact of COVID-19 on the tourism market. However, high numbers of national (from other provinces) and international tourists frequent in Jinkeng and Ping’an Zhuang, with COVID-19 having a higher perceived impact on the tourism industry by farmers, particularly in Ping’an Zhuang.

Jinkeng had the highest score for transforming structures and processes, while Ping’an Zhuang village scored the lowest. Community participation index values of the households in all three villages were poor <0.3. The policy and institution index values of Jinkeng and Ping’an Zhuang noticeably differed, due to the villages having different industrial development measures. The management company of the Longji Terraced Fields Scenic Area constructed a sightseeing ropeway in Jinkeng, which successfully promoted tourism development of this village, attracting large numbers of tourists and making the village a core scenic spot. It indicates that appropriate industrial development measures are of great importance to the livelihood development of local households.

5.1.3. Comparative Analysis of the Livelihood Sustainability Level of Households with Different Types of Livelihood Strategies

Tourism-oriented households had the highest level of livelihood sustainability, followed by balanced households and pure tourism households. The livelihood sustainability indexes of laboring-oriented and agricultural-oriented households are the lowest (

Table 6. Comparative analysis of livelihood sustainability levels of different types of households.

First-Level Indicators	Single Evaluation Value of First-Level Indicator					Second-Level Indicators	Single Evaluation Value of Second-Level Indicator
	Pure Tourism Households	Tourism-Oriented Households	Balanced Households	Laboring-Oriented Households	Agricultural-Oriented Households		Pure Tourism Households	Tourism-Oriented Households	Balanced Households	Laboring-Oriented Households	Agricultural-Oriented Households
Vulnerability context	0.315	0.383	0.396	0.382	0.361	Natural environment vulnerability	0.312	0.396	0.413	0.342	0.294
						Market environment vulnerability	0.322	0.351	0.354	0.480	0.527
Livelihood assets	0.371	0.405	0.336	0.283	0.223	Natural capital	0.102	0.416	0.404	0.393	0.417
						Human capital	0.374	0.367	0.333	0.327	0.297
						Physical capital	0.721	0.656	0.584	0.429	0.307
						Financial capital	0.380	0.371	0.286	0.111	0.063
						Social capital	0.231	0.269	0.180	0.226	0.154
						Cultural capital	0.529	0.618	0.591	0.595	0.542
Livelihood strategies	0.297	0.501	0.525	0.396	0.268	Livelihood diversification	0.148	0.481	0.584	0.422	0.296
						Livelihood freedom	0.535	0.532	0.430	0.354	0.222
Transforming structures and processes	0.327	0.384	0.338	0.226	0.254	Households’ community participation level	0.203	0.272	0.251	0.192	0.150
						Government governance level	0.479	0.494	0.502	0.476	0.361
						Policies and institutions	0.594	0.639	0.500	0.202	0.488
Livelihood outcomes	0.592	0.601	0.566	0.534	0.505	Economic sustainability	0.631	0.643	0.555	0.487	0.474
						Natural resource sustainability	0.589	0.635	0.624	0.634	0.584
						Socio-cultural sustainability	0.549	0.523	0.531	0.505	0.477
Evaluation index of livelihood sustainability of pure tourism households: 0.376 Evaluation index of livelihood sustainability of tourism-oriented households: 0.432 Evaluation index of livelihood sustainability of balanced households: 0.391 Evaluation index of livelihood sustainability of laboring-oriented households: 0.322 Evaluation index of livelihood sustainability of agricultural-oriented households: 0.284

Data sources: the data were calculated by authors.

). Based on the five first-level indicators, livelihood outcome had the highest scores for all household types (>0.5). In-depth analysis found that the economic sustainability index of laboring-oriented households and the economic sustainability and socio-cultural sustainability indexes of agricultural-oriented households were low (<0.5). This result was obtained because rural households that do not participate in tourism or have a low degree of participation in tourism, obtained limited economic benefits from tourism. Consequently, their perceptions of the negative impacts of tourism development were stronger, including the widening gap between the rich and poor and the loss of authenticity of national culture.

