Manager Skill and Risk Budgeting

Abstract

Risk budgeting appears to bridge the gap between two important concepts in modern portfolio management: ex ante total return-risk portfolio optimization and ex post performance measurement relative to a benchmark. The fundamental hypothesis is that a marginal increase in allowed tracking error is associated with an increase in expected alpha. Tests of this hypothesis using monthly returns for 366 mutual funds in three asset classes over 1997–2001 reveal little evidence of a relationship between realized alpha and tracking error. In a sample split according to manager skill (information ratio), however, there is a positive relationship between alpha and tracking error for skilled (high information ratio) specialty managers and a negative relationship for unskilled (low information ratio) managers. This suggests that investors who wish to practice risk budgeting share with all investors the difficulty of identifying skilled managers in advance.