CIPM Level II common mistakes
CIPM Level II · Understand the traps and how to avoid them.
Reporting a time-weighted return computed from beginning and ending market values only, ignoring the timing of external cash flows within the period, and presenting it as the portfolio's TWR.
Candidates memorize the TWR formula for a single sub-period (V1 − V0) / V0 and apply it across a full month or quarter that contains a large contribution or withdrawal. Because the exam usually gives flows mid-period, this shortcut silently embeds the flow's timing into the return, which is exactly what TWR is designed to neutralize. The error is attractive because the arithmetic is simple and produces a plausible-looking number.
What to do instead
Before computing, scan the data for external cash flows and their dates. If any fall inside the measurement period, split the period at each flow date, compute a sub-period return for each segment using market value immediately before the flow as the base, then geometrically link: (1+r1)(1+r2)...−1. Only use the simple (V1−V0)/V0 form when no flows occur inside the period. Practice by re-deriving the same quarter both ways and confirming the naive answer differs whenever a flow is present.
Treating the portfolio's money-weighted return (IRR) as interchangeable with the manager's time-weighted return when judging manager skill, or vice versa, without stating which question each answers.
Both are 'the return', and candidates under time pressure pick whichever is easier to compute from the data given. The distinction-TWR measures the manager's compounding of a unit of capital independent of client flow decisions, while MWR reflects the client's actual dollar experience including the size and timing of their flows-is conceptual, not formulaic, so it is under-rehearsed.
What to do instead
Anchor each metric to a stakeholder question: 'How well did the manager invest?' → TWR; 'What did this client actually earn on their money?' → MWR/IRR. When an exam vignette asks you to evaluate the manager, TWR is the default; when it asks about the client's realized outcome or a private-equity fund with capital calls, MWR is appropriate. Write one sentence for each metric in your notes and check every return question against it before calculating.
Comparing a portfolio's return to a benchmark without first confirming the benchmark matches the portfolio's investment universe, risk level, and style, and without checking whether the benchmark return was calculated on a comparable basis (gross vs. net, same currency, same period).
The arithmetic of active return (portfolio − benchmark) is trivial, so candidates jump to it and spend their effort on attribution math. The comparability checks feel like preliminaries and get skipped, yet exam items frequently plant a mismatched or differently-calculated benchmark precisely to test whether you notice.
What to do instead
Adopt a fixed pre-comparison checklist: (1) Does the benchmark's stated universe and style match the mandate? (2) Same currency and hedging basis? (3) Same period boundaries and valuation timing? (4) Gross or net of fees on both sides? (5) Is the benchmark investable/available to the client? Run the checklist explicitly in your scratch work and note the conclusion; if any item fails, the comparison is invalid and that is likely the answer.
Computing attribution effects that do not reconcile to the total active return-reporting selection and allocation contributions that sum to a different number than portfolio minus benchmark-and not noticing the discrepancy.
Candidates apply allocation and selection formulas mechanically across many sectors and then stop, treating the arithmetic as finished. Residual/interaction terms and the choice of arithmetic vs. geometric attribution are easy to drop, and without a reconciliation step the omission is invisible.
What to do instead
Always finish attribution with a reconciliation: sum allocation + selection (+ interaction, if your method separates it) and confirm it equals total active return to the rounding tolerance. If it does not, you have either omitted the interaction term, mixed arithmetic and geometric methods, or mis-assigned a sector weight. State the method you are using at the top of your answer so the treatment of interaction is explicit.
Using an arithmetic average of sub-period returns, or annualizing a partial-year return by simple multiplication, when the question calls for a compounded or annualized figure.
Averaging and simple scaling are the fastest operations available and produce answers close enough to look right on a multiple-choice list. Candidates also confuse 'average annual return' with 'compound annual growth rate' and pick whichever the answer options reward. The gap only becomes obvious with volatile returns or periods under a year.
What to do instead
Decide first whether the question wants a compound growth rate, a simple average, or a cumulative figure, and label it in your scratch work. For multi-period performance use geometric linking; for annualizing a period shorter than a year, use (1 + cumulative return)^(365/days) − 1 rather than multiplying by the fraction of the year. Sanity-check with a volatile example where the arithmetic mean visibly exceeds the geometric mean.
