When should I use time-weighted rather than money-weighted return?

Use time-weighted return to assess investment performance while removing the effect of external cash-flow timing. Use money-weighted return to measure the investor’s experience, including when money was added or withdrawn. Suppose 100 grows to 110, the investor adds 50, and the resulting 160 grows to 168 over a second equal-length period. The time-weighted return is (110 ÷ 100) × (168 ÷ 160) − 1 = 15.5%. For the money-weighted return, solve 100(1 + r)² + 50(1 + r) = 168. This gives r about 7.00% per period, or about 14.50% over both periods. The later contribution experienced only the weaker second period, so the investor’s return is lower.

How can allocation, selection, and interaction explain active return?

In a two-sector example, the portfolio holds 60% in A at an 8% return and 40% in B at 4%, earning 6.4%. The benchmark holds 50% in each sector, returning 6% in A and 5% in B, for 5.5%. Active return is 0.9 percentage points. Under a benchmark-weight selection convention, allocation contributes 0.1 point from the weight differences relative to the 5.5% benchmark return. Selection contributes 0.5 point from sector return differences at benchmark weights. Interaction contributes 0.3 point because weights and sector returns differ together. The three effects sum to 0.9 point; always check which attribution convention a report uses.

What should I test before choosing a manager with strong past returns?

Start with the mandate: required exposures, benchmark, liquidity, and constraints. Check whether the reported record represents the proposed strategy and whether its composite membership and cash treatment are consistent. Examine attribution and risk to see how returns arose, then assess persistence without treating past results as a promise. Finally, test capacity, fees, trading costs, operational controls, and the practical steps needed to fund the mandate. A manager can have an attractive historical return yet be a poor choice if the strategy cannot be implemented at the client’s size or within its constraints.

CIPM Level I and II: Performance and Manager Analysis

Start with concept 1 Choose a topic

A return is useful only when you know what it measures. In this guide, you will calculate returns around external cash flows, link results across periods, and separate the effects of investment decisions from market movement. You will use attribution to explain active return, then judge whether that return is meaningful relative to risk, a benchmark, and the evidence available. The Level II topics extend that reasoning to composite construction, performance presentations, fixed-income and currency attribution, and manager selection. Worked decisions connect the calculations to client interests: which records belong in a report, what a disclosure must make clear, and whether a manager’s process, capacity, fees, and controls suit a mandate.

Level I - Ethics and professionalism

1. Disclose competing interests

A conflict exists when an analyst has an interest that could reasonably affect, or appear to affect, a recommendation. The concern is the incentive, so a well-supported conclusion does not erase the need for disclosure. Before a client relies on a manager recommendation, identify relevant ownership, compensation, business relationships, or other benefits in clear terms. Explain who holds the interest and how it relates to the manager. Then preserve the evidence behind the assessment: the mandate, comparison criteria, performance data, and reasons for the conclusion. Disclosure gives the client information needed to judge the recommendation; it does not substitute for objective analysis. The CIPM Program covers ethics alongside performance evaluation and manager selection, where such conflicts can directly affect decisions.

Worked example: An analyst recommends Meridian Asset Management while owning shares in Meridian’s parent company. The analyst tells the client about that ownership before the recommendation is used, even though Meridian ranked first on the documented criteria. The analyst retains the scoring and supporting research, and reviews whether the holding affected any judgment. A sound recommendation and a relevant financial interest can coexist.

Watch the distinction: “My recommendation is correct, so my shares are irrelevant” confuses the quality of the conclusion with the incentive behind it. The client needs both the analysis and the disclosure.

Reference: [1]; [2]

2. Protect independent judgment

Independent judgment depends on how evidence is gathered and evaluated. A meeting with a manager can provide useful access to investment staff, holdings explanations, and operational records. A valuable personal benefit offered around a rating decision creates a different incentive: the analyst may feel obliged to reward the provider or may appear to do so. Assess the benefit’s purpose, value, timing, and connection to the decision. Keep the manager assessment tied to a common set of criteria, verify claims against records where possible, and document reasons for the rating. If a benefit threatens objectivity, decline it and continue the review through appropriate professional access. The central test is whether the analyst can reach and defend the same conclusion without special treatment.

Worked example: A manager offers to pay for an analyst’s weekend resort trip two weeks before its rating review. The analyst declines and requests a standard due diligence meeting instead. The review compares the manager’s process, returns, risks, and controls with those of competing firms using the same criteria. The resort offer cannot become an unrecorded advantage in the rating.

Watch the distinction: Calling an expensive trip “manager access” overlooks its personal value and timing. Access can help analysis; a leisure benefit can compromise judgment even without an explicit request for a favorable rating.

Reference: [2]

3. Treat clients fairly

Fair dealing requires a consistent way to deliver material performance information to clients who are entitled to it. When an error or new finding could change a client’s decision, first establish its scope: which reports, periods, accounts, and decisions are affected? Correct the underlying record, prepare a clear explanation, and use a controlled release process so similarly situated clients receive the information without selective advantage. Different clients may need different details because their accounts differ, but priority should follow a defensible service rule rather than revenue or influence. Record when the error was found, what was corrected, and when each affected client was notified. Fairness does not mean every client must make the same decision. It means each receives a reasonable opportunity to act on accurate information.

Worked example: A return calculation error affects reports already sent to three clients. The corrected result is materially lower. The team verifies the revised calculation, identifies all three reports, and sends corrections through its established notification process. It records the release times and offers account-specific explanations. Privately warning the largest client first would give that client an informational advantage.

Watch the distinction: Giving every client the same incorrect report was not enough to satisfy fairness. Once a material error is discovered, fair dealing also governs how the correction reaches affected clients.

Reference: [2]

4. Safeguard client information

Client confidentiality covers information that may identify a client or reveal its holdings, trades, objectives, or results. Removing a name is only the first check. A distinctive position, transaction date, portfolio size, or combination of facts may allow an audience to identify the client from other available information. Before using a case study, ask what a knowledgeable reader could infer and whether the proposed detail is necessary to teach the analytical point. Aggregate or alter nonessential details, and obtain appropriate authorization when identifiable client information must be used. Keep the underlying analysis accurate: changing details must not turn a loss into a gain or distort the lesson. A useful presentation can explain a measurement method with hypothetical figures while leaving the client’s actual positions protected.

Worked example: A conference slide describes a single large position in a thinly traded company. The client’s name is absent, but its public ownership filing makes the position recognizable. The analyst replaces the position and dates with clearly hypothetical figures, checks that the calculation still works, and obtains appropriate authorization for any remaining client-specific details before presenting the case.

Watch the distinction: An unnamed chart is not automatically anonymous. Holdings, dates, and transaction size can identify an account when combined with information the audience already has.

Reference: [2]

5. Maintain trustworthy records

A reported return should be reproducible from dated portfolio values, external cash flows, fee records, and the chosen calculation method. That audit trail matters when a custodian corrects a value or a flow is entered on the wrong date. Recalculate the affected period, then check any linked returns, comparisons, and presentations that depend on it. Retain the original entry, the corrected source, the reason for the change, and the approval history; silently overwriting a number makes later review difficult. Distinguish a market value correction from a deposit or withdrawal, since each affects the return differently. A clean record supports both accurate client reporting and diagnosis of why two systems disagree. The goal is a result another qualified reviewer can trace and reproduce.

Worked example: An account begins the month at 100 units, has no external flows, and was reported at 104 units at month-end. The custodian corrects the ending value to 102. The original return was 104/100 − 1 = 4%; the corrected return is 102/100 − 1 = 2%. The team retains both values and the correction reason, then replaces affected reports.

Watch the distinction: Changing the displayed return from 4% to 2% without keeping the valuation trail fixes the headline but leaves no reliable way to explain or verify the correction.

Reference: [2]

6. Avoid selective performance claims

A performance statement can be numerically true yet misleading if its label suggests a broader result. Ask what population the number describes: one account, a composite, a selected period, or an entire strategy? Then check whether the account was chosen after its strong result was known and whether the stated comparison uses consistent fees, dates, and benchmarks. A single account can illustrate implementation, but its result should be identified as that account’s result and accompanied by context needed to understand the strategy’s wider experience. Distinguish a hypothetical illustration from an actual record as well. The practical diagnostic is to read the sentence as a prospective client would: would that reader reasonably infer that the quoted return represents typical or complete strategy performance?

Worked example: A brochure says a strategy earned 12% last year. The figure belongs to one winning account; the full group of accounts following the strategy returned 5% on the stated basis. The brochure should identify the 12% as that account’s result and present the representative group result with comparable period and fee information. Otherwise, the true account figure creates a false impression.

Watch the distinction: A true return is not automatically a fair strategy claim. Selecting a winner and describing it as the strategy’s result conceals how the other eligible accounts performed.

