How do multifactor models connect to active risk and the information ratio in portfolio management?

This is where the PM curriculum gets practical. Let me build up from factors to the information ratio. ## Multifactor Return Decomposition A multifactor model expresses a portfolio's return as: **R_p = alpha + beta_1 x F_1 + beta_2 x F_2 + ... + beta_k x F_k + epsilon** Where F_i are factor returns (market, size, value, momentum, etc.), beta_i are the portfolio's factor exposures, alpha is the return unexplained by factors, and epsilon is idiosyncratic noise. ## Active Return and Active Risk **Active return** = R_portfolio - R_benchmark This can be decomposed into: - **Factor tilts:** Differences in factor exposures between the portfolio and benchmark, multiplied by factor returns - **Security selection:** Alpha from picking securities that outperform within each factor group **Active risk (tracking error)** = Standard deviation of active returns over time ## Worked Example: Ashford Capital vs. Russell 1000 **Ashford Capital** is an active US equity fund benchmarked to the Russell 1000. Over the past 3 years: | Metric | Value | |---|---| | Annualized portfolio return | 14.2% | | Annualized benchmark return | 11.8% | | Active return | 2.4% | | Tracking error (active risk) | 4.1% | **Information Ratio = Active Return / Active Risk = 2.4% / 4.1% = 0.585** Now decompose using a 3-factor model: | Factor | Portfolio Beta | Benchmark Beta | Active Exposure | Factor Return | Contribution to Active Return | |---|---|---|---|---|---| | Market | 1.05 | 1.00 | +0.05 | 11.8% | +0.59% | | Size (SMB) | 0.30 | 0.10 | +0.20 | 3.2% | +0.64% | | Value (HML) | 0.15 | 0.25 | -0.10 | 4.5% | -0.45% | | **Factor contribution** | | | | | **+0.78%** | | **Security selection (alpha)** | | | | | **+1.62%** | | **Total active return** | | | | | **+2.40%** | ## Interpreting the Results - Ashford's 2.4% active return came mostly from **security selection** (1.62%), not just factor bets - The size tilt (+0.64%) helped, but the value underweight (-0.45%) dragged - An IR of 0.585 is solid — generally, IR > 0.50 suggests genuine skill after accounting for risk ## The Fundamental Law of Active Management Grinold's fundamental law ties this together: **IR = IC x sqrt(BR)** Where: - **IC** (Information Coefficient) = Correlation between manager's forecasts and actual outcomes - **BR** (Breadth) = Number of independent investment decisions per year A manager with modest skill (IC = 0.05) making 200 independent bets per year: IR = 0.05 x sqrt(200) = 0.05 x 14.14 = **0.71** This explains why quantitative managers with many small positions can achieve high IRs despite modest per-bet skill. ## Exam Focus Expect questions that give you factor betas for portfolio and benchmark, ask you to decompose active return, and then compute the information ratio. Also know the limitations — the fundamental law assumes independent bets, which is rarely true in practice. For interactive factor model simulations and IR calculators, visit AcadiFi's CFA Level II Portfolio Management course.