An analysis of trends and cycles of land prices using wavelet transforms

Abstract

This paper analyses the low‐frequency temporal variation of unit land prices in two single localities, the municipalities of Espoo and Nurmijärvi, of the Finnish real estate markets by the use of the wavelet transforms. These transforms are nonparametric orthogonal series estimators, which are capable of providing the necessary time and frequency information of land prices simultaneously in a highly flexible fashion. In the empirical section of this paper both the raw and the quality‐adjusted unit price series are analysed. The estimated cycles and trends are all nonlinear and, in particular, the behavior of cyclical component is highly curvelinear and transient in time. The findings strongly suggest that the sub‐markets in question are not in a steady state, but are continuously evolving in time. It seems that much of the temporal variation present in the untransformed series is, in fact, explained by the quality differences in the attribute variables. The use of quality corrections produced significant improvements in the internal reliability of results.

First Published Online: 18 Oct 2010