Composite Index of Poverty Based on Sustainable Rural Livelihood Framework: A Case from Manggarai Barat, Indonesia

Abstract

Rural poverty in Indonesia remains a complex issue involving various aspects. West Manggarai, East Nusa Tenggara, is a national tourist destination and a significant focus of national development, yet poverty rates remain very high. Therefore, this study developed a Composite Poverty Index (CPI) using the Sustainable Rural Livelihoods Approach (SRLA) to illustrate the complexity of rural deprivation in West Manggarai Regency. The CPI was developed by normalizing eighteen validated indicators across five livelihood capitals—human, social, natural, physical, and financial. These indicators were then classified using a Likert-type scale, and their weights were determined through the Analytic Hierarchy Process (AHP) to produce village-level CIP scores. The results show that most villages fall into the “Moderate” category (CIP: 0.40–0.60), reflecting chronic but not extreme deprivation. Spatial inequalities are evident, particularly in access to education, infrastructure, clean water, financial services, and ecological resources. Remote villages recorded higher CIP scores. Natural and economic capital were weakest, while human and social capital performed relatively well. Therefore, poverty alleviation in West Manggarai requires an integrated strategy tailored to local spatial conditions and livelihood capital.

1. Introduction

Poverty remains a significant challenge facing communities in rural areas. In this study, we define poverty as characterized by income and operationally as multidimensional deprivation across several areas, including limited access to resources, high vulnerability, social exclusion, and marginalization from institutional processes. This creates overlapping deprivations across all dimensions, including human, social, natural, physical, and financial, based on the Sustainable Rural Livelihood Framework (SLRA), exacerbating livelihood fragility and social isolation [1,2]. This goes beyond the World Bank’s monetary poverty definition (income below US $2.15 per day) and aligns with approaches emphasizing overlapping resource deprivation, access, and resilience. This operational definition is particularly relevant in rural Indonesia, where poverty is characterized by income insufficiency and deficits in education, infrastructure, environmental security, and financial inclusion.

In Indonesia, poverty alleviation has long been a governmental priority, pursued through income support programs, infrastructure development, and conditional cash transfers (CCTs). However, poverty reduction has slowed recently, particularly in rural areas. According to the World Bank, between 2007 and 2011, the rural poverty rate declined from 16.6% to 12.5%—a reduction of 4.1 percentage points. In contrast, the period from 2012 to 2018 saw a smaller decline of only 2.1 percentage points. Furthermore, national poverty dropped only marginally from 10.6% in September 2017 to 9.7% in September 2018, averaging a yearly decline of 0.36 percentage points from 2012 to 2018, compared to a 0.97-point annual decrease from 2007 to 2012 [3].

This stagnant situation is common in remote and underdeveloped areas, such as Manggarai Barat Regency, East Nusa Tenggara (NTT) (Figure 1). West Manggarai Regency boasts substantial infrastructure and tourism investment. However, the poverty rate in this region remains above the national average. Between 2015 and 2023, the poverty rate declined from 20.12% to 16.82%, a decrease of only 3.3 percentage points over eight years. This averages out to a reduction of 0.41 percentage points per year. This situation needs more effective poverty alleviation strategies in this region [4]. This slow decline is especially striking when contrasted with the scale of public investment in the region.

Further complicating the picture, research by Gai et al. (2025) [5] reveals that the benefits of tourism-led development in Labuan Bajo have been unevenly distributed. Households with better access to capital, education, and institutional networks disproportionately capture these benefits, while poorer and more remote communities remain economically marginalized and disconnected from the tourism value chain [5]. This difference indicates that income and consumption aspects are not yet adequate to represent the characteristics and complexity of the causes of rural poverty. Therefore, a more comprehensive understanding of rural poverty is needed, considering the various forms of deprivation experienced by rural communities.

An increasing body of literature argues that poverty in rural areas cannot be adequately assessed using income indicators alone. Rural livelihoods are shaped by the dynamic interplay among multiple forms of capital—human, social, natural, physical, and financial—which collectively influence households’ well-being, vulnerability, and ability to respond to shocks [6,7,8]. The Sustainable Rural Livelihoods Approach (SRLA) presents a robust, community-centered framework. It shows how households can utilize and manage their assets, adapt to shocks, and address institutional challenges to achieve more stable and sustainable livelihoods [9,10]. Recent reformulations of SRLA also reflect its adaptation to 21st-century challenges, including climate change, resilience, and technological transitions [11].

While multidimensional poverty indices (MPIs) are widely used globally, this paper tried to combine SRLA-based livelihood capitals with spatial econometric analysis at the village level. The use of spatial statistics, such as Moran’s I and Local Indicators of Spatial Association (LISA), can reveal poverty clustering and regional disparities [12,13]. Moreover, static evaluations are insufficient for capturing long-term changes in livelihoods. System Dynamics Models (SDM) offer an alternative by simulating feedback loops and time-based interactions between assets, policies, and poverty outcomes [14]. This integration allows us to capture deprivation, its spatial clustering, and geographic drivers. Unlike conventional MPIs, our CPI is explicitly asset-based and spatially explicit, offering insights into geographic poverty traps in a tourism-dependent region. This contribution addresses a gap in the literature on how multidimensional poverty intersects with place-based development paradoxes.

This study seeks to answer the following questions: (1) How can a composite poverty index grounded in SRLA be constructed and applied in the Indonesian rural context? (2) What spatial patterns of poverty emerge across villages in Manggarai Barat? (3) How do the results of the composite index extend beyond single-indicator analyses, and what do they reveal about structural poverty drivers in a tourism-focused economy?

This study uses the SRLA to construct a Composite Poverty Index (CPI) to analyze the complexity of rural life using five livelihood capital indicators. This allows the index to measure not just current deprivation but also the sustainability and transformation capacity of rural livelihoods [14,15]. This study aims to provide a broader and more relevant view of rural poverty. The results are expected to help shape more appropriate policies to address overlapping deprivations across human, social, natural, physical, and financial dimensions, exacerbating livelihood fragility and social isolation. Furthermore, the results are expected to improve the design of effective and inclusive poverty reduction strategies in rural Indonesia.

