Conditional Correlations and Real Estate Investment Trusts

Journal of Real Estate Portfolio Management, Apr-Aug 2009 by Chong, James, Miffre, Joëlle, Stevenson, Simon

Executive Summary.

The paper studies the temporal variations in the conditional correlations between real estate investment trust (REIT) returns and equity, bond, and commodity returns. The findings reveal that the correlations between REITs and equity returns rose over the period analyzed, while the correlations with bonds and commodities fell. The findings also reveal that the correlations with REITs rose especially in periods of above average volatility in equity and bond markets; however, for U.S. government securities and the Goldman Sachs Commodity Index, the conditional correlations with REITs fell in periods of high volatility in these markets. This indicates that to reduce the total risk of their portfolio, investors in U.S. government securities and commodities should tilt their asset allocation more towards real estate when they anticipate changes in monetary policy or abnormal fluctuations in commodity prices.

The relationship between real estate investment trusts (REITs) and the broader capital markets is one of ongoing interest. The possible diversification opportunities available from REITs, from the perspective of both equity and fixed income portfolios, have been examined in a wide variety of both academic and industry research. Obviously the relationship between REITs and alternative asset classes is of importance from a fund manager's perspective. However, due to the changing nature of the REIT market, the stability of these relationships and how they have altered during the recent past is also of paramount importance. From the mid-1990s onwards, there has been a marked increase in investor awareness. As Chan, Leung, and Wang (1998) note, institutional investment increased substantially in the 1990s and a positive relationship exists between this and REIT performance. This process has continued and the inclusion of REITs in the Standard & Poor's indices in 2001 was a further step in increased investor acceptance of the sector.

The paper studies the conditional correlation structure in the REIT sector. In particular, the conditional correlations of REITs are estimated against a variety of equity, bond, and commodity series. The paper builds upon recent work in the REIT literature that examines the dynamics of the sector. The data used in this paper consists of daily returns for the period January 1, 1990 through December 30, 2005, totaling 4,000 observations. During this time, the popularity of REITs has expanded dramatically with massive growth in investor awareness and interest focusing on the return and volatility characteristics of the sector. The analysis is conducted using the dynamic conditional correlation model of Engle (2002), whereby the conditional correlations are estimated against a variety of equity, bond, and commodity sectors. Domestic equity sectors representing large capitalization, small capitalization, value and growth stocks are analyzed together with international equity markets. Bonds of different maturities and countries of issuance are also considered. For the sake of completeness and given the growing interest of portfolio managers in commodities, the Goldman Sachs Commodity Index (GSCI) is also used as one of the asset classes against which the correlation with REITs is measured.

The results reveal that over the 16-year sample period, there is a high level of instability in the conditional correlations. As would be expected, the conditional correlations with respect to the domestic equity sectors are largely higher than either the fixed income or foreign stock and bond markets U.S. However, distinct phases in the correlations are observable. While the correlations are generally close to zero, there are periods when substantial increases in the relationships observed are noticed. These time variations in the conditional correlations are subject to two systematic patterns. First, we observe that the conditional correlations with equity returns rose over the period analyzed, while that with bond and commodity returns fell. This indicates that real estate has lost some of its luster for strategic asset allocation to equity portfolio managers but has gained part of its diversification properties to bond portfolio managers and CTAs over time. Second, the correlations with REITs rose especially in periods of above average volatility in equity and bond markets. This is unfortunate as it is precisely in periods of high volatility in the bond and equity markets that investors need the benefits of diversification the most. There are however two noticeable exceptions (for U.S. government securities and the GSCI) where the conditional correlations with REITs fell in periods of high volatility. This indicates that to reduce the total risk of their portfolio, investors in U.S. government securities and commodities should tilt their asset allocation more towards real estate when they anticipate changes in monetary policy or abnormal fluctuations in commodity prices.

The rest of the paper is organized as follows. Section 2 reviews the literature on the relationship between REITs and the broader capital markets. Sections 3 and 4 present the methodological framework and the data, respectively. Section 5 analyzes the temporal variations in the conditional correlations between REITs returns on the one hand and equity, bond, and commodity returns on the other. Section 6 highlights the main conclusions of the paper.