1. Introduction

Poverty has perennially remained one of the most persistent social issues that has yet to be fully resolved. Haughton and Khandker (2009) argue that it exerts a negative impact on societal growth and economic prosperity, posing significant challenges to mitigation efforts. Despite numerous attempts by economists, social scientists, and humanitarians to explain the root causes of global poverty and advocate for necessary reforms, quantifying and predicting poverty at localized levels or within Indigenous communities presents a more difficult challenge, compounded by a scarcity of published literature on the topic.

Indigenous peoples have been an integral part of human civilization since its inception, with poverty intertwined in societal structures from the beginning. These communities, existing since time immemorial, hold a vital place in society. Alongside principles of self-governance, empowerment, social justice, human rights, and cultural integrity, they also possess rights to ancestral territories and lands (Andaya, 2023; Republic Act 8371). It is imperative that native communities are respected by the broader public and afforded official protection. As their rich culture and traditions endure, efforts must be made to address their poverty and foster sustainable growth (Onsay, 2022).

In order to quantify the multidimensional socio-economic deprivations and poverty conditions in their area and develop effective policy interventions that can reduce poverty and advance the economic development of our Indigenous people, researchers have been gathering the data and statistics they need from extensive community-based monitoring systems. The local government owns the community-based monitoring system (cbms), a system for gathering and evaluating data at the community level. Policymaking, local planning, service delivery, and impact evaluation all make use of the cbms’s outputs (Reyes et al., 2014; Sobreviñas, 2017). There is no thorough census or study that is unique to the Indigenous population. Thus, to reduce poverty and advance economic development, we use, process, and produce crucial data on Indigenous people. Data from the social sciences and humanities are crucial for problem assessment and opportunity analysis in contemporary society, both historically and prospectively (Sawkins & Smith, 2023).

Numerous scholars have attempted to measure the unmeasurable in the field of natural sciences (Eid, 2008; Gruijters et al., 2018; Rabajante, 2020; Wu et al., 2022). However, social difficulties and economic problems are hard to measure, difficult to evaluate, and sometimes even considered unmeasurable (United Nations, 2012). A limited amount of research has been conducted on the complex aspects of poverty among Indigenous peoples in the Philippines. No research has been undertaken to quantify poverty using predictive analytics and advanced econometrics in the Bicol region.¹ Therefore, the dataset is crucial for addressing the root causes of multidimensional poverty and promoting economic growth among Indigenous peoples, not only in this locale but also around the world, by leveraging the models and indicators utilized in this study.

To effectively monitor progress towards UN Sustainable Development Goal No. 1 (No poverty) and related objectives, it is crucial to ensure the disaggregation of national poverty measures by subpopulation groups, including Indigenous and non-indigenous communities, ethnicities, and other relevant categories. Therefore, the utilization of community-focused datasets highlighted in this article becomes imperative and valuable (Allen et al., 2018; Onsay & Rabajante, 2024b). This article, along with its associated Indigenous People datasets, enables the extension of previous empirical work and facilitates advanced quantitative analyses, such as econometrics, machine learning, and quantitative social science modeling, for both local and international Indigenous communities.

2. Problem

Poverty among Indigenous peoples varies greatly around the globe, so poverty reduction efforts must be tailored accordingly (Hall & Patrinos, 2012). Currently, there is a lack of research focusing on quantifying multidimensional poverty among Indigenous populations in the Philippines. In contrast, studies in developed nations like Australia have highlighted significant economic development challenges faced by Indigenous peoples (Fuller et al., 2007). Similarly, research has shown that a large proportion of Latin American Indigenous communities live in extreme poverty (Psacharopoulos & Patrinos., 1994). However, existing studies often operate at a macroeconomic level, are susceptible to sampling errors, and primarily inform broad national policies. In the Philippines, research on Indigenous communities is limited and primarily qualitative in nature. For instance, Tindowen (2016) conducted research on 25 Aeta families in the northern Philippines, revealing distinct socio-economic characteristics and limited technological access compared to rural and urban populations. In Southern Luzon, there is a notable absence of comprehensive quantitative assessments for the Agta Tabangnon² indigenous group, who have peacefully inhabited the region since the pre-colonial era, due to data scarcity. Given these gaps, the datasets emphasized in this research are crucial for the socioeconomic advancement of Indigenous communities and can serve as a blueprint for replication by various Indigenous societies in the Philippines and beyond.

