FEAR: Flood Exposure and Asset Repricing

An economics research project investigating how FEMA flood zone reclassifications (Letters of Map Revision) affect local property values. The study uses difference-in-differences and event study methods, treating LOMR effective dates as treatment events and comparing zip codes inside vs. adjacent to reclassified flood zones.

The project includes a full reproducible data pipeline built in Python (pandas, geopandas) that processes data from FEMA's National Flood Hazard Layer, 72 million NFIP insurance policy records, Zillow home value indices, BLS unemployment data, Census boundaries, and election returns. Spatial overlays identify which zip codes are affected by each LOMR, and the panel is exported for econometric estimation in Stata using high-dimensional fixed effects (reghdfe).

Key findings: flood zone reclassifications produce a gradual, significant decline in home values (approximately 2.8% by four quarters post-reclassification), with federal insurance subsidies dampening the repricing and Republican-leaning counties exhibiting smaller price responses. Robustness checks include Callaway and Sant'Anna heterogeneity-robust estimation, Bacon decomposition, leave-one-out by state, and placebo outcome tests.

A companion website built with Next.js, D3.js, and Tailwind CSS presents interactive coefficient plots, treatment maps, and regression tables.