Poverty Status Differentiation in Central Philippines: A Discriminant Analysis of Household Characteristics

Abstract

One of the most intractable socioeconomic issues in Central Philippines is poverty, and a significant number of Filipino households are currently falling below the officially declared poverty line. Although several factors have already been identified as causally linked to poverty in existing studies, these works typically use descriptive analyses and regression techniques-methods capable of determining interrelationships but not of accurately identifying poor households. To bridge this gap, the current study utilized discriminant function analysis to develop a classification model which determines the discriminating attributes that can distinguish poor from non-poor households in Central Philippines. Based on data derived from the 2024 Family Income and Expenditure Survey (FIES), 1,875,723 total households were studied. The selected socioeconomic attributes investigated include the age of household head, level of education achieved, number of household members, number of income earners and household food expenditure percentage. Analysis reveals that all variables selected are highly significant for classification: the percent household food expenditures (β= 0.555) and the educational attainment of the household head (β= 0.449) are the most potent ones. Classification of the household members resulted in 75.8% overall accuracy. The model classified correctly 76.5% of poor households and 74.9% of non-poor households. In essence, observable characteristics of households were able to predict their classification in terms of poverty, even in the absence of actual income data. These findings could aid the government and other policymakers in a more appropriate allocation of their limited resources for poverty reduction interventions and social protection programs.