2. Materials and Methods

2.1. Study Area

This study focuses on Manggarai Barat Regency, a lagging region in East Nusa Tenggara (NTT), Indonesia, despite its designation as a National Strategic Tourism Area (Kawasan Strategis Pariwisata Nasional or KSPN). Nationally, poverty in Indonesia declined from 16.6% (2007) to 9.7% (2018), but rural poverty has been more persistent. In Manggarai Barat, poverty decreased from 20.12% (2015) to 16.82% (2023), a relatively slow pace given the region’s scale of public and private investment. The regency is a representative case for exploring the disconnect between macro-level tourism investments and persistent rural poverty. This condition is further exacerbated by limited institutional capacity, poor infrastructure, and geographic isolation. This context underscores the paradox of rapid tourism development in Labuan Bajo alongside persistent deprivation in rural hinterlands.

2.2. Data Sources

Primary data were obtained through structured household surveys to collect information on income levels, access to basic services, livelihood strategies, asset ownership, and perceived shock vulnerability. Secondary data included official statistics from national and regional agencies (e.g., BPS, TNP2K), spatial datasets (e.g., land use, road network, topography), and programmatic documents such as regional development plans (RPJMD) and strategic tourism policies. These diverse data sources enabled quantitative and spatial analysis of rural poverty conditions.

2.3. Composite Index Construction

The Composite Index of Poverty (CIP) construction followed a rigorous multi-step methodology grounded in the Sustainable Rural Livelihoods Approach (SRLA). The objective was to integrate multiple indicators that reflect the multidimensional and spatially contextual nature of rural poverty into a single composite metric. This process involved: (1) selecting indicators based on five livelihood capitals—human, social, natural, physical, and financial; (2) normalizing data to ensure comparability; (3) assigning weights using the Analytic Hierarchy Process (AHP) based on expert judgment; and (4) aggregating the weighted scores to produce village-level CIP values. Variable classification was context-specific and based on empirical patterns observed in the Manggarai Barat region.

2.3.1. Indicator Selection

Indicators are selected to reflect the five capital assets outlined in the SRLA framework [16,17]. It represents human, social, natural, physical, and financial dimensions of livelihoods. Each indicator was assigned an index score from 1 (very good condition) to 5 (very poor condition), following predefined thresholds (see subcategories below). Initial variables were derived from parameters used in previous research on SRLA. Based on each operational definition of each parameter, data were then derived from PODES data, which is available at the village level in Manggarai Barat (

Table 1. Capitals and Indicators of SRLA, which are Available in PODES Data.

Livelihood Capital	Parameter	Indicators
Human Capital	Education & Skills	Adult literacy program activities; Number of vocational training institutions; Number of kindergartens, primary schools, junior high schools, senior high schools, universities/colleges; Distance to nearest early childhood center, kindergarten, primary, junior high, senior high school, and higher education institution; Availability of public transportation to educational facilities
	Health	Number of health facilities; Distance to nearest health facility; Number of residents suffering from malnutrition; Incidence of extraordinary diseases; Availability of stunting-related service packages for pregnant women and children under two
Social Capital	Social Networks	Existence of community institutions; Number of community institutions in the village; Community habit of cooperation for public purposes; Mutual help activities when residents face misfortunes
	Trust & Reciprocity	Number of village meetings; cooperation habits; Existence of neighborhood security posts; Activation of community-based security systems; Construction/maintenance of security posts
	Institutional Support	Utilization of Village Funds; Existence of village government and village consultative bodies; Number of village meetings conducted; Existence of court offices, police stations, or police posts; Availability of social and health service packages
	Safety & Security	Incidents of crime; Existence of prostitution/localization areas; Existence of neighborhood security posts; Community-based security initiatives; Community participation in safety programs
Natural Capital	Natural Resources	Existence of rivers, irrigation canals, reservoirs, lakes, or water catchments; Existence of excavation/mining sites; Incidents of environmental pollution; Incidents of landslides; Village location in relation to forest areas
	Biodiversity	Existence of mangrove trees; Condition of mangrove ecosystems; Community reforestation/planting activities
	Environmental Sustainability	Waste recycling/management activities by communities; Natural disaster events; Community practice of burning fields/gardens; Existence of village-level waste banks; Participation in social forestry programs
Physical Capital	Infrastructure	Road conditions and types of surface; Electricity access; Housing condition; Main sources of drinking and bathing/washing water; Type of sanitation facilities
	Tools & Equipment	Road quality from/to agricultural production centers; Existence of farm tools and technology
	Transportation	Existence and operational hours of public transportation; Road accessibility; Transportation costs to subdistrict/district offices; Type of transportation infrastructure connecting agricultural production centers to main village roads
Financial Capital	Income & Savings	Primary income sources of village/household population; Existence/accessibility of bank agents; Existence of village flagship products exported abroad; Number of Village-Owned Enterprises, industrial centers, small industry clusters, small industrial estates
	Credit Access	Number of cooperatives (small industry, savings and loans, others); Availability of credit facilities (People’s Business Credit, Food and Energy Security Credit, Small Business Credit, Joint Business Groups); Existence of village-level microfinance institutions
	Financial Services	Existence of commercial banks (government-owned, private), and rural banks; Distance to nearest bank/agent; Number of migrant workers abroad; Existence of recruitment agencies for migrant workers; Utilization of village financial systems

Source: [18,19,20].

). Some potential indicators (e.g., teacher–student ratio, household health expenditure) were excluded due to a lack of reliable or consistent village-level data. Vocational training access was prioritized instead, as it directly reflects skill-building opportunities linked to tourism and agriculture—the two dominant livelihood pathways in Manggarai Barat. This pragmatic selection ensures conceptual alignment with SRLA and data availability for spatial analysis.

Initial variables were processed, and 18 variables were confirmed to be significantly associated with the dependent variable (percentage of the poor population) based on the result of an ANOVA test (

Table 2. Capitals and Significant Indicators of SRLA.

Livelihood Capital	Indicators
Human Capital	Literacy rate, school attendance, number of high schools, distance to nearest university, incidence of extraordinary disease cases, etc.
Social Capital	Number of village-owned enterprises (BUMDes), group participation (proxied through the presence of community security measures), neighbourhood security activation, etc.
Natural Capital	Access to clean water (drinking and bathing), source of drinking water (categorized into five classes), presence of mangrove ecosystem, etc.
Physical Capital	Road accessibility, household toilet type, electricity usage (inverse of percentage of non-users), crime variation, distance to nearest hospital, etc.
Financial Capital	Household income level (via proxy), number of bank agents within a radius, access to a credit institution, etc.

