Fixed-income attribution

Fixed-income attribution

Published September 28, 2026

Fixed-income attribution

On this page

Learning Outcome Statement

Interpret the output from fixed-income attribution analyses.

On the exam: You receive a quarterly fixed-income attribution output with distinct portfolio and benchmark returns plus sector allocation, yield-curve, spread, security selection, and interaction effects. You reconcile active return, then choose or write the conclusion about persistence, benchmark fit, and manager skill.

An attribution report tells you why a bond portfolio beat or lagged its benchmark. It starts with the benchmark return, then adds active effects from sector allocation, yield-curve positioning, spread positioning, security selection, and interaction. Call the portfolio's quarterly total return $R_P$ and the benchmark's quarterly total return $R_B$, then check that their difference equals the sum of the effects. Only after that check do you judge persistence, benchmark fit, and manager skill.

Return or effectNotation
Portfolio return$R_P$
Benchmark return$R_B$
Active return$R_P - R_B$
Sector allocation effect$A$
Yield-curve effect$Y$
Spread effect$S$
Security selection effect$Q$
Interaction effect$I$

Keep the labels separate. A contribution belongs to the portfolio's own return, while an attribution effect explains the difference between $R_P$ and $R_B$.

The arithmetic reconciliation is the check you perform.

$$ R_P - R_B = A + Y + S + Q + I $$

Start with $R_P$ and subtract $R_B$. A positive $A$, $Y$, $S$, $Q$, or $I$ raises active return, while a negative effect lowers it. The right side assigns the return difference to the manager's decisions.

This is arithmetic attribution, so the effects add within the measured period. Geometric attribution compounds returns across periods, so you do not add its effects with this equation. Label the report gross or net because gross returns are before management fees and net returns include them. Separate performance measurement, which records results, from performance appraisal, which judges manager skill.

EffectPositive signNegative sign
Sector allocationPortfolio weights helpedPortfolio weights hurt
Yield-curve positioningCurve stance helpedCurve stance hurt
SpreadSpread positioning helpedSpread positioning hurt
Security selectionBond choices helpedBond choices hurt
InteractionCombined choices helpedCombined choices hurt

Read the signs from the investor's perspective. Sector allocation reflects sector weights, yield-curve and spread effects reflect positioning, security selection reflects bond choices, and interaction reflects their combined result. A positive selection effect is evidence about the decision, not proof that the manager can repeat it.

Use the output in sequence. Reconcile returns before interpreting any effect, because a report that does not sum cannot support a reliable appraisal. Then read the signs from the investor's perspective and identify which decisions raised or lowered active return. Ask whether the pattern persists across periods and whether the benchmark represents the mandate. These checks turn a list of quarterly numbers into evidence you can weigh, while keeping measurement separate from the later judgment about skill.

Loading diagram…

A positive effect earns attention at the decision stage, not a final skill verdict. Read each pattern by asking what it changes in the arithmetic and what it can establish about the manager. A favorable effect can explain active return for this quarter, while persistence and benchmark fit remain open questions. If the effects fail to sum, pause: the active measurement itself needs review before any appraisal. This distinction keeps a correct calculation from being mistaken for a complete conclusion.

PatternArithmetic readingAppraisal caution
Positive contributionPortfolio return roseActive outperformance unknown
Positive selectionActive return rosePersistence still untested
Effects fail to sumAttribution does not reconcileStop before appraisal
Positive active returnPortfolio beat benchmarkFit and persistence still matter

A contribution can raise portfolio return without raising active return. A benchmark that does not fit the core bond mandate can make an effect less persuasive as evidence of skill. Use the two checks in order: first decide whether the returns and effects reconcile, then consider whether the decisions look repeatable and were judged against a suitable mandate. Stronger evidence comes from a pattern that survives both tests; a single favorable quarter or a poor comparison weakens the claim. Keep the measurement result intact while you qualify what it says about manager skill.

CheckStronger evidenceWeaker evidence
Returns reconcileReliable active measurementReview inputs and signs
Effects persistRepeatable decisions possibleSingle-period evidence
Benchmark fits mandateFair manager comparisonSkill claim less persuasive
Selection stays positiveFavorable decision patternIsolated success

Use the reconciliation for performance measurement first. Use persistence, selection pattern, and benchmark fit for performance appraisal.

Worked example

You are handed a quarterly gross arithmetic attribution report for a core bond manager against a distinct benchmark. It gives the benchmark return and active effects, but not the portfolio return. You need to reconstruct the portfolio return, reconcile active return, and decide how much the positive selection result supports a skill claim.

