The Use of MPT for Real Estate Portfolios in an Uncertain World

Executive Summary. This study examines the impact of uncertainty on the asset allocation algorithm. By “bootstrapping” the error terms from OLS-generated equations, multiple frontiers are estimated and “areas of indifference” are created. Since these alternative portfolios are as “efficient” as the original portfolios on the unperturbed frontier, the unique solution provided by traditional mean-variance optimization techniques is called into question. The analysis concludes that ranges, rather than specific investment targets, should be the preferred method of mean-variance optimization. Ranges are not only statistically robust, but also circumvent many of the practical problems associated with owning an illiquid asset, such as valuation issues and lumpiness.Second, the study also examines the role of multifamily housing in a real estate portfolio in the presence of uncertainty. Using econometrically determined ex ante property-type returns, the model's results suggest that most institutional portfolios are likely to be significantly underweighted in apartments using traditional asset allocation techniques.