← Back to Index

Chapter 16: Portfolio Performance Measurement and Evaluation

Module 5 — Return measures, risk measures, risk-adjusted return measures (Sharpe, Treynor, Jensen's Alpha), benchmarking, performance attribution

16.1 Parameters Defining Performance: Risk & Return

The key pitfall in performance measurement is focusing on return alone with little regard for the risk taken to earn it. Proper measurement must recognise both.

16.2 Rate of Return Measures

16.2.1 Holding Period Return (HPR)

Equation 1: HPR = (E − B) / B
Equation 2 (with income): HPR = (I + (E − B)) / B

where E = Ending Value, B = Beginning Value, I = Income (dividends/interest)

Worked Example — HPR

Portfolio value rises from Rs. 1,00,000 (1 Apr 2018) to Rs. 1,20,000 (31 Mar 2019):

HPR = (1,20,000 − 1,00,000) / 1,00,000 = 20%

If the investor also received Rs. 5,000 as dividend/interest income:

HPR = (5,000 + (1,20,000 − 1,00,000)) / 1,00,000 = 25%

16.2.2 Time-Weighted (TWRR) vs Money-Weighted Rate of Return (MWRR)

MWRR = the IRR of the cash flows; depends on the timing of cash flows. TWRR = the compound rate of growth, computed by linking sub-period returns; it is unaffected by cash flow timing (essentially the geometric mean return).

AspectMWRR (= IRR)TWRR
Sensitive to cash flow timing?YesNo
Used byWealth manager / investor reporting (money flows multiple times)SEBI-mandated for discretionary PMS managers (preceding 3 years) for uniform comparison
ReflectsWhat the investor actually experiencedThe fund manager's underlying skill, isolated from cash-flow timing

Worked Example — MWRR (IRR)

Investor contributes at the start of each year: 2015–19 = Rs. 5,000 / 10,000 / 15,000 / 20,000 / 25,000. Portfolio value grows to Rs. 1,05,920.31 by end-2019.

5000 = −10000/(1+r)¹ − 15000/(1+r)² − 20000/(1+r)³ − 25000/(1+r)⁴ + 105920.31/(1+r)⁵

Solving: IRR (MWRR) = 15.15%

Worked Example — TWRR (Step by step)

YearValue at start of year (Rs.)Value at end of period (Rs.)Sub-period returnWealth Relative (1+r)
20155,000.004,750.00−5.00%0.9500
201614,750.0012,508.00−15.20%0.8480
201727,508.0029,736.158.10%1.0810
201849,736.1565,030.0130.75%1.3075
201990,030.011,05,920.3117.65%1.1765

Cumulative Wealth Relative = 0.95 × 0.848 × 1.081 × 1.3075 × 1.1765 = 1.3396

TWRR = (1.3396)(1/5) − 1 = 6.021% (annualized; equals geometric return)

16.2.3 Geometric Mean Return (GMR) vs Arithmetic Mean Return (AMR)

AMR = simple average of period returns  |  GMR = (Product of wealth relatives)(1/n) − 1

Worked Example — AMR vs GMR

Choice 1: −50% (Yr 1), +100% (Yr 2). AMR = (−50%+100%)/2 = 25%; GMR = ((1−0.5)(1+1.0))1/2 − 1 = 0% — Rs. 100,000 returns to Rs. 100,000.

Choice 2: +10% (Yr 1), +10% (Yr 2). AMR = (10%+10%)/2 = 10%; GMR = ((1.1)(1.1))1/2 − 1 = 10% — grows to Rs. 1,21,000.

Despite a lower AMR, Choice 2 leaves the investor better off — GMR is the correct measure for long-run accumulation; AMR is the better single-period forecast.

For a one-year period, AMR = GMR. Over multiple periods, GMR ≤ AMR always (equality only when all yearly returns are identical).

