Performance Evaluation, Governance and Ethics

Introduction

Investment performance evaluation asks a deceptively difficult question: did a portfolio or manager create value, and was the return sufficient for the risk taken? A reliable answer requires more than comparing annual returns. The evaluator must define the investment objective, select an appropriate benchmark, measure risk consistently, distinguish skill from market exposure, and investigate whether governance and ethical standards were observed.

Performance evaluation is therefore both quantitative and institutional. Sharpe, Treynor, Jensen’s alpha and information ratios provide complementary views of risk-adjusted results. Attribution analysis explains the sources of relative return. Governance determines who is accountable for decisions, while fiduciary duty and professional ethics constrain how capital may be managed. These dimensions are inseparable: a high return generated through undisclosed conflicts, excessive risk, benchmark manipulation or unsuitable investments is not an acceptable investment outcome.

Learning Objectives

By the end of this chapter, the learner should be able to:

1. Calculate and interpret Sharpe, Treynor, Jensen alpha and information ratios.
2. Select and evaluate suitable benchmarks.
3. Distinguish performance measurement, attribution and appraisal.
4. Decompose active returns into allocation, selection and interaction effects.
5. Assess manager skill, consistency, style drift and risk-adjusted outcomes.
6. Explain fiduciary duty, stewardship, conflicts of interest and professional ethics.
7. Design governance controls for investment organisations.
8. Critically evaluate performance reports and ethical risks in practical settings.

1. The Architecture of Performance Evaluation

Performance evaluation has three related but distinct purposes:

• Measurement records what the portfolio earned.
• Attribution explains why it earned that return.
• Appraisal judges whether the outcome reflects skill, luck, mandate compliance and acceptable risk.

A basic holding-period return is:

where V₀ is beginning portfolio value, V₁ is ending value and D is cash income received during the period. The formula assumes that external contributions and withdrawals are excluded. If external cash flows occur, time-weighted return is generally appropriate for evaluating the manager, whereas money-weighted return is appropriate for evaluating the investor’s actual experience.

For subperiod returns R₁, R₂, …, Rₙ, the time-weighted return is:

This compounds subperiod returns and removes the effect of investor-controlled cash-flow timing. Annualised return over T years is:

Returns must be compared on a consistent basis: gross or net of fees, before or after tax, and with consistent treatment of dividends, withholding tax and transaction costs.

2. Risk-Adjusted Performance Measures

Let Rₚ denote portfolio return, R_f the risk-free return, R_m the market or benchmark return, σₚ portfolio return volatility, βₚ portfolio beta, and αₚ abnormal return.

2.1 Sharpe Ratio

The Sharpe ratio measures excess return per unit of total risk:

It is appropriate when the investor bears the portfolio’s total variability and when returns are measured over comparable periods. If returns are monthly, both the excess-return numerator and standard deviation must be monthly; annualisation commonly uses:

This approximation assumes reasonably stable periodic returns and is not exact when serial correlation is material. The ratio penalises upside and downside volatility equally and can be misleading for asymmetric or illiquid returns.

2.2 Treynor Ratio

The Treynor ratio measures excess return per unit of systematic risk:

It is most useful for a well-diversified portfolio whose unsystematic risk is not material to the investor. Beta should be estimated against the relevant market index and over a period consistent with the return measurement. A portfolio with a low beta can have a high Treynor ratio despite substantial idiosyncratic risk; hence Treynor should not replace Sharpe when total risk matters.

2.3 Jensen’s Alpha

Under the single-factor Capital Asset Pricing Model, Jensen’s alpha is:

It is the portfolio’s return minus the return predicted by its beta exposure. A positive alpha indicates outperformance relative to the specified factor model, not proof of repeatable skill. Alpha depends critically on benchmark choice, beta estimation, time horizon, portfolio constraints and omitted risk factors. In multi-factor analysis, the general form is:

where Fᵢ is the excess return on factor i and βᵢ is the portfolio’s sensitivity to that factor.

2.4 Information Ratio

The information ratio evaluates active management:

where R_b is benchmark return and σ₍p−b₎ is the standard deviation of active returns, also called tracking error. It measures active return per unit of active risk. The numerator and denominator must use the same frequency and the benchmark must represent the opportunity set and mandate. A high information ratio suggests efficient active risk-taking, but it can be distorted by stale pricing, survivorship bias, benchmark changes or smoothing.

Comparison of Measures

3. Worked Quantitative Example

Assume a portfolio earns 12%, the risk-free rate is 4%, its standard deviation is 10%, beta is 1.1, and the market earns 10%. The benchmark return is 10% and tracking error is 5%.

