What is the Fundamental Review of the Trading Book (FRTB), and how does it change market risk capital?

The Fundamental Review of the Trading Book (FRTB), finalized under Basel III.5 (January 2019, implementation ongoing through 2026-2028), represents the most significant overhaul of market risk capital regulation since the 1996 Market Risk Amendment. For FRM Part II, you need to understand both the motivation and the new framework. **Why FRTB Was Needed:** The 2008 crisis exposed critical flaws in the existing market risk framework: 1. **Trading book / banking book boundary was gameable** — banks moved positions to whichever book required less capital 2. **VaR underestimated tail risk** — 99% 10-day VaR missed catastrophic losses 3. **Diversification benefits were overstated** — correlations spiked in crises, but models assumed stable correlations 4. **Internal models varied wildly** — two banks holding identical positions could report very different capital requirements **FRTB Key Changes:** **1. Stricter Trading Book / Banking Book Boundary** - Clear, rule-based criteria for which positions belong in the trading book - Internal transfers require regulatory approval and "switching" capital charges - Example: Clearmont Securities can no longer shift a losing structured credit position to the banking book to avoid mark-to-market losses **2. Expected Shortfall Replaces VaR** - 97.5% ES replaces 99% VaR for internal models approach - Captures tail risk that VaR ignores - Calculated over different liquidity horizons (10 days for liquid equities, up to 120 days for illiquid structured products) **3. The Standardized Approach — Sensitivities-Based Method (SBM)** This is the most tested FRTB component at FRM Part II. The SBM calculates capital using three components: **A. Delta Sensitivity** - How much does the portfolio value change for a unit move in each risk factor? - Example: Hawthorne Trading holds equity positions. Delta sensitivities to each stock are calculated, then aggregated within risk buckets (e.g., large-cap financials, tech) using prescribed correlations, then aggregated across buckets. **B. Vega Sensitivity** - How much does the portfolio value change for a unit move in implied volatility? - Relevant for all positions with optionality **C. Curvature Sensitivity** - Captures non-linear risk that delta and vega miss - Calculated by repricing under upward and downward shocks and comparing to the delta prediction **Capital = f(Delta) + f(Vega) + f(Curvature) + Default Risk Charge + Residual Risk Add-on** **Worked SBM Example (Simplified):** Allerton Capital has equity delta sensitivities in the "Banking" risk bucket: - Weighted sensitivity to Northgate Financial: +$4.2M - Weighted sensitivity to Granite Trust: -$2.8M - Weighted sensitivity to Pacific Lending: +$1.5M Using prescribed intra-bucket correlation ρ = 0.15: K² = $4.2² + $2.8² + $1.5² + 2(0.15)($4.2)(-$2.8) + 2(0.15)($4.2)($1.5) + 2(0.15)(-$2.8)($1.5) K² = 17.64 + 7.84 + 2.25 + (-3.528) + (1.89) + (-1.26) K² = 24.84 K = **$4.98M** bucket-level capital charge **4. Internal Models Approach (IMA) — Desk-Level Approval** - Banks must obtain regulatory approval at the **individual desk level** (not firm-wide) - Each desk must pass P&L attribution tests and backtesting - Desks that fail revert to the standardized approach - This prevents banks from subsidizing poorly modeled desks with well-modeled ones **5. P&L Attribution Test** - Compares the risk model's predicted P&L to the actual desk P&L - Two metrics: mean ratio and variance ratio - If the model can't explain actual P&L, the desk loses IMA eligibility **6. Non-Modellable Risk Factors (NMRFs)** - Risk factors without sufficient real price data (fewer than 24 observations per year) must be capitalized with stress scenarios - This addresses the problem of banks modeling illiquid instruments with fabricated data **Exam Tip:** FRM Part II focuses on the conceptual differences between old and new frameworks, the SBM calculation methodology, and the desk-level approval process. Know the three sensitivity types (delta, vega, curvature), the concept of NMRFs, and why the P&L attribution test exists. Explore more FRTB topics in our FRM Part II course materials.