Stock Index Performance Attribution: Stock, Sector and FX Contributions
A stock index return can be decomposed into constituent, sector and currency contributions. Learn how attribution works and where it can mislead.
This guide examines the methodology, practical calculations, interpretation risks and a repeatable professional research process using clearly labelled hypothetical examples.
Key takeaways
- A constituent’s contribution is approximately its starting weight multiplied by its return for a simple period.
- A sector with the strongest return need not contribute most index points.
- Portfolio attribution should use exact time series and weights rather than simply adding end-period weights.
- International index returns may combine local equity movements with foreign-exchange effects.
What performance attribution answers
Performance attribution examines which holdings, sectors, countries or factors account for a measured index return. It is not enough to observe that a benchmark gained 3% during a month. An analyst should ask whether gains were broad, concentrated in a few companies, driven by defensive sectors or affected by currency translation.
The answer depends on the exact index: capitalisation weighted, equal weighted, sector capped, local currency, gross total return or net total return. Attribution is a decomposition of an observed number, not a causal proof of why investors traded.
Weighted contribution: the basic equation
For a simplified one-period portfolio with no flows or rebalancing, a constituent’s approximate return contribution equals its beginning weight multiplied by its period return. If a company starts at 6% weight and gains 10%, its contribution is around 0.60 percentage points. If another starts at 3% and loses 8%, its contribution is −0.24 percentage points.
These are contributions in percentage points, not the share’s return in percent. A stock can generate a high raw return but little benchmark impact when its weight is small. Dividing a contribution by the total benchmark return can be misleading when the aggregate return is near zero.
A four-sector numerical example
Imagine a simple index with Technology at 40% and up 5%, Financials at 25% and down 2%, Health Care at 20% and up 1%, and Industrials at 15% and up 4%. The respective contributions are +2.00, −0.50, +0.20 and +0.60 percentage points. The total simplified index return is +2.30%.
Although Industrials rose 4%, Technology dominates the result because it began with a much larger weight. Conversely, a smaller sector could rise much faster yet contribute less. This is why sector rotation claims should report both relative performance and weight-adjusted contribution.
Chain-linking and changing weights
Real indices experience price movement, dividend events, corporate actions, membership changes and periodic rebalances. Starting weights drift within the period. Simply multiplying month-end weights by monthly returns does not generally reproduce benchmark performance, particularly if a company entered mid-month.
Professional attribution often computes contributions over short linked periods using official constituent data and then geometrically links the results. Different attribution conventions allocate interaction or residual effects differently. If the sum of reported effects does not reconcile, check methodology and data alignment before interpreting the unexplained gap.
Currency translation for a UK observer
A GBP-based investor in a foreign equity index is exposed to local equity moves and currency conversion unless a hedge is applied. A useful simplification is (1 + local return) × (1 + currency return for the investor) − 1. If a foreign index gains 8% in its local currency while that currency loses 5% against sterling, the investor’s unhedged GBP return is 1.08 × 0.95 − 1 = +2.6%.
Simply adding +8% and −5% gives +3%, which misses the interaction term. Actual net returns also depend on dividends, fund charges, withholding taxes and valuation timestamps. A GBP dealing currency does not automatically mean that foreign-exchange risk has been hedged.
Attribution is not causal explanation
If technology names contributed most of an index gain, that does not show technology optimism was the only driver. A broad fall in discount rates, earnings surprises, positioning or liquidity can affect several groups simultaneously. Attribution explains where returns were recorded, while a causal analysis tests the underlying economic and market narratives.
Use earnings revisions, duration sensitivity, real yields, sector valuations, breadth, financing conditions and currency movement to test competing explanations. Avoid treating one day’s contribution heatmap as a prediction of the next session’s leadership.
A repeatable attribution dashboard
Record benchmark identifier, return type, currency, start and end dates, sector classification version and constituent weights. Show top positive and negative stock contributions, sector contribution, sector-relative return and FX effects. Prefer consistent source data and include a reconciliation total.
Then ask whether the result was diversified, whether concentration changed, and whether an apparent gain reflected profit growth or multiple expansion. For index ETFs, compare benchmark attribution with any fund implementation difference so portfolio execution is not mistaken for index composition.
Worked comparison table
| Hypothetical sector | Starting weight | Sector return | Return contribution |
|---|---|---|---|
| Technology | 40% | +5% | +2.00 pp |
| Financials | 25% | −2% | −0.50 pp |
| Health Care | 20% | +1% | +0.20 pp |
| Industrials | 15% | +4% | +0.60 pp |
| Total | 100% | Weighted average | +2.30 pp |
Figures are hypothetical, rounded where shown and provided solely to demonstrate the mechanics.
Frequently asked questions
Is the top-performing sector always the biggest contributor?
No. Weight and return both determine contribution.
Why do contributions sometimes fail to sum to the index return?
Possible causes include timing, rebalances, dividends, residual effects, data conventions or inconsistent weight series.
Should I use start or end weights?
A simple one-period calculation uses starting weights; more complex periods require linked attribution methods.
Does a GBP-listed ETF remove currency exposure?
No. Trading currency and currency hedging are separate concepts.
Sources and further reading
Continue learning: Stock index sector rotation and Currency hedged versus unhedged indices.
Educational content only. Illustrations are hypothetical and do not constitute investment advice, forecasts, trading signals or guaranteed outcomes. Financial products and leveraged instruments carry risk of loss.