Reference: [2]

Level I - Performance measurement

7. Calculate time-weighted return

Time-weighted return measures the growth of a unit invested through successive periods, reducing the effect of the size and timing of external contributions or withdrawals. Value the portfolio immediately before a flow, calculate the return on the capital present during that subperiod, then treat the flow as part of the starting value for the next subperiod. Link the subperiod growth factors by multiplication. This sequence separates investment performance from the client’s decision to add or remove capital. It assumes each boundary value and flow date is reliable; a missing valuation can require an approximation and should be understood as such. For focused Level I study guidance, practice drawing a timeline before using the formula. The timeline prevents a contribution from being counted as investment gain.

Worked example: A portfolio starts at 100 units and reaches 110 just before a client contributes 20. Its new invested value is 130, and it ends at 143. The first return is 110/100 − 1 = 10%; the second is 143/130 − 1 = 10%. Linking gives 1.10 × 1.10 − 1 = 21%. The contribution is capital, not profit.

Watch the distinction: Calculating 143/100 − 1 = 43% treats the 20-unit contribution as investment growth. Break the history at the external flow before linking returns.

Reference: [2]

8. Solve money-weighted return

Money-weighted return is the rate that equates dated investor contributions with later distributions and ending value. It is an internal rate of return, so the amount invested at each date affects the answer. Draw a cash-flow timeline from the investor’s perspective: contributions are outflows, while distributions and final proceeds are inflows. Then solve for the discount rate that makes their present values equal, using consistent time units. Equivalently, compound each contribution to the terminal date and compare their sum with terminal proceeds. This measure captures the investor’s actual capital experience, including the timing of decisions to add or withdraw funds. It differs from time-weighted return when cash flows occur around periods with different returns. Multiple sign changes can produce multiple mathematical solutions, requiring careful interpretation.

Worked example: An investor contributes 100 units today and 60 more after one year, then receives 187 at the end of year two. The annual rate r satisfies 100(1 + r)² + 60(1 + r) = 187. At r = 10%, 100 × 1.21 + 60 × 1.10 = 121 + 66 = 187. The money-weighted return is 10% per year.

Watch the distinction: Dividing total profit by the initial 100 ignores the later 60-unit contribution and its shorter investment period. Each cash flow needs its own date in the rate equation.

Reference: [2]

9. Place external cash flows correctly

A cash flow changes the amount of capital exposed to subsequent investment returns. Its date therefore matters even when the opening value, closing value, and contribution amount are unchanged. If a contribution arrives at the start of a period, include it in the opening invested capital. If it arrives at the end, remove it from the closing value before comparing that value with opening capital. A withdrawal receives the opposite treatment. For flows within a period, a valuation at the flow boundary permits a cleaner subperiod calculation; otherwise a stated approximation may be needed. Always specify whether a quoted value is before or after the flow. That small timing label often determines which denominator is correct and whether the resulting return reflects investment performance.

Worked example: A portfolio starts at 100 units, receives a 20-unit contribution, and closes at 132. If the contribution arrived at the start, 120 units were invested throughout, so return is 132/120 − 1 = 10%. If it arrived at the end, investment growth produced 132 − 20 = 112 units, so return is 112/100 − 1 = 12%.

Watch the distinction: Subtracting every contribution from the ending value assumes it arrived at the end. A beginning contribution participated in the period’s return and belongs in opening invested capital.

Reference: [2]

10. Link returns geometrically

Successive percentage returns act on changing capital, so cumulative return comes from multiplying growth factors. Convert each periodic return r into 1 + r, multiply those factors, then subtract one. This works for gains and losses, provided the periods follow one another and the return definitions are consistent. Adding percentages instead gives an arithmetic summary that does not describe the wealth path. A loss also changes the gain needed for recovery: after a 10% loss, capital is 90% of its starting value, so recovering the original amount requires 1/0.90 − 1, or about 11.11%. Geometric linking is central to multi-period performance reporting because it preserves the amount actually available to earn each later return.

Worked example: One unit loses 10% in the first period, leaving 0.90 units. It then gains 20% on that smaller base, ending at 0.90 × 1.20 = 1.08 units. The two-period cumulative return is 1.08 − 1 = 8%. Adding −10% and 20% gives 10%, which overstates the actual growth by 2 percentage points.

Watch the distinction: Periodic returns are not equal-sized cash amounts. Adding them assumes the second return applies to the original balance, although the first period has already changed that balance.

Reference: [2]

11. Value around a major flow

A valuation near a major external flow allows performance to be measured on each side of the change in invested capital. First calculate the return from opening value to the value immediately before the flow. Then add a contribution, or subtract a withdrawal, to establish the next subperiod’s opening value. Calculate the second return and link the two growth factors. The word “before” is essential: adding a contribution to a value that already includes it would count the flow twice. More frequent reliable valuations reduce reliance on estimates when flows are large relative to the portfolio. They also make errors easier to locate, because each subperiod can be checked against its own opening value, closing value, and external flows.

Worked example: A portfolio begins at 100 units and is worth 105 immediately before a 40-unit contribution. Its first return is 105/100 − 1 = 5%. After the contribution, the next opening value is 145; the portfolio ends at 154, giving 154/145 − 1 ≈ 6.21%. The linked return is 1.05 × (154/145) − 1 ≈ 11.52%.

Watch the distinction: Using 105 as the second subperiod’s opening value counts the 40-unit contribution as profit. The post-contribution capital base is 145 units.

Reference: [2]

12. Separate gross and net returns

Gross and net performance answer different questions. A return before investment management fees describes investment results before that fee reduces client wealth. A net return reflects the specified fee deduction and is closer to what the client retains on that basis. State which fees are included, when they are charged, and whether the opening and closing values are measured before or after payment. With no external flows, a period-end management fee can be subtracted from the pre-fee ending value before calculating net return. Other costs or fee timing can require different treatment, so a label alone is insufficient. When comparing managers, align fee conventions; otherwise a gross return from one manager and a net return from another do not support a clean ranking.

Worked example: An account starts with 100 units and grows to 110 before a 2-unit management fee is paid at period-end. With no other flows or costs, gross return is 110/100 − 1 = 10%. After the fee, the ending value is 108, so net return is 108/100 − 1 = 8%. The 2-percentage-point gap follows from the stated fee timing.

Watch the distinction: Calling 110/100 − 1 a client’s net return ignores the 2-unit fee paid from the account. Always identify the return basis and the fees included.

Reference: [2]

Level I - Performance attribution

13. Measure allocation effect

Allocation attribution asks whether portfolio weights differed from benchmark weights in sectors that performed differently from the overall benchmark. In the Brinson–Fachler form, a sector’s allocation effect is (portfolio weight − benchmark weight) × (benchmark sector return − total benchmark return). Overweighting a sector that beats the total benchmark gives a positive effect; overweighting a laggard gives a negative one. This measures a weight decision against the benchmark’s opportunity set, using benchmark returns to avoid mixing in security selection. Apply the same return period and consistently defined sectors throughout. A single sector effect is only one part of a full attribution: other sectors and selection effects must be included before explaining the portfolio’s entire active return.

Worked example: A sector has a 50% portfolio weight and a 40% benchmark weight. Its benchmark return is 8%, while the total benchmark returns 5%. The active weight is 0.50 − 0.40 = 0.10, and the sector’s relative benchmark return is 3 percentage points. Allocation effect is 0.10 × 3% = 0.30 percentage points, a benefit from the overweight.

Watch the distinction: Multiplying the active weight by the sector’s full 8% return measures something different. Brinson–Fachler allocation uses the sector’s return relative to the total benchmark.

Reference: [2]

14. Measure selection effect

Selection attribution asks whether the securities held within a sector beat that sector’s benchmark securities. Under a common three-effect Brinson model, selection for a sector equals benchmark sector weight × (portfolio sector return − benchmark sector return). Benchmark weight holds the sector allocation fixed while assessing the return difference. A positive value indicates favorable within-sector choices on this definition, even if the sector itself had a poor absolute return. Conversely, a portfolio sector can gain money but still produce a negative selection effect when its benchmark sector gains more. Keep sector definitions, return periods, and currency treatment aligned. If portfolio weight also differs from benchmark weight, the extra active-weight contribution to the return difference is captured separately as interaction.

Worked example: A sector represents 30% of the benchmark. Portfolio holdings in it return 9%, while the matched benchmark sector returns 6%. The within-sector excess return is 3 percentage points. Selection effect is 0.30 × 3% = 0.90 percentage points. This is a favorable selection result at benchmark weight; any additional effect from a portfolio overweight belongs in interaction.

Watch the distinction: Using portfolio weight for selection under this three-effect definition folds some weight choice into the result. State the attribution model before interpreting its named effects.

Reference: [2]

15. Account for interaction

Interaction arises when an active sector weight and an active within-sector return occur together. In a three-effect Brinson decomposition, it equals (portfolio weight − benchmark weight) × (portfolio sector return − benchmark sector return). Its sign depends on both differences. Overweighting a sector whose chosen holdings outperform produces positive interaction; overweighting underperforming holdings produces negative interaction. Some attribution presentations combine interaction with selection or allocation, so named effects should always be tied to a stated convention. Interaction is not an unexplained error term. It is the product needed when expanding portfolio weight times portfolio return relative to the benchmark components. Check the full set of sectors before judging its economic importance; one favorable interaction can be offset elsewhere.

