A New Look on Classification of Spanish Mutual Funds Working Paper in progress: First Draft Please, do not quote without permission of authors

In this paper, we apply a nonlinear methodology (Self-Organizing Maps and k-means algorithm) to classify mutual funds. This methodology allows us to identify which mutual funds did not follow the investment objectives established from its official category. We also propose an alternative classification, which has a lower rate of misclassified mutual fund. It achieves better performances in terms of return/risk, with a lower number of assets, which permits increase the final utility of particular investors. We find that in average 60% of Spanish mutual funds are misclassified, according with their financial attributes. It must be taken into account that grouping mutual funds correctly may have several implications from ex-post and ex-ante benchmark and performance measures.