Interest-Rate Risk in the Banking Book

The banking book is not the trading book

Banks hold some positions primarily for customer lending, deposit-taking, treasury liquidity and longer-term balance-sheet management. Those positions sit in the banking book under the applicable regulatory boundary. Other positions held for trading purposes can fall into the trading book and are subject to market-risk capital rules.

The distinction matters because the same interest-rate move can affect both books, but the regulatory measurement architecture is different.

IRRBB therefore asks: what happens to the bank’s ordinary balance sheet when rates move?

Two views matter: earnings and economic value

View	Question
Earnings view	How do changing rates alter net interest income and other rate-sensitive earnings over the measurement horizon?
Economic-value view	How do changing discount rates and expected cash flows alter the present value of assets, liabilities and off-balance-sheet positions?

A bank can look comfortable under one view and vulnerable under the other. Short-term earnings may remain strong while the economic value of long fixed-rate assets has fallen materially.

This is why sound IRRBB management uses both lenses rather than choosing the more flattering one.

Gap risk begins with different repricing clocks

Suppose a bank owns a five-year fixed-rate loan at 3 per cent while funding it with deposits that can reprice within months. If market rates rise, deposit costs can increase while the loan income remains fixed.

funding reprices upward first → asset yield stays fixed → interest margin compresses.

This is gap risk: the timing of rate changes on assets and liabilities does not match.

Read Repricing Risk for the first-principles version of this mechanism.

Yield-curve risk appears when different maturities move differently

Interest rates do not always rise or fall by the same amount at every maturity. Short rates can rise sharply while long rates barely move. Long rates can fall while the front end remains stable. The yield curve can steepen, flatten or change shape.

A bank that hedged against a parallel shift can still lose money when the curve twists instead.

IRRBB therefore is not simply “rates up” versus “rates down.” The structure of the move matters.

Basis risk appears when linked rates stop moving together

A loan can reprice from one benchmark while funding costs follow another. If the two rates normally move together, the bank can appear hedged. If the spread between them changes, the hedge breaks economically.

That is basis risk. Article 52 owns the full mechanism: Basis Risk.

Option risk appears when customers change behaviour because rates changed

Many banking products contain options even when customers do not call them options.

• a mortgage borrower can prepay or refinance;
• a depositor can move money to a higher-yield account;
• a term deposit can be withdrawn early if the contract permits;
• a borrower can draw or not draw a committed facility;
• customers can switch between fixed and floating products where allowed.

When rates move, customer behaviour changes the timing and amount of the bank’s expected cash flows. The risk therefore is partly financial and partly behavioural.

Non-maturity deposits are one of the hardest modelling problems

A current account has no fixed contractual maturity. The customer may withdraw tomorrow, yet part of the deposit base can remain for years. The bank therefore estimates behavioural maturity and how quickly deposit rates respond to market rates.

Those assumptions matter enormously. Treat deposits as too stable and the bank can underestimate funding and rate risk. Treat them as instantly mobile and the model can become unrealistically conservative.

The Basel IRRBB framework specifically emphasises governance around behavioural and modelling assumptions because these assumptions can dominate the result.

Deposit beta measures how much market-rate change reaches deposit pricing

If a benchmark rate rises by 2 percentage points and the bank raises a deposit rate by only 1 percentage point, the deposit rate has moved by half as much in this simplified example.

The speed and size of this pass-through depend on competition, customer type, account features and market conditions.

Deposit beta can therefore change across cycles. A bank that assumes customers remain insensitive can discover that digital comparison and rate competition make deposits reprice much faster than historical experience suggested.

Prepayment risk can shorten long assets exactly when the bank wanted them to stay long

When market rates fall, borrowers with fixed-rate mortgages may refinance. The bank receives principal back earlier and must reinvest at the new lower market rate.

The bank therefore loses some of the high-yield asset just when that asset has become especially valuable.

When rates rise, prepayments can slow, leaving the bank holding low-yield fixed-rate assets for longer. Customer optionality can move against the bank in both directions.

Duration makes economic-value sensitivity visible

Longer-duration cash flows are generally more sensitive to changes in discount rates. A long fixed-rate bond can therefore lose more economic value from a given rate rise than a very short instrument.

Banks compare the duration and timing of assets and liabilities to understand how economic value moves when rates change.

This is not a perfect one-number answer because options, non-linearities and curve shape matter, but duration remains a useful first diagnostic.

A worked miniature: earnings risk

A bank has S$1 billion of fixed-rate loans yielding 3 per cent and funds much of them with deposits costing 1 per cent. The simplified interest spread is 2 percentage points.

Market rates rise. Deposit competition pushes funding cost to 2.5 per cent while the loans remain fixed at 3 per cent.

The spread falls to 0.5 percentage points before operating cost, credit loss and capital cost. No borrower had to default for earnings pressure to appear.

