A method for combining transaction- and valuation-based data in a property price index
(English)Article in journal (Other academic) Submitted
This paper presents a method for combining transaction- and valuation-based data in a property price index. The methodology is devised for a world where observable transaction prices can be used to construct a price index that constitutes a noisy, unbiased signal of the ”true” price index. It is furthermore assumed that valuations can be used to construct a market value index which does not contain noise but that suffers from so called appraisal ”smoothing”. The valuation-based index is thus assumed to lag the ”true” value index and exhibit lower volatility. The model of the valuation-based index follows Geltner (1993). By regressing the transaction-based index on the valuation-based index (contemporaneous and lagged one period) it is possible to filter out the noise in the observable price index thus estimating the ”true” price index. The method may be seen as a way of ”de-smoothing” a valuation-based index without knowing the smoothing parameter beforehand. The methodology may also be used as a way of estimating the smoothing parameter.
real estate, price index, smoothing, valuation, transaction
IdentifiersURN: urn:nbn:se:kth:diva-26950OAI: oai:DiVA.org:kth-26950DiVA: diva2:373127
QS 201203282010-11-302010-11-302012-03-28Bibliographically approved