Validation and comparison of U.S. loss estimates from catastrophe flood models

The study compares and validates seven U.S. catastrophe flood models by analyzing their average annual loss (AAL) estimates against historical National Flood Insurance Program (NFIP) claims data, aiming to assess how accurately these models price flood risk.

Models used in NFIP pricing (Verisk, KatRisk, Moody’s RMS) closely match historical losses (within about 4%), while First Street significantly overestimates flood losses nationally and Karen Clark & Company substantially underestimates them in Florida, highlighting major methodological differences and the need for greater transparency and public validation.