Brinson Performance Attribution: Reading Allocation, Selection, and Interaction Effects

The Brinson attribution table is one of those test-bank traps that punishes pattern-matching: many students see a big positive selection effect and pick that sector, only to discover the question asked for BOTH effects positive. This article covers the Brinson decomposition (the Hood-Beebower and Fachler variants), shows the math behind each row, and explains how to identify which decisions actually added value.

What the Brinson model does

A portfolio manager makes two big decisions: how much to put in each sector (allocation), and which stocks to pick within each sector (selection). The Brinson model separates these so each can be graded independently.

Total active return = Allocation effect + Selection effect + Interaction effect

flowchart TD A[Active Return: Rp - Rb] --> B[Allocation Effect] A --> C[Selection Effect] A --> D[Interaction Effect] B -->|Did I overweight winning sectors?| E[Sum over sectors of Pw - Bw times Rb_sector - Rb_total] C -->|Did my picks beat the sector?| F[Sum over sectors of Bw times Rp_sector - Rb_sector] D -->|Cross term: overweight AND picked winners| G[Sum over sectors of Pw - Bw times Rp_sector - Rb_sector] style B fill:#c9a84c,color:#0a0a0f style C fill:#c9a84c,color:#0a0a0f style D fill:#c9a84c,color:#0a0a0f

(In the Brinson-Hood-Beebower model the selection effect uses $P_w$ instead of $B_w$ ; both variants are tested on the CFA exam.)

Allocation effect: the sign rule

The Brinson-Fachler allocation formula:

\text{Allocation}_i = (P_w - B_w) \times (R_{B,i} - R_{B,\text{total}})

Read the sign carefully:

$(P_w - B_w)$	$(R_{B,i} - R_{B,\text{total}})$	Product	Interpretation
Positive (overweight)	Positive (winning sector)	+	Good — overweighted a winner
Positive (overweight)	Negative (losing sector)	−	Bad — overweighted a loser
Negative (underweight)	Positive (winning sector)	−	Bad — missed a winner
Negative (underweight)	Negative (losing sector)	+	Good — avoided a loser

flowchart TD A[Allocation effect signs] --> B[Overweight winner: +] A --> C[Underweight loser: +] A --> D[Overweight loser: -] A --> E[Underweight winner: -] style B fill:#0d6e3a,color:#ffffff style C fill:#0d6e3a,color:#ffffff style D fill:#7c1d1d,color:#ffffff style E fill:#7c1d1d,color:#ffffff

This is where students get confused. A NEGATIVE allocation effect for a winning sector means the portfolio UNDERWEIGHTED that good sector. A positive allocation effect for a losing sector means the portfolio CORRECTLY underweighted (avoided) it.

Worked example — Technology underweight

Suppose a portfolio has:

Tech portfolio weight $P_w = 35\%$
Tech benchmark weight $B_w = 40\%$
Tech benchmark return $R_{B,\text{Tech}} = 10.10\%$
Total benchmark return $R_{B,\text{total}} \approx 6.00\%$

\text{Allocation}_{\text{Tech}} = (0.35 - 0.40) \times (0.1010 - 0.0600) = (-0.05) \times (0.0410) = -0.205\%

The portfolio UNDERWEIGHTED a winning sector. Result: negative allocation effect. (The −0.505% in some test-bank tables comes from a slightly different total-benchmark assumption — the sign and direction are the point.)

Selection effect: the simpler one

The Brinson-Fachler selection formula:

\text{Selection}_i = B_w \times (R_{P,i} - R_{B,i})

Sign rules are simpler:

$R_{P,i} > R_{B,i}$ (your stocks beat the sector benchmark) → POSITIVE selection
$R_{P,i} < R_{B,i}$ (your stocks lost to the sector benchmark) → NEGATIVE selection

Selection grades stock-picking WITHIN a sector, independent of how much you allocated to that sector.

Worked example — Technology selection

Tech portfolio return $R_{P,\text{Tech}} = 11.20\%$
Tech benchmark return $R_{B,\text{Tech}} = 10.10\%$
Tech benchmark weight $B_w = 40\%$

\text{Selection}_{\text{Tech}} = 0.40 \times (0.1120 - 0.1010) = 0.40 \times 0.0110 = +0.44\%

Your Tech stocks beat the Tech benchmark by 110 basis points, and the resulting selection effect is +0.44% on the overall portfolio.