The Sharpe ratio is:

The Treynor ratio is:

Thus, excess return is 7.27 percentage points per unit of beta. Jensen’s alpha is:

The alpha is 1.4%. The information ratio is:

The portfolio therefore produced a 0.40 unit of active return per unit of tracking error. These results should not be ranked mechanically: each answers a different question.

4. Benchmarking and Manager Evaluation

A benchmark is a reference portfolio against which performance and risk are assessed. A valid benchmark should be:

• Unambiguous: constituents and calculation rules are known.
• Investable or replicable: it represents a feasible alternative.
• Appropriate: it matches the portfolio’s asset class, geography, duration, currency and style.
• Measurable: returns are available at suitable frequency.
• Specified in advance: it is not selected after observing results.
• Accountable: the manager understands the mandate and deviations permitted.

A blended benchmark may be appropriate for a multi-asset portfolio. For example, a strategic benchmark could combine 60% global equities and 40% high-quality bonds, with weights rebalanced according to a documented rule. A peer-group median is useful for context but is not necessarily a true opportunity-cost benchmark because peers may have different risks, fees and constraints.

Manager evaluation should consider:

1. Net and gross returns.
2. Absolute and benchmark-relative performance.
3. Volatility, drawdown, downside risk and liquidity exposure.
4. Tracking error and active share where relevant.
5. Consistency across market regimes.
6. Fees, turnover and tax efficiency.
7. Compliance with mandate, risk limits and responsible-investment policies.
8. Evidence of repeatable process rather than one favourable episode.

A manager should not be rewarded for taking hidden beta, leverage or illiquidity that was not part of the mandate. Evaluation periods should match the strategy’s horizon: a long-horizon private-market strategy cannot be judged solely by one quarter, although valuation discipline and cash-flow reporting remain essential.

5. Performance Attribution

Attribution explains active return. In a simplified multi-asset setting, active return is:

For asset class i, let wₚᵢ and w_bᵢ be portfolio and benchmark weights, and Rₚᵢ and R_bᵢ their respective returns. A common Brinson-style decomposition is:

Summing across sectors gives active return, subject to the chosen Brinson convention and treatment of cash, fees and residuals. Allocation measures the effect of overweighting or underweighting a sector relative to the benchmark’s overall return. Selection measures security selection within the benchmark weight. Interaction captures the joint effect of weight and selection decisions.

Attribution cautions

Attribution requires reliable holdings, prices, classifications, cash-flow treatment and benchmark data. Results may change when a security is classified differently or when currency effects are separated. Attribution is descriptive, not automatically causal: a positive selection effect may result from an unintended factor exposure. Robust reporting should include residuals and reconciliation to the audited portfolio return.

6. Governance, Fiduciary Duty and Stewardship

Investment governance is the system through which objectives are set, authority is allocated, decisions are monitored and accountability is enforced. A typical governance structure includes the asset owner or investment committee, management, portfolio managers, risk and compliance functions, operations, custodian and external service providers.

A fiduciary must act loyally and prudently for beneficiaries or clients, place their interests ahead of personal interests, exercise care and competence, diversify appropriately, control costs and disclose material conflicts. The exact legal standard varies by jurisdiction and mandate, but the core principle is avoidance of self-dealing and faithful execution of the client’s purpose.

Stewardship extends fiduciary practice beyond buying and selling. Institutional investors may vote shares, engage issuers, monitor climate and social risks, and participate in collaborative initiatives. Stewardship should be linked to investment objectives, documented, proportionate and reported transparently. It should not become a vehicle for undisclosed political preferences or publicity detached from financial and beneficiary interests.

Governance controls

Effective controls include:

• written investment policy statements;
• delegated authorities and segregation of duties;
• pre-trade and post-trade compliance testing;
• independent valuation and performance verification;
• best-execution and transaction-cost monitoring;
• conflict-of-interest registers and personal-account dealing rules;
• gifts, entertainment and outside-business-activity policies;
• whistleblowing and escalation channels;
• cybersecurity, access and data-retention controls;
• periodic manager, custodian and benchmark reviews.

A control is credible only when breaches are recorded, investigated, remediated and reported to an independent oversight body.

7. Professional Ethics and Conflicts

Professional ethics concerns conduct that may be legal yet inconsistent with client trust. Core obligations include integrity, competence, diligence, fair dealing, confidentiality, independence and accurate representation. Examples of misconduct include presenting gross returns as if they were net, selectively showing successful accounts, changing benchmarks after poor performance, front-running client orders, allocating attractive trades preferentially, or concealing valuation uncertainty.

Conflicts are not eliminated merely by disclosure. A robust response follows a hierarchy: identify the conflict, assess its materiality, avoid it where possible, manage it through independent review and controls, disclose it clearly, and decline or withdraw when impartiality cannot be preserved. Compensation structures should be reviewed for incentives to increase leverage, turnover, valuation optimism or short-term risk.