16.2.4 Gross vs Net Return

Gross Return = (Gross Value − Capital Contribution) / Capital Contribution
Net Return = (Net Value − Capital Contribution) / Capital Contribution

Worked Example — Gross vs Net Return (Rs. 100 lacs portfolio, 1 yr)

Capital Contribution / AUM1,00,00,000
Profit @ 20%20,00,000
Gross Value of Portfolio1,20,00,000
Less: Other Expenses (0.50%)60,000
Less: Fixed Mgmt Fee (1.5% of avg of opening/closing)1,65,000
Hurdle (Required) Value @ 10% = 1,00,00,000 × 1.101,10,00,000
Less: Performance Fee (20% of profit over hurdle)1,55,000
Less: Exit Load (2%)2,32,400
Net Value of Portfolio1,13,87,600

Gross Return = (1,20,00,000 − 1,00,00,000)/1,00,00,000 = 20%

Net Return = (1,13,87,600 − 1,00,00,000)/1,00,00,000 = 13.88%

16.2.5 Pre-tax vs Post-tax Return

Post-tax return = Pre-tax return × (1 − tax rate)  |  Pre-tax return = Post-tax return / (1 − tax rate)

Worked Example

Pre-tax return = 5%, capital gains tax = 15%: Post-tax return = 5% × (1 − 0.15) = 4.25%

16.2.6 Compounded Annual Growth Rate (CAGR)

CAGR = (A / P)(1/t) − 1   [A = closing wealth, P = opening wealth, t = years]

Worked Example — CAGR

Rs. 1,00,000 grows to Rs. 1,33,960 over 5 years (returns: −5%, −15.2%, 8.1%, 30.75%, 17.65%):

CAGR = (1,33,960 / 1,00,000)(1/5) − 1 = 6.02%

16.2.7 Annualizing Returns

Worked Example — 3-Year Annualized Return

Returns: 2017 = 15.5%, 2018 = 9.5%, 2019 = −6.9%. Rs. 100 grows: 100 → 115.5 → 126.47 → 117.7459

Compounded Annualised Return = (117.7459/100)(1/3) − 1 = 5.60%

16.2.8 Cash-Drag Adjusted Return

Worked Example — Cash Drag

Rs. 100 lacs invested; manager deploys Rs. 75 lacs in equity (return Rs. 7.5 lacs) and keeps Rs. 25 lacs in liquid funds (4% return).

Ignoring cash drag: Return = 7.5/75 = 10%

Adjusting for cash drag: (10% × 75 lacs) + (4% × 25 lacs) = Rs. 8.5 lacs ⇒ 8.5% on Rs. 100 lacs — the figure that must actually be reported.

16.2.9 Alpha and Beta Return (CAPM Decomposition)

Required Return (CAPM) = Rf + β (Rm − Rf)
Alpha (Jensen's Alpha) = Portfolio Return − [Rf + β(Rm − Rf)]

Rf = risk-free return; β = market beta; Rm = return on market portfolio. Beta return rewards bearing market (systematic) risk; Alpha return rewards bearing non-market (unsystematic) risk via manager skill.

Worked Example — Decomposing Portfolio Return into Rf + Beta + Alpha

Portfolio return = 25%, Market return = 15%, Beta = 1.5, T-bill (risk-free) yield = 5%

Risk-free return = 5%

Beta return = 1.5 × (15% − 5%) = 15%

Jensen's Alpha = 25% − {5% + 1.5 × (15% − 5%)} = 25% − 20% = 5%

(Note: simple "excess return over benchmark" = 25% − 15% = 10% — some professionals call this "alpha," distinguishing it from Jensen's alpha.)

16.2.10 Portfolio Return (Weighted Average)

Portfolio Return = Σ (Weighti × Returni)

Worked Example

SecurityReturnWeight
A15%30%
B10%20%
C12%20%
D18%30%

Portfolio Return = (15%×30%)+(10%×20%)+(12%×20%)+(18%×30%) = 14.30%

16.3 Risk Measures

Risk TypeDescription / Measure
Total riskVariability in expected return — measured by Variance / Standard Deviation
Downside riskLosses or worse-than-expected outcomes — measured by Semi-variance/Semi-SD or Target semi-variance
Market riskArises from market-wide price fluctuations affecting all investments; cannot be diversified away (can be hedged); measured by Beta
Tracking errorStandard deviation of the difference between portfolio and benchmark total returns; lower = closer to benchmark; calculated against the Total Returns Index
Systematic riskRisk from common factors — interest rates, exchange rates, commodity prices; cannot be diversified away (can be hedged); measured by Beta
Unsystematic riskSector/company-specific; CAN be diversified away; alpha return rewards bearing this risk
Liquidity riskUncertainty from difficulty of converting an asset to cash near its economic worth (e.g., T-bills = low; art/illiquid assets = high)
Credit riskRisk that a borrower fails to repay on time — arises in debt instruments
Portfolio risk ≠ weighted average of individual security risks (except under perfect positive correlation, practically impossible). Portfolio risk depends on weights, individual SDs, AND correlation across investments. Covariance = Correlation × SD1 × SD2; poorer correlation ⇒ lower portfolio risk.

