Changing Tastes : Estimating Changing Attribute Prices in Hedonic and Repeat Sales Models

Repeat sales indices are commonly used to estimate the trend in house prices. Two assumptions are made when estimating these indices: 1) that the house characteristics are unchanged between observed sales and 2) that the underlying shadow prices for those attributes are constant or change proportionally. The second assumption is the focus of this paper. First, using data from the American Housing Survey, we show that certain attribute prices in Chicago changed idiosyncratically between 1985 and 2011. These results confirm earlier reports of changing tastes and attribute prices. We show that such changes can be approximated in a hedonic model with a low order polynomial functional form. We find that a simple extension of the standard repeat sales methodology can be used to test for an estimate changes in characteristic prices attributable to changing tastes. Our results are of interest to all users of repeat sales indices, but especially those portfolio investors interested in estimating current values of homes securing loans in their portfolio to help manage delinquencies and defaults