GivenValue
Benchmark return0.0400
Sector allocation effect0.0010
Yield-curve effect0.0040
Spread effect0.0020
Security selection effect0.0030
Interaction effect-0.0020
Report basisArithmetic, gross of fees

The bars show cumulative quarterly total return in decimal form as each effect is added to the benchmark starting point. Begin at the benchmark return of 0.0400, then follow the effects in the same order as the report. Each bar shows the running total, so the last bar represents the portfolio return after interaction is included. This view makes the bridge visible, but it does not replace the arithmetic check: add the effects separately and confirm that their total equals portfolio return minus benchmark return. Only then interpret selection as evidence for appraisal.

Loading diagram…
  1. Add the listed effects because their sum is the active return above the benchmark.

$$ A + Y + S + Q + I = 0.0010 + 0.0040 + 0.0020 + 0.0030 - 0.0020 = 0.0080 $$

  1. Walk through the bridge in report order because the final cumulative value is the portfolio return.

$$ \begin{aligned} R_B + A &= 0.0410 \ R_B + A + Y &= 0.0450 \ R_B + A + Y + S &= 0.0470 \ R_B + A + Y + S + Q &= 0.0500 \ R_B + A + Y + S + Q + I &= 0.0480 \end{aligned} $$

  1. Subtract the benchmark from that portfolio return because the independent active-return measure must match the attribution sum.

$$ R_P - R_B = 0.0480 - 0.0400 = 0.0080 $$

The report reconciles to a portfolio return of $0.0480$ and an active return of $0.0080$, but the positive selection effect of $0.0030$ supports review of the selection process rather than a finding of repeatable skill.

Changed condition

Keep the mandate and report basis unchanged, but weaken security selection.

MeasureOriginal reportWeaker selection
Security selection effect0.00300.0010
Active return0.00800.0060
Portfolio return0.04800.0460

Predict the direction before looking at the calculations. A lower selection effect lowers active return and portfolio return because the benchmark and the other effects do not move. The remaining positive active return is weaker evidence about skill, so persistence and benchmark fit deserve more weight. Before reading on, state whether the weaker selection changes the skill conclusion or only weakens its evidence.

Common error

You see a positive contribution or positive selection effect and may call the manager skilled. That feels right because the manager’s own portfolio choices produced the result. But a failed reconciliation, or a single-period result without evidence that it persists and fits the benchmark, does not support that conclusion.

Check yourself

Practice

A weaker-selection report leaves the benchmark and other effects unchanged. Which conclusion best follows?

Revision snapshot

Reconcile active return before appraisal with the equation below.

$$ R_P - R_B = A + Y + S + Q + I $$

Keep portfolio return, benchmark return, contribution, and attribution as separate labels. Fixed-income effects commonly include sector allocation, yield-curve positioning, spread, security selection, and interaction. Arithmetic attribution adds within the period, while geometric attribution compounds across periods. Label gross or net and distinguish performance measurement from manager appraisal. A positive selection effect shows a favorable decision, not repeatable skill. Check persistence and benchmark fit. The exam may ask you to review a quarterly gross core bond report against a distinct benchmark and compare the weaker-selection condition before judging skill.

Unlock the full lesson and continue reading

Signup to continue reading this lesson and unlimited access questions, flashcards, AI notes, and more

Scan to download app

Scan to download
UNLOCK FREE ACCESS

Practice Questions: Fixed-income attribution

  1. 1. Meridian Asset Consulting is reviewing the annual performance of a US small-cap equity fund for a pension client. The fund is managed against a broad US total-market index, although its investment mandate permits only securities within the US small-cap opportunity set. The pension committee wants performance measurement to reflect the manager's investment decisions rather than the timing of contributions determined by the plan sponsor. The sponsor made one external contribution during the year. The fund was valued immediately before the contribution, and no other external cash flows occurred. Exhibit 1 provides all values needed to calculate the annual return. Returns are not affected by taxes, and the fund values are after transaction costs but before investment management fees. The consultant also conducts a sector attribution analysis using beginning-of-period weights. Exhibit 2 reports sector weights and returns for both the fund and its current benchmark. The consultant uses the Brinson–Fachler model, under which allocation is measured relative to the total benchmark return. Selection and interaction effects are reported separately. During a committee meeting, one trustee argues that the current broad-market index should be retained because it is widely recognized. Another trustee proposes using the median return of a database of small-cap managers. The consultant notes that the database median cannot be purchased or replicated and is affected by changes in the managers included in the database. The consultant instead considers an investable index representing the securities available under the fund mandate. Finally, the committee asks whether the manager has generated active return through security-specific skill after controlling for systematic factor exposures. Exhibit 3 reports the fund's annualized active return, tracking error, and residual risk. The consultant explains that the information ratio uses tracking error, whereas the appraisal ratio uses residual risk. All risk measures were estimated over the same period as the annualized active return.

