A Stochastic Programming Approach for the Asset and Liability Management of a Real Estate Investment Trust

Not only insurance and banking companies, or pension funds and trusts, but many more examples can be found where stochastic programming has been both theoretically and practically proven to help the management to make superior decisions in pursuing shareholder wealth and hedging against risk. This paper deals with applying this approach to strategic asset and liability management of UK based real estate investment trusts (REITs). A multistage stochastic program is used to represent the multi-period sequential decision process of the REITís management which pursues an optimal strategic composition of different asset and liability classes. The model that is proposed in this paper contains classes and subclasses composed of assets and liabilities that are homogenous in terms of cash flow and risk structure or maturity and payback specification. All assets and liabilities are interrelated through the main components of the model which are the income statement, a cash flow balance and the balance sheet of the REIT. Over the planning horizon of the model different scenarios will be considered to reflect the uncertainty inherent in parameters like rent and acquisition costs per square meter, occupancy, or interest rates. Beside the construction of future scenarios the challenge in the formulation lies in specific regulatory requirements for REITs and the indivisibility of real investment projects, latter leading to a mixed integer mathematical program. Beside the discussion of a REIT specific model and the generation of scenarios, the paper will also provide a numerical example and the choice of software for its solution.