). The ANOVA results showed that all 18 final indicators were statistically significant (p < 0.05), confirming their empirical relevance for inclusion in the Composite Poverty Index. Subsequently, each variable was classified according to its value range (for interval-scale data) or was based on its original category (nominal-scale data converted into ordinal classes), and then mapped to a five-point Likert-type scale:

Example: Classification of Selected Indicators

1.

Number of high school institutions:

• 5 = Worst condition (0.00–0.80).
• …
• 1 = Optimal condition (3.20–4.00).

2.

Source of drinking water:

• 1 = Branded bottled water.
• …
• 5 = River, rainwater, etc.

3.

Access to a financial credit institution:

• 1 = presence.
• 5 = absence.

4.

Distance to nearest higher education institution (km):

• 1 = Optimal condition (3.00–22.38).
• …
• 5 = Worst condition (80.52–99.90).

A variable index table maintains a complete classification rubric (see Appendix A).

2.3.2. Normalization

After classification, all indicators were normalized to a [0, 1] scale to ensure comparability across dimensions with different units and ranges. The Min-Max normalization formula was applied:

$$X^l={\displaystyle \frac{X-X_{min}}{X_{max}-X_{min}}}$$

where

• $X$ is the raw class score (1 to 5);
• $X^l$ is the normalized value;
• $X_{min}-1, X_{max}-5$.

This results in a scale where 0 = optimal condition and 1 = worst condition, aligning the directionality across all indicators for composite aggregation.

2.3.3. Weighting

Indicator weights are determined using the Analytical Hierarchy Process (AHP), a structured decision-making method that facilitates prioritizing weights based on expert judgment and pairwise comparisons [21]. This method allows for consistently evaluating multiple criteria and sub-indicators by quantifying their relative importance in influencing poverty outcomes.

To ensure contextual relevance and methodological rigor, AHP questionnaires were administered to a panel of 4 experts with professional experience in poverty alleviation and rural development in Indonesia. The panel included two academics, one local government agency’s policymakers, and one NGO practitioner. Experts were purposively selected based on their professional expertise. Respondents assessed the relative importance of each livelihood capital and indicator using Saaty’s 9-point scale.

The consistency ratio (CR) was calculated for each respondent’s matrix, with all values below 0.10, confirming logical consistency. Final weights were obtained by aggregating the normalized eigenvectors of the pairwise comparison matrices, reflecting the collective judgment of all experts.

This AHP-based approach provides a transparent and participatory weighting structure that aligns with both empirical priorities and local development perspectives, thereby strengthening the validity of the composite index used in the spatial poverty analysis.

2.3.4. Aggregation

The Composite Index of Poverty (CIP) was calculated using a linear weighted aggregation model:

$${CIP}_j=\sum _{i=1}^n{w}_i·{X\prime }_{ij}$$

where

• ${CIP}_j$ = composite index for unit $j$ (e.g., village);
• ${X\prime }_{ij}$ = normalized score of indicator $i$ or unit $j$;
• $w_i$ = final weight for indicator $i$.

Referring to the general model above, the equation for CIP can be written explicitly as:

$$\begin{array}{cc}\mathrm{CIP} = \hfill & {CW}_1\cdot \left[\right({IW}_1\cdot {\mathrm{X}}_1) + ({IW}_2\cdot {\mathrm{X}}_2) + ({IW}_3\cdot {\mathrm{X}}_3) + ({IW}_4\cdot {\mathrm{X}}_4\left)\right]\hfill \\ & + {CW}_2\cdot \left[\right({IW}_5\cdot {\mathrm{X}}_5) + ({IW}_6\cdot {\mathrm{X}}_6) + ({IW}_7\cdot {\mathrm{X}}_7\left)\right]\hfill \\ & + {CW}_3\cdot \left[\right({IW}_8\cdot {\mathrm{X}}_8) + ({IW}_9\cdot {\mathrm{X}}_9) + ({IW}_{10}\cdot {\mathrm{X}}_{10}) + ({IW}_{11}\cdot {\mathrm{X}}_{11}) + ({IW}_{12}\cdot {\mathrm{X}}_{12}\left)\right]\hfill \\ & + {CW}_4\cdot \left[\right({IW}_{13}\cdot {\mathrm{X}}_{13}) + ({IW}_{14}\cdot {\mathrm{X}}_{14}) + ({IW}_{15}\cdot {\mathrm{X}}_{15}\left)\right]\hfill \\ & + {CW}_5\cdot \left[\right({IW}_{16}\cdot {\mathrm{X}}_{16}) + ({IW}_{17}\cdot {\mathrm{X}}_{17}) + ({IW}_{18}\cdot {\mathrm{X}}_{18})]\hfill \end{array}$$

where

${CW}_1$: Human capital weight; ${CW}_2:$ Social capital weight; ${CW}_3:$ Physical capital weight; ${CW}_4:$ Natural capital weight; ${CW}_5:$ Financial capital weight; ${IW}_n$: Weight for indicator $n$; X1: Number of high schools; X2: Number of epidemic disease cases; X3: Distance to the nearest higher education; X4: Number of vocational training institutions; X5: Community Initiative on local security system; X6: Local security infrastructure;

X7: Variety of crime cases; X8: Main source of drinking water; X9: Main source of clean water for washing/bathing; X10: Type of sanitation facility; X11: Road access to production centers; X12: Percentage of households without on-grid electricity; X13: Frequency of disaster events; X14: Presence of mangrove vegetation; X15: Variety of surface water sources; X16: Access to microcredit; X17: Number of village-owned enterprises; X18: Distance to the nearest bank agent.