    • 2. Northbridge Pension Plan is reviewing the performance of an externally managed domestic equity fund. The investment policy statement gives the manager discretion to select securities from the domestic mid-cap value universe. The manager is not responsible for determining the timing or amount of contributions and withdrawals initiated by the plan sponsor. Accordingly, the investment committee wants a return measure that isolates the manager's investment decisions from sponsor-controlled external cash flows. During the most recent quarter, the sponsor made a large contribution immediately after the portfolio was valued at the end of Month 1. Exhibit 1 provides the valuations and cash-flow timing. All portfolio values are stated after investment management fees except where otherwise indicated, and no other external cash flows occurred during the quarter. The committee asks its performance analyst to calculate the quarter's time-weighted return by valuing the portfolio at the time of the external cash flow and geometrically linking the resulting subperiod returns. The analyst also performs sector attribution. Exhibit 2 contains the complete portfolio and benchmark sector data; there were no other sectors. The analyst uses the Brinson–Fachler framework, under which each sector's allocation effect is calculated as the difference between portfolio and benchmark sector weights multiplied by the difference between that sector's benchmark return and the total benchmark return. Selection and interaction effects are calculated separately. The committee next reviews annualized statistics in Exhibit 3. One member states that the information ratio should use the fund's total standard deviation because this measure captures all uncertainty experienced by beneficiaries. The analyst responds that the appropriate denominator depends on whether performance is assessed relative to the risk-free asset or relative to the designated benchmark. Finally, the committee questions whether its current broad domestic equity benchmark remains appropriate. The manager's mandate has consistently required investment in domestic mid-cap value securities, and the manager has not used large-cap or growth securities as strategic allocations. Investable, rules-based broad-market, large-cap value, and mid-cap value indices are all available, transparent, and measurable. The committee wants the benchmark that provides the most valid assessment of decisions within the manager's assigned opportunity set.

      • 3. A portfolio earns a cumulative return of 21.00% over two years. The calculated annualized geometric return is closest to:

        • A. 9.53%
        • B. 10.00%
        • C. 10.50%
        Show answer

        Answer: B. 10.00%

        10.00% - A 21.00% cumulative return gives an ending wealth factor of 1.21 after two years. - The annualized geometric return is \(1.21^{1/2}-1=\) 10.00%, which compounds back to 21.00% over two years. 9.53% - This is the continuously compounded annual rate: \(\ln(1.21)/2 \approx 9.53\%\). - It is not the annualized geometric return, which must be expressed as an effective annual rate. 10.50% - Dividing the cumulative return equally between two years gives 21.00% ÷ 2 = 10.50%. - That simple average ignores compounding: \(1.105^2\) exceeds the given ending wealth factor of 1.21.

      • 4. An asset manager markets one pooled fund as “GIPS compliant,” although the manager does not claim firm-wide compliance. Which conclusion is most accurate?

        • A. The firm's compliance claim applies only to the pooled fund
        • B. The pooled fund may claim compliance independently of the firm
        • C. The firm must claim compliance on a firm-wide basis
        Show answer

        Answer: C. The firm must claim compliance on a firm-wide basis

        The firm must claim compliance on a firm-wide basis - Under the GIPS standards, the investment management firm—not an individual pooled fund—claims compliance. - Because the manager does not claim firm-wide compliance, it cannot market the pooled fund as “GIPS compliant.” The firm's compliance claim applies only to the pooled fund - A firm cannot limit a GIPS compliance claim to the pooled fund while leaving the rest of the firm outside the claim. - The manager’s fund-specific marketing statement therefore does not establish a valid compliance claim. The pooled fund may claim compliance independently of the firm - The pooled fund is a product managed by the firm, not an entity that can make an independent GIPS compliance claim. - Describing the fund itself as “GIPS compliant” does not substitute for compliance claimed by its manager.

      • 5. An evaluator should most likely emphasize a time-weighted return when which condition applies?

        • A. External cash flows are controlled by the client
        • B. Portfolio valuations are available only at year-end
        • C. The manager determines all contribution dates
        Show answer

        Answer: A. External cash flows are controlled by the client

        External cash flows are controlled by the client - Time-weighted return removes the effect of external cash-flow timing by linking returns for the periods between those flows. - Because the client, not the manager, controls the flows, this measure better isolates the manager’s investment performance. Portfolio valuations are available only at year-end - Accurate time-weighted returns generally require valuations at external cash-flow dates to calculate returns for the periods between flows. - Year-end-only valuations do not establish when cash flows occurred or allow their timing effects to be removed accurately. The manager determines all contribution dates - When the manager controls contribution timing, that timing is part of the manager’s decisions rather than a client-imposed effect to remove. - A money-weighted return is more relevant because it reflects the impact of those contribution decisions on the investor’s return.

