Correlation structure of real estate markets over time

Abstract

PurposeThe purpose of this paper is to examine the time‐varying correlation structure of international real estate stock markets and its implications for portfolio management.Design/methodology/approachThe analysis focuses on real estate markets only and examines the appropriateness of the Markowitz approach based on mean‐variance‐optimization. Therefore, the properties of the return distributions are analyzed first. Afterwards, the stability of the correlation and covariance structure over time is analyzed and statistically tested by the Jennrich test.FindingsBecause of low correlation among real estate markets worldwide there exists diversification benefits from broadening the investment horizon to international real estate markets. However, using correlation coefficients as a measure for diversification benefits is limited by empirical findings: Returns are not normally distributed, correlations increase in downward moving phases and neither the correlation nor the covariance structures are statistically stable over time. Therefore, mean‐variance optimization is inherent with misleading results.Originality/valueThe study provides some interesting and valuable insights into the correlation structure of real estate stock markets over time and the limitations for portfolio management based on mean‐variance‐optimization.