Flood risk is an increasingly important factor influencing housing markets, particularly in the U.S. Gulf Coast region where hurricanes, sea level rise, and urban flooding pose serious threats to residential communities. To address this limitation, this study integrates Average Annual Loss (AAL), a dollar-denominated metric representing the long-term expected annual damage from flooding, into a hedonic pricing framework to quantify how households value flood risk in Louisiana, Mississippi, and Alabama.

AAL is combined with structural, geographic, and socioeconomic characteristics and evaluated across ten modeling approaches: five statistical models (Linear Regression, Generalized Linear Models, Mixed-Effects Models, Spatial Lag Models, and Geographically Weighted Regression) and five machine learning algorithms (Random Forest, Gradient Boosting, XGBoost, Support Vector Regression, and Neural Networks). Using R², RMSE, MAE, MAPE, and spatial cross-validation, we assess whether incorporating AAL improves the explanatory and predictive performance of hedonic models compared to traditional flood zone indicators.

This research advances understanding of how objective risk metrics (AAL) versus categorical designations (FEMA zones) influence real estate markets, with implications for insurance pricing, mortgage underwriting, climate adaptation financing, and equitable housing policy in high-risk regions.