Abstract
Executive Summary. By moving asset allocation models from the classroom to the boardroom, institutional real estate investors are confronted with a number of practical problems. Many of these difficulties are directly related to both data quality and availability and have received much attention in the current literature. While important, these concerns are perhaps overshadowed by the impact of uncertainty on the allocation process. In this study, uncertainty is introduced by bootstrapping expected returns and standard deviations in a traditional Markowitz framework. The results show that the efficient frontier is not singular, but “fuzzy.” That is, this study illustrates that numerous statistically dissimilar weighted portfolios can be equally attractive for any given combination of expected risk and return and that allocation ranges rather than specific targets are, more realistic.