Heterogeneous Foreclosure Discounts of Homes

Abstract

When the foreclosure crisis hit the U.S. housing market, there was little consensus on which homeowners were affected the most by home value impairment. The goal of this study is to flexibly estimate house-specific foreclosure discounts and to explore the merits of heterogeneous foreclosure discounts across market segments. I use a comprehensive dataset that encompasses home transactions from 2000 to 2020 in Florida and Indiana. Summary statistics show that foreclosures are realized across the entire home value and home size distributions. I estimate a structural model that builds on Rosen (1974) and Bajari and Kahn (2005) and estimates a price function using a weighted least least-squares regression approach. The estimation results show that foreclosure discounts in Indiana are higher than in Florida. In Indiana, foreclosed homes lost the most value at the lower part of the house value distribution. Moreover, owners of foreclosed large houses experienced immense value losses, and this applies to every city. In Indiana, houses at the lower part of the house size distribution also suffered from large foreclosure discounts, while Floridian houses lost significantly less value in this market segment. I also found that homes in neighborhoods with higher mortgages, urbanization, median incomes, and education rates realize higher foreclosure discounts. Neighborhoods with smaller Asian, Black, and Hispanic populations experienced higher foreclosure discounts.