Worked Example — Portfolio Beta (weighted average)

Stock A: β = 1.2 (60% weight); Stock B: β = 1.1 (40% weight)

βp = 0.60 × 1.2 + 0.40 × 1.1 = 1.16

β > 1 ⇒ more volatile than benchmark; β < 1 ⇒ less volatile than benchmark.

16.4 Risk-Adjusted Return Measures

Risk-Adjusted
Return Measures
Sharpe Ratio
(total risk — SD)
Treynor Ratio
(systematic risk — Beta)
Sortino Ratio
(downside risk — semi-SD)
Information Ratio
(active risk — tracking error)
M² Measure
(risk-equalised comparison)

16.4.1 Sharpe Ratio (Reward to Variability)

Sharpe Ratio = (Rp − Rf) / σp

Worked Example — Sharpe Ratio

Annualized Return Rp = 10.50%, Risk-free Rate Rf = 5.50%, SD σp = 6.50%

Sharpe Ratio = (10.50% − 5.50%) / 6.50% = 0.7692

Interpretation: 0.7692 percentage points of excess return generated for each percentage point of standard deviation (total risk).

Limitation: Ignores the diversification potential of the portfolio (uses total risk, not just systematic risk).

16.4.2 Treynor Ratio

Treynor Ratio = (Rp − Rf) / βp

Worked Example — Treynor Ratio

Using the same Rp = 10.50%, Rf = 5.50%, and Beta = 1:

Treynor Ratio = (10.50% − 5.50%) / 1 = 0.05

Interpretation: 0.05 percentage points of excess return per unit of systematic risk.

16.4.3 Sharpe vs Treynor — Comparison

AspectSharpe RatioTreynor Ratio
Risk measure usedStandard deviation (total risk)Beta (systematic risk only)
Best suited forInvestors who have NOT achieved adequate diversification (evaluating standalone/mutually exclusive portfolios)Investors with well-diversified wealth, where unsystematic risk is minimal
Ranking vs each otherIdentical for a fully-diversified portfolio (total risk = systematic risk); for a poorly-diversified portfolio, Treynor ranking can exceed Sharpe ranking because Treynor ignores unsystematic risk

16.4.4 Sortino Ratio

Sortino Ratio = (Rp − Rf) / Semi-Standard Deviation of the portfolio

Adjusts excess return for downside risk only — appeals to investors who define risk as "chance of losing money" rather than overall variability. Higher Sortino = superior performance.

16.4.5 Information Ratio (Appraisal Ratio)

IR = (Rp − Rb) / Stdev(p − b)

Rb = benchmark return; Stdev(p−b) = standard deviation of (portfolio − benchmark) differences = tracking error (active risk). Numerator = "active return" (manager skill vs. benchmark). Helps determine whether observed alpha is due to skill or chance.

16.4.6 Modigliani & Modigliani (M²) Measure

Adjusts the portfolio's risk (via levering/de-levering with T-bills) to match the market portfolio's risk, then compares the resulting return with the market return.

Worked Example — M²

Portfolio return = 35%, σportfolio = 42%; Market return = 28%, σmarket = 30%; T-bill rate = 6%

Weight in Portfolio (to match market vol) = 30/42 = 0.714; Weight in T-bills = 1 − 0.714 = 0.286

r* = (0.714 × 0.35) + (0.286 × 0.06) = 26.7%

Since 26.7% < market's 28% (by 1.3%), the managed portfolio underperformed on a risk-adjusted basis.

16.5 Benchmarking & Peer Group Analysis

GIPS defines a benchmark as "an independent rate of return (or hurdle rate) forming an objective test of effective implementation of investment strategy."

Criteria for a Good Benchmark

  • Constituents and weights are clearly defined
  • Investable (passive exposure possible)
  • Consistent with the portfolio's investment approach/style (e.g., blue-chip vs. blue-chip; value vs. value)
  • Same risk-return profile as the portfolio
  • Performance is measurable

Customized benchmarks are used when market indices don't fit a manager's style — advantage: a valid fit; disadvantage: high construction/maintenance cost.