      • 6. A firm has a discretionary, fee-paying portfolio that has not been included in any of its composites. Under the GIPS standards, the firm should most likely:

        • A. Include the portfolio in at least one composite
        • B. Exclude the portfolio if it is immaterial to firm assets
        • C. Report the portfolio's performance separately outside any composite
        Show answer

        Answer: A. Include the portfolio in at least one composite

        Include the portfolio in at least one composite - GIPS requires every actual, fee-paying discretionary portfolio to be included in at least one composite. - Including this portfolio prevents the firm from selecting only favorable results for its composite performance. Exclude the portfolio if it is immaterial to firm assets - Immateriality does not exempt an actual, fee-paying discretionary portfolio from composite inclusion. - The requirement applies even if the portfolio represents a small share of firm assets. Report the portfolio's performance separately outside any composite - Separate reporting does not satisfy the requirement to include this portfolio in at least one composite. - Its performance may be reported separately, but that does not replace composite inclusion.

      • 7. A portfolio earned a cumulative time-weighted return of 12.00% over 18 months. The calculated annualized return is closest to:

        • A. 7.85%
        • B. 8.00%
        • C. 12.00%
        Show answer

        Answer: A. 7.85%

        7.85% - Annualize the 18-month cumulative return by compounding over 12 months: \((1+0.12)^{12/18}-1 \approx 7.85\%\). - The exponent 12/18 converts the 18-month growth factor into a one-year growth factor. 8.00% - Multiplying \(12.00\%\times12/18\) produces 8.00%. - That arithmetic annualization ignores the compounding required for a time-weighted return. 12.00% - 12.00% is the cumulative return over the full 18 months, not an annual return. - Reporting it unchanged omits the 12/18 adjustment for the measurement period.

      Flashcards: Fixed-income attribution

      1. 1. Micro attribution

        Show answer

        Analyzes the portfolio at the portfolio manager’s level.

      2. 2. Returns-Based Style Analysis

        Show answer

        The returns on the manager's fund are regressed against the returns for various security indices. The regression coefficients, which represent the exposure, are constrained to be nonnegative and sum to one. The security indices used in the regression should be mutually exclusive of one another, be exhaustive and represent distinct, uncorrelated sources of risk. R2 measures style fit and one minus this indicates the amount unexplained by style and due to security seleciton. Helps to determine if the manager's reported style and actual style are the same. A series of regressions can be used to check the style consistency of a manager.

      3. 3. The appraisal ratio

        Show answer

        Measures the ratio of active return, α, to the volatility of the residual term, σε, both derived from a factor-based regression.\

      4. 4. Performance Attribution

        Show answer

        - Macro Attribution - Conducted at the fund sponsor level - Micro Attribution - Carried out at the investment manager level

      5. 5. Benchmark Properties

        Show answer

        - Specified in advance - Appropriate - Measurable - Unambiguous with clearly defined securities and weights - Reflective of the manager's current investment opinions - Accountable: manager and client agree it is applicable - Investable SAMURAI

      6. 6. Validity of Using Manager Universes as Benchmarks

        Show answer

        Finishing in the top half of all managers in the same style might be a good performance objective, it has a number of drawbacks as a benchmark: - Apart from being measurable it fails all of the other properties of a valid benchmark - Have to rely on compiler's classifications - Subject to survivor bias

      7. 7. Information Ratio

      8. 8. Portfolio Return Components

        Show answer

        A portfolio can be broken up into three components: Market, Style and Active Management P = M + S + A S = B - M\ A = P - B

      9. 9. Hedge Fund Index Issues

        Show answer

        - Biases in index creation - Value weighted often become skewed to best performer - Relevance of past data - Survivorship bias - Stale price bias - Backfill bias

      10. 10. Fixed-Income Attribution

        Show answer

        Factors that contribute to the total return of a fixed income portfolio are the External IR Effect and the Management Effect. External IR Effect - Is estimated using term structure analysis - Manager has no control over external IR environment - Can be seperate into expected (forward curve) and unexpected (what actually happened) Management Effect - Interest rate management effect - Sector/quality management effect - Security-selection effect - Trading effect

      Everything you need for CFA prep

      Get full Oncourse access with lessons, practice questions, flashcards and AI study tools.

      GET STARTED FOR FREE
      75% OFFLimited time offer
      GET 75% OFF