The livelihood strategy index values of balanced households and tourism-oriented households exceeded 0.5, whereas the other three types of households performed poorly. In particular, the livelihood strategy index of pure tourism households and agricultural-oriented households was low (<0.3). These two household types had lower scores for different reasons. Pure tourism households were able to choose their livelihood activities freely; however, the overall score of the livelihood strategy index was low because of their single livelihood activity. The livelihood freedom and livelihood diversity of agricultural-oriented households were both poor. At the same time, because pure tourism households and agricultural-oriented households have a single livelihood activity, the livelihood development of these two types of households was restricted by the natural environment. In addition, the impact of COVID-19 resulted in pure tourism households facing higher market vulnerability, due to their strong dependence on the tourism industry. Therefore, pure tourism households and agricultural-oriented households had the lowest vulnerability context index scores, with strong livelihood instability.

In terms of livelihood assets, the total livelihood assets indexes of various types of households are tourism-oriented households (0.405) > pure tourism households (0.371) > balanced households (0.336) > laboring-oriented households (0.283) > agricultural-oriented households (0.223). The comparison of different types of livelihood capital of various types of households is shown in Figure 4. All types of livelihood capital of tourism-oriented households ranked in the top two for the five types of households. However, for tourism-oriented households, social capital and human capital limited livelihoods. Out of the five types of households, human capital index, physical capital index, and financial capital index were highest for pure tourism households. However, pure tourism households had the fewest natural resources, coupled with social capital constraints, resulting in the overall level of livelihood capital not being high. Both the physical capital and cultural capital indexes of balanced households were greater than 0.5, while social capital, financial capital, and human capital were low, especially social capital (0.180). Social capital, financial capital, and human capital were also the main factors restricting the level of livelihood assets of laboring-oriented and agricultural-oriented households. In particular, the financial capital index of agricultural-oriented households was only 0.063, demonstrating the large gap between rich and poor farmers in ethnic tourism villages, and demonstrating that the dividends of tourism development are not being shared effectively.

Tourism-oriented households had the highest scores for transforming structures and processes, whereas laboring-oriented households had the lowest scores. The community participation index values of all types of households were low (<0.3), indicating the need to improve the participation of households in local affairs and tourism development decision making. Agriculture-oriented households were particularly dissatisfied with the allocation of various poverty alleviation subsidies; therefore, this type of household had the lowest government governance level index. Regarding industrial development policies and support and welfare policies, more policies and measures benefitted households that participated in the tourism industry compared to other types of households.

5.2. Livelihood Risks Faced by Households

The rural households in the surveyed areas were exposed to various livelihood risks. As shown in

Table 7. The main livelihood risks faced by households in ethnic tourism villages.

Risk Category	All Households		Jinkeng Village Households		Ping’an Zhuang Village Households		Chengyang Eight Village Households
	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk
Market risk	62.45%	35.62%	71.77%	54.84%	74.11%	58.04%	51.74%	14.35%
Financial risk	35.84%	21.46%	48.74%	25.00%	48.21%	20.54%	23.91%	20.00%
Social risk	43.56%	5.36%	58.06%	4.84%	71.43%	7.14%	22.17%	4.78%
Employment risk	24.03%	16.09%	10.48%	4.84%	10.71%	8.93%	37.83%	25.65%
Natural risk	22.10%	5.58%	18.55%	2.42%	18.75%	1.79%	25.65%	9.13%
Education risk	41.42%	9.87%	25.00%	4.84%	28.57%	0.89%	56.52%	16.96%
Pension risk	6.44%	0.43%	4.03%	1.61%	5.36%	0.00%	8.26%	0.00%
Health risk	23.61%	3.86%	17.74%	0.00%	12.50%	0.89%	32.17%	7.39%
Policy risk	15.45%	1.07%	12.90%	1.61%	20.54%	0.89%	14.35%	0.87%
Marriage risk	7.08%	0.64%	1.61%	0.00%	14.29%	0.89%	6.52%	0.87% 1

¹ The percentages in the table are the proportion of corresponding households in total households. Data sources: the data were calculated by the authors.

