The Alpha and Beta of Risk Attribution

Abstract

Davis and Menchero present a methodology for attributing the alpha and beta components of tracking error. They define an intuitive and precise decomposition of tracking error into alpha and beta components. Their approach is completely general and is applicable to security-, sector-, and factor-based investment processes. The authors show that by following this methodology, drilldowns into tracking error at any level of granularity can be neatly decomposed into parallel alpha and beta drilldowns. They also show that factor-based risk attribution represents an important special case of the alpha and beta risk decomposition, and they demonstrate that the residual exposure-based analysis of residual risk is prone to counterintuitive attribution effects.

TOPICS: Factor-based models, volatility measures, statistical methods