While Indigenous people are represented in Philippine national living standards surveys such as The Family Income and Expenditure Survey (fies), they are underrepresented due to reliance on sample selection. Additionally, these surveys lack comprehensive information about the standard of living of Indigenous communities. The Annual Poverty Indicator Survey (apis) includes some information about Indigenous communities using a sampling design, but it lacks comprehensive health information. The Labor Force Survey (lfs) also suffers from underrepresentation and focuses solely on labor-related data. The current surveys in the Philippines are not very useful for Indigenous communities. More often, Indigenous communities are underrepresented, and they lack economic, demographic, and social data. The current surveys conducted by the National Statistics Agency are not designed to cater specifically to Indigenous communities. In fact, there is no survey specifically designed to capture the overall conditions of Indigenous people (Onsay & Rabajante, 2024a). The current census conducted by the Indigenous People Office is carried out manually, with information written on bond paper using a pen. These records are susceptible to fading, loss, and damage from moisture, and can be easily overlooked or misrepresented. This survey only captures the name of the household head and their members, the age of each member, and their address. It does not include data pertaining to sex, income, and other vital social and economic statistics. Thus, this dataset cannot be used for modeling, inferential analysis, or advanced analysis. However, the datasets we present here are unique, highly reliable, and relevant. Obtaining data for Indigenous communities is particularly difficult, especially in mountainous or remote areas. Nevertheless, the authors have displayed innovation and resourcefulness by filtering, cleaning, and transforming the extensive community-based monitoring system to generate data exclusively for Indigenous communities. The data we provide here are comparable to other datasets in the Philippines as the Community-Based Monitoring System (cbms) is being implemented nationwide (Republic Act 11315 [2018]). Furthermore, the cbms network is implemented in more than 20 countries worldwide, allowing for the generation of similar datasets by other researchers (cbms Network, 2020). These datasets are valuable for achieving the UN Sustainable Development Goals (sdg s) by 2030 and can be used in conjunction with other datasets to assess progress towards those goals. The data we produce and the work we put forth can be utilized by other researchers and replicated in different regions that have Indigenous communities.

3. Data

1. –Measuring the Unmeasurable Multidimensional Socio-Economic Deprivations and Poverty Predictions deposited at Harvard Dataverse – doi: www.doi.org/10.7910/DVN/QSZKUP
2. –Temporal coverage: 2018–2020

The datasets in the repository contain 8 components, namely: the multidimensional poverty-comprehensive data; the population dynamics; the Indigenous people analytics; the poverty computational data; the data dictionary; the graph of poverty results; the table of poverty indices; and the prescriptive policy analytics. The multidimensional poverty variables are contained in an Excel file with 64 indicators from 1,660 Households of Indigenous People and 8,022 Individuals from 12 tribes of the 3 sectors (Isarog, Ranggas, and Salog). The application of indicator deprivation cutoffs and poverty cutoffs will facilitate the assessment of multidimensional socio-economic deprivations and poverty predictions. The cbms Simple Composite Indicator (sci) can be adopted (Sobreviñas, 2017). The sci consists of four clusters: Severely poor (out of 15 indicators, deprived in at least 7 basic needs); Poor (deprived in at least 4 basic needs); Near Poor (deprived in 1–3 basic needs); and Not Poor (not deprived in any of the basic needs). Income poverty and food poverty, as measured by income and expenses, are based on the country (province) poverty line (threshold), usually determined by the governing national statistics agency of a country. It is important to remember that different countries have different poverty lines. Thus, when applying the models and utilizing the data, it is vital to establish the poverty line. Moreover, other deprivation cut-offs can be applied. Localized multidimensional composite cut-offs, as developed by the authors, have five groups: Extremely poor (deprived of 13–15 basic needs); Very Poor/Not well-off enough standard of living (deprived of 10–12 basic needs); Poor/barely enough standard of living (deprived of 7–9 basic needs); Moderately Poor/A not well-to-do standard of living (deprived of 4–6 basic needs); Slightly poor/A well-to-do standard of living (deprived of 1–3 basic needs); and Not Poor or a Wealthy Life (not deprived of any basic needs) (Onsay & Rabajante, 2024b).

