Performance attribution and hidden family concentration

Establish the return record before explaining it

Agree on the portfolio boundary, valuation dates, currency, income treatment, cash flows, and fee basis. Reconcile the account or consolidated record before producing explanatory charts. A missing holding, inconsistent exchange rate, or stale valuation can create an apparent contribution that is really a data problem. Keep any residual visible rather than allocating it to a favorable category to make the report balance.

CFA Institute distinguishes performance measurement, attribution, and appraisal. These are related but different tasks: calculating a result, explaining its sources, and judging the process that produced it. Start with the first two before discussing skill. Label hypothetical study attribution separately from attribution of actual investment executions and funded-account holdings. A live research archive alone does not supply account-level economics.

Match the method to the decision architecture

An allocation committee needs to distinguish policy choices from manager implementation. An equity selector may need sector allocation and within-sector selection. A bond portfolio can require duration, curve, spread, and issuer explanations. A return-only factor regression answers a different question from a holdings-based decomposition. Choose the method before interpreting the results and identify what the available data cannot resolve.

Benchmark specification matters. Match the relevant universe, currency, return type, and decision horizon. MSCI's methodology resources distinguish price, net, and gross return variants; the committee should record the exact variant used. A mismatched benchmark can make ordinary market exposure look like skill. Benchmark changes should be documented prospectively rather than selected after observing which comparison gives the most favorable excess return.

Use a transparent arithmetic bridge

The following is a deliberately hypothetical single-period arithmetic example, not provider or customer performance. Assume gross portfolio return of 8.0%, benchmark return of 6.0%, and portfolio costs of 0.3 percentage points. The active gross difference is 2.0 points and the active net difference is 1.7 points. The displayed components are assumed outputs of a compatible allocation-selection method.

Look through manager names to shared exposures

Aggregate holdings where permissions and transparency allow. Identify repeated issuers, sectors, currencies, duration, and economically similar derivatives. Do not add notional exposures indiscriminately: the appropriate measure depends on the instrument and question. For funds without holdings transparency, state the limitation and use clearly labeled proxies or scenarios rather than presenting estimated exposure as observed fact.

In a hypothetical family portfolio, a public-equity manager, a private fund, and the operating business all depend on the same industrial spending cycle. Strong recent contributions from each do not demonstrate independent sources of return. Examine a joint downturn scenario and distinguish priced investment returns from an estimated business-value shock. The operating business may lack a reliable periodic valuation, which limits numerical consolidation but does not remove its economic relevance.

Separate contribution, risk, and implementation effects

A large positive return contribution is not a risk contribution. Risk attribution depends on a model, horizon, and assumptions about relationships between exposures. A position can have contributed little historically while dominating a forward stress. Report realized contribution and estimated risk side by side with clear labels rather than implying that one directly measures the other.

For model-driven decisions, bridge the research portfolio to the actual account through timing, sizing, cash, restrictions, financing, and costs. Preserve the frozen research-reference convention and identify actual executions separately. The SEC's performance guidance makes cost inclusion an important comparison question. If an office override improved results, record the actual override and rationale; do not retrospectively attribute that outcome to the untouched model.

Turn explanations into bounded follow-up questions

Ask whether the observed drivers were intended, authorized, and consistent with the original thesis. A manager may outperform through an exposure the committee already owns elsewhere. The response could be a smaller allocation, a mandate clarification, or more transparency rather than termination. A negative contribution can also be consistent with a deliberately purchased hedge, so judge it against its role and costs.

Failure cases include attributing residuals to skill, summing single-period arithmetic effects across time without an appropriate linking method, and changing factor models until the narrative looks favorable. Preserve the method and report sensitivity to reasonable alternatives. Attribution informs challenge and concentration management; it cannot by itself distinguish persistent skill from chance, especially with short records, overlapping decisions, or stale private-asset valuations.

Sources and further reading

1. CFA Institute: Portfolio Performance Evaluation
2. MSCI: Index Methodology resources
3. SEC Investor.gov: Investor Bulletin, Performance Claims

The linked sources provide conceptual background. The worked examples and planning assumptions on this page are this publication's own illustrations.

Continue the research