Worked example: A sector weighs 60% in a portfolio and 50% in its benchmark. Portfolio holdings return 12%; the benchmark sector returns 10%. Active weight is 0.10 and active sector return is 2 percentage points. Interaction is 0.10 × 2% = 0.20 percentage points. The portfolio benefited because its overweight coincided with favorable within-sector selection.

Watch the distinction: Treating interaction as a rounding discrepancy hides a real joint effect. It reflects the return difference earned on the weight above or below the benchmark allocation.

Reference: [2]

16. Reconcile active return

Attribution is useful only if its effects reconcile to the return difference it claims to explain. Begin with portfolio return minus benchmark return for the same period, currency, fee basis, and valuation convention. Under a stated Brinson–Fachler three-effect setup, calculate allocation, selection, and interaction for each sector, sum across sectors, then compare their total with active return. A mismatch can reveal omitted holdings, inconsistent weights, different benchmarks, stale prices, or a model convention that assigns an effect differently. Small rounding differences should be separated from material unexplained residuals. Reconciliation checks arithmetic and data alignment; it does not prove that each effect represents skill. Interpretation still requires understanding the mandate, sector definitions, and whether the observed exposures were deliberate.

Worked example: A hypothetical two-sector analysis reports allocation of 1.0, selection of 0.5, and interaction of 0.3 percentage points after summing sector effects. Their total is 1.8 points. The portfolio returned 6.8% and the benchmark 5.0%, so active return is 6.8% − 5.0% = 1.8 points. The stated attribution reconciles; the separate sector inputs still need review.

Watch the distinction: A plausible story about good sector choices cannot replace reconciliation. If the reported effects do not sum to active return, inspect the data and model convention before drawing conclusions.

Reference: [2]

17. Read sector contributions

A sector’s contribution to total portfolio return is its portfolio weight multiplied by its portfolio sector return, assuming consistent beginning weights for the period. This describes how much the sector added to the portfolio’s absolute return. Active contribution answers another question: how much did the portfolio’s sector exposure and results add relative to the benchmark? One direct sector comparison is portfolio weight × portfolio sector return minus benchmark weight × benchmark sector return. A formal attribution model can then split that difference into allocation, selection, and interaction. A sector may add positively to total return while detracting from active return because its benchmark counterpart did better. Always identify which question a chart answers before interpreting a positive bar as outperformance.

Worked example: A sector has a 40% weight in both portfolio and benchmark. Portfolio sector return is 5%, giving total portfolio contribution of 0.40 × 5% = 2.0 percentage points. The benchmark sector returns 7%, contributing 0.40 × 7% = 2.8 points. The sector’s direct active contribution is 2.0 − 2.8 = −0.8 points despite its positive absolute contribution.

Watch the distinction: A positive contribution to total return does not imply successful relative performance. Compare with the matched benchmark sector and its weight before calling the sector an active winner.

Reference: [2]

18. Link attribution across periods

Single-period attribution effects are usually expressed as percentage-point differences for that period. Across periods, portfolio and benchmark wealth compound on different bases. Adding the period effects therefore need not equal the difference between their cumulative returns. First link portfolio returns and benchmark returns separately. Subtract the resulting cumulative returns to obtain cumulative active return in percentage points. Then use a consistent multi-period linking method to scale or allocate period effects so their total reconciles. The choice of method affects how the cumulative difference is assigned across periods and effects, so disclose the convention when comparing reports. A simple sum remains useful for seeing the direction of period decisions, but it should not be labeled the exact compounded active result.

Worked example: A portfolio returns 10% and then −5%; its benchmark returns 8% and then −4%. Portfolio cumulative return is 1.10 × 0.95 − 1 = 4.50%. Benchmark cumulative return is 1.08 × 0.96 − 1 = 3.68%. Cumulative active return is 0.82 percentage points, while period differences of +2 and −1 points sum to +1 point.

Watch the distinction: Adding period active returns ignores the different wealth bases used in later periods. Link portfolio and benchmark returns separately before interpreting cumulative active performance.

Reference: [2]

Level I - Performance appraisal

19. Measure total variability

Standard deviation describes how widely returns vary around their mean. For a complete, equally weighted set of observations, subtract the mean from each return, square the deviations, average them, and take the square root. Squaring prevents gains and losses from canceling. If the observations are treated as a sample used to estimate a wider return process, the familiar sample calculation divides by one fewer than the observation count instead. State which convention is used, along with the return frequency, because monthly and annual standard deviations are not directly comparable. This is a measure of total variability: it includes movements shared with the benchmark and movements specific to the portfolio. Tracking error instead applies standard deviation to portfolio-minus-benchmark returns.

Worked example: Consider three equally weighted period returns of −2%, 2%, and 6%. Their mean is (−2 + 2 + 6)/3 = 2%. Deviations are −4, 0, and +4 percentage points; squared deviations are 16, 0, and 16. Population standard deviation is √[(16 + 0 + 16)/3] = √(10.6667) ≈ 3.27 percentage points per period.

Watch the distinction: Calling portfolio standard deviation “active risk” confuses total movement with movement relative to a benchmark. Calculate tracking error from return differences when benchmark-relative variability is the question.

Reference: [2]

20. Interpret the Sharpe ratio

The Sharpe ratio divides a portfolio’s excess return over a risk-free rate by its total return standard deviation. Its numerator rewards return beyond the chosen reference rate; its denominator measures variability of the whole portfolio, including market and portfolio-specific movements. Align the return, risk-free rate, and volatility to the same period and currency before dividing. A larger ratio indicates more excess return per unit of measured total volatility within the stated comparison, but it does not identify why the result arose. Its interpretation can be strained by short samples, unstable volatility, or return distributions with substantial tail risk. When comparing strategies, check the observation window and the return basis as well as the ratio. CIPM practice questions can help reinforce the calculation and its limits.

Worked example: A portfolio earns 12% over a stated year, the comparable risk-free rate is 2%, and the portfolio’s annual standard deviation is 20%. Excess return is 12% − 2% = 10 percentage points. The Sharpe ratio is 10%/20% = 0.50. That means 0.50 units of excess return per unit of measured total volatility for the stated inputs.

Watch the distinction: A Sharpe ratio of 0.50 does not mean a 50% return or guarantee future performance. It is a ratio of excess return to measured volatility for a specified period.

Reference: [2]

21. Interpret the Treynor ratio

The Treynor ratio divides a portfolio’s return above the risk-free rate by its market beta: (portfolio return − risk-free return) / beta. Beta measures sensitivity to movements in the chosen market benchmark. A ratio of 5 percentage points means five points of excess return per unit of beta over the measured period. Compare ratios only when returns, the risk-free rate, beta estimates, and periods are compatible. This measure is most informative for a well-diversified portfolio, where market exposure is the main risk relevant to the investor. A concentrated portfolio can have substantial security-specific volatility that beta does not capture; its Treynor ratio may look attractive despite uncomfortable total risk. A beta near zero also makes the ratio unstable, and a negative beta makes its usual ranking difficult to interpret. The CIPM curriculum factsheet places performance appraisal within Level I.

Worked example: Suppose a diversified portfolio earns 9% over a year, the matching risk-free return is 3%, and its estimated market beta is 1.2. Excess return is 9% − 3% = 6 percentage points. Treynor ratio is 6% / 1.2 = 5 percentage points per unit of beta. Interpret that alongside the beta estimate and the portfolio’s total volatility before comparing managers.

Watch the distinction: A higher Treynor ratio does not prove lower overall risk. The denominator excludes security-specific variation, so a poorly diversified portfolio can appear strong on this measure.

Reference: [2]

22. Estimate Jensen alpha

Jensen alpha asks whether a portfolio earned more or less than a simple beta-adjusted market model predicts. First calculate the model return: risk-free return + beta × (benchmark return − risk-free return). Subtract that result from the portfolio return. Positive alpha is performance above the model’s prediction for the observed market return and estimated beta; negative alpha is below it. Use returns from the same period and currency, and choose a benchmark that represents the market exposure being modeled. The calculation is sensitive to the beta estimate, benchmark choice, and omitted sources of risk. For example, a portfolio tilted toward small companies may show positive alpha against a broad market model because that model does not account for the tilt. Alpha is a diagnostic residual, so investigate exposures and measurement choices before attributing it to skill.

Worked example: Assume a portfolio returns 11%, its benchmark returns 8%, the risk-free return is 2%, and beta is 1.1 for the same year. The benchmark excess return is 8% − 2% = 6%. The model predicts 2% + 1.1 × 6% = 8.6%. Jensen alpha is 11% − 8.6% = 2.4 percentage points, conditional on that benchmark and beta.

Watch the distinction: Alpha is not simply portfolio return minus benchmark return. That difference ignores the portfolio’s market sensitivity and the risk-free return used in the model.