, most households (62.45%) faced market risks. COVID-19 severely damaged the tourism industry in all three villages, and beyond. The tourism market of all three villages fluctuated significantly, especially in Jinkeng and Ping’an Zhuang, which support high international tourism, coupled with the seasonal characteristics of the industry. Consequently, households participating in tourism development had a stronger perception of market risk. Households facing social risk and education risk came next, accounting for 43.56% and 41.42% of the total surveyed households, respectively. Social risks included conflicts of farmers with village committees or scenic spot management companies regarding the distribution of interests and development of tourism. Examples of conflicts included the distribution of ticket income for scenic spots where villages are located, compulsory land acquisition, construction of new houses by households, and the allocation of poverty alleviation subsidies. Conflicts also arose between local households, due to the competition for tourists. The widening gap between the rich and poor, due to tourism development, has reduced harmony among households. As Putnam [60] has pointed out, economic inequality is likely to be socially divisive. Education risk was also a concern of households. Due to the limitations of history, economic development level, and natural conditions, the education level of farmers in the three villages was generally low, with few opportunities to receive vocational skills training. Educational resources were limited, with only primary schools being distributed, and teachers were not of high quality. Bandyopadhyay and Tang [61] highlighted that the inequality of educational opportunity adversely affects economic growth and human capital. Furthermore, rural households with many children receiving education were widespread in ethnic areas; therefore, education expenditure represented an important economic expenditure of households.

However, when enquired about the livelihood risks that worried the households the most, market risk (35.62%), financial risk (21.46%), and employment risk (16.09%) ranked the highest. Due to the relatively good level of tourism development in the surveyed areas before the COVID-19 outbreak, especially in Jinkeng and Ping’an Zhuang, households participating in tourism operations were able to obtain bank loans easily. Therefore, to enhance tourism development, many households obtained large loans from several banks with high interest rates. However, after the COVID-19 outbreak, the banks did not provide interest relief policies to farmers, with continued high interest rates, which significantly increased the pressure on household livelihoods. Consequently, financial risk became a livelihood risk that many tourism operators most worry about. Because of the low level of education of many farmers, they could only work on construction sites or perform odd jobs when working outside their hometowns. As a result, the employment of these households was unstable, with low income levels and higher employment risks. Because employment also directly determines the total income level of households, employment risk was the livelihood risk that most concerned these households.

The livelihood risks that worried the households in Jinkeng village and Ping’an Zhuang village are similar. When enquired about the livelihood risks that worried them, market, financial, and social risks ranked the highest in these two villages; when enquired about a single risk that worried them the most among these risks, market, financial, and employment risks ranked the highest. In Chengyang Eight village, market, education, and employment risks ranked the highest among the livelihood risks that were worrisome; meanwhile, when enquired about a single risk that worried them the most among these risks, employment, financial, and education risks ranked the highest. In terms of different types of households (

Table 8. The main livelihood risks faced by different types of households in ethnic tourism villages.

