Correctly Specifying Environmental Assets in Spatial Hedonic Pricing Models

Abstract

In hedonic pricing models, researchers attempt to capture the marginal benefits of environmental assets using various classifications to differentiate them, such as natural parks versus playground parks. This improves the accuracy of the estimation, but the trade-off is the increased time and cost of data acquisition. Emphasis is given to classifying the environmental asset of interest as well as possible, while other assets of “minor” interest are given less importance. Often, conventional techniques are applied to estimate the benefits of assets of “minor” interest, such as estimating distance to the nearest site only, or applying a weighted distance measure to all sites. In this study we demonstrate the errors associated with using the conventional techniques to deal with poor data quality. By assuming homogeneity of all sites, researchers inadvertently place too much value on some assets and too little value on others, which could result in misleading policy recommendations.