Factor-adjusted alpha is the return you generated that cannot be explained by known investment factors like market, value, size, or momentum. It isolates genuine stock-picking skill from factor exposures.
Beginner
What It Means
Regular alpha can be misleading - a “value fund” might show high alpha just from value stock exposure, not manager skill. Factor-adjusted alpha removes all known factor effects to reveal true skill.
Portfolio Example
- Your portfolio returned 16% this year
- S&P 500 returned 10% (6% apparent excess return)
- But your portfolio has heavy value exposure, and value beat growth by 4%
- Factor-Adjusted Alpha = 16% - (10% market + 4% value) = 2%
Your real skill added 2%, not the apparent 6%.
Why It Matters
Factor-adjusted alpha reveals whether a manager has genuine stock-picking ability or is simply riding factor exposures that could be obtained cheaply through factor ETFs.
Advanced
Mathematical Definition
Factor-Adjusted Alpha (α) = Rp - [Rf + Σ(βi × Fi)]
Fama-French Five-Factor Model:
α = Rp - [Rf + βM(RM-Rf) + βSMB×SMB + βHML×HML + βRMW×RMW + βCMA×CMA]
Where:
- RM-Rf = Market risk premium
- SMB = Small Minus Big (size)
- HML = High Minus Low (value)
- RMW = Robust Minus Weak (profitability)
- CMA = Conservative Minus Aggressive (investment)
Comparison Example
Two portfolios both returned 15% vs. market’s 10%:
| Portfolio | Jensen’s Alpha | Factor Exposures | Fama-French Alpha |
|---|
| Value Fund | 5% | High value, high size | 1.5% |
| Stock Picker | 5% | Neutral all factors | 4.8% |
Common Factor Models
| Model | Factors | Use Case |
|---|
| CAPM | Market | Basic alpha |
| Fama-French 3 | Market, Size, Value | Traditional |
| Carhart 4 | + Momentum | Include trends |
| Fama-French 5 | + Profitability, Investment | Current standard |
| Custom | Industry-specific factors | Specialized analysis |
Historical Context
Eugene Fama and Kenneth French (1993) developed the three-factor model showing that size and value explain returns beyond market beta. They expanded to five factors in 2015. This revolutionized how we evaluate active management.
What It Reveals
| Finding | Implication |
|---|
| High factor-adjusted alpha | Genuine skill, justifies fees |
| Low factor-adjusted alpha | Factor exposure, not skill |
| Negative factor-adjusted alpha | Destroying value vs. factors |
Data Requirements
| Requirement | Details |
|---|
| Minimum | 36 months for basic estimate |
| Preferred | 60+ months for stable estimates |
| Factor Data | Need factor portfolio returns |
| Statistical Test | t-statistic should exceed 2.0 |
Factor loadings (betas) are time-varying. Use rolling estimation and be cautious about conclusions from short periods.
Limitations
| Limitation | Description |
|---|
| Model Dependency | Different factor models give different alphas |
| Missing Factors | If true factors not in model, alpha still contaminated |
| Data Intensive | Requires factor return data and regression |
| Time-Varying | Factor exposures change over time |
Modern Extensions
Beyond Fama-French, sophisticated analysis includes:
- Betting Against Beta (BAB): Low-beta stock premium
- Quality Minus Junk (QMJ): Quality factor
- Liquidity Factor: Compensation for illiquidity
- Momentum: Trend persistence
Managers are now evaluated against 8-10 factor models.