A worked miniature: economic-value risk

Imagine the same bank holds long fixed-rate assets whose present value falls substantially when market yields rise. If its liabilities are shorter or less rate-sensitive, the economic value of equity can decline even if the accounting carrying value of some assets does not immediately show the same movement.

Economic-value measurement therefore reveals interest-rate exposure that short-term accounting earnings can miss.

Accounting treatment does not remove economic risk

An asset held at amortised cost may not show every market-price movement through current profit. The bank can still suffer economic loss if the asset must be sold, if its funding cost rises, or if the present value of its cash flows has fallen.

Accounting tells the bank how to report. Risk management asks what reality can do to value and earnings.

Hedging reduces risk by creating offsetting sensitivity

Banks can use interest-rate swaps, futures, options and balance-sheet structure to reduce IRRBB. A fixed-rate asset can be partly offset with a swap that receives floating and pays fixed, for example.

The hedge itself introduces basis, counterparty, collateral and operational considerations. The goal is not to make every line item individually neutral; it is to manage the combined balance-sheet sensitivity within risk appetite.

Hedges can fail because the instrument and the exposure are not identical

A swap can hedge a benchmark rate while the actual deposit cost follows customer behaviour. A futures contract can hedge one maturity while the loan book reprices across several maturities. A hedge can also expire before the underlying exposure.

Residual mismatch is normal. The bank must know how much remains and how it behaves in stress.

IRRBB interacts with credit risk

Higher rates can improve asset yield and simultaneously make floating-rate borrowers less able to repay. Lower rates can help borrowers while reducing bank margin.

A rate scenario therefore can change both the bank’s interest income and its expected credit loss.

Read Expected Credit Loss for the credit pathway.

IRRBB also interacts with liquidity risk

If rising rates reduce the market value of securities that the bank expected to sell for liquidity, using those assets can crystallise losses. At the same time, depositors can move toward higher-yield alternatives, accelerating outflows.

One rate shock can therefore hit earnings, economic value and liquidity together.

This is why High-Quality Liquid Assets and IRRBB should not be managed as unrelated topics.

Stress scenarios should include more than one rate path

The Basel framework uses standardised interest-rate shock scenarios and expects banks to assess IRRBB across a wide and appropriate range of shocks and stresses.

Useful internal scenarios can explore:

• parallel rate increases;
• parallel rate decreases;
• steepening and flattening curves;
• short-rate shocks;
• changes in deposit behaviour;
• faster or slower mortgage prepayment;
• basis widening between benchmarks.

The aim is to find where the bank’s assumptions stop cooperating.

Credit-spread risk in the banking book is related but separate

A banking-book security can lose value because its credit spread widens even if the risk-free interest-rate curve does not move. Basel supervision therefore also considers credit spread risk in the banking book—CSRBB—as a related risk requiring monitoring and assessment.

Article 51 will follow the broader market-risk pathway through securities.

IRRBB governance begins with risk appetite

The board and senior management need to decide how much earnings and economic-value sensitivity the bank is willing to carry. Treasury can then hedge or structure the balance sheet within those limits.

Without risk appetite, measurement becomes descriptive rather than controlling. The bank knows the exposure but has not decided when it is too large.

Data quality matters because IRRBB is built from cash-flow timing

A wrong reset date, incorrect prepayment assumption or stale deposit classification can distort the model. Small data errors repeated across millions of accounts can materially change reported sensitivity.

IRRBB therefore depends on accurate product data, model validation, reconciliation and independent review—not only clever financial mathematics.

The deepest risk is assuming yesterday’s behaviour will survive tomorrow’s rates

Customer behaviour is not fixed. Digital banking can speed deposit movement. New competitors can change pricing. Mortgage refinancing can accelerate. A new rate cycle can reveal behaviour never observed in the model’s historical data.

That is why the current Basel IRRBB guidance emphasises behavioural assumptions, model governance and multiple measures rather than one mechanical gap report.

Four misconceptions to remove

Misconception	Better model
“Interest-rate risk belongs only to traders.”	Ordinary loans, deposits and treasury positions create material rate risk in the banking book.
“If net interest income is stable, IRRBB is low.”	Economic value can deteriorate even when short-term earnings remain comfortable.
“Deposits have no maturity, so they cannot be modelled.”	Banks use behavioural assumptions, but those assumptions require strong governance and stress testing.
“A hedge removes interest-rate risk.”	Hedges can leave basis, timing, option, counterparty and model risk.

A mastery test

1. What makes IRRBB different from trading-book market risk?
2. Why must banks measure both earnings and economic-value effects?
3. How do deposit behaviour and mortgage prepayment create option risk?
4. Why can a hedge still leave basis risk?
5. How can one rate shock affect credit, liquidity and capital at the same time?

If those answers connect, IRRBB becomes visible as a balance-sheet timing problem with human behaviour inside it: the bank has made promises across years, but market rates can change every day.

---