Foreclosure Spillovers within broad Neighborhoods

Most evidence on foreclosure spillovers identifies localized effects that are modest in magnitude, but these effects could multiply to larger aggregate effects across broad neighborhoods. We test this proposition developing a proxy for the fraction of mortgages in negative equity during the foreclosure crisis and estimating a difference-in-differences model for foreclosure. This proxy exploits the timing of foreclosures in each tract, and this within tract variation is not predicted by mortgage attributes, housing attributes or sales prices. Our estimates suggest that 61 percent of the increase in across tract dispersion in foreclosure filings can be explained by these spillover effects.