Study guides / CFA L2: CFA Program Level II / Active portfolio management
Active portfolio management
CFA L2: CFA Program Level II · quick review
Measuring value added by active management, the information ratio versus the Sharpe ratio, the fundamental law of active management and its terms (transfer coefficient, information coefficient, breadth, active risk), and using them to select managers, set active risk and compare strategies.
The rules
- Value added is active return: portfolio return − benchmark return = Σ active weight × asset return, where active weights (portfolio weight − benchmark weight) sum to zero. It splits into asset allocation (active weights on asset classes × each class's benchmark return) and security selection (portfolio weights × each class's return above its benchmark).
- Sharpe ratio versus information ratio: the Sharpe ratio is excess return over cash per unit of total risk, and is unchanged by adding cash or leverage. The information ratio (IR) is active return per unit of active risk (tracking error), and is unchanged by scaling active weights up or down, for example by mixing the active portfolio with the benchmark. Ex post IR uses realized results; ex ante IR uses expectations.
- Combining with the benchmark: the best possible Sharpe ratio comes from the active portfolio with the highest IR; the squared Sharpe ratio of the best mix = benchmark Sharpe ratio squared + IR squared. The optimal active risk = (IR ÷ benchmark Sharpe ratio) × benchmark risk; taking more or less active risk than that lowers the Sharpe ratio.
- Manager selection: choose the manager with the highest IR, then reach the investor's target active risk by splitting money between that manager and the benchmark: weight on the active fund = target active risk ÷ the fund's active risk.
- Fundamental law: expected active return = TC × IC × √BR × active risk, so IR ≈ TC × IC × √BR. The information coefficient (IC) is the correlation between forecast and realized active returns: skill. Breadth (BR) is the number of independent decisions a year. The transfer coefficient (TC) is the correlation between optimal and actual active weights; constraints such as no short selling push it below 1.
- Constraints: with TC below 1, optimal active risk = TC × (unconstrained IR ÷ benchmark Sharpe ratio) × benchmark risk, and the best squared Sharpe ratio = benchmark Sharpe squared + TC squared × unconstrained IR squared. Ex post, roughly TC squared of the variance of active return comes from the forecasts and 1 − TC squared is noise from constraints.
- Market timing versus security selection: a market timer makes few independent bets a year, so breadth is low and each bet needs a high IC; for a timer, IC = 2 × proportion of correct calls − 1. A stock picker has high breadth but usually a low IC. The fundamental law lets the two be compared on equal terms.
- Evaluating strategy changes: a change raises the IR only if it raises the IC, adds genuinely independent breadth or relaxes constraints (raises TC). More frequent rebalancing or more securities help only if the extra bets are independent of the existing ones.
- Limitations: the IC must be estimated in advance and is uncertain; decisions are rarely fully independent, so breadth tends to be overstated; and the law ignores trading costs and assumes skill is the same across all decisions.
Formulas
- Active return: R_A = R_P − R_B = Σ Δw_i × R_i
- Return decomposition: R_A = Σ Δw_j × R_Bj (asset allocation) + Σ w_Pj × (R_Pj − R_Bj) (security selection)
- Sharpe ratio: SR = (R_P − R_F) ÷ σ_P
- Information ratio: IR = (R_P − R_B) ÷ σ(R_P − R_B)
- Best Sharpe ratio: SR_P^2 = SR_B^2 + IR^2
- Optimal active risk: σ_A* = (IR ÷ SR_B) × σ_B
- Fundamental law: E(R_A) = TC × IC × √BR × σ_A; IR ≈ TC × IC × √BR
- With constraints: σ_A* = TC × (IR* ÷ SR_B) × σ_B; SR_P^2 = SR_B^2 + TC^2 × IR*^2, IR* = unconstrained IR
- Market timing IC: IC = 2 × (proportion of correct calls) − 1
- Weight on the active fund: Weight = target active risk ÷ active risk of the fund
Common traps
- Adding cash leaves the Sharpe ratio unchanged; scaling active bets leaves the IR unchanged. Don't swap the two.
- Breadth counts independent decisions, not securities; correlated bets add little.
- IR rises with the square root of breadth: doubling breadth raises IR by about 41%, not 100%.
- Taking more active risk than the optimum lowers the portfolio's Sharpe ratio even with a positive IR.
How it's tested
Vignettes give IC, breadth, TC and risk figures and ask for expected active return, the IR, optimal active risk, the best manager, or whether a change in strategy adds value.
Practise it
The full lesson, worked examples and practice questions are in Tickmark, where every answer choice is explained.
Open the full lesson