This produces a continuous index ranging from 0 (least poor) to 1 (most poor), which can be classified into quintiles or categorical classes: Very Low Poverty (0.00–0.20); Low Poverty (0.20–0.40); Moderate Poverty (0.40–0.60); High Poverty (0.60–0.80); and Very High Poverty (0.80–1.00) (Appendix

ANOVA Test Result a
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
18	(Constant)	3.313	8.890		0.373	0.710
	Main family toilet waste disposal site	2.408	0.638	0.223	3.776	<0.001
	Number of Senior High Schools (SMA), Islamic Senior High Schools (MA), and Vocational High Schools (SMK)	−4.597	1.679	−0.167	−2.738	0.007
	Activation of community-based security system (initiative-based)	−4.333	2.339	−0.123	−1.852	0.066
	Percentage of households without electricity	0.074	0.035	0.121	2.131	0.035
	R1208C—Small Business Credit (KUK) facility received by village residents	9.842	3.849	0.140	2.557	0.012
	R709AK–IK3—Number of people suffering from diseases	−0.331	0.083	−0.223	−4.000	<0.001
	R1402A—Number of Village-Owned Enterprises (BUMDes)	−2.853	1.087	−0.149	−2.625	0.010
	R1304A—Construction/maintenance of community security posts	−4.413	2.827	−0.111	−1.561	0.121
	R1209GK3—Distance to nearest bank agent	0.215	0.095	0.129	2.255	0.026
	Number of crime types (variation)	−3.951	1.791	−0.130	−2.206	0.029
	Trend in number of disaster events (2020–2021)	−0.065	0.022	−0.165	−2.950	0.004
	R701KK4—Distance to nearest Academy/Higher Education Institution	0.097	0.044	0.139	2.224	0.028
	Total number of vocational education institutions	−3.449	1.518	−0.126	−2.272	0.025
	Number of surface water sources (variation)	−1.878	0.917	−0.120	−2.048	0.042
	Main household drinking water source (1 = piped water utility/PDAM, 2 = wells, etc.)	2.291	0.818	0.301	2.799	0.006
	R308B2—Presence of mangrove trees (1 = present, 2 = absent, 3 = non-permanent)	3.447	1.562	0.139	2.207	0.029
	R507A—Main household drinking water source	−2.400	1.174	−0.236	−2.045	0.043

^a Dependent Variable: Poor households.

). These classifications can be spatially mapped using GIS tools to visualize geographic disparities and support spatially targeted interventions [22,23].

3. Results

3.1. Spatial Disparities in Multidimensional Poverty

This study adopts a multidimensional poverty approach grounded in the Sustainable Rural Livelihood Framework (SRLF) to provide a more comprehensive assessment of poverty conditions at the village level. The approach utilizes 18 indicators that represent the five core livelihood capitals, namely:

• Human Capital: Access to senior high schools (A), prevalence of communicable diseases (B), access to higher education (C), access to vocational training (D).
• Social Capital: Neighborhood safety initiatives (E), maintenance of local security systems (F), and variations in crime rates (G).
• Natural Capital: Trends in disaster occurrences (H), presence of mangrove ecosystems (I), and access to surface water sources (J).
• Physical Capital: Access to drinking water (K), access to clean water (L), sanitation facilities (M), road infrastructure quality (N), and access to electricity (O).
• Financial Capital: Access to financial credit (P), presence of village-owned enterprises (BUMDes) (Q), and access to banking services (R).

Each indicator was assessed at the village/urban village level across multiple subdistricts in Manggarai Barat Regency, including Komodo, Boleng, Sano Nggoang, Mbeliling, Lembor, Welak, Lembor Selatan, Kuwus, Ndoso, Kuwus Barat, Macang Pacar, and Pacar (Figure 2). The indicators were scored based on actual conditions using a five-point ordinal scale: very good (1), good (2), moderate (3), poor (4), and very poor (5). These scores were then aggregated to construct a Multidimensional Poverty Index (MPI) for each village.

The indicator-level results highlight disparities in drinking water, sanitation, road quality, and financial access. These individual patterns are important because they identify specific deprivations. This approach reveals significant spatial disparities in poverty levels across Manggarai Barat Regency. Key determinants such as access to clean water, road infrastructure, education, and financial services critically influence rural livelihoods and poverty outcomes. Although progress has been made in sanitation, electrification, and the development of BUMDes, persistent spatial inequalities—particularly in remote inland villages—underscore the urgent need for targeted, multi-sectoral interventions to ensure inclusive and sustainable poverty reduction.

The spatial analysis reveals clear patterns and contrasts across Manggarai Barat Regency. Remote inland subdistricts such as Welak, Ndoso, and Lembor Selatan consistently rank highest in deprivation across physical and financial capitals. These areas face poor road access, limited credit facilities, and weak market integration, making them particularly vulnerable to chronic poverty. In contrast, coastal and tourism-adjacent villages—especially around Labuan Bajo and Komodo—record lower poverty scores, largely due to better infrastructure, stronger access to services, and economic spillovers from tourism.

Human capital shows relative strength across much of the regency, with widespread access to vocational education and secondary schools. However, significant inequalities remain in access to higher education, where southern and western villages are disadvantaged by their distance from universities. Outliers such as Labuan Bajo, with strong educational infrastructure but poor health outcomes due to high disease prevalence, illustrate the uneven distribution of human capital benefits.

Physical infrastructure also demonstrates pronounced disparities. Villages such as Ndoso and Welak face “Very Poor” road conditions, creating isolation and restricting access to markets, schools, and health facilities. Conversely, peri-urban areas closer to Labuan Bajo and the coastal corridor show “Very Good” road access, reflecting concentrated investment in tourism-related infrastructure.

Financial inclusion remains weak overall. Despite widespread presence of banking agents in more accessible areas, many interior villages such as Kuwus and Welak continue to struggle with very limited access to microcredit and financial institutions. This exclusion constrains livelihood diversification and resilience. A few outliers, such as Golo Mori and Komodo, perform better than expected given their location, benefiting from strong village-owned enterprises (BUMDes) and integration into the tourism value chain.

Natural capital is another domain marked by strong contrasts. Coastal villages with mangrove ecosystems, such as Komodo and Boleng, enjoy natural protection and livelihood opportunities tied to fisheries and ecotourism. In contrast, mangrove-deficient areas remain highly exposed to erosion and flooding, intensifying vulnerability.

Access to safe drinking water and sanitation highlights one of the starkest divides. Remote hinterland villages such as Welak, Ndoso, Kuwus, and Lembor Selatan, depend on untreated wells, rivers, and rainwater, leaving them in the “Poor” and “Very Poor” categories. These communities are vulnerable to contamination, waterborne diseases, and prolonged drought conditions. Meanwhile, few peri-urban settlements like Labuan Bajo and Batu Cermin, benefit from piped or refillable water systems and score “Good” or better. This spatial contrast emphasizes persistence of water insecurity as a structural poverty driver in inland areas.

