This month we are taking a different approach for our Risk Focus publication.  I had the pleasure of attending Campbell Harvey’s presentation on New Approaches to Performance Evaluation last fall, at the International Centre of Pension Management’s Montreal discussion forum. Campbell discussed a range of issues and new approaches to evaluating the performance of portfolio managers, and so continuing our partnership with ICPM we wanted to pass on his research. We are also pleased that Campbell has agreed to present at the GRI conference this fall, where his topic will be  an Introduction to Blockchain.The materials this month come in two parts. First, below we reproduce the slides from Campbell’s Montreal presentation, which effectively walks through the concept of Type I (picking a poor performing asset manager by mistake) and Type II errors (rejecting a strong performing asset manager). Campbell explains how the propensity to add multiple explanatory factors to models, and the propensity to perform multiple tests, causes models to overstate the claimed performance of portfolio models; he then sets out to rigorously test performance assertions and offers three approaches to separate luck from skillful management (explicitly adjusting for multiple tests; bootstrapping; and noise reduction (i.e. examining why past performance does such a poor job of indication future performance.))So we feel that a read through the slide show below will provide you with interesting insights and perspectives on the topic of performance evaluation of portfolio performance.  For those members with a deeper interest in the topic we are also attaching links to five research papers by Campbell and his colleague Yan Liu, which walk through these issues systematically:

1. Rethinking Performance Evaluation, by Campbell R. Harvey and Yan Liu
2. Evaluating Trading Strategies, by Campbell R. Harvey and Yan Liu
3. Backtesting, by Campbell R. Harvey and Yan Liu
4. Lucky Factors, by Campbell R. Harvey and Yan Liu
5. …and the Cross-Section of Expected Returns, by Campbell R. Harvey & Yan Liu

We would like to thank Campbell and Yan for allowing us to reproduce their work here for our members, and we would also like to acknowledge ICPM for their ongoing support.You can view Campbell’s presentation below: