Time‐varying correlations between stock and direct real estate returns
| DOI | https://doi.org/10.1108/14635781311302591 |
| Published date | 01 March 2013 |
| Date | 01 March 2013 |
| Pages | 179-195 |
| Author | Tien Foo Sing,Zhuang Yao Tan |
| Subject Matter | Property management & built environment |
Time-varying correlations
between stock and direct real
estate returns
Tien Foo Sing
Department of Real Estate, National University of Singapore,
Singapore and
Institute of Real Estate Studies, National University of Singapore,
Singapore, and
Zhuang Yao Tan
Department of Real Estate, National University of Singapore, Singapore
Abstract
Purpose – Understanding correlations between stock and direct real estate returns, which is the
key factor that determines diversification benefits in a portfolio, helps formulate and implement
better investors’ asset allocation and risk management strategies. The past studies find that direct
real estate returns have a low unconditionally (long-run) correlation with the returns of equities.
However, assuming that such correlation is constant throughout all periods is implausible.
The purpose of this study is to test the time-varying correlations of returns between general stocks and
direct real estate.
Design/methodology/approach – This study uses the dynamic conditional correlation (DCC)
model, which is a simplified version of the multivariate generalised autoregressive conditional
heteroskedasticity (GARCH) model, proposed by Engle to test the time-varying correlations between
stock and direct real estate returns in six markets, which include the USA, the UK, Ireland, Australia,
Hong Kong and Singapore.
Findings – The empirical results show significant time-varying effects in the conditional covariance
between stock returns and direct real estate returns. The results vary across different real
estate sub-sectors, and across different countries. It is observed that the conditional covariance
increases in the boom markets, but becomes weaker in the post-crisis periods. The authors
observed significant jumps in the conditional covariance between the two asset markets in Singapore
and Hong Kong in the post-1977 Asian Financial crisis periods and in the post-2007 US Sub-prime
crisis periods.
Originality/value – The past studies find that direct real estate returns have a low unconditionally
(long-run) correlation with the returns of equities. However, assuming that such correlation is constant
throughout all periods is implausible. This study fills in the gap by using the dynamic conditional
correlationmodelsto allow fortime-varyingeffects in thecorrelationsbetween stockand real estatereturns.
Keywords Constant correlation, Time-varyingrisks, Dynamic conditional correlations,Stock returns,
Direct real estatereturns, Real estate, United States of America,United Kingdom, Ireland, Australia,
Hong Kong, Singapore
Paper type Research paper
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1463-578X.htm
The authors would like to thank Graeme Newell, Alistair Adair and other participants at the 6th
International Real Estate Symposium (IRERS), 24-25 April 2012, Malaysia, for their comments.
The paper has been awarded the Best International Paper at the above conference organized by
National Institute of Valuation (INSPEN), Malaysia.
Received March 2011
Accepted July 2012
Journal of Property Investment &
Finance
Vol. 31 No. 2, 2013
pp. 179-195
qEmerald Group Publishing Limited
1463-578X
DOI 10.1108/14635781311302591
Stock and direct
real estate
returns
179
1. Introduction
Stocks and real estate are two important asset classes in investors’ portfolios. The
aggregate market value of equities in the Morgan Stanley Capital International-All
Country World Index (MSCI ACWI), which covers 85 percent of the equity values of
49 developed and emerging markets, was estimated at more than $19 trillion as of the
fall of 2004. The size of global real estate investible universe was estimated at
approximately $6.2 trillion as of the end of 2003 (UBS Real Estate Research, 2006). Real
estate accounts for more than 10 percent of the total value of an efficient mixed-asset
portfolio. The importance of the two asset classes warrants a thorough investigation
into the correlations of the two asset returns over time.
Understanding correlation, which is a key factor in determining optimal portfolio
composition, helps better implement investors’ asset allocation and risk management
strategies. The past studies though find that direct real estate returns have a low
unconditional correlation with equity returns. It is, however, impausible to assume that
the correlations are constant across all sample periods (Lee, 2002). The mean variance
portfolio optimization model that makes a constant correlation assumption is
susceptible to errors when short-run conditional heteroskedasticity are observed in
the asset returns. The restrictive assumption should be relaxed to better model the
dynamics of the two asset markets (Elton and Grubber, 1973).
If common macroeconomic factors significantly explain variations in the returns of
real estate and stocks, correlations between the two asset classes are likely to be
time-variant (Liang and Schulz, 2008). Few studies have thus far attempted to explicitly
model time-varying correlations between direct real estate and stock returns. This
research attempts to fill in the gap in the literature by empirically testing the
time-varying correlations of the two assets across six international markets over sample
periods that go as far back to 1977 (for the US market).
The remainder of the paper is organized as follows: Section 2 reviews the literature
examining long-run and short-run correlations between direct real estate and general
stocks, and the impact of different growth cycles, inflation, and monetary policies.
Section 3 discusses data and empirical methodology. Section 4 analyses the empirical
results of dynamic correlations between returns of stocks and real estate. Section 5
concludes the study.
2. Literature review
Real estate offers significant diversification benefits to investors because of its low
correlation with general stock price changes (Ibbotson and Siegel, 1984; Hartzell, 1986;
Worzala and Vandell, 1993; Fu and Ng, 1997). However, there are some empirical results
that go against the proposition that direct real estate is a good diversifier for equity
portfolios. For examples, some studies show that correlations between direct real estate
and general stocks are positive (Quan and Titman, 1999), and non-static (Brown and
Matysiak, 2000); and Liow (2006) shows that the two asset markets are integrated.
Generalized autoregressive conditional heteroskedasticity (GARCH) models have
been widely used by researchers to examine the stochastic behavior of financial time
series (Bollerslev et al., 1992). From a univariate analysis, the GARCH family of models
is extended to a multivariate GARCH (MGARCH) structure that allows modeling of
covariance of two or more asset classes. The VEC-GARCH model of Bollerslev et al.
(1988), the constant conditional correlation (CCC) model proposed by Bollerslev (1990)
JPIF
31,2
180
To continue reading
Request your trial