2019 V00 00: pp. 1–33 DOI: 10.1111/1540-6229.12281 REAL ESTATE ECONOMICS Foreclosure Externalities: Have We Confused the Cure with the Disease? Yishen Liu* and Anthony M. Yezer** Foreclosure externalities, in which recent foreclosures proximate to a housing unit depress its sales price, are well accepted in the literature. Past research on foreclosure externalities implicitly assumes that both the partial and total derivatives of the outcome (house value) with respect to the treatment (fore- closure) are equal and constant. This article relaxes these assumptions about functional form. Results reported here show that the causes of foreclosure, mortgage delinquency and negative equity also lower nearby housing prices. Because foreclosure eliminates delinquency and negative equity, it appears to be the cure for the externality rather than the cause. Delay in the foreclosure process prolongs the externality problem and slows price recovery just as delay in application of a cure makes the disease worse. These results demonstrate that functional form matters in testing for overall effects of policy changes. Introduction The high rate of foreclosure on single-family mortgages in the period follow- ing 2006 prompted substantial research into the possibility that the foreclosure process generates local externalities. With few exceptions, papers have found a negative relation between recent foreclosures and the current sales price of housing proximate to those foreclosures. These papers have recognized that local unobserved heterogeneity in housing markets may generate the “treatment” of foreclosure and have adopted clever differencing techniques to eliminate the effects of this correlation on the relation between foreclosure and house price. The issues raised in this article have nothing to do with identification problems of selection into treatment that motivate or mandate differencing to identify causal effects. 1 *Fannie Mae, Washington DC or yishen_liu@fanniemae.com. **Department of Economics, George Washington University, Washington DC or yezer@gwu.edu. 1 In the case of the research results presented in this article, bias due to selection into treatment is not a concern for two reasons. First, the identification strategy is based on the approach used in previous papers. Second, and most important, the direction of any simultaneous equation bias is negative, i.e., in favor of the alternative hypothesis. C  2019 American Real Estate and Urban Economics Association