Downside risk optimization in securitized real estate markets

Optimization of international securitized real estate portfolios has been a key topic for several decades. However, most previous analysis has focused on regional diversification by applying the traditional mean-variance (MV) framework suggested by Markowitz (1952) even if the limitations of this approach are well-known. Thus, we focus on a more suitable and appealing downside risk (DR) framework suggested by Estrada (2008), which applies a similar optimization algorithm as the MV framework. The analysis covers the eight largest securitized real estate markets from January 1990 to December 2009 and thus captures a more global perspective. The main findings are as follows: first, the return distributions are non-normally distributed and negatively skewed. Second, optimal portfolio weights differ substantially between the MV and DR approach. Third, portfolio weights are shifted from the U.S. and Australian market to the Dutch and the French market when applying the DR framework instead of the MV framework. Fourth, the dominance of the DR framework is well-documented by comparing out-of-sample performance. The empirical results are remarkable and emphasize the practical merit of the presented DR framework for investors and portfolio managers.