Space-time approach to commercial property prices valuation

Abstract

There exists three ways of approaching real estate prices: the cost approach, the market data approach and the income capitalization approach. In this article, we propose an improvement of the market data approach that takes into account the spatial component. In particular, we propose a modified market data approach based on interpolation, being the structure of the spatial correlation between the prices of properties the main factor to obtain the weights. Interpolation methods have been widely used for estimating real estate prices, but they do not take into account the structure of their spatial dependence. Although this drawback is overcome by kriged estimation, in the case of the prices of commercial properties, they do not provide good estimates because the scarceness of the market information. This is why auxiliary information is needed and cokriging methods are used to obtain estimates that are more accurate. The aim of this article is the comparison of cokriged estimation of premises prices in two different temporal moments in the emblematic old part of Toledo city (Spain), using housing prices as an auxiliary random function due to their strong correlation with the main one. Cokriging, kriging and inverse distance weighting results are compared.

Keywords:

• spatial correlation
• cokriging, premises prices
• house prices
• variogram

JEL Classification::

• C13
• C10
• R00

Acknowledgements

This research has been supported by the Spanish MICINN through the project CSO2009-11246.

Notes

¹ We use the term ‘house prices’ throughout, in accordance with the quoted literature, even though the data only include flatted properties.

² We are extremely grateful to Imagil Gestión Inmobiliaria, Zocopiso, Amian Inmobiliaria, Imperial Inmabel S.L., Agencia Inmobiliaria Gudiel, Inmobiliaria Castaño, Agrufinca, Acrópolis, Albatros, Teleinmobiliaria, Inmobiliaria Época, Inmobiliara Ábaco, Simar, Agencia Inmobibliaria Orgaz and Fondo Piso Toledo for their first-rate help in providing the detailed data to the Department of Statistics at the University of Castilla-La Mancha (Spain).

³ It does not significantly affect results.

⁴ The analysis has been conducted in terms of price per square meter.

⁵ Adjusted for housing and premises mix, in Fotheringham et al. (2002) terminology, although these authors also use the expression ‘equivalent houses’.

⁶ See http://www.geovariances.fr