Reference: [2]

23. Use the information ratio

An information ratio compares average active return with the variability of active returns. For each period, subtract benchmark return from portfolio return. Estimate the average of those differences and divide it by tracking error, their standard deviation. The numerator indicates the size and direction of benchmark-relative performance; the denominator indicates how consistently it was earned. Both must use compatible time units. A positive ratio can help compare managers pursuing similar mandates against appropriate benchmarks, but its estimate depends on the observation window. A strategy with one exceptional period may have a high recent ratio without a dependable advantage. Check whether the benchmark reflects the mandate and whether fees, changing exposures, and short histories affect the comparison. When tracking error is close to zero, a small measurement difference can make the ratio unstable.

Worked example: Suppose a strategy has an annualized active return of 3 percentage points and annualized tracking error of 4 percentage points, measured against its mandate benchmark. Its information ratio is 3 / 4 = 0.75. That is 0.75 units of estimated active return per unit of active-return variability. Examine the sample period and fee basis before treating it as evidence of repeatable performance.

Watch the distinction: Dividing excess return over cash by tracking error mixes two reference points. The information ratio uses return relative to a specified benchmark in both its numerator and risk calculation.

Reference: [2]

24. Calculate tracking error

Tracking error measures how much a portfolio’s return difference from its benchmark varies across periods. Calculate one active return for each period, find their average, then take the standard deviation of those active returns. State whether the calculation uses a sample or population standard deviation and whether it has been annualized. A portfolio may be volatile yet have low tracking error if it moves closely with its benchmark. Conversely, a low-volatility portfolio can have meaningful tracking error against a much riskier benchmark. Tracking error measures the dispersion of active returns, not whether their average is positive. Interpret it with the mandate: a tight index-tracking assignment calls for a different range of active variation than a flexible strategy. A benchmark mismatch can make the number look large for reasons unrelated to implementation skill.

Worked example: For two equally weighted periods, suppose portfolio returns exceed benchmark returns by 1 percentage point in the first and fall short by 1 point in the second. Mean active return is (1 + −1) / 2 = 0 points. Using the population convention, tracking error is √[(1² + (−1)²) / 2] = 1 percentage point per period. Two observations give a fragile estimate.

Watch the distinction: Tracking error is not the standard deviation of portfolio returns. Calculate portfolio return minus benchmark return period by period, then measure the dispersion of those differences.

Reference: [2]

Level I - Manager selection

25. Check mandate fit first

Begin a manager search with the client’s investment objective, permitted assets, risk limits, liquidity needs, and reporting currency. Translate them into an investable opportunity set and a benchmark that represents the work the manager is being asked to do. Then compare each candidate’s actual authority and typical holdings with those requirements. A strong historical return cannot repair a basic mismatch: the candidate may have earned it by taking risks the client cannot accept. Separate essential constraints from preferences so a screen does not discard a suitable manager for a minor difference. Examine whether the candidate’s published record used the same strategy and constraints proposed for the client. The Level I CIPM study guide can provide further practice with performance measures used later in the search.

Worked example: A client’s mandate permits only investment-grade bonds, while a candidate may place 30% of assets in below-grade debt. That authority violates an essential constraint even if the candidate has the highest reported return. Remove the candidate from this search, or evaluate a separately constrained proposal with a relevant record and benchmark. Do not compare its unrestricted return as though it were investable for this client.

Watch the distinction: Choosing the manager with the best past return assumes every candidate pursued the same investable mandate. Check permitted risks and constraints before ranking performance.

Reference: [2]

26. Design quantitative screens

A quantitative screen narrows a manager universe using measures tied to the mandate. Define each filter before examining candidates: relevant history, assets managed in the strategy, benchmark-relative risk, fees, liquidity, or other observable requirements. Specify the measurement period and data source. A five-year record is useful only if those years represent the proposed process; tracking error is useful only against a suitable benchmark. Apply the same definitions to every candidate and flag missing or noncomparable data. Decide which conditions are hard eligibility rules and which are prompts for investigation. Screens save effort, but they can reward stale data, favorable reporting choices, or a strategy that changed during the sample. After screening, test process, personnel, operations, and capacity before making a selection.

Worked example: Suppose a hypothetical search requires at least five years of relevant history and tracking error below 6% against the mandate benchmark. Candidate A has six years and 4% tracking error, so it passes both filters. Candidate B has eight years and 8%, so it fails the risk rule despite a longer record. Next verify A’s history and benchmark before advancing it.

Watch the distinction: Passing a numerical screen is not a manager endorsement. The figures may describe a different process, an unsuitable benchmark, or a firm with weak operational controls.

Reference: [2]

27. Test qualitative evidence

Qualitative due diligence tests whether a manager’s stated method describes its actual decisions. Start with the investment belief: what opportunity does the manager claim to exploit, and why might it persist? Follow that claim through research files, buy and sell decisions, position sizing, risk limits, and portfolio holdings. Look for dated evidence created when decisions were made, since an explanation written after returns are known is less persuasive. Interview the people who make and challenge decisions, and examine both successful and unsuccessful trades. A difference between stated and observed behavior may have a sound explanation, such as a documented change in assumptions, but it requires investigation. The aim is to judge whether the process is coherent, repeatable, and suited to the client’s mandate, not whether its story sounds polished.

Worked example: A manager says it buys securities only below its documented estimate of value. Review three dated purchases: the recorded prices were 102, 108, and 95 units, while contemporaneous value estimates were 100, 105, and 100. Two purchases exceeded the estimates. Ask for the decision records and any revised assumptions before accepting the claimed discipline; the third purchase alone does not resolve the pattern.

Watch the distinction: A convincing interview does not establish process discipline. Compare claims with dated research, actual trades, portfolio weights, and sell decisions, including cases that performed poorly.

Reference: [2]

28. Examine process consistency

A track record is useful evidence only to the extent that its generating process resembles the proposed mandate. Map changes in investment universe, concentration, valuation method, portfolio construction, risk limits, and decision makers across the record. Determine when each change occurred and whether it was deliberate, documented, and explained to clients. Some changes are ordinary refinements; others alter the strategy enough that older returns answer a different question. Examine performance before and after a material change without treating the shorter current period as automatically conclusive. Also ask whether a shift was made after poor results, which can create misleading comparisons. The decision is about the relevance of historical evidence for today’s process. A manager may still be investable after a change, but its full history should not be presented as uninterrupted proof of the new approach.

Worked example: A ten-year record comprises five years of concentrated value investing followed by five years of diversified growth investing. If the client is considering the present growth strategy, the first five years describe a different opportunity set and decision method. Assess the later five-year period separately, check when client communications and portfolio holdings changed, and compare the current process with a suitable growth benchmark.

Watch the distinction: An unbroken return series does not prove an unbroken strategy. Material changes in holdings, constraints, or decision rules can reduce the earlier record’s relevance.

Reference: [2]

29. Assess organizational stability

A manager’s record reflects people, decision rights, incentives, and supporting resources. Identify who originated ideas, approved positions, controlled risk, and executed trades during the reported period. Then compare that structure with the current organization. Departures matter most when knowledge or authority was concentrated in the people who left. Review succession plans, analyst coverage, ownership changes, compensation, and whether investment staff have incentives to remain. A departure is a warning sign, not automatic disqualification: a documented team process may continue effectively. Conversely, a familiar brand may conceal a substantially different team. Speak with current decision makers and inspect recent investment decisions to see whether the claimed process still operates. Treat the historical return as evidence about the present team only to the degree that the relevant capabilities have continued.

Worked example: A five-year record names three portfolio decision makers, but two leave immediately afterward. Review committee minutes to see whether the remaining person drove past decisions or mainly implemented them. Check who now supplies research, who approves trades, and whether recent portfolios follow the same rules. The five-year return remains a historical fact, yet its relevance to the new team is uncertain until those links are established.

Watch the distinction: A firm name does not carry investment skill by itself. When key personnel leave, trace the decisions and resources behind past results to the people working today.

Reference: [2]

30. Review operational resilience

Investment judgment and operational control address different risks. A manager may choose securities well yet misstate holdings, price illiquid assets poorly, or fail to protect client assets. Review how trades move from order to confirmation, how positions and cash are reconciled with independent records, who can change valuations, and how errors are escalated. Examine custody arrangements, segregation of duties, access controls, business continuity, and oversight of outside service providers. Seek evidence that controls operate, such as reconciliation logs and resolved exceptions, rather than accepting a diagram of intended procedures. The seriousness of a gap depends on the assets and processes involved: hard-to-value holdings and complex cash flows may demand closer review. Resolve material weaknesses before funding, since attractive returns do not offset uncertainty about what the portfolio owns or what it is worth.

Worked example: A manager reports strong returns but reconciles its holdings only to an internal trading file. A custodian independently records positions, yet nobody compares the two. Request dated reconciliations, investigate any differences, and confirm who reviews exceptions. Until the gap is addressed, reported returns may rest on inaccurate positions or valuations. The historical return cannot establish that assets and records are controlled.

Watch the distinction: A return audit alone does not show that custody, valuation, and record controls are sound. Operational due diligence follows the assets and transaction records independently.

