Controlling for cross-sectional and spatial heterogeneity in the hedonic property value model using entropy econometrics

Abstract

Hedonic pricing models that apply data from many different metropolitan areas are subject to considerable cross-sectional and spatial heterogeneity. When disregarded, this heterogeneity can have important consequences for estimation and inference. In the present paper we develop a flexible entropy-based estimator that allows controlling for several sources of (unobserved) heterogeneity in the hedonic price function. In the empirical application, we examine the performance of the new estimator applying data from the seminal Harrison and Rubinfeld paper on hedonic pricing. Our empirical results show how failure to control for cross-sectional and spatial heterogeneity may result in biased coefficients and a loss of explanatory power. Furthermore, analysis of the individual sensitivities reveals important dynamics regarding the operation of the housing markets.