A multivariate repeat-sales model for estimating house price indices

In the Case-Shiller repeat-sales housing price index method, the effect of age is embedded into the model's estimates of the time effect. This paper develops a multivariate repeat-sales model that is able to separately control for the effects of age and time, as well as other assets with changing attributes in the construction of price indices. The ability to control for an asset's pattern of depreciation should lead to a superior price index when compared to a traditional repeat-sales price index. The degree of bias between the traditional and multivariate repeat-sales models is dependent on the level of price change with respect to age. With minimal additional assumptions, a pure time (time-varying age-constant) price index and a depreciation (time-constant age-varying) price index can be constructed using the multivariate repeat-sales framework. (c) 2004 Elsevier Inc. All rights reserved.