Risk Measurement Choice in Selecting REITs: Evidence from the U.S. Market

Abstract

Executive Summary. The returns of real estate investments trusts (REITs) in the United States are not normally distributed and should therefore not be analyzed using the standard mean-variance approach. In this paper, we compare REIT performance rankings based on the Sharpe ratio with those using risk-adjusted performance (RAP) measures that do not assume the normality of the return distribution. The results demonstrate that the rankings obtained by different risk measures are significantly correlated with the Sharpe measure but some of them can show a high degree of time persistence with respect to the Sharpe index. We also study the usefulness of RAP rankings in selecting REITs by comparing the results of a naïve diversified portfolio with those of portfolios concentrating on only the top REITs identified using different RAP measures. The results demonstrate that past information is useful in improving REIT portfolio performance and the choice to consider a more complex RAP measure can increase the yearly performance of investment strategy by around 1%–2%, especially for value-weighted portfolios and/or multiple-year horizons.