The widespread availability of passive products have investors asking: Why pay for active management if I can potentially do just as well with a passive product? Active managers must respond, demonstrating the value of their products and substantiating that value by demonstrating consistent and repeatable investment skill. Performance attribution is a natural starting point for assessing the value of an investment product. One of the issues encountered when using standard attribution methods is that the structure used to decompose returns—a classification scheme for Brinson-based attribution and a linear returns model for factor-based attribution—may not match the investment process of the portfolio manager. In this paper, we contend that investment managers should tailor attribution methods to match their investment processes, highlighting sources of portfolio returns that correspond to their intended bets and showing that these sources are statistically significant overtime. The goals of this paper are two-fold: first, we identify issues commonly encountered with standard, off-the-shelf performance attribution methodologies, and, second, we suggest ways to customize and extend these methodologies to yield attribution results that better match the investment process at hand.
By Robert Stubbs, PhD and Esther Mezey, PhD