Closely linked to water access is the availability of infrastructure for bathing and washing. Over 70% of villages fall into the “Fairly Poor” category, often relying on rivers or unprotected wells. This again underscores serious hygiene-related health risks, particularly for children and vulnerable populations, and exacerbates the intergenerational cycle of poverty. Conversely, urban and tourism-adjacent communities such as Labuan Bajo enjoy “Very Good” infrastructure, highlighting the uneven distribution of public investments.

Road infrastructure is a key factor in poverty, especially in remote areas like Ndoso, Kuwus Barat, Pacar, and Welak. Villages here are labeled as having “Very Poor” access, and many routes become impassable during the rainy season. This isolation limits access to markets, schools, health facilities, and government services, which deepens poverty and stagnation. In contrast, coastal and peri-urban areas such as Macang Tanggar and Wae Kelambu have “Very Good” road conditions, often due to investments related to tourism. These improvements boost connectivity and economic mobility. Villages with intermediate road quality, categorized as “Good” or “Poor,” are precarious. Their infrastructure allows for some access but is still at risk of disruption from environmental factors.

Education plays a vital role in breaking the cycle of poverty. Access to senior high school education is relatively strong in many parts of Manggarai Barat, including urban and rural areas. Villages such as Labuan Bajo, Golo Keli, and Wae Lolos are categorized as “Very Good,” reflecting the presence of more than three senior high schools. However, access is still uneven. Villages like Batu Cermin and Golo Bilas are classified as “Poor” or “Moderate,” highlighting barriers to human capital development that can inhibit long-term socio-economic mobility.

Vocational education access shows similar patterns. Over 95% of villages are categorized as “Very Good,” indicating high availability of vocational institutions supporting skills training for employment in agriculture, tourism, and service sectors. Yet, outliers such as Desa Golo Mbu and Desa Bari fall into “Moderate” or “Very Poor” categories, exposing pockets of exclusion that risk being overlooked in broader regional development plans.

Financial inclusion presents a complex picture. Although banking agent access is classified as “Very Good” in over 85% of villages, thanks partly to national programs like Laku Pandai and BRILink, true financial inclusion remains elusive. Credit access remains critically low across most of the regency. Most villages are categorized as “Poor,” particularly in inland areas like Welak, Kuwus, and Ndoso, where banks, cooperatives, and microfinance services are largely absent. This financial exclusion limits the ability of residents to invest in livelihoods, manage crises, and escape poverty.

Higher education access presents perhaps the most significant spatial inequality. While villages near Labuan Bajo benefit from strong proximity to universities and academies, only 17.8% of villages fall into the “Very Good” accessibility category. In contrast, nearly 30% of villages, particularly in interior subdistricts such as Macang Pacar, Lembor Selatan, and Welak, have “Poor” or “Very Poor” access, located over 60 km from the nearest institution. These distances create barriers to higher education enrollment, reduce opportunities for social mobility, and perpetuate generational poverty.

Access to diverse surface water sources like rivers, springs, and lakes also shapes poverty dynamics. Villages like Desa Komodo and Pasir Panjang benefit from “Very Good” access, supporting agriculture and domestic needs. However, many villages, particularly in Komodo, Sano Nggoang, and Lembor, fall under “Poor” or “Very Poor” classifications, exposing them to seasonal shortages and reduced livelihood options. These disparities reinforce the importance of water resource diversification in reducing rural vulnerability.

Electricity access is one area in which Manggarai Barat has made substantial progress. Approximately 86% of villages have “Very Good” electrification, enhancing education, health, and economic activity. Nonetheless, villages such as Loha (Pacar), Watu Galang (Mbeliling), and Golo Poleng (Ndoso) remain without reliable grid access, classified as “Poor” or “Very Poor.” Without targeted investment, these areas risk falling further behind.

Sanitation access paints a more optimistic picture, with over 95% of villages using private household latrines and being categorized as “Very Good.” This success, primarily attributed to the STBM (Sanitasi Total Berbasis Masyarakat) initiative, demonstrates how targeted public health programs can have a widespread impact. However, villages such as Papagarang and Golo Keli still fall into the “Very Poor” category, indicating persistent challenges in specific geographic or densely populated areas.

Village-Owned Enterprises (BUMDes) are also a strong point in Manggarai Barat’s development landscape. More than 90% of villages are classified as “Very Good,” hosting 4–5 active enterprises that stimulate local economies and generate village revenue. Yet a handful of villages—Golo Kempo, Golo Pua, and Tanjung Boleng—remain in the “Poor” or “Very Poor” categories, often due to limited human capital, weak governance, or geographic remoteness. Strengthening these institutions could help bridge economic gaps.

Environmental poverty, measured through mangrove presence, also plays a role. Although only 27 of the regency’s villages are coastal, those with healthy mangrove ecosystems—such as parts of Komodo and Boleng—benefit from natural protection and support for fisheries-based livelihoods. Coastal villages like Labuan Bajo and Watu Waja without mangroves face higher exposure to erosion and flooding, amplifying their economic vulnerability.

Public health indicators, particularly disease prevalence, show generally positive trends. About 97.2% of villages are classified as “Very Good” with low epidemic rates. This suggests effective rural disease control and supports reduced healthcare costs for households. However, due to its urban density and tourism influx, Labuan Bajo is an outlier with “Very Poor” classification, highlighting the need for targeted health interventions in growing urban centers.

Finally, low crime variation across 85% of villages classified as “Very Good” supports a secure environment that can attract investment and foster community well-being. However, isolated hotspots in Batu Cermin, Komodo, and Lembor subdistricts point to pockets of socio-economic stress that may require integrated social and security policies.

Overall, the results show a dual geography of poverty in Manggarai Barat. Remote inland villages face persistent deprivation across multiple capitals, while coastal and tourism-adjacent settlements enjoy relative advantages. At the same time, several outlier cases demonstrate that targeted investments in education, water systems, and local enterprises can significantly improve poverty outcomes, even in challenging locations.

3.2. Composite Poverty Index Distribution

The study began by identifying a wide range of variables previously recognized in the literature as relevant to rural poverty to construct the Composite Livelihood Capital Index. These candidate indicators were then statistically tested using multiple linear regression analysis, with the poverty rate in each village as the dependent variable. This step ensured empirical relevance by retaining only variables significantly associated with poverty levels. Following this, a multi-criteria decision-making approach was employed to determine the relative importance of each capital type and its associated indicators. Specifically, the Analytic Hierarchy Process (AHP) was applied, involving experts who evaluated and weighted the significance of each component based on its perceived contribution to livelihoods and vulnerability (

Table 3. Weights of Livelihood Capital and Their Indicators Based on AHP Result.