What does "optimal decision" mean?

In a typical Brinson attribution table you might see:

Sector	Allocation	Selection	Interaction	Optimal?
Technology	−0.505%	+0.44%	−0.055%	NO — selection good but allocation bad
Basic Materials	+0.64%	+0.06%	+0.04%	YES — both positive
Health Care	0.00%	−0.16%	0.00%	NO — neither
Utilities	+0.055%	−0.275%	−0.055%	NO — selection bad

For a decision to be "optimal," BOTH the allocation and the selection effects must be positive. The manager made both correct calls: overweighted (or correctly underweighted) the right sector AND picked the right stocks within it.

The Technology trap

Technology shows the classic conflict:

Tech benchmark return is high (winning sector → $R_{B,\text{Tech}} > R_{B,\text{total}}$ )
Portfolio UNDERWEIGHTED Tech ( $P_w < B_w$ ) → allocation effect NEGATIVE
Tech portfolio stocks beat the Tech sector benchmark ( $R_{P,\text{Tech}} > R_{B,\text{Tech}}$ ) → SELECTION effect POSITIVE

The manager picked great Tech stocks (good selection) but did not own enough of them (bad allocation). It is a half-right decision.

How to read an attribution table fast on the exam

flowchart TD A[Attribution table given] --> B[Read the question carefully] B --> C{What does the question ask?} C -->|Best allocation decision| D[Find largest positive allocation column] C -->|Best selection decision| E[Find largest positive selection column] C -->|Best/optimal decision| F[Find rows where BOTH allocation AND selection positive] C -->|Highest total contribution| G[Sum allocation + selection + interaction; find max row] C -->|Worst decision| H[Find rows where BOTH allocation AND selection negative] style F fill:#c9a84c,color:#0a0a0f style H fill:#7c1d1d,color:#ffffff

"Optimal" almost always means both effects positive. "Worst" means both negative. A mixed-sign row (one positive, one negative) is neither optimal nor worst — it is a half-right or half-wrong decision.

Brinson-Hood-Beebower vs Brinson-Fachler

Two formulations differ on the allocation reference and the selection weighting:

Component	Brinson-Hood-Beebower (BHB)	Brinson-Fachler (BF)
Allocation	$(P_w - B_w) \times R_{B,i}$	$(P_w - B_w) \times (R_{B,i} - R_{B,\text{total}})$
Selection	$P_w \times (R_{P,i} - R_{B,i})$	$B_w \times (R_{P,i} - R_{B,i})$
Interaction	embedded in selection	$(P_w - B_w) \times (R_{P,i} - R_{B,i})$

The Brinson-Fachler form is the version the current CFA curriculum emphasizes because the allocation effect is intuitive: overweighting a sector that beat the OVERALL benchmark gives a positive allocation, regardless of whether that sector had positive or negative absolute return. In BHB, allocation depends on absolute sector return (so overweighting a sector with positive absolute return — even if it underperformed the benchmark — gives a positive allocation effect, which can mislead).

Interaction effect — usually small, sometimes diagnostic

The interaction effect is the cross-term: $(P_w - B_w) \times (R_{P,i} - R_{B,i})$ . It is positive when both decisions go the SAME direction (overweighted AND picked winners, OR underweighted AND avoided losers) and negative when they go opposite directions.

flowchart TD A[Interaction effect signs] --> B[Overweight + Good picks: ++ -> +] A --> C[Underweight + Bad picks: -- -> +] A --> D[Overweight + Bad picks: +- -> -] A --> E[Underweight + Good picks: -+ -> -] style B fill:#0d6e3a,color:#ffffff style D fill:#7c1d1d,color:#ffffff

For the Tech row in our example: (P_w − B_w) = negative, (R_P − R_B) = positive → interaction NEGATIVE (−0.055%). This makes sense: the manager picked great Tech stocks but did not own enough of them — so the SECOND-ORDER effect (the cross-term) cost something.

Practice and dig deeper

Test your attribution skills in our CFA Level III question bank or discuss real attribution scenarios with peers in the community. The Brinson model is one of those topics that pays off enormously once you can read a table fast — practice on a few different layouts and the sign rules become automatic.

Brinson Performance Attribution: Reading Allocation, Selection, and Interaction Effects (CFA Level III)