Performance advertising deserves particular care. Composite construction should follow documented inclusion rules, and simulated or back-tested results must be clearly labelled. Survivorship bias, backfill bias and look-ahead bias can materially overstate apparent skill.

8. Technology, Data and Contemporary Practice

Automated performance systems improve timeliness but do not remove judgement. Data lineage should show the source, transformation and approval of every material input. Artificial intelligence may assist in anomaly detection, document review and factor analysis, but models can reproduce biased data, infer confidential information or generate unsupported explanations. Human accountability, model validation, access controls and explainability remain necessary.

Contemporary evaluation increasingly incorporates climate transition risk, biodiversity exposure, private assets, derivatives and operational resilience. These additions create measurement challenges: private-asset appraisal values may be smoothed, ESG ratings may disagree, and derivative exposures may be nonlinear. Analysts should report uncertainty rather than imply false precision.

9. Illustrative Case: The Meridian Balanced Fund

Background

Meridian Balanced Fund has a 60% equity and 40% bond strategic allocation. During a year of falling bond yields and strong technology shares, the fund returned 11.2% net of fees. Its policy benchmark returned 10.0%. The investment committee initially concluded that the manager had generated 1.2% alpha.

Problem

A review found that the fund had temporarily increased equities to 72%, held a technology overweight, and used derivatives to reduce reported cash drag. The benchmark had not been rebalanced as required by the policy. In addition, the performance report used gross returns for the fund but net returns for a peer comparison.

Analysis

Once the policy benchmark was correctly rebalanced and all costs included, relative performance fell to 0.4%. Attribution showed that most of the excess return came from equity allocation and technology exposure, not broad security selection. The derivative position also created collateral and liquidity risk that was not prominent in the original report. The manager had not breached a numerical limit, but the temporary allocation was outside the documented decision authority.

Approach and outcome

The committee required a corrected report, revised delegated-authority limits, independent derivative-risk monitoring and a conflict review concerning the performance fee. It retained the manager but changed appraisal criteria to include mandate compliance, downside behaviour, net-of-cost results and quality of attribution. The fund also adopted a benchmark governance policy requiring changes to be approved before the evaluation period.

Learning points

High return is not equivalent to skill. Benchmark integrity, net performance, attribution, authority limits and ethical disclosure are jointly necessary for a defensible evaluation. Governance often determines whether quantitative evidence is interpreted honestly.

Challenges and Limitations

No performance measure is universally superior. Short samples produce unstable estimates of volatility and beta. Non-normal returns undermine simple ratio interpretations. Illiquid assets may show artificially low volatility and inflated Sharpe ratios. Factor models can misclassify exposures, and a positive alpha may disappear after fees, trading costs or multiple-testing adjustments. Manager selection is also affected by survivorship and selection bias. These limitations require triangulation: use several measures, examine drawdowns and exposures, validate data, and combine quantitative evidence with process and governance assessment.

Best-Practice Evaluation Process

1. Confirm mandate, objective, horizon and benchmark before measurement.
2. Reconcile holdings, cash flows, prices, fees and corporate actions.
3. Calculate time-weighted and, where relevant, money-weighted returns.
4. Report Sharpe, Treynor, alpha and information ratio with assumptions.
5. Analyse factor exposures, drawdowns and liquidity.
6. Reconcile attribution to total active return.
7. Review compliance, conflicts, stewardship and operational controls.
8. Assess persistence across independent periods and market regimes.
9. Present uncertainty, residuals and limitations to the oversight body.
10. Document decisions and remedial actions.

Chapter Summary

Performance evaluation combines return measurement, risk adjustment, attribution and professional judgement. Sharpe uses total volatility; Treynor uses systematic risk; Jensen’s alpha compares return with a factor-model prediction; and the information ratio assesses active return relative to tracking error. Their usefulness depends on consistent data, valid benchmarks and appropriate assumptions.

Attribution separates allocation, selection and interaction effects but does not by itself prove skill. Manager appraisal must consider net returns, costs, risk, consistency, mandate compliance and the possibility of hidden exposures. Governance translates investment objectives into authority, oversight and controls. Fiduciary duty requires loyalty, prudence, care, competence and conflict management. Ethical performance reporting must be accurate, fair and transparent. Stewardship, technology and AI can strengthen investment practice only when supported by accountability, data governance and human oversight.

Key Terms

Fiduciary duty: Obligation to act loyally and prudently for the client or beneficiary.

Sharpe ratio: Excess return divided by total return volatility.

Treynor ratio: Excess return divided by portfolio beta.

