Directional heterogeneity in distance profiles in hedonic property value models

In environmental economics, hedonic property value (HPV) models have often employed distance from a localized environmental disamenity as a proxy for perceived risk. The magnitude of a distance effect on housing prices, however, may depend upon the direction in which it is being measured. We generalize conventional distance models to allow for continuously varying directional effects by converting from Cartesian to polar coordinates. A simple empirical example (for housing prices around Woburn, Massachusetts, between 1988 and 1996) illustrates how failure to allow for directional heterogeneity can obscure otherwise statistically significant distance effects in HPV.