Title: Accounting for non-normality and luck in fund peer performance evaluation
Authors: David Ardia - HEC Montréal (Canada)
Kris Boudt - Vrije Universiteit Brussel (Belgium) [presenting]
Abstract: Delegating the investment decision to a group of carefully chosen fund managers requires to evaluate how well the manager performs compared to his/her peers. The question then is not only whether the manager is individually talented, but also whether he/she is more talented than the others. We discuss the statistical tools in the R package PeerPerformane that provide an answer to this question based on the analysis of historical fund returns. In particular, we stress that such a peer performance evaluation requires to take the non-normality of the fund return series into account, and also needs to correct for the possibly large number of false positives, when comparing the fund's performance with a universe of peer funds. We illustrate this for various performance measures, including the fund's alpha residual performance measure and the modified Sharpe ratio.