A Stylized Agent-Based Framework for Exploring Flood–Poverty Dynamics Among Rice Farmers: A Case Study of Candaba, Philippines

Abstract

Flood disasters increasingly shape rural livelihoods in Southeast Asia, not only through crop losses but also through long-term socioeconomic effects. This study develops a stylized agent-based framework (ABMAAF) to analyze flood–poverty dynamics in Candaba, Philippines. The model integrates spatially heterogeneous flood risk, endogenous rice market responses, and household wealth trajectories. Flood sequences were generated stochastically to highlight structural mechanisms rather than reproduce historical events. Results reveal that medium floods often generate larger inequality effects than rare large floods, as localized impacts create sharp disparities. Price elasticity of demand is critical: with low elasticity, price surges partly offset losses from large floods but amplify inequality during medium floods; with high elasticity, large floods dominate. Regression analysis further reveals that sequential shocks matter: medium floods following a large flood significantly increase inequality, as demonstrated by counterfactual comparisons and the interaction term in the regression model. These findings highlight that flood impacts depend not only on magnitude, but also on market functioning and event sequence. Methodologically, the ABMAAF links hydrological shocks with inequality trajectories, offering a framework for exploring disaster–poverty traps. Policy implications include strengthening safety nets during recovery and designing food market interventions that mitigate distributional risks under recurrent floods.