Real estate performance measurement in markets with thin information

Historically, different index construction methodologies have been used to represent the behaviour of real estate markets. They can be grouped into four main categories: valuation based indexes and transaction-based ones, synthetic measures ( e.g . created by using prime rents and yields) and vehicle-based performances (property companies and Real Estate Investment Trusts). Each measure requires a different set of data. When we consider markets with thin information, data availability plays a major role in defining the applicability of these construction methodologies. Moreover, if the aim of an index is to show long-term performances in such markets, individual property data (e. g. periodic valuations) used by main index providers may not be retrievable historically. This work describes main index construction methodologies used in the property industry or suggested in relevant finance literature. Three new methodologies are applied to the UK market and their ability to represent a "true" estimate of market performance is tested by comparing these new figures with the current valuation-based index. The first methodology employs purchase prices and last valuations to create repeated-measure regression returns. We find this index to behave more similarly to an unsmoothed version of the valuation based index than to its original series. Secondly, we obtain an estimate of market performance from vehicle-based in formation by adopting a weighted average cost of capital framework. Finally, we apply a capital asset pricing model net of illiquidity costs to public real estate returns and find an improvement in correlation coefficients even at a monthly frequency. All these three methodologies may be used to create historical series in markets where information are not easily available. They all represent a good proxy for unsmoothed real estate returns, The choice between these three methodologies should be data driven since there is no theoretical a-prior to prefer one to another.

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

City University London

Year dc:date.issued

2005

Author dc:creator

• Marcato, G.

• HG Finance