Risk Category	Pure Tourism Households		Tourism-Oriented Households		Balanced Households		Laboring-Oriented Households		Agricultural-Oriented Households
	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk	Worry about the Risk	Most Worried about the Risk
Market risk	90.70%	48.84%	82.74%	55.95%	79.81%	44.23%	21.21%	3.03%	28.13%	6.25%
Financial risk	37.21%	37.21%	42.26%	30.95%	49.04%	20.19%	23.23%	11.11%	18.75%	0.00%
Social risk	58.14%	6.98%	54.17%	8.93%	48.08%	4.81%	24.24%	1.01%	40.63%	3.13%
Employment risk	9.30%	4.65%	10.71%	4.17%	11.54%	10.58%	63.64%	48.48%	46.88%	21.88%
Natural risk	4.65%	0.00%	23.21%	1.19%	11.54%	0.96%	29.29%	12.12%	65.63%	34.38%
Education risk	20.93%	0.00%	28.57%	5.95%	42.31%	9.62%	66.67%	18.18%	81.25%	25.00%
Pension risk	9.30%	2.33%	4.17%	0.60%	4.81%	0.00%	4.04%	0.00%	31.25%	0.00%
Health risk	11.63%	0.00%	22.62%	4.17%	22.12%	7.69%	38.38%	1.01%	18.75%	6.25%
Policy risk	6.98%	0.00%	8.93%	0.00%	16.35%	0.00%	35.35%	5.05%	6.25%	0.00%
Marriage risk	0.00%	0.00%	5.36%	0.00%	6.73%	1.92%	11.11%	0.00%	18.75%	3.13% 1

¹ The percentages in the table are the proportion of corresponding households in the total households. Data sources: the data were calculated by authors.

), the number of households who worried about market, financial, and social risks were the highest among the pure tourism and tourism-oriented households; these three risks also ranked the top when enquired about a single risk that worried them the most. This implies that for households whose main livelihood activity is tourism, these risks are the main factors restricting their livelihood stability and need to be studied closely. Farmers whose main livelihood activity is tourism have the strongest perception of the impact of COVID-19 on the tourism market. Although these farmers have higher financial capital, they also bear higher financial risks. They need to repay bank loans with high-interest rates even if they cannot earn much money during COVID-19. Moreover, compared with non-tourism operators, farmers participating in tourism development have a stronger perception of social risks and conflicts caused by tourism. In terms of balanced households, although most balanced households are concerned about market, financial, and social risks such as pure tourism and tourism-oriented households, employment risk was included when enquired about a single risk that worried them the most among these risks. While participating in tourism development, the family members of balanced households also choose to go out to work (or work nearby); thus, employment risk is one of the livelihood risks that worries them the most.

Among the livelihood risks that laboring-oriented households worry about, education, employment, and health risks ranked the highest. The financial, social, and human capital values of laboring-oriented households are low, which makes them lack the funds, skills, and social connections necessary to participate in tourism operations. They could only work on construction sites or perform odd jobs outside or nearby their hometowns. Therefore, they encounter higher employment risks. Furthermore, children’s education expenditure is a great burden for laboring-oriented households due to their low level of financial capital. Hence, they have a strong perception of education risks. In addition to education and employment risks, the health risks that such households worry about mainly originate from costs of daily medicines that are required for the elderly in their households. Furthermore, such households are engaged in high-intensity physical labor throughout the year. Hence, family members of laboring-oriented households exhibit high incidences of waist and leg pain. Among the livelihood risks that worried these households the most, employment, education, and natural risks ranked the highest. In addition to engaging in industrial or engineering work, laboring-oriented households also tend to carry out small-scale agricultural cultivation; thus, their perception of local natural risks is relatively strong. As far as agricultural-oriented households are concerned, most agricultural-oriented households worried about education, natural, and employment risks; when enquired about a single risk that worried them the most among all the risks that worried them, the above-mentioned risks ranked the highest as well. The largest expenditure in most agricultural-oriented households is children’s education such as laboring-oriented households; thus, most agricultural-oriented households are worried about education risks. However, for agricultural-oriented households, they have relatively high natural capital; thus, planting agriculture is the main livelihood activity. They are most worried about the natural risks that significantly affect agricultural development.