Common benchmarking errors: choosing a non-representative benchmark, an un-investable/hard-to-replicate benchmark, one with frequently-changing components, or one lacking available (Total Return Index) data.

Managers' Universe / Peer Group Analysis: ranks portfolios with similar investment characteristics and risk-return profiles on a risk-adjusted return measure — the median portfolio serves as the yardstick.

16.6 Performance Attribution Analysis

Attribution dissects portfolio return into: (a) return driven by the benchmark and (b) the differential return — then determines whether the differential was driven by manager skill or random factors (luck).

Asset & Sector Allocation

Differential return from being overweight outperforming sectors or underweight underperforming sectors/categories vs. the benchmark. The "allocation effect" quantifies this.

Selection Effect

Differential return from picking winning securities within a sector/category, or avoiding poor performers, relative to the benchmark.

Market Timing vs Selectivity

Market timing = anticipating market moves and shifting exposure accordingly; selectivity = picking securities that generate extra returns. Most studies show timing rarely adds value; results on selectivity are mixed.

Net Selectivity (Fama)

An absolute measure: annualised fund return minus risk-free yield, further reduced by (standardised expected market premium × total portfolio risk). Shows the excess return earned that could not have been earned by simply holding the market portfolio.

16.6.5 Local vs Foreign Currency Returns (Currency Risk)

Worked Example — Currency-Adjusted Return

Indian investor invests Rs. 50 lacs in a US equity fund at ₹70/$. Fund earns 15%; INR appreciates to ₹65/$.

Capital in USD = 50,00,000 / 70 = $71,428.57

Value after 15% gain = 71,428.57 × 1.15 = $82,142.86

Value in INR = 82,142.86 × 65 = Rs. 53,39,285.62

Rupee return = (53,39,285.62 − 50,00,000)/50,00,000 ≈ 6.79% (not 15%) — the gap is explained by INR's appreciation from Rs. 70/$ to Rs. 65/$.

Note: in most cases INR depreciates against major currencies, which then ADDS to the rupee-return (e.g., part of gold's INR returns reflect INR depreciation, not just LBMA gold price moves).

Key Takeaways

  • HPR = (E−B)/B or (I+(E−B))/B; MWRR = IRR (sensitive to cash-flow timing); TWRR links sub-period wealth relatives and equals geometric return (SEBI mandates TWRR for PMS managers).
  • GMR ≤ AMR for multi-period returns; GMR is the right measure for long-run accumulation; AMR is the best single-period forecast.
  • CAGR = (A/P)^(1/t) − 1; Post-tax return = Pre-tax return × (1 − tax rate); cash drag must be incorporated in reported returns.
  • CAPM: Required Return = Rf + β(Rm−Rf); Jensen's Alpha = Actual Return − Required Return; Beta return rewards systematic risk, Alpha rewards manager skill/unsystematic risk.
  • Risk types: total (SD), downside (semi-SD), systematic/market (Beta), unsystematic (diversifiable), tracking error, liquidity, credit risk.
  • Sharpe = (Rp−Rf)/SD (total risk; best for non-diversified portfolios); Treynor = (Rp−Rf)/Beta (systematic risk; best for well-diversified portfolios); Sortino uses semi-SD (downside risk); Information Ratio uses tracking error.
  • A good benchmark is well-defined, investable, style-consistent, risk-matched and measurable; peer group/managers' universe analysis ranks similar portfolios.
  • Performance attribution splits return into benchmark-driven vs. differential (allocation effect + selection effect), and assesses skill vs. luck (incl. market timing vs. selectivity, net selectivity, and currency effects).

Self-Test Quiz

1. A portfolio has an annualized return of 10.50%, a risk-free rate of 5.50%, and an annualized standard deviation of 6.50%. What is its Sharpe Ratio?

2. Using the same portfolio (Rp = 10.50%, Rf = 5.50%) but with Beta = 1, what is the Treynor Ratio?

3. A portfolio earned 25% return when the market returned 15%, the portfolio's beta is 1.5, and the risk-free rate is 5%. What is the Jensen's Alpha?

4. Which risk-adjusted return measure is most appropriate for an investor whose overall wealth is already well-diversified (so unsystematic risk is minimal)?

5. An investment grows from Rs. 1,00,000 to Rs. 1,33,960 over 5 years. What is its CAGR (approximately)?