Reference: [2]

Level II - Ethics and professionalism

31. Put client interests first

Putting clients first requires evaluating a recommendation against each client’s agreed objective, risk tolerance, restrictions, and costs. Identify who benefits from a proposed action, including the manager, adviser, and client. A financial incentive does not automatically make a recommendation unsuitable, but it creates a conflict that needs disclosure and disciplined analysis. Compare reasonable alternatives on expected role in the portfolio, downside exposure, liquidity, fees, and implementation burden. Document why the chosen course serves the client rather than merely increasing revenue. If the proposal exceeds an agreed constraint, obtain a properly considered change in mandate before acting; a promising return forecast cannot silently override the existing limit. The Level II CIPM study guide covers the broader set of advanced topics around this decision.

Worked example: A cautious client has an agreed limit against leverage. A manager would earn a larger fee by moving the client into a leveraged strategy. Even if its projected return is higher, the proposal conflicts with the current mandate and adds downside risk. Retain an eligible approach while assessing alternatives. Any change requires an informed client decision about the revised risks, costs, and objective.

Watch the distinction: Disclosing a fee incentive does not make an unsuitable recommendation suitable. The investment must still serve the client’s objective and comply with agreed constraints.

Reference: [2]

32. Control nonpublic information

Information is potentially material if a reasonable investor would consider it important to a trading decision or if it could meaningfully affect a security’s price. Information is nonpublic while it has not been made broadly available to investors. When both conditions may apply, contain the information before trying to profit from it: stop affected trading, limit access to those who need it, preserve the facts, and involve the firm’s designated review function. Consider the source, confidentiality, specificity, and whether a public release has actually occurred. A rumor, market price change, or private circulation among several analysts does not by itself establish public availability. Controls such as restricted lists and information barriers help prevent the information from influencing client or personal transactions while its status is assessed.

Worked example: An analyst receives a confidential briefing that a portfolio holding is the target of an impending merger. The transaction could change the target’s price, and the briefing has not reached the market. The analyst limits sharing, records how the information arrived, and stops affected trading while the firm reviews it. A subsequent price rise does not turn the private briefing into public information.

Watch the distinction: A moving share price does not prove confidential facts are public. Assess whether the substance has been broadly disseminated before lifting trading restrictions.

Reference: [2]

33. Manage personal trading

Personal trading creates conflicts when employees can act before clients, benefit from planned client orders, or use information obtained through their work. Preclearance allows a check before an employee trade; a restricted list blocks trading in securities subject to particular concerns; post-trade monitoring compares actual activity with approvals and client transactions. These controls work together because an approval request may omit later changes, while monitoring alone discovers a problem after execution. Apply them to relevant accounts over which an employee has beneficial control, according to the firm’s rules. Review order timing, security identity, and the employee’s access to pending client decisions. Client priority is the practical test: a personal purchase should not compete with, anticipate, or disadvantage a planned client purchase.

Worked example: An employee requests approval to buy a security that the client desk plans to purchase that afternoon. Preclearance identifies the pending client order. The personal request is held or denied, and the reason is recorded under the applicable controls. Monitoring then checks that no related personal account traded anyway. The client order receives appropriate priority even if the employee first proposed the idea.

Watch the distinction: Preclearance is not a permanent license to trade. A newly planned client order or restriction can change the conflict before the employee’s order executes.

Reference: [2]

34. Reveal referral incentives

A referral arrangement can influence which adviser or manager someone recommends. Identify compensation or benefits tied to an introduction, including direct payments, reciprocal referrals, and other economic advantages. Disclose the arrangement and its nature before the client relies on the recommendation, so the client can judge its independence. Then evaluate the recommended manager on its merits using the client’s mandate, comparable evidence, and fees. A sound due diligence score does not erase the incentive; likewise, the existence of an incentive does not by itself prove the recommendation is poor. Keep the referral terms and the recommendation rationale clear enough for review. Ask whether the payment changes with the client’s choice or assets, since that detail may affect the strength of the conflict.

Worked example: A consultant recommends Manager A and would receive a payment if the client appoints A. Manager A also scores well on process and operations. Before the decision, the consultant explains the payment arrangement and when it is earned, then presents the same comparison criteria used for other candidates. The client can weigh both the evidence and the incentive.

Watch the distinction: A good manager score does not make a referral payment irrelevant. The client needs to know about the incentive before deciding whether to rely on the recommendation.

Reference: [2]

35. Supervise through effective controls

Effective supervision pairs clear rules with evidence that people follow them. Define who approves composite changes, who checks source data, and who can release corrected performance. Monitor exceptions rather than relying only on annual attestations. When a warning sign appears, preserve the affected records, establish what changed and why, determine which reports are affected, and correct errors promptly. Then address the cause through training, access limits, independent review, or revised approval steps. The response should match the risk: repeated after-the-fact portfolio reassignment can distort reported returns and deserves closer investigation than a documented administrative correction. A supervisor cannot delegate away responsibility merely by assigning the calculation to someone else. Review evidence that the new control operates before treating the matter as closed.

Worked example: A supervisor notices three changes to composite membership after quarterly returns became known. For each change, compare the portfolio’s mandate and eligibility dates with the written membership rules. Recalculate affected composite returns, correct presentations if needed, and require documented independent approval for future changes. Repeated changes without support indicate a control failure, even before intent is determined.

Watch the distinction: A written policy does not supervise itself. Repeated unexplained exceptions call for investigation, correction of affected results, and evidence that the strengthened review works.

Reference: [2]

Level II - Performance measurement

36. Apply composite membership rules

A composite groups portfolios managed according to a defined investment strategy so its return represents that strategy rather than a handpicked account. Write inclusion rules that identify the strategy, discretion, relevant portfolio characteristics, entry and exit timing, and treatment of changes in mandate. Apply those rules consistently using information available at the classification date. Performance outcomes must not determine whether an otherwise eligible account stays in the group. When a portfolio genuinely changes strategy or ceases to meet a stated condition, document the effective date and reason. Reconcile the eligible portfolio list with account records to detect omissions. This discipline matters because a composite can be biased upward by excluding weak accounts even when every individual return is calculated correctly. Review the current standard and the firm’s documented approach for detailed presentation requirements.

Worked example: Two discretionary portfolios follow the same defined strategy throughout a quarter. One earns 8% and the other loses 4%. Both satisfy the established membership rules, so both belong in the composite for that period. The losing return is not grounds for removal. If its mandate changes next quarter, record the change date and apply the documented exit rule prospectively.

Watch the distinction: Composite membership is not a reward for good results. Decide eligibility from strategy and documented conditions, then calculate performance using all eligible accounts.

Reference: [2]

37. Weight composite returns by assets

A composite return aggregates the returns of eligible portfolios using an asset weighting method appropriate to the stated calculation period. For a simple single period with no complicating flows, beginning values provide weights: multiply each portfolio return by its beginning value, add the products, and divide by total beginning value. Larger portfolios therefore contribute more to the composite return. Keep portfolio returns on a comparable basis, including currency, fees, and valuation dates. Cash flows or portfolio entries during the period can require additional period calculations rather than applying one starting weight to the whole span. Check the result against a reasonableness bound: with positive weights, a simple weighted return should lie between the lowest and highest included returns. CIPM practice questions can help reinforce this calculation and its assumptions.

Worked example: Two eligible portfolios begin a period with 60 and 40 currency units. With no external flows, they return 10% and 5%. The beginning-asset weighted composite return is (60 × 10% + 40 × 5%) / (60 + 40) = (6 + 2) / 100 = 8%. That lies between 5% and 10%, as expected; the equal-weighted 7.5% would understate the larger portfolio’s influence.

Watch the distinction: A simple average gives each account equal influence regardless of size. Use the stated asset weighting method and check the relevant values and dates.

Reference: [2]

38. Evaluate record portability

When an investment team moves to a new firm, its earlier returns may be relevant, but the record does not travel automatically with one person. Investigate whether substantially the same decision-making capability and strategy continue, whether the former process can be reconstructed, and whether complete supporting records are available. Identify who conducted research, approved trades, managed risk, and supplied operational support at the predecessor firm. Compare those resources with the new setting and disclose any material break. A lead manager’s presence matters, yet the old results may have depended on a team or tools that stayed behind. Portability is therefore an evidence question about continuity and verifiability. Apply the relevant current performance reporting requirements before presenting predecessor results as part of a new firm’s record.

Worked example: A lead manager joins a new firm after producing a strong six-year record, but the former research team and trading platform remain at the prior firm. The manager brings trade summaries but lacks complete source records. Compare who made decisions before and after the move and whether the history can be supported. On these facts, simply attaching the six years to the new firm would overstate continuity.

Watch the distinction: A named portfolio manager does not personally own every result generated by a former organization. Test continuity of strategy, people, resources, and records.