6. Discussion

From the results, it was found that the overall livelihood sustainability level of households in ethnic tourism villages was relatively low. The sustainable livelihood of farmers in ethnic tourism villages deserves special attention. The calculation and comparison of the five aspects of the DFID sustainable livelihood analysis framework indicated that households perform better in livelihood outcomes and livelihood strategies. However, they experience intense livelihood vulnerability. Consistent with the findings of Su et al. [6] and Su et al. [12], tourism development generates a variety of economic opportunities and increases family income for local households, which has improved livelihood outcomes. Local government and farmers are paying more attention to the protection of the rural ecological environment due to the tourism development, which has a definite positive effect on the improvement of the rural environment [3,56]. Besides direct tourism employment, the development of tourism induced development of related businesses has triggered the generation of various local employment opportunities, thus improving the livelihood diversity of farmers [3,6,62]. The households participating in tourism enjoy more freedom in choosing and moving among potential economic strategies, while other households cannot enjoy livelihood freedom due to restricted assets [6]. Furthermore, the COVID-19 pandemic caused serious damage to the travel and tourism industry, making the livelihood of local farmers participating in tourism industry more vulnerable [63,64]. Meanwhile, the households who participate in single livelihood activity are more vulnerable to the external risk and shocks [4].

Livelihood assets and transforming structures and processes were the most significant factors restricting livelihoods. Out of the six types of livelihood capitals, cultural and physical capital were important for improving the overall assets of household livelihoods, whereas financial and social capital were key limiting factors. Consistent with the findings of Ma et al. [24] and Wang et al. [25], cultural resources play an important role in the development of rural tourism and are an important part of farmers’ livelihood, which should not be ignored in the analysis of livelihood capital. The households realize the high commercial value of local traditional culture and modern tourists’ preferences for ancient buildings and traditional customs. Local households can effectively apply their traditional culture to rural tourism management [24]. After participating in rural tourism, local households own more fixed assets for providing better tourism reception; thus, their physical capital value has risen [3]. Consistent with the findings of Huang et al. [4] and Paul et al. [65], there are significant differences in financial capital and social capital between different types of farmers, which suggest that financial and social capitals are important factors affecting livelihood choice and bringing sustainability. Low financial capital limits the idea of tourism-participating farmers to expand the scale of tourism business. It also makes it difficult for agricultural-oriented farmers and laboring-oriented households to engage in tourism business. Due to the contradictions between farmers caused by the development of tourism, the weakening of traditional ethical concepts, and strong homogeneity of the economic status of rural households with relatives/friends, the social network support that households can obtain is limited. Moreover, low participation of households in community activities limited transforming structures and processes. With the rapid growth of rural tourism in China, local governments have implemented policies to encourage resident participation [6]. However, these policies have little effect, and the degree of farmers’ participation in tourism decision making and local affairs is still low [11].

In terms of households with different types of livelihood strategies, the livelihood sustainability levels were highest in tourism-oriented households, which indicated that tourism fits well into the participating households’ livelihood strategies, and tourism-engaged households had livelihood advantages, which is consistent with finding of Su et al. [32]. Livelihood outcomes and livelihood assets were highest for tourism-oriented households and pure tourism households. Tourism development has improved the overall living standards of farmers involved in tourism industry. Tourism participants possess more assets when compared with non-participants, particularly due to requirements of initial investments for renovation and purchasing necessary facilities and supplies [6]. As a result, those with low asset levels find it difficult to participate in tourism, thus receiving fewer benefits [6]. Tourism-oriented households and balanced households had the highest livelihood strategies, vulnerability context, and transforming structures and processes. These households have a high level of livelihood diversity and livelihood capital advantage, which makes them have a strong ability to deal with risks. The natural capital index of pure tourism households was low, which limits the livelihood diversity of such farmers to a certain extent. Moreover, the financial capital indexes of agricultural-oriented households and laboring-oriented households were also very low. As King and Levine [66] have pointed out, financial institutions evaluate prospective entrepreneurs and they fund the most promising ones. Financial institutions intend to fund projects with relatively good chances of success. Therefore, local financial institutions are more inclined to provide loans to tourism-engaged households who can obtain high income.

