The standardised IRRBB framework and CSRBB · RiskAtoms Academy

Know these first

L3.6.08: Optionality in loans
L3.6.09: Non-maturity deposits
L3.6.12: Credit spread risk in the banking book and the SVB case

Math Atoms

By the end you will be able to

1Explain what the standardised framework fixes: the buckets, the caps on non-maturity deposits, and the prepayment and redemption multipliers.
2Build the add-on for automatic options from a Bachelier caplet, and say why a sold option adds to the loss.
3Run the framework on the bank: £270.01m, against Module 3.6's £257.23m, and explain the gap.
4Describe CSRBB, what is in and out of scope under the EBA guidelines, and read an illustrative figure.
5Set out how the UK and EU versions differ from Basel, with dates.

Why it matters

Every bank's ΔEVE rests on its own models: how long deposits stay, how fast mortgages prepay. A supervisor who doubts them needs a yardstick that does not. The standardised framework is that yardstick: the same buckets, caps and multipliers for every bank. Run on RiskAtoms Bank it gives £270.01m, more than the bank's own £257.23m, though the framework is cruder in places. The reason is a small cap on £1bn of tracker mortgages.

Core explanation

A yardstick in five steps

Illustrative dataRiskAtoms Bank plc's banking book, under the rules

The same made-up UK bank as Level 3, at 31 December 2025, in £ millions: the banking book of M3.6: Interest rate risk in the banking book and ALM with its cash flows, its £19.2bn macro hedge and £16bn structural hedge, and its assumptions for deposits and prepayment. Module 3.6's changes in economic value (ΔEVE) and net interest income (ΔNII) stay the bank's own figures. Here the rules of Basel, the EU and the UK are applied to them one by one; they change only the shock sizes, the floor on post-shock rates and the capital the loss is compared with.

Tier 1. Every test on these pages uses the bank's Tier 1 of £3,419.9m from M4.1: The Basel framework (CET1 plus £400m of AT1). Module 3.6 took £3,500m as a round figure, so its percentages are slightly lower. Anything a page adds to the book is labelled illustrative.

This page adds two labelled assumptions: the tracker cap's terms for the option add-on, and illustrative credit spread widenings for CSRBB. Both are tagged illustrative where they appear.

When it is used

SRP31.94 lets a supervisor mandate the framework and lets a bank choose it; a supervisor that finds a bank's own system deficient can require the bank to use it (31.86). Its positions are the banking book: assets not deducted from CET1 (other than fixed assets and equity exposures), liabilities other than CET1, and off-balance sheet items (31.97). Commercial margins may be kept in the cash flows or left out, as the bank chooses (31.100). SRP31.95 sorts positions into three groups: amenable to standardisation (fixed and floating flows), less amenable (non-maturity deposits, loans that can be prepaid, term deposits that can be redeemed early) and not amenable (automatic options). The steps follow from that.

ExploreThe standardised framework, step by step (SRP31.95)

Click a box to see what happens there.

1. Slot

Amenable positions: every repricing cash flow goes to one of 19 buckets, from overnight to over 20 years (31.96). Floating positions reprice at their next reset.

Non-maturity deposits: caps, not models

The framework accepts the bank's split of non-maturity deposits into stable and less stable, and core and non-core, but caps the result. The non-core part reprices overnight (31.111). The core part can be slotted further out, up to a share and an average maturity that depend on the depositor (31.112, Table 4). Module 3.6's assumptions sit within every cap, so the framework takes them as they are.

Deposit	Basel category	Core share	Core cap	Avg maturity	Maturity cap
Personal current accounts	retail transactional	81%	90%	2.5y	5.0y
Instant-access savings	retail non transactional	30%	70%	1.5y	4.5y
Operational corporate deposits	wholesale	50%	50%	1.5y	4.0y
Corporate call accounts	wholesale	10%	50%	1.0y	4.0y

Operational corporate deposits sit exactly on the 50% wholesale cap: the bank's own estimate, 60%, was capped in Module 3.6 (L3.6.09: Non-maturity deposits). The EU is stricter still: wholesale deposits from financial customers are non-core, and its guidelines cap the weighted average repricing date of modelled non-maturity deposits at five years.

Prepayment and early redemption

The bank supplies a base prepayment rate (CPR) and a base share of term deposits redeemed early (TDRR). The standard scales them by scenario (L3.6.08: Optionality in loans). A term deposit with no legal right to early withdrawal, or only with a significant penalty, can be treated as fixed if the supervisor is satisfied (31.122): in effect, a base TDRR of zero.