Evaluating a manager's track record on raw return alone, without risk-adjusting (Sharpe, information ratio, Treynor) or checking whether the excess return is consistent with the stated strategy and within the mandate's risk budget.
Higher return reads as 'better manager', and risk-adjusted ratios require extra inputs (risk-free rate, tracking error, beta) that candidates skip when the raw numbers are supplied. The exam rewards the habit of asking 'return per unit of what risk?' and many candidates never form it.
What to do instead
For any manager comparison, compute or request at least one risk-adjusted measure and interpret it: Sharpe for total risk, information ratio for active risk against the benchmark, Treynor for systematic risk. Pair the number with a consistency check-does the return come from the strategy the manager says they run, and is realized volatility/tracking error inside the mandate? A high return with tracking error far outside the mandate is a failure, not a win.
Writing a manager search specification that lists only return objectives and asset class, omitting risk tolerance, liquidity needs, tax status, fee budget, regulatory/ESG constraints, and the decision timeline.
Candidates treat the search specification as a short wish-list and reproduce the client's headline goal. The specification is actually the screening instrument, and every omitted constraint becomes a manager that should have been excluded or a shortlist that misses qualified candidates. Under exam time pressure the soft constraints are the first to be dropped.
What to do instead
Build the specification from a fixed template covering: return objective, risk tolerance and maximum acceptable volatility/drawdown, liquidity and cash-flow profile, time horizon, tax and regulatory constraints, fee budget, any values-based or ESG mandates, and governance/decision timeline. Treat it as a checklist and confirm each line is populated before moving to screening. If the vignette omits a constraint, note the assumption explicitly.
Screening candidates on past performance ranking alone and dropping managers whose process, team stability, or capacity are strong but whose recent numbers are mediocre-or the reverse, retaining a manager on a hot three-year number with an unexplained process change.
Performance is quantifiable and easy to rank, so it dominates the shortlist. Qualitative due diligence (process repeatability, key-person risk, ownership and compensation, asset growth versus capacity) is harder to score and gets deferred to a stage that never arrives. Candidates also anchor on recency, overweighting the last few years.
What to do instead
Separate screening into quantitative and qualitative gates and require a manager to pass both. Quantitatively, look at risk-adjusted returns across a full market cycle, not just the recent window. Qualitatively, document the investment process, team turnover, ownership/compensation alignment, capacity versus assets, and any strategy drift. A shortlist entry should carry a written rationale for why the process is expected to repeat, not just a return rank.
Concluding a manager search by naming a single recommended manager without documenting the shortlist, the scoring criteria and their weights, the reasons other finalists were rejected, and the fee/terms negotiated.
The 'answer' feels like the name of the winner, so the selection rationale is treated as paperwork. But the defensibility of a search rests on the documented comparison, and exam items often ask what should accompany the recommendation or what a fiduciary must be able to demonstrate. Candidates also forget that fees and terms are part of the decision, not an afterthought.
What to do instead
Structure every recommendation as: criteria and weights → scores per finalist → qualitative notes → rejection reasons for non-selected finalists → fee and terms comparison → recommendation with conditions for monitoring and termination. State that the file must support the decision if challenged. Practise writing a two-paragraph selection memo so the habit becomes automatic.
Setting no pre-agreed monitoring triggers or termination criteria at hire, then judging the manager later on short-term underperformance alone or never revisiting the mandate at all.
The search is treated as ending at selection, so monitoring is out of scope in the candidate's mind. Without pre-committed criteria, any subsequent review becomes either inertia (never firing) or performance-chasing (firing after one bad quarter). Both are fiduciary failures, and the exam tests whether you know the criteria must be set in advance.
What to do instead
At selection, define explicit monitoring triggers: tracking error or style-drift thresholds, key-person departures, process or ownership changes, capacity breaches, and a performance evaluation window long enough to be meaningful. Distinguish 'underperformance within the expected range' (monitor) from 'breach of process or mandate' (act). Write the termination conditions into the recommendation so the review standard is fixed before results are known.