Because of COVID-19 and the seasonal characteristics of tourism industry, house-holds participating in tourism development had a stronger perception of market risk. Most households also expressed concern about social risk and education risk. Social risks included conflicts of farmers with village committees or scenic spot management. Conflicts also arose between local households. Education risks came from the low level of education of farmers themselves and the limited educational resources of local villages. Moreover, when households were enquired about a single risk that worried them the most, a large number of farmers chose market risk, financial risk, and employment risk. All three risks are related to COVID-19 pandemic, which severely damaged the tourism industry in all three villages. The tourism market of all three villages fluctuated significantly. The income of farmers involved in tourism development dropped considerably. However, most households have obtained large loans from several banks with high interest rates. Therefore, these households are most concerned about financial risk. Moreover, because of the low level of education of many local farmers and the impact of COVID-19 on employment, many households cannot find stable jobs.

7. Conclusions and Implications

7.1. Conclusions

This study quantitatively evaluated the sustainability of household livelihoods in three typical ethnic tourism villages in China (Jinkeng, Ping’an Zhuang, and Chengyang Eight) and identified key livelihood risks faced by these households. The results showed that the overall livelihood sustainability index value was 0.387. Out of the five first-level indicators, livelihood outcomes performed best (0.572), followed by livelihood strategies (0.457), with households experiencing intense livelihood vulnerability (0.380). The lowest scores were obtained for livelihood assets and transforming structures and processes, with index values of 0.350 and 0.329, respectively. Out of the various livelihood capital index values, the highest were obtained for cultural capital and physical capital, followed by natural capital and human capital. Financial capital and social capital were the lowest. The lack of community participation of households adversely impacted the score of transforming structures and processes. As for different villages, the level of livelihood sustainability was highest in Jinkeng, followed by Ping’an Zhuang and Chengyang Eight. Chengyang Eight had the lowest livelihood outcomes, livelihood strategies, and livelihood assets of the three villages, whereas Ping’an Zhuang village performed the worst in terms of vulnerability context and transforming structures and processes. As for different types of households, the livelihood sustainability levels were highest in tourism-oriented households (0.432), followed by balanced households (0.391) and pure tourism households (0.376). Laboring-oriented households (0.322) and agricultural-oriented households (0.284) performed the worst.

When enquired about the livelihood risks that worried the households, market, social, and education risks ranked the highest; when enquired about a single risk that worried them the most among these risks, market, financial, and employment risks ranked the highest. Households in Jinkeng and Ping’an Zhuang had similar concerns. When enquired about the livelihood risks that worried them, market, financial, and social risks ranked the highest in these two villages; when enquired about a single risk that worried them the most among these risks, market, financial, and employment risks ranked the highest. Most households in Chengyang Eight were concerned about market, education, and employment risk, while employment, financial, and education risks ranking the highest when enquired about a single risk that worried them the most among these risks. In terms of different types of households, the main livelihood risks of households participating in tourism development were market, financial, social, and employment risks, whereas the key risks of non-tourism oriented households were education, employment, health, and natural risks.

7.2. Theoretical Contribution

This study makes the following theoretical contributions: First, based on the sustainable livelihood framework proposed by DFID [19], this study innovatively constructed a quantitative evaluation index system of the sustainability of rural household livelihoods in ethnic tourism villages. These indicators expand the specific content of the DFID sustainable livelihoods framework [19]. This study also addressed the knowledge gap on sustainable tourism livelihoods, which, previously, only systematically analyzed changes to farmer livelihoods after the development of rural tourism [6,12], or only evaluated the livelihood capital of farmers in rural tourism destinations [9,14]. The results of this study are expected to provide a point of reference and guidance toward qualitatively analyzing livelihood sustainability of tourism. Second, when constructing the evaluation index system of livelihood sustainability of rural households, this study added “cultural capital” to “livelihood assets” and added “household community participation level” to “transforming structures and processes.” These new components helped to overcome current limitations in the sustainable livelihood framework proposed by DFID [19] when analyzing household livelihoods in rural tourism areas. This extra information enriches the theory of sustainable livelihoods. Third, according to the DFID sustainable livelihood framework [19], farmers face a variety of risk shocks resulting in their livelihood vulnerability. However, the livelihood risks faced by rural households have been rarely explored in relation to rural tourism areas. This study identified potential risks to the sustainability of household livelihoods in tourism areas, providing new knowledge for households in ethnic tourism areas.

