A portfolio manager's returns are generally compared with the returns to some reference or benchmark portfolio. The difference between the portfolio's return and benchmark return is called as manager's excess return. Portfolio performance attribution systems separate this excess return into three categories an allocation effect, a selection effect and a cross product, or interaction, effect.
These measures of the determinants of portfolio performance will be distorted, however, if managers systematically choose portfolios whose risk levels differ from their appropriate benchmarks. For many managers these distortion may be small. Research suggests, however, that for as much as one-third of certain style managers, risk distortion could amount to 160 to 240 basis points a year.
A simple risk-adjustment procedure can compensate for the distortion in each of the attributes. Performance attributes can thus be separated from normal expected returns to risk.