This dissertation, which is based on a still progressing research path, presents an innovative approach for evaluating the performance of mutual funds which extends the standard Sharpe’s return-based analysis. The goal is to try to explain the sources of the extra-performance of the funds which remain unexplained by the style analysis, the so called Jensen’s Alpha of the fund’s performance.
We apply the proposed methodology to a large set of simulated mutual fund returns constructed using different degrees of prediction accuracy on the historical realizations of stocks, bonds and monetary indices over a 30-year window. The results of this simulation exercise show that the proposed approach is able to recover the correct degree of prediction accuracy in all the different settings considered and it thus seems to be a promising starting point for future extensions and applications.