A multidimensional poverty index (mpi) can be computed using various approaches, such as via theoretical frameworks or modifications (Alkire & Foster, 2011; Foster et al., 1984; Sen, 2006; Onsay, 2022; Sobreviñas, 2017). This index encompasses three key dimensions at the household level: health, education, and standard of living. To identify individuals living in poverty, a dual cutoff method is employed, involving cutoff levels within each dimension and a cutoff based on the number of dimensions in which a person must be deprived to be considered multidimensionally poor. Assigning weights to indicators is a viable option (Alkire & Foster, 2011). In this approach, each dimension—health, education, and standard of living—is assigned an equal weight of 1/3. Individuals are classified as multidimensionally poor when their family experiences a weighted sum of 0.3 or more in terms of deprivation across the dimensions. It is worth noting that these cutoffs and the index itself can be adapted and modified when applied to other local contexts or marginalized groups, such as Indigenous populations. In this study, the deprivation indicators are confined to widely used metrics found in global and national Multidimensional Poverty Indices (mpi s). This decision stems from the similarities observed between the situations of Indigenous communities and the broader population in the Philippines. Despite their distinct cultures, Indigenous groups are integrated into the same societal framework, subjected to comparable policies, encountering similar economic hardships, and engaging in political processes akin to other households. Therefore, certain critical indicators unique to Indigenous communities, such as land area, ancestral domains, clothing design, and arts, have been omitted from the analysis. This exclusion is influenced by their shared land ownership, particularly evident in the Mount Isarog barangays and Natural Park. The primary focus of the analysis revolves around economic and social science aspects, deliberately sidelining cultural and artistic considerations due to their more abstract and subjective nature. This decision has been made to uphold the precision and integrity of future modeling and analytical endeavors. Furthermore, the datasets also included Indigenous people analytics, both descriptive and diagnostic, covering 12 tribes with measurements of magnitude and proportion. These datasets were curated, mined, cleaned, filtered, transformed, coded, analyzed, and processed by researchers within the extensive community-based monitoring system. The dataset contained numerous variables that were not pertinent to Indigenous policy-making. Thus, researchers utilized various software tools such as R, stata, Python, spss, and ms Excel to efficiently generate the required data for economic development. Additionally, they employed descriptive, diagnostic, predictive, and prescriptive models to extract and analyze data, providing valuable insights for economic advancement.

