A generalized performance attribution technique for mutual funds

In this contribution we propose a model which can be used to define a measure of the relative performance of mutual funds that takes into account all the different aspects considered by the traditional performance indexes. This model adopts a data envelopment analysis approach, which is an operational research technique that enables to measure the relative efficiency of a set of "decision making units" and can be applied to the measurement of mutual fund performance. The model proposed synthesizes the results of the traditional risk adjusted indexes in a global performance measure which takes into account also the subscription and redemption costs of the investment. An average cross efficiency measure is also suggested, in order to give a measure of the rating of a fund from the point of view of the styles of all the funds under consideration. The generalized performance attribution techniques proposed are tested using data of the Italian mutual funds market