Fundamental Review of the Trading Book: FRTB for Bankers (IIBF RM 2026)

The Fundamental Review of the Trading Book is the single biggest change to market risk capital rules since Basel II, and every CAIIB Risk Management candidate needs a working grasp of it. FRTB rewrites how banks draw the line between the trading book and the banking book, and replaces the old Value-at-Risk capital engine with a stricter, more granular framework built on sensitivities, expected shortfall, and desk-level model approval. This article breaks the framework into three exam-ready pieces: the boundary rules, the Standardised Approach (SA), and the Internal Models Approach (IMA), with the practical detail you need to answer scenario-based questions confidently.

📊 Trading Book vs Banking Book: Where's the Boundary?

Before FRTB, banks had wide discretion to classify positions as "trading" or "banking" book, and that discretion was exploited during the 2008 crisis — instruments were moved between books to capture lower capital charges. FRTB tightens this boundary sharply. A position now qualifies for the trading book only if it is held with genuine trading intent: for short-term resale, to profit from price movements, to lock in arbitrage, or to hedge other trading book positions.

The framework also introduces a presumptive list. Instruments such as positions in the correlation trading portfolio, equities listed on a recognised exchange held for trading, and instruments arising from market-making generally belong in the trading book by default. Conversely, unlisted equity, real estate holdings, and positions intended to be held to maturity presumptively sit in the banking book. A bank can rebut the presumption, but only with clear evidence and prior supervisory approval — it is no longer a unilateral desk decision.

The strictest change is around switching. Once a position is assigned to a book, moving it to the other book is permitted only in extraordinary circumstances (such as a change in trading strategy approved by senior management), and any capital benefit from the switch is disallowed — the bank must hold the higher of the pre-switch and post-switch capital charge. This closes the regulatory arbitrage loophole that motivated the entire FRTB project. Candidates should also revisit related capital-adequacy concepts in Regulatory Capital And Capital Adequacy, since the trading book capital charge sits inside the same Pillar 1 structure.

📌 Remember: Under FRTB, a book-switch is not a way to save capital — the higher of the two charges always applies.

🧮 Standardised Approach (SA) Under FRTB

The revised Standardised Approach, sometimes called SA-TB, is no longer a simple fallback for banks that cannot build internal models — it is now a mandatory parallel calculation for every bank, run alongside IMA where IMA is used, and reported to the regulator. This is a deliberate design choice: it gives supervisors a consistent, comparable floor across the industry and prevents internal models from drifting too far below a standardised benchmark.

SA-TB rests on three components. The first and largest is the sensitivities-based method (SBM), which calculates a capital charge across seven prescribed risk classes — general interest rate risk (GIRR), credit spread risk (CSR) for non-securitisations, CSR for securitisations, equity risk, commodity risk, foreign exchange risk, and (for the correlation trading portfolio) a separate treatment. Each risk class requires the bank to compute delta, vega, and curvature sensitivities and aggregate them using regulator-prescribed risk weights and correlations, including three correlation scenarios (high, medium, low) to capture correlation breakdown in stress.

The second component is the Default Risk Charge (DRC), a jump-to-default measure that captures the risk of sudden issuer default separately from ordinary spread widening — something the old VaR-based approach handled poorly. The third is the Residual Risk Add-On (RRAO), a simple charge for exotic and hard-to-model instruments (like those with gap risk or correlation risk) that the SBM framework does not adequately price. Together these three pieces give supervisors a risk-sensitive yet mechanically consistent capital number that any examiner can replicate from a bank's own trade data.

📈 Internal Models Approach (IMA) and the P&L Attribution Test

The Internal Models Approach lets a bank use its own risk engine for capital, but FRTB makes IMA approval far harder to earn and far easier to lose than under Basel 2.5. Two structural changes stand out. First, the risk measure itself changes: IMA now runs on Expected Shortfall (ES) at a 97.5% confidence level instead of the old 99% Value-at-Risk. ES better captures tail risk because it averages losses beyond the confidence threshold rather than just marking the cut-off point — a lesson drawn directly from the crisis, where VaR understated tail losses.