Livelihood Capital	Capital Weight	Indicator	Indicator Weight	IW × CW
	(CW)		(IW)
Human Capital	0.371	Number of senior high schools (SMA/MA/SMK)	0.368	0.137
		Number of Epidemic Disease Cases	0.14	0.052
		Distance to the nearest higher education institution	0.153	0.057
		Number of Vocational training institutions	0.339	0.126
Social Capital	0.239	Activation of neighbourhood security system initiated by residents	0.11	0.026
		Construction/maintenance of local security systems	0.309	0.074
		Number of crime type variations	0.581	0.139
Natural Capital	0.123	Trend in the number of disaster events	0.261	0.032
		Presence of mangrove vegetation	0.411	0.051
		Number of surface water source types available	0.328	0.04
Physical Capital	0.167	Main drinking water source	0.189	0.032
		Primary water source for bathing/washing	0.257	0.043
		Predominant type of sanitation facility used	0.276	0.046
		Road access from production areas to trade centers	0.095	0.016
		Percentage of households not using electricity	0.183	0.031
Financial Capital	0.1	Availability of small business credit facilities (KUK)	0.548	0.055
		Number of village-owned enterprises (BUMDes)	0.241	0.024
		Distance to the nearest banking agent	0.211	0.021

).

According to the result shown in Table 3, human Capital emerged as the most influential category with a weight of 0.371, driven particularly by the availability of senior high schools and vocational training institutions. Social Capital followed, highlighting the significance of crime variation and local security infrastructure. Physical Capital and Natural Capital were weighted moderately, reflecting the importance of water, sanitation, electricity, road access, and environmental conditions such as mangrove presence and disaster trends. Although relatively lower in weight, Financial Capital emphasized access to credit and local economic institutions such as BUMDes. These weighted components were used in subsequent index construction to assess village-level livelihood resilience and poverty risk.

To assess multidimensional deprivation across villages in the study area, a Composite Index of Poverty (CIP) was constructed using a weighted linear aggregation of normalized indicators. CIP synthesizes the multiple dimensions of SLRA into a single measure, providing added value beyond indicator-level analysis. While individual indicators identify sector-specific weaknesses, the CIP reveals how these deprivations overlap spatially and cumulatively. It distinguishes between isolated sectoral problems and compounded, multidimensional poverty risks. The resulting index, scaled from 0 (least poor) to 1 (most poor), captures relative poverty intensity across villages and suburbs in Manggarai Barat Regency. The index was subsequently classified into five categories: Very Low, Low, Moderate, High, and Very High poverty, following quintile-based thresholds (Figure 3).

The Composite Index of Poverty (CIP) (Figure 4), constructed through weighted linear aggregation, confirms these patterns. The majority of villages (over 75%) fall within the “Moderate” poverty category (CIP: 0.40–0.60). This reflects chronic, spatially distributed deprivation rather than extreme poverty. None of the villages exceed a CIP of 0.60, suggesting the absence of very high poverty clusters.

Villages near Labuan Bajo, Komodo, and parts of Mbeliling stand out with lower CIP values (<0.35), benefiting from stronger infrastructure and services. Conversely, upper-moderate poverty scores concentrate in inland subdistricts such as Welak, Lembor Selatan, and Pacar. Outliers (approximately 15% of villages) like Seraya Maranu (0.29), Pasir Putih (0.32), and Liang Ndara (0.33), show unexpectedly low poverty levels due to robust community enterprises and stronger education services, while Golo Ronggot (0.55), Golo Lajang Barat (0.51), Rancang Welak (0.51), and Wae Mowol (0.52) are among the most deprived despite not being the most geographically isolated (Figure 5).

These spatial patterns highlight a narrow poverty band across the district. Deprivation is spread out moderately, without extreme differences. However, the steady presence of villages at the upper end of the moderate range shows the need for focused help to stop poverty from worsening. This is especially important in remote or peri-urban areas, where access to services or job options may be limited.

These findings also show that both spatial inequalities and the role of targeted assets in shaping poverty outcomes. They suggest that while deprivation is widespread, strategic interventions in water, infrastructure, education, and finance can shift poverty trajectories, especially in high-risk inland subdistricts.

The CIP model’s output gives a solid foundation for spatial visualization and prioritization. This helps local policymakers allocate resources to moderately poor clusters while learning from the best-performing villages. Using this composite index supports recent studies that promote detailed and precise poverty diagnostics as a tool for targeted planning and regional fairness [22,23].

4. Discussion

4.1. Linking Tourism Investment and Persistent Poverty

A central paradox in Manggarai Barat is that despite massive tourism investment, multidimensional poverty remains widespread. The results of this study confirm that tourism benefits are spatially concentrated around Labuan Bajo and other peri-urban coastal areas. These locations enjoy stronger infrastructure, service delivery, and financial linkages, which contribute to lower CIP scores. However, remote inland subdistricts such as Welak, Ndoso, and Lembor Selatan remain largely excluded from the tourism economy due to weak connectivity, limited financial services, and low integration into value chains.

This uneven distribution reflects a lack of spillover effects from tourism. Infrastructure development is often concentrated in tourism corridors, leaving hinterland villages isolated. Moreover, there is a persistent skills mismatch between the labor demand in tourism (hospitality, language, digital skills) and the skills available in rural villages, which remain oriented toward agriculture and subsistence livelihoods. As a result, tourism investment reinforces, rather than resolves, existing inequalities.

4.2. Drivers of Spatial Inequalities Across Multidimensional Capitals

This study underscores the critical role of spatial heterogeneity in shaping multidimensional poverty across Manggarai Barat Regency. The uneven distribution of natural, physical, human, and financial capitals demands targeted, context-specific interventions to foster sustainable rural livelihoods and equitable development outcomes. Drawing on the sustainable livelihoods framework [17,24], we argue that addressing disparities in asset endowments is essential to poverty alleviation and long-term resilience.

The spatial distribution of mangrove ecosystems reveals significant environmental and economic implications. Villages with intact mangrove cover benefit from vital ecosystem services, including coastal protection, biodiversity support, and fishery productivity [25,26]. These areas present opportunities for ecosystem-based livelihood programs such as community-based conservation, ecotourism, and sustainable fisheries [27,28]. Conversely, mangrove-deficient coastal zones require reforestation efforts integrated into broader climate adaptation and disaster risk reduction strategies [29].

