Measurements and determinants of multifaceted poverty: absolute, relative, and subjective poverty in Indonesia

Abstract

The notion of ‘poverty’ is diversified and dynamic. It varies across countries with different socio-economic norms. It may also change over time even in the same society, with different stages of social and economic development. A country may be struggling with absolute poverty at the early stages of development, while it may well be more concerned with relative and/or subjective poverty as its average per-capita income increases. This article intends to conduct an exploration of multiple poverty measures by looking into the absolute, relative and subjective poverty incidence in Indonesia. Using the 2005 National Socio-Economic Survey (Susenas), we observed that there was a roughly 28 percentage-point difference in the poverty headcount ratios computed by applying absolute (14.47%) and subjective (42.03%) poverty. There were virtually no correlations among the poverty rankings in the provinces of Indonesia obtained by five poverty metrics. Results of logit model and ordered logit model estimations of the possible determinants of poverty indicate that the main determinants of poverty are educational attainment, number of household members, physical assets (land and house ownership), existence of migrant workers (possible remittances), negative shocks of layoffs and/or health problems, development of public services, and the availability of road infrastructure. A higher educational attainment increases the probability of never being poor in any of the five poverty metrics by almost 11 percentage points. This study also confirmed that households having less than society’s averages in terms of the physical asset of land and consumption of durable goods and fashion tended to subjectively asses themselves as poor. The study suggests that any poverty alleviation programs should consider relative impacts among beneficiaries and non-beneficiaries within each locality and across provinces.