Home / Wiki / B / Brinson Attribution Model

Brinson Attribution Model

Investment and Capital MarketsUpdated 17 Nov 2026
Share

The Brinson attribution model is a method of performance attribution that explains why a portfolio's return differed from its benchmark by dividing the difference into parts linked to the manager's decisions. It asks two main questions for each sector, country or asset class: did the manager hold the right amount of it, and did the manager pick the right securities within it? The answers are reported as an allocation effect, a selection effect and, in the original form, an interaction effect.

The model is named after Gary Brinson and co-authors. Brinson and Nimrod Fachler described a version in a 1985 article on measuring non-US equity performance, and Brinson, Randolph Hood and Gilbert Beebower published their study of pension fund returns, "Determinants of Portfolio Performance", in the Financial Analysts Journal in 1986. The two papers gave rise to the Brinson-Fachler and Brinson-Hood-Beebower (BHB) variants, which remain the standard framework for equity and multi-asset attribution in asset managers, consultants and pension funds.

01

The three effects

The model starts from the benchmark-relative return, the portfolio return minus the benchmark return, and assigns it to segments:

  • Allocation. The allocation effect measures the value added by overweighting segments that did well and underweighting those that did poorly.
  • Selection. The selection effect measures the value added by choosing securities within a segment that beat that segment of the benchmark.
  • Interaction. The interaction effect captures the joint result of being overweight or underweight in a segment where selection was also good or bad.
02

Formula and worked example

For each segment ii, the Brinson-Fachler allocation effect is:

Ai=(wp,i−wb,i)×(Rb,i−Rb)A_i = \left(w_{p,i} - w_{b,i}\right) \times \left(R_{b,i} - R_b\right)

The selection and interaction effects are:

Si=wb,i×(Rp,i−Rb,i)S_i = w_{b,i} \times \left(R_{p,i} - R_{b,i}\right)
Ii=(wp,i−wb,i)×(Rp,i−Rb,i)I_i = \left(w_{p,i} - w_{b,i}\right) \times \left(R_{p,i} - R_{b,i}\right)
  • wp,iw_{p,i} and wb,iw_{b,i} are the weights of segment ii in the portfolio and the benchmark.
  • Rp,iR_{p,i} and Rb,iR_{b,i} are the returns of segment ii in the portfolio and the benchmark.
  • RbR_b is the total benchmark return.

Take a benchmark of 60 per cent equities returning 10 per cent and 40 per cent bonds returning 4 per cent, so Rb=7.6%R_b = 7.6\%. The portfolio holds 70 per cent equities returning 12 per cent and 30 per cent bonds returning 3 per cent, so it returns 9.3 per cent, an active return of 1.7 percentage points. Allocation is 0.10×2.4=0.240.10 \times 2.4 = 0.24 for equities and −0.10×−3.6=0.36-0.10 \times -3.6 = 0.36 for bonds, a total of 0.60. Selection is 0.60×2=1.200.60 \times 2 = 1.20 for equities and 0.40×−1=−0.400.40 \times -1 = -0.40 for bonds, a total of 0.80. Interaction is 0.10×2=0.200.10 \times 2 = 0.20 plus −0.10×−1=0.10-0.10 \times -1 = 0.10, a total of 0.30. The three totals add to 1.70, matching the active return exactly.

03

Brinson-Fachler and BHB compared

The BHB version measures allocation against the segment's own return, (wp,i−wb,i)×Rb,i\left(w_{p,i} - w_{b,i}\right) \times R_{b,i}, rather than its return relative to the whole benchmark. In the example, BHB gives equities 1.00 and bonds -0.40. The total allocation effect is the same, 0.60, because the weight differences sum to zero, but the sector figures differ. Brinson-Fachler is usually preferred for reporting by segment, because under BHB an underweight in a segment with a positive but below-average return is shown as a loss, which is counter-intuitive. Many practitioners also fold the interaction effect into selection, since a manager rarely makes a separate interaction decision.

04

Multi-period results and limitations

The effects add up exactly in a single period, but arithmetic effects for successive months do not compound to the cumulative active return. Practitioners use smoothing or linking algorithms to spread the difference, or switch to geometric performance attribution, which defines excess return as a ratio and links cleanly over time. Other limitations include:

  • Sensitivity to the choice of segments, which should mirror how the manager actually takes decisions.
  • A poor fit for fixed income portfolios, where returns are driven by duration, curve and spread decisions rather than sector weights.
  • Distortions from cash flows, transaction costs and currency effects unless they are measured separately.

Risk-based alternatives such as a factor attribution model explain returns by exposure to factors like size, value or momentum instead of by sector.

05

Use in practice

Brinson attribution is a standard part of manager reporting. Consultants and pension trustees use it to see whether an active manager adds value through asset allocation, through stock picking or through neither. A portfolio manager uses it internally to check that results come from intended bets. It is commonly reported alongside tracking error and the information ratio, which measure how much risk was taken to earn the active return.

Last updated 17 Nov 2026
Share this entry

Ref 1180001300