Scenario	Prepayment γ (Table 5)	Redemption u (Table 6)
Parallel up	0.8	1.2
Parallel down	1.2	0.8
Steepener	0.8	0.8
Flattener	1.2	1.2
Short rates up	0.8	1.2
Short rates down	1.2	0.8

Automatic options

An automatic option is one the holder exercises whenever it pays, such as a cap on a mortgage rate. The framework values them separately (KAO). The book values one (the zero floors on its deposit rates stay inside the deposit treatment): £1bn of tracker mortgages pay Bank Rate + 0.75% but never more than 6% (Module 3.6's extra), so the bank has sold its borrowers a cap on Bank Rate struck at 5.25%. With Bank Rate at 4.00% the cap is out of the money and worth £2.61m. Raise rates by 275bp and lift volatility by a quarter and it is worth £43.38m: a loss of £40.77m to the bank. illustrative

Credit spread risk in the banking book

IRRBB uses the risk-free curve. A bond's yield also carries a spread for credit and liquidity, and that spread moves too. CSRBB is that risk: Basel defines it (SRP31.3) and asks banks to monitor and assess it (31.9), but sets no shocks and no test. The EBA's guidelines (EBA/GL/2022/14) set the scope, applying from 31 December 2023: changes in the market credit spread and the market liquidity spread (para 120), not changes in the issuer's own credit quality (121) or non-performing exposures (122). Nothing is left out because of its accounting, and assets at fair value are never left out (124). The method is the bank's own (154).

Sort itCSRBB or not?

Sort each loss: is it credit spread risk in the banking book under the EBA guidelines?

BBB corporate bonds in the buffer lose value as spreads on all BBB issuers widen.
Covered bonds lose value because buyers ask more to hold a less liquid bond.
Gilts held at amortised cost lose market value when the gilt-SONIA spread widens.
A bond falls because its issuer is downgraded.
A non-performing loan is marked down.
Every bond falls because SONIA rises by 100bp, spreads unchanged.

An illustrative stress on the bank's buffer bonds, with Module 3.6's widenings of 20bp on gilts, 50bp on covered and AA bonds and 125bp on BBB bonds illustrative:

Line	Accounting	Amount	CS01	Widening	Loss
Gilts in the liquidity buffer	FVOCI	5,000	1.511	20bp	30.08
Gilts held to collect	Amortised cost	2,000	0.604	20bp	12.03
Covered and AA corporate bonds	FVOCI	2,000	0.580	50bp	28.68
BBB corporate bonds	FVOCI	600	0.197	125bp	23.93
Total		9,600	2.892		94.72

The loss is £94.7m, or 2.77% of Tier 1, with a CS01 of £2.89m. £82.7m goes through other comprehensive income into CET1 at once; £12.0m sits unseen in gilts at amortised cost. The LCR counts every buffer bond at market value, so after haircuts the same move takes £69.4m off HQLA and the ratio falls from 157.4% to 156.6%. Double the widening and the loss is £187.0m (5.47%).

The formal version

Each bucket is discounted with $e^{-Rt}$ (M-A05: The number e and continuous growth) and the buckets are summed (M-A08: Summation notation). For scenario $i$ , bucket $k$ with midpoint $t_k$ , and currency $c$ :

\begin{aligned} \text{EVE}_{i,c} &= \sum_{k=1}^{19} CF_{i,c,k}\; e^{-R_{i,c}(t_k)\, t_k} \\ \Delta\text{EVE}_{i,c} &= \text{EVE}_{0,c} - \text{EVE}_{i,c} + \text{KAO}_{i,c} \\ \text{measure} &= \max_{i \in \{1,\dots,6\}} \max\Big(0,\ \sum_{c:\ \Delta\text{EVE}_{i,c} > 0} \Delta\text{EVE}_{i,c}\Big) \end{aligned}

The option add-on uses the normal distribution (M-C06: The normal distribution). Under the Bachelier model the rate fixing at $t$ is normal around its forward $F$ with standard deviation $\sigma\sqrt{t}$ :

\begin{aligned} c &= (F - K)\, \Phi(d) + \sigma \sqrt{t}\, \varphi(d), \qquad d = \frac{F - K}{\sigma\sqrt{t}} \\ V(R, \sigma) &= \sum_j N\, q\, e^{-R(t_{j+1})\, t_{j+1}}\, c_j \\ \text{KAO}_i &= \sum_{\text{sold}} \big[V(R_i, 1.25\,\sigma) - V(R_0, \sigma)\big] - \sum_{\text{bought}} \big[V(R_i, 1.25\,\sigma) - V(R_0, \sigma)\big] \end{aligned}

$\Phi$ is the standard normal distribution function and $\varphi$ its density; $K$ the strike; $q$ the accrual period (a quarter); $N$ the notional.
The EU's text (2024/857 Art 13) writes the add-on as bought less sold, with the opposite sign convention for ΔEVE; the result is the same.

