Performance Attribution — What Drives Returns?
Performance attribution splits the gap between a fund's return and its benchmark's return into the decisions that produced it. This lesson explains the allocation and selection effects, the interaction term, and what attribution can and cannot show.
The question attribution answers
Knowing that a fund finished ahead of or behind its benchmark says nothing about why. Performance attribution takes the difference between the two returns, the excess return, and splits it into the decisions behind it.
It works by comparison. For each sector or asset class, the fund's weight and return are set beside the benchmark's weight and return for the same sector.
Allocation and selection
The allocation effect comes from holding more or less of a sector or asset class than the benchmark does. If a fund holds more of a sector that goes on to do better than the benchmark as a whole, the gain from that overweight is the allocation effect.
The selection effect comes from the particular securities chosen within a sector doing better or worse than that sector in the benchmark. It measures the fund's holdings against the sector, not against the whole market.
The two overlap: a fund can hold more of a sector and also pick securities within it that do better. An interaction term covers that overlap. Timing is sometimes measured separately.
Other things that move the return
Style analysis is a related tool: it looks at a portfolio's tilts, such as growth versus value, or large versus small companies.
Costs such as the expense ratio reduce a fund's return. They are part of the gap between fund and benchmark, but they are not one of the two main effects.
What attribution cannot show
Attribution explains a past period only. An effect can reflect luck as well as skill: a positive selection effect is often read as a sign of stock-picking, but one period cannot separate the two.
The result is also specific to the benchmark it is measured against. It says nothing certain about future results.
Naming the effect (illustrative)
A fund holds 20% in automobile shares when its benchmark holds 12%, in a year when the automobile sector does better than the benchmark as a whole. The gain from holding 8 percentage points more of the sector is an allocation effect.
In the same year the fund's banking shares return less than the banking sector in the benchmark. That shortfall is a selection effect, and it is negative. One fund can show a positive effect of one kind and a negative effect of the other in the same period.
A two-sector attribution (illustrative)
- Assumed figures, for arithmetic only. The benchmark holds 50% in sector X and 50% in sector Y; over the year sector X returns 20% and sector Y returns 10%. Benchmark return = (0.5 × 20) + (0.5 × 10) = 10 + 5 = 15%.
- The fund holds 70% in X and 30% in Y. Its X shares return 20%, the same as the sector; its Y shares return 14%. Fund return = (0.7 × 20) + (0.3 × 14) = 14 + 4.2 = 18.2%. Excess return = 18.2 − 15 = 3.2 percentage points.
- Allocation: with the fund's weights but the benchmark's sector returns, the return would be (0.7 × 20) + (0.3 × 10) = 14 + 3 = 17%. Allocation effect = 17 − 15 = 2.0 percentage points.
- Selection, at benchmark weights: the fund's Y shares did 14 − 10 = 4 percentage points more than sector Y, and its X shares matched sector X. Selection effect = 0.5 × 4 = 2.0 percentage points.
- Interaction: the fund held 30% in Y, not 50%, so the weight difference is 0.3 − 0.5 = −0.2. Interaction = −0.2 × 4 = −0.8 percentage points.
- Check: 2.0 + 2.0 − 0.8 = 3.2 percentage points, the excess return.
Result. Of the 3.2 percentage points of excess return, 2.0 came from allocation, 2.0 from selection and −0.8 from the interaction of the two.
Key points
- Attribution splits a fund's excess return over its benchmark into the decisions behind it.
- The allocation effect comes from holding more or less of a sector or asset class than the benchmark.
- The selection effect comes from the securities chosen within a sector doing better or worse than that sector in the benchmark.
- An interaction term covers the overlap of the two; timing is sometimes measured separately.
- Costs reduce return but are not one of the two main effects.
- An effect can reflect luck as well as skill, and attribution describes a past period only.
Common misunderstandings
- A gain from holding more of a sector is not a selection effect: it is the allocation effect.
- The expense ratio is not one of the two main effects: costs reduce return, but allocation and selection are the main split.
- A positive selection effect does not prove skill: it can reflect luck, and it does not guarantee future results.
Questions people ask
What is the selection effect compared against?
The same sector in the benchmark, not the benchmark as a whole.
Why is there an interaction term?
Because the allocation and selection decisions overlap: the effect of the securities chosen in a sector depends on how much of that sector the fund held.
Does attribution show how a fund will do next year?
No. It explains a past period only and says nothing certain about future results.
What this lesson relies on
- SEBI Master Circular for Mutual Funds, 20 March 2026 — benchmarks against which scheme performance is compared
- Standard definitions of the allocation, selection and interaction effects in performance attribution (plain arithmetic)
This lesson was reviewed independently against these sources on 8 October 2026. Rules change: check the current regulation, scheme document or policy wording before relying on any figure. This is education, not advice.