Access to clean water emerges as another critical dimension of spatial inequality. Villages classified as “Very Poor” and “Fairly Poor” in drinking water availability face heightened health risks and reduced livelihood productivity. Targeted interventions such as rainwater harvesting, deep-well construction, and piped water systems have proven effective in similar rural contexts [30]. In highly vulnerable areas, interim solutions like mobile water delivery and subsidized refillable containers are essential for preventing disease and ensuring dignity. Access to clean water underpins improvements in health, education, and productivity, key levers for breaking intergenerational poverty cycles [31].

While sanitation services appear to be a relative strength in Manggarai Barat, with most villages classified as “Very Good”, persistent gaps in locations such as Papagarang and Golo Keli highlight the need for continued investment. Achieving Open Defecation Free (ODF) status requires not only infrastructure support but also behavior change campaigns and community-based financing mechanisms [32]. The regency’s progress offers a model for rural sanitation success in Eastern Indonesia, yet sustaining these gains necessitates investment in maintenance, training, and wastewater management.

Transportation infrastructure exhibits stark spatial disparities. Poor road conditions in interior areas such as Ndoso and Welak inhibit access to markets, schools, and healthcare facilities, reinforcing multidimensional poverty. These findings align with previous research highlighting road quality as a determinant of rural development and economic integration [33,34]. Strategic investments in road rehabilitation, all-weather corridors, and innovative solutions like riverine transport or cableway systems in mountainous zones are needed to overcome physical isolation and support inclusive growth.

Financial capital remains one of the most unevenly distributed livelihood assets in Manggarai Barat. Although programs like Laku Pandai and BRILink have extended banking access, financial exclusion persists in many interior and agriculture-dependent communities. This limits households’ capacity to invest in income-generating activities and buffer economic shocks [35,36]. Expanding mobile banking services and incentivizing agent presence in underserved areas are critical for inclusive financial development.

Human capital indicators, particularly access to vocational and senior secondary education, show encouraging coverage but reveal lingering qualitative and spatial disparities. In remote or sparsely populated villages, limited access to qualified teachers, learning materials, and relevant vocational tracks curtails educational outcomes. Policy solutions should include targeted expansion of institutions, alignment with local economies (e.g., tourism, agriculture, fisheries), and integrated youth employment strategies [37].

Electricity access remains a crucial determinant of quality of life and development potential. The spatial variation in electrification reflects uneven infrastructure rollout, particularly in peripheral villages. As a proxy for broader service delivery, electricity availability impacts educational attainment, business development, and household well-being [38]. Investing in localized, renewable energy solutions like solar mini-grids could bridge energy gaps and accelerate inclusive development.

The pattern of surface water access highlights ecological and livelihood vulnerabilities. Inadequate access undermines daily consumption, irrigation, and resilience to climate shocks. Strengthening watershed protection and water infrastructure—particularly in “Poor” and “Very Poor” villages—is essential for sustaining natural capital and mitigating spatial environmental inequality [39].

Institutional capital, represented by the proliferation of Village-Owned Enterprises (BUMDes), reflects rural governance and innovation progress. However, poor-performing villages reveal gaps in local capacity, underscoring the importance of tailored training and technical assistance. Strengthening BUMDes can enhance village autonomy, stimulate rural entrepreneurship, and anchor local economic development [40,41].

In public health, the low prevalence of epidemic diseases in rural areas contrasts sharply with higher rates in urban Labuan Bajo. This discrepancy may reflect both effective rural health governance and potential underreporting. Labuan Bajo’s status as a tourism hub necessitates stronger urban health infrastructure, disease surveillance, and preventive services [42]. Ensuring robust data collection systems across settings will be critical for monitoring spatial health inequalities.

Higher education access remains limited for residents of western and southern subdistricts. Strategies to promote inclusion include satellite campuses, distance learning platforms, and mobility subsidies. The integration of higher education planning with spatial poverty data can guide multisectoral interventions and support regional human capital development [43,44].

Finally, the success of agent-based banking models in Manggarai Barat offers lessons for other service sectors. While gaps remain in hard-to-reach areas, initiatives like mobile banking units, subsidized agent deployment, and infrastructure co-location with BUMDes could expand coverage. These innovations exemplify how cross-sectoral learning can enhance rural service delivery and poverty alleviation.

Overall, the spatially explicit analysis presented in this study provides actionable insights into the distribution of livelihood assets and deprivation in Manggarai Barat. Bridging these gaps requires integrated, multi-capital strategies aligned with local contexts. By investing in infrastructure, institutions, and ecosystem services, the regency can move closer to achieving sustainable, inclusive, and resilient rural development.

4.3. Interpretation of the Composite Poverty Index and Policy Implications

The findings from the Composite Index of Poverty (CIP) underscore the complexity of rural poverty in Manggarai Barat Regency, where deprivation manifests in a spatially moderate but persistent manner. The prevalence of villages in the “Moderate” poverty category (CIP: 0.40–0.60) suggests a condition of chronic deprivation that lacks the dramatic indicators of extreme poverty. Nonetheless, it represents a significant barrier to well-being and development. This aligns with Chambers and Conway’s (1992) [24] seminal definition of sustainable rural livelihoods, which emphasizes survival and the ability to cope with and recover from stress and shocks. Villages that hover near the upper boundary of the “Moderate” category may be particularly vulnerable to shocks, such as climate variability, economic downturns, or health crises, that could rapidly exacerbate their poverty conditions [24].

The use of a composite index based on livelihood capitals is consistent with the Sustainable Livelihoods Framework (SLF), which posits that poverty is multidimensional and driven by unequal access to human, social, physical, natural, and financial capital [17]. In this study, the high weight of Human Capital, particularly secondary education and vocational training, reinforces existing evidence that educational access and quality are pivotal in rural poverty alleviation [45,46]. The substantial contributions of vocational education infrastructure resonate with recent findings emphasizing the role of skills development in enhancing employability and entrepreneurship in rural contexts [47,48].

