A home is typically thought of as a bundle of land and structure. Land is supplied inelastically and is non-reproducible. Land values are therefore affected by a number of demand factors. Conceptually, structures are easily produced, and thus are supplied elastically. Under elastic supply, it is reasonable to assume that replacement cost should be equivalent to the value of the structure for new properties. Here, we examine how particular structure characteristics may introduce heterogeneity into these demand and supply relationships. In other words, how do structure vintages influence price dynamics. Older vintages are not easily reproducible leading to the value of an older vintage to potentially diverge from its replacement cost. To test our hypotheses, we employ a nonlinear model in a Bayesian structural timeseries approach that explicitly disentangles structure and land values to identify vintage effects separately from physical deterioration and land values. We find large differences in price dynamics between four distinct vintages of Amsterdam old city center apartments. Between 1999 “ 2016, new construction had an average return of 1.7%, with a standard deviation of 2.4%. On the other hand, properties build in the 19th century had an average return of 3.6% with a standard deviation of 6.1%, during the same period.
Keywords: land and structure prices, depreciation, new construction, structural time series.
JEL Classifications: R21, R31, C11.
Working paper no. 617