Performance attribution and appraisal
In this chapter: Brinson attribution (allocation, selection, interaction) · Multi-period geometric attribution · Sharpe, Treynor, Jensen alpha, IR · M² (M-squared) · Style analysis
Attribution decomposes active return into sources. Appraisal ratios measure risk-adjusted returns. Together they evaluate manager skill.
**Brinson-Hood-Beebower (BHB) attribution**: For multi-asset portfolio: - Allocation effect: sum_i (w_p,i – w_b,i) × (R_b,i – R_b) - Selection effect: sum_i w_b,i × (R_p,i – R_b,i) - Interaction effect: sum_i (w_p,i – w_b,i) × (R_p,i – R_b,i) - Total active = R_p – R_b = Allocation + Selection + Interaction. BHB tells you: was return from picking right asset classes (allocation) or right securities within (selection)? **Sharpe ratio** = (R_p – R_f) / σ_p. Total risk basis. **Treynor ratio** = (R_p – R_f) / β_p. Systematic-risk basis. **Jensen alpha** = R_p – [R_f + β(R_m – R_f)]. CAPM-implied excess. **Information ratio** = α / TE. Active-management measure. **M² (M-squared)** = R_f + Sharpe × σ_m. Levers/de-levers portfolio to match benchmark vol.
Multi-period attribution: - Geometric linking handles compounding correctly. - Arithmetic linking sums attributions but misses compounding effects over long periods. GIPS-compliant performance: - Time-weighted returns mandatory for composites. - Quarterly+ when significant external cash flows. - Currency: report in firm's base currency or composite definition. - Disclose: gross vs net, fees, benchmark. Factor-based attribution (BARRA-style): - Decompose into factor exposures (size, value, momentum, etc.) + specific. - Says "30bps from value tilt, 20bps from momentum, –10bps from size, 60bps stock-specific."
L3 essay traps: - Confusing TWR with money-weighted (IRR). TWR for manager skill (timing-controlled), MWR for client experience. - Allocation vs selection effect signs flip with benchmark choice. Define benchmark first. - Style drift: manager claims value but factor analysis shows growth tilt. Affects attribution interpretation. Manager evaluation framework: 1. Performance: TWR vs benchmark over multiple periods. 2. Risk-adjusted: Sharpe, IR. 3. Style consistency. 4. Process integrity. 5. People (key-person risk). 6. Operations (compliance, ops infrastructure).
- GIPS Standards
- CFA Institute PM curriculum
- AMFI Disclosure Norms
- Mixing TWR and MWR — different purposes, different stories.
- Comparing manager vs wrong benchmark (style mismatch).
- Annualising short-period attribution without geometric linking.
Frequently asked
Sharpe vs Treynor — which to use?
Why does BHB allocation effect depend on benchmark return?
Practice questions
Click each question to reveal the answer and explanation.
Q 1Sharpe ratio uses:- (a)Beta
- (b)Total volatility (σ)
- (c)Tracking error
- (d)R²
- (a)Beta
- (b)Total volatility (σ)
- (c)Tracking error
- (d)R²
Q 2Treynor ratio uses:- (a)σ
- (b)Beta
- (c)TE
- (d)Alpha
- (a)σ
- (b)Beta
- (c)TE
- (d)Alpha
Q 3Information ratio uses:- (a)σ
- (b)β
- (c)Tracking error
- (d)Variance
- (a)σ
- (b)β
- (c)Tracking error
- (d)Variance
Q 4BHB attribution decomposes active return into:- (a)α and β
- (b)Allocation, selection, interaction
- (c)Sharpe components
- (d)Risk components
- (a)α and β
- (b)Allocation, selection, interaction
- (c)Sharpe components
- (d)Risk components
Q 5TWR is preferred over MWR for:- (a)Client returns
- (b)Manager skill evaluation
- (c)Tax reporting
- (d)IRR calculation
- (a)Client returns
- (b)Manager skill evaluation
- (c)Tax reporting
- (d)IRR calculation