The spatial analysis reveals that certain villages consistently underperform across multiple indicators even within a relatively narrow band of poverty levels. These findings align with the growing body of literature arguing that poverty is both place-based and path-dependent [49,50]. In the case of Manggarai Barat, several subdistricts—such as Welak, Kuwus, Lembor Selatan, Mbeliling, and Ndoso—are not among the poorest regions when viewed through the lens of income or expenditure. However, their persistently high Composite Index of Poverty (CIP) values indicate structural disadvantages beyond conventional poverty measurements. These disadvantages may stem from geographic isolation, inadequate infrastructure, limited access to markets and services, and weak institutional support. Such conditions reflect what the literature refers to as geographic poverty traps or development traps, in which poverty becomes self-perpetuating due to spatial and systemic constraints [51]. As such, policy responses in these areas must go beyond income transfers or economic assistance, instead tackling the entrenched barriers that limit mobility and opportunity.

The relatively low weight assigned to Financial Capital in the AHP analysis reflects broader trends in rural Indonesia, where formal financial inclusion remains uneven despite policy advances [36]. Access to financial services, including credit, savings, and insurance, is essential for smoothing consumption and enabling productive investment [52,53]. The finding that financial indicators contributed less to the overall index suggests a lack of infrastructure and a potential mismatch between existing financial services and local livelihood needs.

From a natural capital perspective, the inclusion of mangrove presence and surface water diversity as key indicators reflects their role in both ecosystem service provision and climate resilience [54,55]. Villages with poor scores on these indicators are environmentally vulnerable and less able to buffer livelihood shocks, particularly those reliant on fisheries, agriculture, or ecotourism. This supports the argument by Adger (2003) that environmental degradation and poverty are tightly interlinked, particularly in peripheral rural economies [56].

The relatively strong performance of Social Capital, highlighted through indicators such as crime variation and community-based security systems, offers an essential insight into informal institutional resilience. As Woolcock & Narayan (2000) and Putnam (2000) have argued, social cohesion and local governance can substitute for weak formal institutions and play a critical role in poverty mitigation, especially in areas lacking consistent state presence [57,58].

Moreover, the CIP’s utility lies in diagnosis and strategic planning. The index provides a spatially explicit template for place-based policy targeting, echoing Alkire, Roche, and Vaz (2017) [59] call for multidimensional poverty indices to be used in subnational planning. Identifying high-CIP villages within an otherwise moderate poverty landscape enables targeted interventions that are both equitable and efficient. This spatially nuanced approach is essential for achieving the Sustainable Development Goals (SDGs), particularly SDG 1 (No Poverty), SDG 10 (Reduced Inequalities), and SDG 11 (Sustainable Cities and Communities) [59].

Finally, the absence of villages in the “High” or “Very High” poverty categories may indicate some degree of basic service reach and policy success, yet it also raises questions about data sensitivity and indicator calibration. As noted by Alkire and Foster (2011) [46], the ability of a multidimensional poverty index to distinguish severity is crucial for effective policy response. Future refinements could incorporate temporal dynamics or shocks (e.g., COVID-19, climate extremes) to test the robustness of current classifications [46].

Even though this composite approach offers a robust analytical lens to examine rural poverty in its multidimensional and spatially differentiated forms, it confirms the critical role of human capital development, the persistent lag of financial and natural capital, and the importance of integrated, localized interventions. Addressing rural poverty in Manggarai Barat will require coordinated action across sectors, guided by empirically grounded, spatially explicit diagnostics such as the CIP.

Overall, these findings call for a shift from generic poverty reduction strategies to targeted, place-based interventions that address structural inequalities. By addressing the capital-specific drivers, policies can move beyond income transfers toward systemic poverty alleviation. More importantly, integrating inland villages into the benefits of tourism and national investment is essential to resolving the paradox of “growth without inclusion” in Manggarai Barat.

5. Conclusions

This study contributes to multidimensional poverty research by integrating the Sustainable Rural Livelihoods Approach (SRLA) with spatial analysis. The novelty lies in demonstrating how livelihood capitals can be operationalized into a composite index that not only measures deprivation but also maps its geographic clustering. This approach is especially valuable in contexts where economic growth and investment—such as tourism in Manggarai Barat—coexist with persistent poverty.

This study provides a detailed look at poverty in Manggarai Barat Regency by creating a Composite Index of Poverty (CIP) based on the Sustainable Rural Livelihoods Approach (SRLA). The CIP combines indicators from five types of livelihood capital: human, social, natural, physical, and financial. This gives a better view of rural deprivation compared to traditional income-based measures. The results show that while absolute poverty is low in most villages, many communities face moderate deprivation. This situation makes them vulnerable to structural problems and risks from outside events.

The spatial analysis revealed significant inequalities in the distribution of assets. This is especially clear in natural capital, such as mangrove coverage and access to surface water, physical infrastructure like road quality and electricity, and financial access. These differences highlight challenges related to geographic isolation, local environmental issues, and institutional barriers that hinder adaptation and the ability to diversify livelihoods. On the other hand, stronger performance in human and social capital indicates areas of resilience that could be used to support broader strategies for reducing poverty.

By making the CIP both multidimensional and spatially explicit, this study offers policymakers a practical tool for diagnosing and addressing poverty. The findings confirm existing evidence on spatial inequality while also highlighting the limits of income-based measures. They suggest that effective poverty alleviation must target overlapping deprivations through place-based strategies. Future research should continue refining indicator sets and thresholds to enhance sensitivity, ensuring that the most vulnerable communities are not overlooked.

While the CIP has demonstrated utility as a diagnostic and policy design tool, this study has several limitations that warrant consideration. First, although the Composite Index of Poverty captures multidimensional deprivation, it is based on cross-sectional data and thus cannot reflect temporal dynamics or household-level changes. Second, reliance on secondary data may underrepresent informal or undocumented aspects of rural livelihoods. Third, although expert judgment contributed to contextual weighting, the modest sample size of experts may not fully reflect the diversity of perspectives. Lastly, the absence of villages categorized as “High” or “Very High” poverty suggests that indicator calibration may require refinement using longitudinal and shock-responsive data. These limitations highlight areas for further development in applying composite indices for rural poverty analysis.

The CIP framework presented in this study offers a strong foundation for spatially and contextually integrated strategies to resolve rural poverty in West Manggarai. It enables policymakers and practitioners to identify where and how deprivation is most entrenched and which forms of capital most need support. Only through integrated, data-driven, and locally tailored interventions can sustainable, inclusive, and resilient rural development be realized in this and similar contexts across Indonesia.