Peer Group Rankings and Their Biases

A common way to judge a fund manager is to rank their return against a "peer group" of similar funds — say, all large-cap US equity mutual funds — rather than against a single index. The appeal is real: peer rankings implicitly control for a shared environment, since every fund in a badly performing category was fighting the same headwinds. But the ranking is only as good as the peer group's construction, and several quiet biases creep in.

Survivorship bias is the biggest: peer-group databases are typically built from funds that exist today, so funds that closed or merged away due to poor performance have already been dropped from the comparison set, silently raising the average return of the surviving peer group and making an ordinary fund look worse by comparison than it actually was relative to its true original peer set. Style drift compounds this: a fund manager who quietly shifted from value to growth stocks over the measurement period is still compared against a "value fund" peer group, even though it no longer holds a comparable portfolio, making the ranking compare apples to a partially-grown orange. Peer groups are also often self-reported and voluntary, so managers with disappointing recent performance can simply stop reporting, further inflating the group's apparent average.

None of this means peer rankings are useless — they still correct for a real, shared market environment that a single benchmark can't fully capture — but a ranking should always be read alongside the peer group's survivorship policy and each fund's actual holdings-based style, not taken as skill measured in a vacuum.