Dynamic Correlations Among Asset Classes: REIT and Stock Returns

Abstract

We use the Dynamic Conditional Correlation model with Generalized Autoregressive Conditional Heteroskedasticity (DCC-GARCH) developed by Engle (Journal of Business & Economic Statistics 20(3):339–350, 2002) to examine dynamics in the correlation of returns between publicly traded REITs and non-REIT stocks. The results suggest that REIT-stock correlations form three distinct periods. During the first period, ending in August 1991 with the start of the modern REIT era, correlations were high and without trend, never dipping below 59%. During the second period, ending in September 2001 with the inclusion of REITs in broad stock market indexes, correlations declined precipitously to 30%, enabling substantially higher portfolio allocations to both high-return asset classes and therefore higher portfolio returns without increasing portfolio volatility. During the third period, since September 2001, correlations increased steadily but only reached 59% in late 2008. A simple portfolio optimization suggests that asset managers would be willing to pay 20 basis points per year, plus the difference in transaction costs, for the ability to use DCC-GARCH modeling of dynamic correlations in place of rolling 24-month asset correlations.

Keywords: DCC-GARCH, REITs, Correlation, Volatility, Portfolio optimization, Asset allocation