Go to page 2, 3, 4, Next       

Ron Surz provides one explanation of how peer group composition explains the underperformance of value managers in 2006, followed by their outperformance in 2007.

So, what constitutes a valid peer group, and how can advisors assess the reliability of peer group analysis?

Peer Group Composition

In a perfect world, all funds in a peer group would be restricted to investing in the same group of securities, and these securities would have common characteristics (e.g., they would all lie within one of the nine conventional style boxes).   Such homogeneity would facilitate an easy comparison within the peer group; funds would be ranked by risk-adjusted returns, and the ones at the top would be the clear winners. 

But peer groups are not homogenous, and in some cases they do not represent the complete universe of funds for comparative purposes. Three primary biases - backfill bias, survivorship bias, and classification bias – lead to non-homogenous or incomplete peer groups.  Other biases, particularly in hedge fund databases, compromise the utility of peer group analysis.

Backfill bias occurs when a fund does not begin reporting results until it has achieved success, omitting poor performance which may have occurred in the early stages of the fund.  Survivorship bias occurs when funds stop reporting results because of poor performance or in the event of failure and dissolution.  In the tightly regulated world of mutual funds, backfill and survivorship biases are typically ignored.  In the unregulated world of hedge funds, backfill and survivorship biases are far more problematic. It is critical to understand the collection process used by hedge fund data providers to assess the impact of any backfill or survivorship biases.  Both biases have the effect of artificially inflating the results of the remaining universe.

Display article as PDF for printing.

Would you like to send this article to a friend?