Key Takeaways — Chapter 14: Market Risk and the Basel Framework in Practice
Core Definitions
Market risk is the risk of losses arising from adverse movements in market prices — interest rates, equity prices, exchange rates, commodity prices, credit spreads, and implied volatility.
Trading book positions are held for short-term trading and marked to market daily. Market risk capital (FRTB) applies to the trading book.
Banking book positions are held to maturity. Interest Rate Risk in the Banking Book (IRRBB) governs rate risk in the banking book — separately from market risk capital.
Basel Market Risk Capital: The Evolution
| Regime | Period | Key Metric | Confidence | Holding Period |
|---|---|---|---|---|
| 1996 Market Risk Amendment (IMA) | 1996–2023 | VaR | 99% | 10 days |
| FRTB — Standardized Approach (SA) | 2025+ | Sensitivity-based | N/A | Risk-class specific |
| FRTB — Internal Models Approach (IMA) | 2025+ | Expected Shortfall | 97.5% | Liquidity-horizon adjusted |
Value at Risk (VaR): Three Approaches
| Approach | Method | Key Assumption | Strength | Limitation |
|---|---|---|---|---|
| Historical Simulation | Replay actual historical returns | None (non-parametric) | Captures actual fat tails | Depends entirely on historical window |
| Parametric (Delta-Normal) | σ × z × √t | Normal distribution | Fast; analytically tractable | Fails for fat tails and options |
| Monte Carlo | Random scenario generation | Distribution specified | Handles non-linearities | Computationally intensive; model-dependent |
VaR formula (parametric, 1-day): VaR = z₀.₉₉ × σ_portfolio = 2.326 × σ
10-day scaling: VaR₁₀ = VaR₁ × √10 (square root of time; assumes i.i.d. returns)
Expected Shortfall (ES) vs. VaR
| Metric | VaR (99%) | ES (97.5%) |
|---|---|---|
| Definition | Maximum loss not exceeded with 99% probability | Average loss in worst 2.5% of scenarios |
| Tail information | None — blind beyond the threshold | Captures average magnitude of tail losses |
| Basel use | Pre-FRTB IMA | FRTB IMA |
| Regulatory rationale | Established standard | More risk-sensitive; addresses VaR's tail blindness |
FRTB Key Concepts
P&L Attribution (PLA) Test: Each trading desk must demonstrate that its risk factor model explains actual trading P&L. Desks that fail PLA must use the SA — typically a higher capital charge.
Non-Modellable Risk Factors (NMRFs): Risk factors with insufficient market price observations. NMRF capital = stress scenario loss (not ES from history). Can significantly increase capital for illiquid positions.
Liquidity Horizons: Capital charges are scaled to the time needed to exit/hedge positions: - 10 days: liquid equity, G10 FX, G10 interest rates - 20 days: investment grade credit spreads, non-G10 FX - 40 days: high yield credit, investment grade corporates - 60 days: equity volatility, EM FX - 120 days: illiquid credit, physical commodity
FRTB Standardized Approach (SA): Sensitivity-based method using delta (linear risk), vega (volatility risk), and curvature (option non-linearity) inputs. More risk-sensitive than pre-FRTB SA.
VaR Backtesting: Basel Traffic Light System
| Breaches in 250 days | Zone | Implication |
|---|---|---|
| 0–4 | Green | Normal; no additional capital |
| 5–9 | Yellow | Warning; potential capital multiplier increase |
| 10+ | Red | Model concerns; regulatory attention; capital multiplier likely increased |
IRRBB: Key Concepts
Economic Value of Equity (EVE): Net present value of all bank cash flows. EVE falls when rates rise (for a bank with more long-term assets than liabilities).
Net Interest Income (NII): Sensitivity of interest income to rate changes. A bank funded by variable-rate deposits and holding fixed-rate loans will see NII fall if rates rise.
Standard shocks (Basel IRRBB): Parallel ±200bp, short rates ±250bp, flattener, steepener. Supervisors compare EVE sensitivities across institutions.
IRRBB capital: Not a Pillar 1 charge — managed through Pillar 2 / ICAAP. PRA SS31/15 (UK) and ECB guidance require IRRBB capital commensurate with exposure.
Market Risk Technology Stack
Pricing Systems (Murex / Calypso / Bloomberg)
↓ Mark-to-market values + Greeks
Risk Calculation Engine (MSCI / Axioma / Numerix)
↓ VaR / ES / sensitivity measures
Risk Data Warehouse (Snowflake / cloud platform)
↓ Historical P&L / position data / backtesting
Regulatory Reporting Layer (FRR / FRTB platform)
↓ COREP C24–C32 templates + XBRL submission
COREP Market Risk Templates (Key)
| Template | Content |
|---|---|
| C 18.00 | RWA overview — market risk line items |
| C 24.00 | Market risk SA — total |
| C 25.00–28.00 | Interest rate, equity, FX, commodity (SA) |
| C 31.00 | FRTB SA — sensitivity-based method |
| C 32.00 | FRTB IMA — ES by liquidity horizon and risk class |
Common Practical Failure Modes
- Position data incompleteness: New instruments without risk factor mapping fall out of calculations
- Stale market data: Illiquid positions valued at outdated prices
- P&L attribution residuals: Risk engine P&L diverges from actual — FRTB PLA test failure
- Historical window gaps: Market data feed missing history needed for historical VaR
- NMRF identification failures: Non-modellable risk factors missed, undercharging capital
Practitioner Checklist: Market Risk Data Quality
- [ ] Position completeness: all trading desk positions in risk engine daily
- [ ] Market data coverage: all instrument risk factors have current and historical prices
- [ ] Stale price monitoring: alert on positions unchanged for >2 days (for actively traded instruments)
- [ ] P&L attribution: daily comparison of risk engine P&L to actual trading P&L; residuals < threshold
- [ ] VaR backtesting: monthly count of breaches; report to risk committee quarterly
- [ ] FRTB PLA test: calculated monthly for each IMA-approved desk
- [ ] NMRF register: quarterly update of non-modellable risk factor list
- [ ] IRRBB shocks: calculated monthly; reported to ALCO
- [ ] COREP market risk templates: validated against risk engine outputs before submission