Journal of Real Estate Finance and Economics, 28:2/3, 179±208, 2004
# 2004 Kluwer Academic Publishers. Manufactured in The Netherlands.
The Hierarchical Trend Model for Property Valuation
and Local Price Indices
MARC K. FRANCKE
Vrije Universiteit Amsterdam, The Netherlands
GERJAN A. VOS
University of Amsterdam, The Netherlands
This paper presents a hierarchical trend model (HTM) for selling prices of houses, addressing three main
problems: the spatial and temporal dependence of selling prices and the dependency of price index changes on
housing quality. In this model the general price trend, cluster-level price trends, and speci®c characteristics play a
role. Every cluster, a combination of district and house type, has its own price development. The HTM is used for
property valuation and for determining local price indices. Two applications are provided, one for the Breda
region, and one for the Amsterdam region, lying respectively south and north in The Netherlands. For houses in
these regions the accuracy of the valuation results are presented together with the price index results. Price indices
based on the HTM are compared to a standard hedonic index and an index based on weighted median selling
prices published by national brokerage organization. It is shown that, especially for small housing market
segments the HTM produces price indices which are more accurate, detailed, and up-to-date.
Key Words: Hedonic models, Kalman ®lter, real estate price indices, thin markets
This paper concerns the modeling of selling prices of houses by hedonic price models.
Besides the size and the location of a house, the selling date is an important characteristic
to explain selling prices in a time of rapid price movements. A hierarchical trend model
(HTM) is presented, addressing the spatial and the temporal dependence of selling prices.
In the literature on hedonic modeling the temporal dependence of selling prices is
addressed in several ways. Case and Quigley (1991) present a model in which information
on repeated sales is combined with that of single sales. Discrete and continuous time
varying locational and structural parameters are considered. In the discrete case for every
time period an extra parameter is added to the model. In the continuous case, the model
speci®cation is more parsimonious, but a linear trend for parameter evolution is imposed.
Fleming and Nellis (1992) propose repeated regressions for every time period, so
regression parameters may vary over time. A closely related approach is provided by
Knight et al. (1995). In this approach only one regression is performed with varying
parameters over time and correlation between multiple sales is considered.