Modeling and Measuring Protective Action Decisions under Flood Hazards in Brazil

Gilvan R. Guedes, Cedeplar, UFMG
Rodrigo Raad, Cedeplar, UFMG
Lucélia Vaz, Universidade Federal do Rio de Janeiro, UFRJ

This paper performs an econometric analysis based on the Protective Action Decision Model (PADM), using Seemingly Unrelated Regression models on a 1,200 probabilistic sample for Governador Valadares, Brazil. The site was chosen because river floods are recurrent in the area, reaching thousands of households every other year. Building on models of private insurance, we provide a theoretical framework stating that risk aversion is a key factor affecting the directions of influence given by the price and resource effectiveness on the probability of adopting preventive behavior. We show that, under certain conditions on risk aversion, hazard-related attributes (HRA) positively affect propensity to take protective actions. In contrast, resource-related attributes (RRA) positively or do not affect propensity to adopt protective actions. Finally, the impact of HRA was empirically higher than RRA on preventive behavior since the effective cost of preventive actions is low relative to a person's budget constraint.