Second, and more consequential for implementation, IMA approval is granted at the trading desk level, not bank-wide. Each desk must independently pass two tests every quarter: the P&L Attribution (PLA) test, which compares the desk's risk-theoretical P&L (from the risk model) against its hypothetical P&L (from actual pricing) and flags a desk if the two diverge beyond set thresholds; and a backtesting requirement using both 1-day 99% and 1-day 97.5% VaR exceptions. A desk that fails either test is pushed back onto the (more punitive) Standardised Approach until it can demonstrate model quality again — this is why market risk teams now watch PLA metrics as closely as they watch actual trading P&L. Candidates should cross-reference model validation and governance practices here, since PLA failure is fundamentally a model-validation trigger.

FRTB also introduces Non-Modellable Risk Factors (NMRFs) — any risk factor without enough real, verifiable price observations (the standard requires at least 24 observations in the prior 12 months, with no gap longer than a month) cannot sit inside the ES model at all. Instead it attracts a separate, punitive stressed capital charge calculated individually. This provision was designed specifically to stop banks from claiming diversification benefit on illiquid or thinly-traded instruments inside their VaR-style models, and it has proved to be one of the most capital-expensive parts of the entire framework in practice.

⚠️ Common Mistake: Candidates often assume IMA approval is bank-wide. It is desk-by-desk — one desk can run IMA while another sits on SA.

🌍 Implementation Status and Why It Matters for Indian Banks

Internationally, the Basel Committee finalised FRTB in January 2019, and jurisdictions have phased in the reporting and capital requirements at different speeds — several major regulators pushed their effective dates back multiple times because of the sheer data and systems burden the framework places on trading desks. For Indian banks, the practical starting point has been supervisory reporting and readiness: building the sensitivities data infrastructure, running SA-TB in parallel with existing VaR models, and preparing desk structures for eventual PLA testing, ahead of the framework becoming a binding capital requirement.

Why does this matter beyond the exam? FRTB capital charges are materially higher than the old VaR-based charges for most banks, particularly for less liquid credit and equity exposures, because of the DRC, RRAO, and NMRF add-ons. Treasury and market risk functions at Indian banks are already re-architecting their sensitivities calculation engines and trade data pipelines to be FRTB-ready, which is why RM examiners test this topic heavily — it is live regulatory change, not historical theory. You can track how this connects to the broader supervisory rationale in Why Do Banks Need Regulation, and the official position on Basel market risk standards is published by the Reserve Bank of India.

A quick side-by-side of the two approaches helps fix the differences before the exam:

💡 Exam Tip: If a question describes a desk failing quarterly backtesting or P&L attribution, the answer is almost always "reverts to the Standardised Approach."

Related exam topics worth revisiting together include VaR backtesting techniques for banks, since PLA and backtesting share the same statistical logic, and risk appetite framework, which governs how a board sets the limits that trading desks operate within under FRTB. On the credit side, candidates preparing RM alongside RFS should also look at portfolio credit risk measurement, which shares the correlation and tail-risk concepts FRTB applies to market risk.

✅ Conclusion: Locking In FRTB for the Exam

FRTB is best remembered as three linked ideas: a stricter, non-negotiable trading book/banking book boundary; a mandatory Standardised Approach built on sensitivities, default risk, and residual risk; and a harder-to-earn, desk-level Internal Models Approach built on Expected Shortfall, P&L attribution, and NMRF charges. Together they close the capital-arbitrage gaps that VaR-era rules left open. Browse more explainers on the risk management tag hub, work through related chapters like Operational Risk And Management Framework for the full Module D-E picture, and if you're preparing for the CAIIB elective, explore the full CAIIB course track. Ready to test yourself? Start a free RM mock test now →

🧠 Practice MCQs: Fundamental Review of the Trading Book

Want chapter-wise mock tests with 100+ MCQs? Start practising free →

❓ FAQs on FRTB