CVaR optimization of real estate portfolios in an ALM context
||CVaR optimization of real estate portfolios in an ALM context
||19th Annual European Real Estate Society Conference in Edinburgh, Scotland
||In this paper we show the result of an optimization study of sector and region allocation within real estate portfolios. The optimizations are performed in an ALM context. That is, we try to determine optimal allocations based on a funding ratio risk measure for Dutch pension funds. The risk measure we use is the Conditional Value at Risk (CVaR). In our optimizations, we assume a fixed portfolio for the non-real estate part and optimize the composition of the part allocated to non-listed real estate. We conclude that within Europe, the number of countries and sectors that appear in the optimal portfolio is limited. So the diversification gain from investing in a large number of sectors and countries is relatively limited within Europe. For countries outside Europe, the results are sensitive to changes in the input and assumptions. Furthermore, we conclude that high leverage is only acceptable when the underlying real estate market is very stable. Finally, although the optimal weights differ per type of pension fund, the countries and sectors that appear in the optimal portfolios are quite stable.
||file.pdf (800,698 bytes)
Post discussion ...
||Parallel Session H2
These pages are best viewed with any standards compliant browser (e.g. Mozilla).