Mean-variance vs. full-scale optimization: broad evidence for the U.K.

In the Full-Scale Optimization approach the complete empirical financial return probability distribution is considered, and the utility maximising solution is found through numerical optimization. Earlier studies have shown that this approach is useful for investors following non-linear utility functions (such as bilinear and S-shaped utility) and choosing between highly non-normally distributed assets, such as hedge funds. We clarify the role of (mathematical) smoothness and differentiability of the utility function in the relative performance of FSO among a broad class of utility functions. Using a portfolio choice setting of three common assets (FTSE 100, FTSE 250 and FTSE Emerging Market Index), we identify several utility functions under which Full-Scale Optimization is a substantially better approach than the mean variance approach is. Hence, the robustness of the technique is illustrated with regard to asset type as well as to utility function specification.