Reference: [2]

39. Include cash in portfolio results

Cash held inside a managed portfolio is part of the capital whose performance the client experiences. If the manager controls the investment decision, the cash weight and its return belong in the total portfolio calculation. At a common valuation point, total return is the sum of each sleeve’s starting weight times its return, assuming no external flows that require further adjustment. Cash may reduce gains in a rising market and cushion losses in a falling one. Reporting only invested securities can conceal a decision to hold cash or delay deployment. Distinguish portfolio cash from an external client contribution: the former is an asset within the account, while the latter changes the capital base and may require a return break or flow adjustment. Examine cash policy when comparing managers, especially where mandates allow different cash ranges.

Worked example: A portfolio starts with 80% in equities and 20% in cash. Equities earn 10% over the period, cash earns 0%, and there are no flows or fees. Portfolio return is 0.80 × 10% + 0.20 × 0% = 8%. Quoting the equity sleeve’s 10% as the account return would omit the effect of the manager’s cash position.

Watch the distinction: Cash is not automatically outside performance measurement. If it belongs to the managed account, include its weight and return when calculating the account result.

Reference: [2]

40. Test composite construction bias

Composite construction can change the reported result even when every account return is accurate. Review which accounts entered, left, or moved between composites, the effective dates, and the documented reasons. A performance-driven exclusion, delayed inclusion of a weak account, or retroactive reassignment can bias the history. Recalculate the composite with the full eligible population and compare it with the reported figure. Asset weights matter, so the effect of an omitted portfolio depends on its size as well as its return. Look for changes made after returns became known and for patterns that favor one presentation repeatedly. Valid membership changes can occur when mandates or discretion genuinely change; those need dated evidence and consistent treatment. This review links performance measurement with governance over the account population.

Worked example: Two equally sized portfolios meet the same strategy rules for a period. One returns +10% and the other −10%, so the beginning-asset weighted composite return is (10% + −10%) / 2 = 0%. Removing the losing account after seeing its result would make the reported figure 10%, an upward distortion of 10 percentage points despite unchanged account returns.

Watch the distinction: Accurate individual returns do not guarantee an unbiased composite. Check eligibility and reassignment decisions, especially those made after performance is known.

Reference: [2]

Level II - Performance presentation

41. Assess compliance claims

A claim of compliance with the Global Investment Performance Standards concerns the firm’s performance reporting as a whole. A calculation method is only one part of that system. First define the firm and its discretionary strategies, then assign eligible portfolios to composites under consistent rules, maintain supporting records, and prepare presentations with the required information. A missing composite can make a firm-wide claim unsupportable even when every reported return is calculated correctly. When assessing a claim, ask what population of portfolios it covers, whether exclusions follow documented rules, and whether the records support the presentations. Independent verification can examine firm-wide processes, but it does not convert partial adoption into compliance. The CIPM program covers performance reporting based on GIPS; a specific claim still needs evidence from the firm.

Worked example: Firm A calculates returns using its documented method and reports its large equity strategy. It also manages an eligible small-cap strategy but has left that strategy out of its composite records. Accurate equity returns do not cure the omission. Firm A must determine the proper composite treatment, reconstruct supported records where possible, and resolve any affected presentations before making a firm-wide compliance claim.

Watch the distinction: It is tempting to treat a correctly calculated return as proof of GIPS compliance. The claim also depends on firm boundaries, composite coverage, records, and presentation practices.

Reference: [1]; [2]

42. Read performance disclosures

A reported percentage has little meaning until you know what was measured. Establish the valuation dates, return period, cash-flow treatment, and whether returns are gross or net of investment management fees. Then check the benchmark and material facts that could change how a reader interprets the result, such as a strategy change or a short operating history. Gross returns help isolate the investment result before specified fees; net returns show the effect of those fees under the stated calculation. Other costs may have different treatment, so the labels and methodology matter. Compare periods and fee bases consistently before ranking managers. If two presentations use different net-return assumptions, their figures are not directly comparable. Read disclosures as part of the result, rather than as optional fine print.

Worked example: A portfolio begins a period at 100 monetary units. With no external flows, it gains 10 before a management fee of 1 paid at the end. Gross return is 10/100 = 10%. Ending value after the fee is 109, so net return is (109 − 100)/100 = 9%. A report showing only 10% would leave the fee effect unclear.

Watch the distinction: A gross return is not necessarily the return a client kept. Check which fees were deducted, when they were charged, and whether comparison figures use the same basis.

Reference: [1]; [2]

43. Choose and explain a benchmark

A benchmark should reflect the mandate the manager was asked to pursue, including its asset mix, market exposure, currency, and investable opportunity set. Define it before judging results, because choosing an index after seeing returns can make ordinary performance appear exceptional. For a blended benchmark, state its component indexes, target weights, rebalancing convention, and return basis. Match the portfolio and benchmark over identical periods and in the same currency. A benchmark can be useful even when imperfect, provided its limitations are visible. Examine both the total difference and its sources: a manager can beat a poorly matched benchmark by taking exposures the mandate never intended. For more work on Level II performance reporting, keep the benchmark definition beside the return series.

Worked example: A hypothetical mandate targets 60% equities and 40% bonds, rebalanced to those weights at the period start. Equity and bond indexes return 10% and 2%. The blended benchmark earns 0.60 × 10% + 0.40 × 2% = 6.8%. A portfolio return of 7.5% exceeds it by 0.7 percentage points, subject to checking comparable fees, currency, and dates.

Watch the distinction: The highest available index return is not automatically the right comparator. A benchmark tests the agreed mandate; changing it after observing performance distorts the comparison.

Reference: [2]

44. Understand verification limits

Verification addresses whether a firm has established and applied firm-wide policies and procedures for performance reporting under GIPS. Its scope differs from checking every individual number in every client report. A separately scoped performance examination may address a particular composite, but neither exercise promises future accuracy or investment results. When a discrepancy appears, trace the source valuation, cash flows, classifications, and calculations; determine which periods and reports were affected; and correct the record. Ask what period the verification covered and read the verifier’s actual opinion before drawing conclusions. An assurance label cannot replace reconciliation of a specific disputed figure. This distinction matters when evaluating evidence: firm-wide controls reduce reporting risk, while a return claim still needs its own supporting data and calculations.

Worked example: A verified firm discovers that one account’s month-end value was overstated. The verification status does not establish that the client’s reported return was correct. The firm should obtain the corrected valuation, recalculate affected periods and any composite using that account, identify recipients of incorrect presentations, and issue corrections where needed. It should also investigate why its controls missed the error.

Watch the distinction: Verification is sometimes read as a guarantee for each return in every report. Its stated scope must be checked; a newly found valuation error still requires investigation and correction.

Reference: [1]; [2]

45. Define the reporting firm

The reporting firm is the entity whose performance record and compliance claim are being presented. Its boundary should reflect how the investment business is organized and held out to clients, with enough consistency that weak results cannot be separated from strong ones merely for marketing. Start with the legal and operational structure, then examine branding, shared investment decision makers, client relationships, and the ability to maintain complete records. Document changes such as acquisitions or reorganizations and assess their effect on historical presentations. Once the boundary is set, identify the portfolios and composites within it. A defensible definition lets readers understand what the claim covers; selective definition can change the apparent performance record without changing a single portfolio return.

Worked example: A group markets two desks under one investment brand and shares client reporting systems. After one desk performs poorly, management proposes describing only the stronger desk as the reporting firm. The analyst should test whether that narrower boundary matches how the business has actually been held out and operated. If it merely removes weak results, the proposed presentation gives clients a distorted record.

Watch the distinction: A firm boundary cannot be chosen solely to improve reported performance. Examine the business presented to clients and apply the resulting definition consistently across portfolios and periods.

Reference: [1]; [2]

Level II - Performance attribution

46. Separate fixed-income effects

Bond performance can reflect several mechanisms at once. Carry is income earned while holding the bonds; changes in government yield curves alter prices through rate sensitivity; credit-spread changes alter compensation for credit risk; and security or trading decisions add further effects. A key-rate duration estimates sensitivity to a move at a specified maturity, holding the other modeled rates fixed. For a small yield change, approximate price return from that key rate is negative duration times the yield change in decimal form. Add the key-rate contributions for an approximate curve effect, then examine carry, spread, and residual effects under the chosen attribution model. Duration is a first-order approximation: large moves, convexity, changing holdings, and interactions can prevent exact reconciliation without additional terms.

Worked example: Suppose a bond portfolio has illustrative key-rate durations of 4 at five years and 2 at ten years. The five-year yield falls 20 basis points, or 0.002, adding about −4 × (−0.002) = 0.8%. The ten-year yield rises 10 basis points, or 0.001, subtracting about 2 × 0.001 = 0.2%. Approximate curve effect is +0.6% before carry, spreads, and other effects.

Watch the distinction: A positive bond return after rates fall is not entirely a rate effect. Income, spread moves, and security choices may also matter; key-rate estimates are approximations.

