The Brinson-Hood-Beebower (BHB) model, introduced in the 1986 paper Determinants of Portfolio Performance, decomposes the difference between a portfolio's return and its benchmark into three effects:
- Allocation effect — the contribution from over- or under-weighting asset classes relative to the benchmark
- Selection effect — the contribution from choosing securities that outperform the benchmark within each asset class
- Interaction effect — the combined effect of tilting toward asset classes where selection was also positive (or negative)
Together: Allocation + Selection + Interaction = Total Active Return.
BHB attribution is standard in institutional portfolio reporting and provides a clean decomposition of whether alpha came from tactical asset allocation or from stock (or bond) picking within each segment.
In quantitative signal research
BHB-style attribution extends naturally to factor models: active returns can be decomposed into the contribution of each signal rather than each asset class, identifying which forecasting inputs drove performance and which were noise over the evaluation period.