The Philippines is home to hundreds of distinct Indigenous groups. However, it is crucial to note that the data presented in this article focuses specifically on the population and households of Agta Tabangnon residing in Mount Isarog, Bicol Region. The dataset comprises 66 columns and 1,279 rows (66 × 1,279), which have been coded and transformed into easily comprehensible cross-sectional data by referring to the data dictionary. Through the codes prepared by the researchers, data analytics and econometric modeling can be easily conducted. The dataset is hosted in the Harvard Dataverse Repository (https://dataverse.harvard.edu) under the reference www.doi.org/10.7910/DVN/QSZKUP, Harvard Dataverse, V2. The data included in this database were collected between 2018 and 2020 through the Community-Based Monitoring System, conducted once every three years. This system, under the ownership of the Local Government Unit of Goa, Camarines Sur, Philippines, is publicly accessible, particularly for academia, which has been instrumental in data analysis and policy formulation. Data collection and analysis in social sciences and economics are often resource-intensive and challenging endeavors. Public enumerators, supervised by the local government, are responsible for collecting this data. While the basic analysis and processing are overseen by the planning office, the more advanced analyses are entrusted to academia, specifically researchers, to extract valuable insights from the vast dataset. The cbms serves as a valuable instrument for assessing local poverty levels, promoting transparency, accountability, and efficient resource allocation by providing essential data for research purposes. Aligned with the guidelines of Republic Act 11315 (2018), the cbms acts as a blueprint for crafting impactful policies and programs that bolster community progress, alleviate poverty, and identify its underlying causes. While the survey is mandatory for all residents, the primary focus is not on Indigenous populations. However, researchers have demonstrated innovation, creativity, and resourcefulness in generating data that benefit Indigenous communities. Consequently, we selectively identified Indigenous individuals from 12 distinct tribes by filtering through extensive datasets. The evaluation encompasses 14,021 households comprising 63,749 members to ascertain the total count of Indigenous populations and their constituents. To facilitate comprehensive analysis, we established 12 distinct groups, each representing a unique tribe, forming the basis for our multi-tribal examination. The researchers functioned as data analysts and statisticians, extracting critical insights, generating predictive summaries, conducting advanced econometric analyses, and formulating strategies for economic advancement. This pioneering initiative marks the first comprehensive effort of its kind in the Philippines and within academia, championing the welfare of Indigenous communities through quantitative and empirical means. This groundbreaking work represents a significant advancement in the academic sector’s commitment to supporting the welfare of Indigenous populations. Table 1, sourced from the dataset, illustrates the 12 tribes of the Agta Tabangnon, providing an in-depth breakdown of the population dynamics within these indigenous communities.

View Table

The datasets are valuable for conducting poverty analytics and can serve as a replicable model for researchers studying Indigenous communities in different regions globally. Through the utilization of these datasets, one can extract key metrics including the watt index, gap index, severity statistics, and poverty incidence. Table 2 illustrates the poverty analytics centered on Indigenous populations in Southern Luzon, Philippines.

View Table

Table 3 presents a summary of the multidimensional poverty deprivation profiles of 12 indigenous tribes. The metrics outlined above serve as tools to gauge poverty across various dimensions. Given the multifaceted nature of poverty, it is imperative to employ multidimensional variables for its assessment (Onsay, 2022, Onsay & Rabajante, 2024b; Reyes et al., 2011; Sobreviñas, 2020). With access to this dataset, the ability to measure and forecast poverty for the advancement of economic development within Indigenous communities becomes both plausible and practical. These variables encompass livelihood and income, employment, housing and settlement, water and sanitation, basic education, health and nutrition, as well as peace and order.

View Table

4. Research Potential and Conclusion

Through utilizing comprehensive datasets encompassing all Indigenous households and populations in this work, we circumvent known limitations like sampling errors, thereby reinforcing both internal and external validity assumptions. The high-quality dataset, free of biases, provides a foundation for economic development strategies in developing nations and impoverished regions inhabited by Indigenous communities. Researchers can leverage this dataset to access raw baseline data and indicators of multidimensional poverty, enabling the calculation and simulation of various poverty metrics specific to Indigenous tribes. Utilizing advanced data analytics techniques, subsequent investigators can extract valuable insights into population dynamics, health, education, livelihood, and other critical aspects affecting Indigenous communities. Predictive and prescriptive analytics, along with econometric models, offer empirical evidence to support poverty theories and inform policy decisions. These datasets not only facilitate research endeavors but also enable policy mapping and targeting efforts aimed at addressing vulnerabilities and promoting economic advancement among Indigenous populations globally. The outcomes of policy targeting can guide resource allocation and the formulation of tailored strategies to combat poverty effectively. By adapting to local conditions and utilizing rigorous methodologies, policymakers can develop flexible plans that align with the evolving needs of Indigenous communities. Collaborations between government authorities, academic institutions, and commercial organizations can further enhance poverty reduction initiatives and economic development efforts. This integrated approach as hereby recommended by researchers, supported by comprehensive datasets and advanced analytics, serves as a catalyst for interdisciplinary research, policy development, and community engagement to uplift Indigenous populations and foster sustainable economic growth.