Poverty is the oldest social problem that ever existed and is difficult to reverse. It is multidimensional and unmeasurable. Thus, measuring by decomposing rural multidimensional poverty is critical. Most poverty studies are usually generic, exposed to large sampling errors, and intended for macroeconomic decisions. Thus, measuring poverty for a specific locality with various configurations (15) is critical for economic development. The paper combines predictive analytics and advanced econometrics to decompose poverty at the micro-level by utilizing the Community-Based Monitoring system at complete enumeration (L = 34, S = 4). Logistic Regression (78) Models with 19 Independent Variables and 12 Intervening Variables were fitted. Headcount Analysis (0.2138–0.9845), Poverty Gap (0.2228–0.0502), Severity statistics (0.0723–0.0168) and Watts Index (0.2724–0.0618) are scrutinized. Poverty levels vary by location; a significant fraction of the population (P0i = 68.50%, P0f = 55.80%) and households (P0i = 63.70%, P0f = 50.70%) live below the poverty line and food threshold. It has been revealed that poverty is extreme in Isarog (i = 0.7793), moderate in Poblacion (p = 0.4019), intense in Ranggas (r = 0.6542), and severe in Salog (s = 0.6353). Multidimensional variables (13VAR) significantly predict poverty outcomes (p-value = 0.0000, PseudoR2 = 0.75). Moreover, intervening variables have been impacting poverty across all locals. All models tested are significant across all sectors and correctly predicted by the model classifications (Estat = 73.29–74.12%). Poverty is multifaceted; thus, it requires different interventions. Finally, policy proposals (54) were outlined to alleviate poverty and promote local economic development.
Read full abstract