Reference: [2]

47. Attribute currency returns

For an unhedged overseas investment, the base-currency value changes with both the asset’s local return and the exchange rate. Express the currency return as the change in base-currency value of one unit of foreign currency. Then compound: base-currency return equals (1 + local return) × (1 + currency return) − 1. Expanding the product reveals a local term, a currency term, and their interaction. An attribution system may assign that interaction by convention, so check definitions before comparing reports. For a hedged position, forward contracts and their costs or gains require separate treatment; the simple unhedged formula no longer tells the whole story. Currency attribution is most informative when exposures, hedge ratios, and benchmark currency treatment are stated alongside the results.

Worked example: An unhedged overseas asset gains 8% in local currency. Over the same period, that currency loses 5% against the investor’s base currency. One base-currency unit invested grows by 1.08 × 0.95 = 1.026, giving a 2.6% base-currency return. The interaction is 8% × −5% = −0.4 percentage points, explaining why 8% − 5% overstates the answer.

Watch the distinction: Simply adding local and currency percentages drops their interaction. Multiply the two growth factors, and check whether the reported investment was hedged before interpreting currency effects.

Reference: [2]

48. Decompose alternative returns

An illiquid investment can distribute cash while losing value overall. Start with the opening value, add income received and the closing value, subtract the opening value and period costs, then divide by the opening value when there are no external capital flows. Separate recurring operating income from valuation changes and transaction or management costs. This decomposition helps identify whether a result came from cash generation, appraisal assumptions, leverage, or expenses. Appraised values may be infrequent and sensitive to assumptions, so a smooth reported series need not imply low economic risk. If capital contributions or distributions occur during the period, their timing must be handled explicitly using a suitable return method. State the valuation basis and return horizon before comparing alternative assets with liquid benchmarks.

Worked example: An asset opens at 100 monetary units, distributes 6 of income, closes at an appraised value of 94, and incurs 2 of period costs. With no other flows, net return is (6 + 94 − 100 − 2)/100 = −2%. The 6% income yield was outweighed by a 6% valuation decline and 2% in costs.

Watch the distinction: Cash yield alone does not measure total performance. Include the change in asset value and relevant costs, then inspect whether the closing valuation is reliable.

Reference: [2]

49. Reconcile cumulative attribution

Single-period active return is the portfolio return minus the benchmark return for that period. Across multiple periods, each series compounds on its own base, so the sum of period active returns generally differs from the difference between cumulative portfolio and benchmark returns. To reconcile an attribution report, first calculate each cumulative return by multiplying its period growth factors. Next subtract cumulative benchmark return from cumulative portfolio return. Finally apply a stated linking method to allocate that total active result among period effects. Different valid linking conventions can assign interaction terms differently, but the linked effects should reconcile to the reported cumulative active return. Watch for changes in portfolio weights and benchmark composition, since these alter the economic meaning of period effects as well as the arithmetic.

Worked example: A portfolio earns 20% and then loses 10%; its benchmark earns 10% and then 0%. Period active returns are +10 and −10 percentage points, summing to zero. Portfolio cumulative return is 1.20 × 0.90 − 1 = 8%; benchmark cumulative return is 1.10 × 1.00 − 1 = 10%. Cumulative active return is therefore −2 percentage points.

Watch the distinction: Period active effects cannot usually be added to obtain cumulative active return. Compound the portfolio and benchmark separately, then use a linking convention to reconcile attribution.

Reference: [2]

50. Interpret factor attribution

Factor attribution estimates how much return is associated with measured exposures, such as market, size, value, duration, or credit risk. Compare the portfolio’s factor exposure with the benchmark’s, multiply each active exposure by the corresponding factor return, and sum the modeled contributions. The difference between observed active return and modeled contributions is a residual. That residual may contain security selection, unmeasured factors, estimation error, trading costs, or interactions. It is therefore a question to investigate, not immediate proof of skill. Check that the factor model matches the assets and period, that exposures were measured at appropriate times, and that factors are not so overlapping that their separate contributions become unstable. A useful interpretation connects the arithmetic to actual portfolio positions and decisions.

Worked example: A market factor returns 5%. The portfolio’s estimated market beta is 1.2 and the benchmark’s is 1.0, so active market exposure contributes approximately (1.2 − 1.0) × 5% = 1 percentage point. If observed active return is 2 points and this is the only modeled factor, the residual is 1 point. Investigate its sources before calling it skill.

Watch the distinction: An unexplained residual is not automatically manager alpha. It can reflect omitted exposures, model error, costs, or timing; review those possibilities against holdings and trades.

Reference: [2]

Level II - Performance appraisal

51. Measure downside deviation

Downside deviation measures the size of returns below a stated target, often called the minimum acceptable return. For each period, calculate return minus target; retain negative differences as shortfalls and replace nonnegative differences with zero. Square the shortfalls, average them using the stated denominator, and take the square root. This measure does not let an exceptionally strong period cancel a weak one. State whether the calculation uses all periods or only shortfall periods, because those conventions give different answers. Also align the target’s frequency with the return observations: a monthly target belongs with monthly returns. Downside deviation describes shortfall dispersion around a chosen threshold, not the worst loss or the probability of a particular future outcome.

Worked example: Take three equally weighted period returns of −4%, 2%, and 6%, with a 0% target. The shortfalls are −4, 0, and 0 percentage points. Using all three periods, downside deviation is √[(16 + 0 + 0)/3] = √(5.333) = 2.31 percentage points per period. Positive returns contribute zero rather than offsetting the loss.

Watch the distinction: Do not average only the losing periods unless that convention is explicitly specified. The denominator changes the measure, so state the target and calculation convention.

Reference: [2]

52. Trace maximum drawdown

Maximum drawdown measures the largest fall from a previous wealth peak to a later trough within an observed path. Move chronologically through the wealth index, updating the running peak whenever a new high occurs. At each observation, compute current value divided by that peak minus one. The most negative result is maximum drawdown. The order of returns matters: the same set of periodic returns in a different order can produce a different drawdown. After the trough, a gain does not erase the drawdown until wealth regains the earlier peak. Report the observation frequency and window, since monthly data can miss deeper declines visible in daily data. Drawdown describes one realized path, so it should complement, rather than replace, broader risk analysis.

Worked example: A wealth index follows 100, 120, 90, and 108. The running peak reaches 120. At 90, drawdown is 90/120 − 1 = −25%, the deepest observed loss. The final value of 108 is 108/120 − 1 = −10% below that peak, so the portfolio has partly recovered but has not set a new high.

Watch the distinction: A rebound from the trough is not the same as recovery. Compare the new value with the preceding peak, and calculate drawdowns in chronological order.

Reference: [2]

53. Compare market capture

Market capture compares a manager’s return with a benchmark’s return over periods selected by the benchmark’s direction. Upside capture uses periods when the benchmark rose; downside capture uses periods when it fell. Define the sampling frequency and compound returns within each subset before taking the manager-to-benchmark ratio. A result above 100% for upside capture means the manager gained more than the benchmark during the selected up periods. A downside capture below 100% often indicates a smaller loss, provided both compounded subset returns are negative and the ratio is interpreted consistently. Capture figures can become unstable when the benchmark subset return is close to zero. They also omit the sequence and number of selected periods, so inspect the underlying returns before treating one ratio as a complete risk assessment.

Worked example: In two periods when the benchmark rose, a manager earns 8% and 7%; the benchmark earns 10% and 5%. Compounded manager gain is 1.08 × 1.07 − 1 = 15.56%. Benchmark gain is 1.10 × 1.05 − 1 = 15.50%. Upside capture is 15.56%/15.50% = 100.39%, a slight edge over those two up periods.

Watch the distinction: Capture is calculated on periods classified by benchmark direction, not by whether the manager rose or fell. Mixing those classifications changes the question being answered.

Reference: [2]

54. Check performance persistence

Persistence asks whether managers with strong relative results remain strong in later, independent periods. Define a peer group and ranking rule, sort managers in the first period, and observe their ranks in the next. Compare the repeat rate with a plausible baseline, such as 25% for an even four-quartile split under a simple random-ranking model. A small difference from that baseline may reflect chance, especially with few managers. The analysis also depends on survivorship, changes in mandates, and whether the peer group remains comparable. Repeated top ranks can motivate investigation of process and risk exposures; they do not by themselves establish skill. Use multiple nonoverlapping observations where possible, and ask whether apparent persistence survives fees and a consistent benchmark.

Worked example: Suppose 20 managers were in the top quartile last year and four are again in the top quartile this year. The observed repeat rate is 4/20 = 20%. An even-quartile random-ranking baseline predicts 20 × 25% = 5 repeaters, or 25%. Four rather than five offers little evidence of persistence, especially from one transition and a small sample.

Watch the distinction: A few repeat winners are expected even with random rankings. Compare their frequency with a baseline and check peer-group and survivorship effects before inferring durable skill.

Reference: [2]

55. Test excess-return evidence

A positive average active return is an estimate, not a certainty about a manager’s underlying ability. With independently distributed observations, estimate its standard error as sample standard deviation divided by the square root of the number of observations. Divide the sample mean active return by that standard error to form a t statistic for a zero-mean comparison. Larger dispersion or fewer observations weaken the evidence. Check whether active returns are serially correlated, whether the benchmark and strategy remained stable, and whether the manager was selected after seeing strong results. Those conditions can make a simple standard error too optimistic. Statistical evidence also differs from economic value: fees, capacity, and implementation may consume an apparent edge. Use CIPM practice questions to rehearse the arithmetic and its assumptions.