Jensen’s alpha: Return above that predicted by a specified asset-pricing model.

Information ratio: Active return divided by tracking error.

Benchmark: Reference portfolio used to evaluate performance and risk.

Tracking error: Standard deviation of active returns.

Attribution: Decomposition of performance into identifiable sources.

Allocation effect: Active return arising from portfolio weights relative to benchmark weights.

Selection effect: Active return arising from security choices within categories.

Interaction effect: Joint effect of allocation and selection.

Time-weighted return: Compounded subperiod return that removes external cash-flow timing effects.

Money-weighted return: Investor-specific return affected by the timing and size of cash flows.

Stewardship: Responsible ownership through monitoring, voting and engagement.

Conflict of interest: Situation in which personal or institutional interests may impair impartiality.

Style drift: Movement of a portfolio away from its stated investment style.

Tracking portfolio: Portfolio representing the benchmark opportunity set.

Composite: Group of portfolios managed under a consistent strategy for performance presentation.

Governance: Structure of decision rights, oversight, accountability and controls.

Review Questions

Short Answer Questions

1. Distinguish performance measurement, attribution and appraisal.
2. Why does the Sharpe ratio use total volatility?
3. State two characteristics of an appropriate benchmark.
4. What is tracking error?
5. Explain why disclosure alone may not resolve a conflict of interest.

Long Answer Questions

1. Compare the Sharpe, Treynor, Jensen alpha and information ratios, including their assumptions and limitations.
2. Explain how a governance framework supports fiduciary duty in an institutional investment organisation.
3. Describe the construction and interpretation of Brinson-style allocation, selection and interaction effects.
4. Discuss the principal sources of bias in manager performance reporting.
5. Explain how stewardship and ethical conduct should be incorporated into manager appraisal.

Analytical / Application Questions

1. A fund earns 9%, the risk-free rate is 3%, volatility is 8%, beta is 0.9, the market earns 8%, and tracking error is 4%. Calculate its Sharpe ratio, Treynor ratio, Jensen alpha and information ratio, and interpret the differences.
2. A manager outperforms because of a large unapproved sector overweight. Explain how attribution, governance review and fiduciary analysis should affect the manager’s evaluation.
3. Design a performance-reporting control system that prevents benchmark changes, gross/net inconsistency, survivorship bias and undisclosed conflicts.

Multiple Choice Questions with Answers

Question 1

Which ratio measures excess return per unit of total volatility?

A. Sharpe ratio

B. Treynor ratio

C. Information ratio

D. Jensen alpha

Correct Answer: A. Sharpe ratio

Question 2

A portfolio earns 12% and its benchmark earns 10%, with tracking error of 5%. What is its information ratio?

A. 0.20

B. 0.40

C. 2.50

D. 5.00

Correct Answer: B. 0.40

Question 3

The Treynor ratio uses which risk measure in its denominator?

A. Standard deviation

B. Tracking error

C. Beta

D. Maximum drawdown

Correct Answer: C. Beta

Question 4

In Jensen’s alpha, the benchmark-model expected return is subtracted from:

A. The risk-free rate

B. The portfolio return

C. The tracking error

D. The portfolio beta

Correct Answer: B. The portfolio return

Question 5

Which benchmark characteristic means that its rules are known and clearly defined?

A. Unambiguous

B. Survivorship-free

C. Leveraged

D. Subjective

Correct Answer: A. Unambiguous

Question 6

Which attribution effect reflects the result of overweighting or underweighting a category relative to the benchmark?

A. Selection effect

B. Allocation effect

C. Fee effect

D. Valuation effect

Correct Answer: B. Allocation effect

Question 7

Which return measure is generally most suitable for evaluating a manager when the investor makes external cash flows?

A. Money-weighted return

B. Time-weighted return

C. Coupon return

D. Dividend yield

Correct Answer: B. Time-weighted return

Question 8

Which action is most consistent with fiduciary duty?

A. Changing the benchmark after observing poor results

B. Prioritising the manager’s personal account

C. Acting prudently in the client’s best interest

D. Hiding material conflicts from beneficiaries

Correct Answer: C. Acting prudently in the client’s best interest

Question 9

A positive Jensen alpha necessarily proves that a manager has repeatable skill.

A. True in every evaluation period

B. True when beta is positive

C. False because alpha depends on model, benchmark and sample

D. False because alpha is always negative

Correct Answer: C. False because alpha depends on model, benchmark and sample

Question 10

Which control most directly addresses unauthorised personal trading by employees?

A. Personal-account dealing policy and monitoring

B. Increasing portfolio leverage

C. Changing the strategic benchmark

D. Removing independent valuation

Correct Answer: A. Personal-account dealing policy and monitoring