A state-space model for residential real estate valuation

All property in the Netherlands has to be appraised yearly. Yearly valuation has only been made possible with the help of models. The number of real estate appraisers is simply too small to value the more than 7 million residential properties. This paper describes the statistical model that is used by Ortec Finance to value residential real estate for the local government, housing corporations, and mortgage providers. Transaction prices are explained by housing characteristics, location and time in a hierarchical trend model. In this state-space model the impact of time on transaction prices is modeled in an advanced and flexible manner. Estimation results are provided for Amsterdam.