Boundary discontinuity, unobserved spatial heterogeneity, and flood risk analysis: application to a rural housing market

Increasing precipitation variability has led to growing concerns about the risk of flood in many parts of the U.S. and worldwide. This study adopts a spatial regression analysis framework based on floodplain boundary discontinuities to control for unobserved spatial heterogeneities in flood risk assessment. The empirical strategy combines spatial matching with the regression discontinuity design. Using parcel-level data from Juniata County, Pennsylvania, we find an over six percent property value reduction due to additional one-standard-deviation flood exposure as defined by the percentage of property area inside the flood zone. Specifically, given that the study region’s average single-family residential housing price is $114,684 (in 2015 USD), the estimated flood risk impact is $7056 (in 2015 USD, or 6.15%) for a one-standard-deviation increase in flood exposure. The estimate (for a rural housing market) is of a similar magnitude to several recent estimates from urban and suburban housing markets. This paper contributes to the literature as one of the first quantitative studies of flood risk impacts on rural housing markets.