Is There A Common Trend In Securitized Real Estate Market Correlations With The Regional Stock Market And The World Stock Market?

This paper searches for a common trend in the securitized real estateñregional stock market and securitized real estate-global stock market correlations. Although previous studies have reported that correlations between international stock returns, and to a lesser extent, real estate securities returns have increased in the context of globalization and financial integration, there has been little work on the common behavior of the real estate securities market correlations with the regional stock market as well as with the global stock market themselves. This is where this study intends to contribute. Using weekly local return data from the S&P Property database over the period January 2000 ñ January 2010, we examine long and short-term behavior of return co-movements between 11 European securitized real estate markets (Austria, Belgium, Finland, France, Germany, Italy, Netherlands, Spain, Sweden, Switzerland and UK) and the regional stock market as well as the global stock market after controlling for volatility spillovers across the markets. We first use the Asymmetric-Bivariate BEKK-GARCH methodology to estimate the time-varying correlation between each of the real estate securities markets and the regional stock market and between each of the real estate securities markets and the global stock market after controlling the effect of own- and cross-volatility spillovers. Then we apply the Autoregressive Distributed Lag (ARDL) linear cointegration as well as the Gregory-Hansen (GH) approaches to check whether the estimated regional and global conditional correlation series share at least a common trend in the respective long-term relationships. One key advantage of the ARDL approach is that it can be applied to the conditional correlation series regardless of whether they are I(0) and I(1), and this avoids pre-testing problems associated with the Johansen FIML test which requires the classification of the correlation variables into I(1) and I(0). The GH methodology allows for a structural break in the correlation relationship especially in light of the current global financial crisis which is covered by the present study. Under this approach, the timing of the structural change is estimated endogenously. Finally, the short-term adjustment will be investigated via a standard vector-error correction model. Additionally, we also repeat the analyses by including the US (regional and global) correlations into the cointegrating space and both set of findings (i.e. with and without the US real estate securities markets) will be compared . Our study is important as investors continue to search for diversification in international real estate securities and mixed-asset portfolios.