Alternative Approaches to Traditional Mean-Variance Optimisation

The use of Mean-Variance Optimisation as of tool of strategic asset allocation is affected by estimation errors of the inputs required for portfolio construction and the extreme sensitivity of Markowitz portfolios to estimation errors. In recent years, having become aware of these problems, academics, practitioners, and institutional investors have re-evaluated existing strategies or promoted new approaches to portfolio construction that no longer strive to optimise the trade-off between expected return and risk, and consequently no longer resort to the application of Mean-Variance Optimisation. The common feature of these strategies is that they remove expected returns from the set of inputs required for portfolio construction, which explains their being called μ-free strategies. Their development only involves estimating and modelling risk parameters, and their focus is exclusively on the risk dimension of portfolios. For this reason, they are commonly known as risk-based asset allocation approaches or strategies. This chapter illustrates two examples which are representative of this class of asset allocation methodology: the global minimum-variance strategy and the optimal risk parity strategy.