Performance measurement, attribution and investment governance
This domain tests whether you can measure what a manager actually delivered, explain where it came from, and show that the decision-making and oversight around it are sound.
Return measurement
Time-weighted return chain-links sub-period returns, removing the effect of cash-flow timing; it is the right measure for judging a manager who does not control flows. Money-weighted return (the portfolio IRR) reflects the investor's actual experience and suits cases where the client controls contributions. GIPS require composites, accrual and trade-date accounting, time-weighted total returns, valuation at least quarterly, and five years of history building to ten.
Attribution
Brinson attribution splits active return into allocation, selection and interaction:
- allocation = (w_p − w_b) × (r_b,segment − r_b,total)
- selection = w_b × (r_p − r_b)
- interaction = (w_p − w_b) × (r_p − r_b) Fixed income attribution instead decomposes duration, curve, spread and selection effects.
Risk-adjusted measures
Sharpe = (r_p − r_f)/σ_p, suitable for a standalone portfolio. Treynor = (r_p − r_f)/β and Jensen's alpha suit portfolios held within a diversified fund. Information ratio = active return / tracking error, the natural measure against a benchmark.
Worked example: Brinson
Benchmark: 60% equities returning 10%, 40% bonds returning 2%, total 6.8%. Portfolio: 70% equities returning 11%, 30% bonds returning 3%, total 8.6%. Allocation = 0.1 × (10 − 6.8) + (−0.1) × (2 − 6.8) = 0.32 + 0.48 = 0.80%. Selection = 0.6 × 1 + 0.4 × 1 = 1.00%. Interaction = 0.1 × 1 + (−0.1) × 1 = 0. Total 1.80%, matching 8.6 − 6.8. Overweighting the better sector and underweighting the worse both added value here.
Governance
An explicit investment strategy statement, delegated mandates with clear benchmarks and risk budgets, independent custody, conflict-of-interest management, stewardship and voting, and climate reporting under TCFD/ISSB. Documented decision trails and evidence of challenge are what examiners reward.
Traps
Using money-weighted returns to judge a manager. Mixing arithmetic attribution effects so the residual is unexplained. Quoting Sharpe for a sleeve of a larger fund. Claiming GIPS compliance without the required disclosures.