Using clustering methods for a practicable real estate portfolio allocation process

Abstract

By moving asset allocation models from the scientific theory to the investment reality, institutional real estate investors are confronted with a number of practical problems. Starting with a diversified international real estate portfolio by using the portfolio theory, this research paper shows a pragmatic approach to transform the calculated results into a real allocation process for finding a practical target portfolio for direct investments. However, the literature of the last 20 years shows a high level of diversification effects by using the portfolio theory also for real estate portfolios. But there is still a dilemma to transform these results into a daily allocation process. The new transformation process ìas a result of this paperî is using clustering methods and different return calculations, offering more possibilities for choosing the suitable components for an existing real estate portfolio. With a top down to bottom up procedure according to a counter-current principle, it is possible to find properties which are suitable in practice. The paper gives an outlook on the practicable application of the results using the Markowitz theory in consideration of the uncertain and imperfect real estate markets. In practical consideration, there are still problems regarding the properties of direct real estate investments and their markets, for example the characteristics of properties, real estate market situations and sizes. On the basis of empirical statistical tests based on real estate total return indices, this paper aims to find a solution to overcome these difficulties.