7.3. Practical Implication

First, improving the diversity of farmers’ livelihoods reduces the vulnerability and instability of farmers’ livelihoods. Livelihood diversification is an important precursor to household wellbeing and sustainability [65]. In particular, it is necessary to encourage and support laboring-oriented and agricultural-oriented households to participate in tourism development, sharing the dividends of tourism development. Local government should also support pure tourism households with lower income levels to participate in other livelihood activities. Second, focusing on the short board of farmers’ livelihood capital improves farmers’ livelihood ability. This study found that financial and social capitals were key factors limiting local households’ livelihood. King and Levine [66,67] pointed out that the development of financial systems can promote productivity growth, economic development, and physical capital accumulation. Financial development can also improve the efficiency with which economies employ physical capital. Therefore, the local government should take multiple measures to promote the development of the local financial system and encourage financial institutions to provide local farmers with financial support, especially for laboring-oriented and agricultural-oriented households. These measures include building a free and open financial market, supporting financial institutions to mobilize sufficient resources for tourism livelihood projects, helping financial institutions evaluate and select farmers or tourism projects with high developmental potential, and facilitating risk management to diversify the risks associated with these livelihood activities. Furthermore, Putnam [60] proved that social capital can increase educational performance, bring down violent crime rates, improve health conditions, and narrow economic inequality and civic inequality. Hence, it is necessary to strengthen the communication and cooperation between farmers and intensify the link between farmers and the outside world by relying on the Internet, thus expanding the social network owned by farmers and improving farmers’ social capital. Moreover, the human capital index of local farmers is also not high. It is important to enhance the quality and equity of education in minority areas. The local government should adopt various technologies and channels to ensure that local children have access to national and global knowledge and information to enhance the accumulation of human capital [61]. Moreover, mitigating unequal access to knowledge increases the economy’s growth potential [61].

Third, innovating the assistance mechanism and development model helps the tourism livelihood farmers cope with the impact of COVID-19. For instance, the local government should cooperate with banks and other financial institutions to appropriately reduce or exempt loan interest for farmers, reducing the pressure on farmers to repay. Fourth, improving the level of farmers’ community participation promotes farmers’ participation in local governance and industrial development decision making. The local government should take various measures to ensure that rural tourism cooperatives and other community organizations can play a role and encourage more farmers to join these organizations. Relying on these community organizations ensures that farmers truly participate in local governance and industrial decision making to protect their own interests in a better way. Finally, building a sound social security system and early warning and intervention mechanism improves farmers’ ability to resist livelihood risks. Poor families seek out economic opportunities. However, they tend not to become too specialized. This lack of specialization makes it difficult for them to obtain high-paying jobs [68]. Therefore, the local government should improve the employment training service system and form a market-oriented employment mechanism. It is necessary to develop public health care and medical insurance in ethnic villages, implement precise targeting and refined management, and reduce the impact of health risks on the sustainability of farmers’ livelihoods. Moreover, according to Banerjee and Duflo [68], when farmers are unable to obtain government public services, their social network can provide them with certain help, such as informal insurance and informal loans, to effectively resist risks. Therefore, as mentioned above, attention should be paid to expanding the social network of farmers to reduce the burden of the government to a certain extent.

7.4. Limitations and Future Research

This study had some limitations. First, the study duration was short. Long-term follow-up surveys and diachronic comparative studies are required in the future. Second, this study only selected three ethnic tourism villages in Guangxi, China, due to time and funding constraints. Thus, a larger sample of ethnic villages is required to validate the representativeness of our study. Third, the evaluation index system of livelihood sustainability of households in ethnic tourism destinations proposed here could be expanded to include more livelihood indicators and improve the reliability of the results.