International Articles: Securitized Real Estate and Its Link with Financial Assets and Real Estate: An International Analysis

Article excerpt

Abstract

This paper provides cross-country evidence of the link between securitized real estate and stocks, bonds, and direct real estate. First, the behavior of betas in sixteen countries is examined and then the causes of their variation are identified. Second, securitized real estate returns are regressed on "pure" stock, bond, and real estate factors. The betas are generally found to decrease over the 1990-2004 period, but the causes for such decline differ across countries. Securitized real estate returns are found to be positively associated with stock and direct real estate returns, but negatively related to bond returns. Financial assets contribute greatly to the variance of securitized real estate, while the impact of direct real estate is limited. However, a large fraction of the variance is not accounted for by these factors, especially in the United States, which suggests that other factors are at play.

(ProQuest: ... denotes formulae omitted.)

With a market capitalization estimated at USD 800 billion as of the end of 2005 (Brounen, Ling, and Eichholtz, 2006), the importance of securitized real estate as an asset class has grown considerably. Accordingly, much research has been devoted to studying the behavior and to unmasking the driving factors of publicly traded real estate.1 As for any other asset class, there are many important reasons for studying the return generating process of securitized real estate, such as to examine whether returns can be forecasted (Liu and Mei, 1992; Bharati and Gupta, 1992; and Brooks and Tsolacos, 2003), to ascertain the diversification benefits of the asset class (Mull and Soenen, 1997; and Gordon, Canter, and Webb, 1998), and to analyze the inflationhedging effectiveness of such vehicles (Liu, Hartzell, and Hoesli, 1997; and Adrangi, Chatrath, and Raffiee, 2004).

The aims of this paper are twofold. The first objective is to expand the investigation done by Khoo, Hartzell, and Hoesli (1993) on the behavior of real estate investment trust (REIT) betas in the United States by employing a similar procedure to securitized real estate in sixteen countries from 1990 to 2004 and by providing a cross-country analysis of the causes underlying the variations in the betas. Second, this research expands Clayton and MacKinnon's (2003) work on the importance of stock, bond, and real estate ''pure factors'' in explaining securitized real estate returns by doing the analysis for five countries. In sum, this research seeks to expand the literature by depicting the dynamics that underpin different markets and analyze if the previous findings for the U.S. can be generalized for other markets.

The paper is organized as follows. First, there is a discussion of related work on securitized real estate's betas and on the hybrid nature of such securities. Second, there is a presentation of the methodologies employed and the data, respectively, followed by a discussion of the results. Finally, the paper closes with concluding remarks.

Literature Review

An important question addressed by the literature concerning securitized real estate has been whether or not this asset class outperforms other asset classes. Chan, Hendershott, and Sanders (1990), Glascock (1991), and Peterson and Hsieh (1997) find that REITs do not display higher risk-adjusted returns with respect to stocks. On an international basis, this result is confirmed by Ling and Naranjo (2002). Contrasting evidence for the U.S. is provided by Liu and Mei (1992) who show that on a riskadjusted basis, EREITs outperform large caps and small caps, but not bonds. Lying in between and probably settling the differences, Chen, Hsieh, and Jordan (1997) derive a Jensen measure for each EREIT using excess returns and conclude that EREITs outperform other investments during some periods but not always.

To better understand the asset class, the analysis of the time-varying nature of securitized real estate betas constitutes another stream of research. …