Estimating a price surface for vacant land in an urban area

This paper examines urban land prices within a non-parametric framework with specific interest in the land price gradient with respect to distance from the city center. The land price surface is estimated over a 900-square mile portion of the Chicago metropolitan area and is based on transaction level data of vacant land parcels. The price surface is estimated using piecewise parabolic multiple regression, proposed by Colwell 1998. Given the flexibility of this approach, a rather complex price surface is revealed. The estimated price surface provides some support for the standard urban economic doctrine of falling land prices with distance.