Constructing Peer Benchmarks for Mutual Funds

Abstract

Mutual fund managers are evaluated primarily on their ability to outperform their peers (rather than outperform a stated index), although managers have no ex ante information on investment decisions made by their competitors. A new approach for constructing benchmarks of peer managers uses return-based style analysis, deriving a peer group's exposures to a predetermined set of market factors from past performance using an optimization algorithm. The objective of the optimization is to find weights for factors that result in the best possible replication of the mutual fund returns. Understanding the aggregate exposures of a manager's peers (and the dispersion among them) allows managers to better assess the trade-off between the risk taken relative to their peers and the benefit expected (outperformance). The approach offers the advantages of more timely comparisons, the capacity to observe intraperiod shifts, and the ability to construct historical time series more readily than enabled by data on the actual holdings of mutual funds.

TOPICS: Mutual fund performance, statistical methods, interest-rate and currency swaps