Worked example: Thirty-six monthly active returns average 0.4% per month and have a 2% monthly sample standard deviation. Assuming independent observations, standard error is 2%/√36 = 0.333% per month. The t statistic is 0.4%/0.333% ≈ 1.2. That is weak evidence against a zero mean under this simple model; serial dependence or manager selection would warrant further caution.

Watch the distinction: A positive sample mean does not prove repeatable excess return. Its uncertainty depends on variation, sample size, dependence between observations, and how the manager was selected.

Reference: [2]

Level II - Manager selection

56. Structure a manager search

Begin a search with the client’s objective, permitted assets, risk limits, liquidity needs, benchmark, and implementation constraints. Convert firm requirements into eligibility screens before reviewing returns. Then request comparable evidence: mandate-specific history, holdings and risk exposures, investment process, people, fees, and operational controls. Compare eligible candidates on measures that answer the client’s question, using the same periods and definitions. Follow quantitative screening with qualitative due diligence and reference checks, then test whether the mandate can actually be funded and monitored. An attractive return from an ineligible strategy cannot cure a conflict with the client’s risk limit. For a related perspective on investment manager evaluation, see the CIMA investments study guide.

Worked example: A client seeks global equities with a hypothetical tracking-error limit of 5 percentage points annually. Candidate A has 8 points of annual tracking error and a strong five-year return; Candidate B has 4 points and a lower return. A fails the stated risk screen unless the mandate changes. Compare B with other eligible managers, then examine process, fees, controls, and implementation before selection.

Watch the distinction: Screening solely for the highest historical return reverses the decision sequence. Establish the mandate and eligibility first, then assess return quality among candidates that fit.

Reference: [1]; [2]

57. Evaluate RFP responses

A request for proposal produces comparable answers only when candidates receive the same questions, definitions, dates, and evidence requirements. Set weights for criteria before reading submissions, using the client’s priorities. Score documented process, reporting capability, operations, fees, and other relevant areas with a common scale. Multiply each score by its criterion weight and sum the results. A weighted total helps organize a decision; it does not erase a serious control failure or prove that a self-reported answer is accurate. Investigate gaps, inconsistencies, and claims unsupported by records through interviews, demonstrations, and independent checks. Record why any judgment departs from the numeric ranking. This makes the selection traceable while leaving room for material facts that a simple score cannot capture.

Worked example: Using a five-point scale, weight process 40%, reporting 35%, and operations 25%. Candidate A scores 4, 5, and 2: 0.40 × 4 + 0.35 × 5 + 0.25 × 2 = 3.85/5. Candidate B scores 4, 4, and 4: 1.60 + 1.40 + 1.00 = 4.00/5. B leads, while A’s operations weakness needs specific investigation.

Watch the distinction: A weighted score is not an automatic appointment rule. Verify the evidence behind responses and treat a material operational weakness as a decision issue even when the total looks competitive.

Reference: [1]; [2]

58. Test strategy capacity

Capacity is the amount a strategy can manage before added assets are likely to erode its expected advantage or impede execution. Compare current strategy assets plus the proposed mandate with a reasoned capacity estimate, then examine the assumptions behind that estimate. Liquidity, typical position size, turnover, market impact, concentration limits, and available investment opportunities all matter. Capacity is strategy-specific: an approach trading large liquid securities may absorb more capital than one buying thinly traded names. A capacity estimate is not a guarantee or a fixed regulatory limit. Ask how the manager would deploy new money, whether it would dilute existing positions, and what happens in stressed markets. A proposed mandate can be reduced, staged, or rejected if execution would undermine the track record being evaluated.

Worked example: A hypothetical strategy estimates capacity at 1 billion monetary units and already manages 850 million. Adding a proposed 250 million mandate would bring assets to 1.10 billion, exceeding the estimate by 100 million, or 10% of estimated capacity. The allocator should test liquidity and expected market impact, then consider a smaller or staged allocation rather than assuming historical returns scale unchanged.

Watch the distinction: Past returns do not establish that a strategy can accept unlimited assets. More capital can raise trading costs, dilute opportunities, and change the portfolio that produced the record.

Reference: [1]; [2]

59. Compare fee structures

Compare manager fees under the same assets, period, return assumption, and definition of the fee base. A flat fee is assets multiplied by its rate. A performance fee adds a percentage of the eligible gain, often subject to a hurdle and other contract terms. Clarify whether the hurdle is absolute or benchmark-relative, whether it compounds, how losses carry forward, when fees crystallize, and whether fees are assessed before or after other charges. These details can materially change the cost. A performance fee may align pay with gains, but it can also reward market exposure or encourage risk taking if losses are treated asymmetrically. Model several return paths, not just one favorable year, and compare expected net outcomes and incentives for the client.

Worked example: On hypothetical assets of 50 million units, a flat annual fee of 0.6% costs 50 million × 0.006 = 300,000 units. Another contract charges 0.3% plus 10% of the gain above a 5% hurdle. If return is 10% and the fee base is opening assets, cost is 150,000 + 0.10 × 0.05 × 50 million = 400,000 units, before other terms.

Watch the distinction: The lower base fee is not necessarily the cheaper contract. Calculate the performance charge under the same return scenario and inspect hurdle, loss-recovery, and crystallization terms.

Reference: [1]; [2]

60. Check implementation feasibility

A manager’s published record matters only if the proposed client mandate can be implemented under its own constraints. Translate the intended portfolio into required trades, then compare trade sizes with market liquidity, participation limits, settlement needs, and the client’s funding deadline. Estimate a minimum trading period from position size divided by feasible daily execution, while allowing for price impact and days with less liquidity. Also check account opening, custody, derivatives permissions, currency conversion, cash management, and reporting readiness. Theoretical portfolio weights are not enough when operational setup or market access delays investment. If the client needs exposure sooner than the strategy can build it, consider a smaller initial mandate, a transition portfolio, or a different manager, and show how those choices alter cost and risk.

Worked example: A hypothetical mandate requires a 60 million-unit position. Under its participation limit and available liquidity, the manager can buy at most 1 million units per trading day. Even with full daily execution, minimum build time is 60 million/1 million = 60 trading days. A requirement to complete investment within 10 trading days is infeasible on these assumptions; faster trading would breach the stated limit.

Watch the distinction: A strategy’s track record does not prove the client can enter it promptly. Test trade size, liquidity, participation limits, and operational readiness against the actual deadline.

Reference: [1]; [2]

Official references and further reading

All study guides

Choose your exam version

CIPM Level 1

CIPM Level 2

Level I - Ethics and professionalism

1. Disclose competing interests

2. Protect independent judgment

3. Treat clients fairly

4. Safeguard client information

5. Maintain trustworthy records

6. Avoid selective performance claims

Level I - Performance measurement

7. Calculate time-weighted return

8. Solve money-weighted return

9. Place external cash flows correctly

10. Link returns geometrically

11. Value around a major flow

12. Separate gross and net returns

Level I - Performance attribution

13. Measure allocation effect

14. Measure selection effect

15. Account for interaction

16. Reconcile active return

17. Read sector contributions

18. Link attribution across periods

Level I - Performance appraisal

19. Measure total variability

20. Interpret the Sharpe ratio

21. Interpret the Treynor ratio

22. Estimate Jensen alpha

23. Use the information ratio

24. Calculate tracking error

Level I - Manager selection

25. Check mandate fit first

26. Design quantitative screens

27. Test qualitative evidence

28. Examine process consistency

29. Assess organizational stability

30. Review operational resilience

Level II - Ethics and professionalism

31. Put client interests first

32. Control nonpublic information

33. Manage personal trading

34. Reveal referral incentives

35. Supervise through effective controls

Level II - Performance measurement

36. Apply composite membership rules

37. Weight composite returns by assets

38. Evaluate record portability

39. Include cash in portfolio results

40. Test composite construction bias

Level II - Performance presentation

41. Assess compliance claims

42. Read performance disclosures

43. Choose and explain a benchmark

44. Understand verification limits

45. Define the reporting firm

Level II - Performance attribution

46. Separate fixed-income effects

47. Attribute currency returns

48. Decompose alternative returns

49. Reconcile cumulative attribution

50. Interpret factor attribution

Level II - Performance appraisal

51. Measure downside deviation

52. Trace maximum drawdown

53. Compare market capture

54. Check performance persistence

55. Test excess-return evidence

Level II - Manager selection

56. Structure a manager search

57. Evaluate RFP responses

58. Test strategy capacity

59. Compare fee structures

60. Check implementation feasibility

Official references and further reading

Frequently Asked Questions

Related Study Guides

Investment Management Certificate (IMC)

Certificate in ESG Investing (CFA Institute)

Certified Commercial Investment Member (CCIM)