How does factor-based risk decomposition work for market risk management?
FRM Part II covers risk decomposition using factors. I understand that total risk can be broken into systematic and specific components, but how does a factor-based approach work in practice for a trading portfolio?
Factor-based risk decomposition breaks portfolio risk into components driven by identifiable market factors (equity indices, interest rates, currencies, commodities, volatility, etc.) and a residual component. This is far more actionable than looking at total VaR alone.
The factor model framework:
Rp = Σ βᵢFᵢ + ε
Where:
- Rp = portfolio return
- Fᵢ = factor returns (e.g., S&P 500, 10Y Treasury, EUR/USD, VIX)
- βᵢ = portfolio sensitivity to factor i
- ε = idiosyncratic/residual return
Risk decomposition:
Total Variance = Factor Variance + Specific Variance
σ²p = β'Σ_F β + σ²_ε
Where Σ_F is the factor covariance matrix.
Practical example — Sentinel Trading Desk:
Sentinel runs a multi-asset portfolio. Factor risk decomposition reveals:
| Risk Factor | Factor VaR Contribution | % of Total VaR |
|---|---|---|
| Equity market (S&P 500) | $3.2M | 40% |
| Interest rates (10Y) | $1.6M | 20% |
| Credit spreads (IG index) | $1.2M | 15% |
| FX (EUR/USD, JPY/USD) | $0.8M | 10% |
| Volatility (VIX) | $0.4M | 5% |
| Idiosyncratic | $0.8M | 10% |
| Total Portfolio VaR | $8.0M | 100% |
Actionable insights:
- 40% of risk is equity market exposure — if Sentinel wants to reduce risk, hedging equity beta is the most impactful action
- Interest rate exposure is significant (20%) — duration hedging with futures would help
- Idiosyncratic risk is only 10% — the portfolio is driven mostly by systematic factors, not individual positions
Why factor decomposition > position-level decomposition:
| Approach | Advantage | Limitation |
|---|---|---|
| Position-level (CVaR) | Identifies risky positions | Doesn't show WHY they're risky |
| Factor-level | Shows underlying risk drivers | Requires factor model estimation |
| Combined | Best of both worlds | More complex to implement |
Common factor models:
- Equity: Fama-French factors (market, size, value, momentum, profitability)
- Fixed income: Key rate durations (2Y, 5Y, 10Y, 30Y), spread factors
- Multi-asset: Principal component analysis (PCA) to extract statistical factors
- Macro factors: GDP growth, inflation, credit conditions
Benefits for risk management:
- Stress testing: Shock individual factors and see the impact
- Hedge identification: Know exactly which factor exposure to hedge
- Limit setting: Set limits per factor (e.g., max $2M of equity factor VaR)
- Performance attribution: Attribute returns to factor exposures vs. alpha
- Concentration monitoring: Detect hidden factor concentrations
Exam tip: FRM Part II tests the framework (factor variance vs. specific variance), interpretation of factor risk reports, and how to use decomposition for hedging and limit-setting decisions.
Learn more about market risk on AcadiFi.
Master Part II with our FRM Course
64 lessons · 120+ hours· Expert instruction
Related Questions
Why is DV01 so much smaller than dollar duration if both are supposed to measure rate risk?
When should I stop using modified duration and switch to effective duration?
How should I think about the relationship between Macaulay duration and modified duration instead of memorizing two separate definitions?
Why do hedge calculations often use dollar duration or DV01 instead of just modified duration?
When should I prefer historical simulation VaR over delta-normal VaR?
Join the Discussion
Ask questions and get expert answers.