Risk-based performance attribution is a method of explaining a portfolio's results relative to its benchmark in terms of exposures to risk factors, such as market beta, size, value, momentum, industries, countries, currencies or interest-rate duration. It is a form of performance attribution built on a factor model: the active return is divided into each factor's exposure multiplied by that factor's return, plus a stock-specific residual.
The same model can decompose active risk, so that a manager's tracking error is traced to its sources, both before the period (forecast) and after it (realised). This differs from the Brinson attribution model, which groups results by allocation and selection decisions across sectors or regions. Risk-based analysis shows whether returns came from intended bets or from unintended factor exposures, and is used to check that active management matches a stated strategy. Its results depend heavily on the factor model chosen.
Ref 1180001910