Try it

Build the bank's standardised ΔEVE from its parts. Leave out the option add-on and the measure drops below Module 3.6's; put it back and it rises above. Then assume some term deposits are redeemed early: when rates rise, more of that fixed-rate funding leaves overnight, and the loss grows.

Try itBuild the standardised ΔEVE

Term deposits redeemed early, TDRR₀

Rise in option volatility

Option add-on (KAO)

Standardised measure£270.01m

Share of Tier 17.90%

Against Module 3.6+£12.78m

Worst: Parallel up, £229.24m on the slotted flows plus £40.77m for the tracker cap. Module 3.6 valued each flow at its own date and had no add-on: £257.23m. Limit at 15% of Tier 1: £513.0m.

Worked example

The standardised measure for RiskAtoms Bank

£ millions, sterling only, Basel sizes (275/425/250bp) and Module 3.6's 0% floor so the result can be set against Module 3.6's; loss positive; TDRR₀ = 0.

From the book to one number

Step 1 of 5

Slot the flows and value them today

Bucket	Assets	Liabilities	Hedges	Net
Overnight	7,500.8	−19,681.3	0.0	−12,180.5
O/N–1M	12,783.5	−4,267.9	1,335.3	9,850.8
1M–3M	10,703.9	−3,545.0	806.3	7,965.3
3M–6M	3,648.2	−4,074.5	−992.2	−1,418.5
6M–9M	2,485.8	−1,557.5	−912.5	15.8
9M–1Y	3,816.2	−3,632.6	−835.4	−651.8
1Y–1.5Y	4,435.1	−3,053.5	−1,461.5	−80.0
1.5Y–2Y	6,597.8	−5,872.9	−1,206.3	−481.4
2Y–3Y	3,540.5	−5,752.1	1,025.0	−1,186.6
3Y–4Y	8,192.0	−7,789.7	1,294.1	1,696.4
4Y–5Y	2,306.0	−3,678.9	1,521.1	148.2
5Y–6Y	117.1	−106.0	0.0	11.1
6Y–7Y	3,167.1	−2,106.0	0.0	1,061.1

Nothing falls beyond seven years. The overnight bucket holds −£12,180.5m: the non-core deposits and reserves. Discounting each bucket at its midpoint gives $\text{EVE}_0 = 4{,}365.82$ , against 4,308.92 in Module 3.6 with each flow at its own date.

Regulatory lens

Basel: SRP31.94–31.129, in force from 1 January 2026 with the July 2024 shock sizes; CSRBB in 31.3 and 31.9. Statements on jurisdictions are as at 3 November 2026 as at 3 Nov 2026.

In the UK

UK The framework is in the PRA Rulebook, ICAA 9.13–9.43, with the Basel tables: NMD caps (9.36), prepayment (9.37), redemption (9.39) and options at +25% volatility (9.43). A firm may use it only where it suits its nature, size and complexity (9.1B) and must tell the PRA (9.1C). From 1 January 2027, SS31/15 says the PRA may require a firm to use it (2.9A) and expects ten years of data behind non-maturity deposit assumptions (2.12B). On CSRBB, ICAA 9.1A asks firms to monitor and assess it; the PRA assesses CSRBB as part of IRRBB in Pillar 2A (PS15/26, May 2026, effective 1 January 2027).

In the EU

EU Delegated Regulation (EU) 2024/857 (Official Journal, 24 April 2024; in force since 14 May 2024) sets a standardised approach and a simplified one for small and non-complex institutions. It differs from Basel in places: wholesale deposits from financial customers are non-core; core deposits shrink by 20% in scenarios where short rates rise and grow by 20% where they fall (Art 8); gains across currencies count at 50% (Art 17(6)); it adds an earnings measure over at least one year (Art 18) and a basis risk add-on (Art 23). The simplified approach uses lower core shares and scales option pay-offs by 1.1 (Art 25). EBA/GL/2022/14 applies from 30 June 2023, its CSRBB sections from 31 December 2023, with internal reports on CSRBB at least quarterly (para 144).

In the US

We found no standardised IRRBB framework or CSRBB rule in US regulation. Banks use their own models, and the 2010 interagency advisory (SR 10-1) expects them to test assumptions such as deposit behaviour and prepayment; examiners judge the results.

Spark Atoms

Spark AtomHow much does the 25% matter?

With no rise in volatility the add-on under parallel up is £39.81m; with 25% it is £40.77m; with 50%, £42.06m. After a 275bp rise the cap is deep in the money, so it behaves almost like a swap and volatility hardly moves it. Most of the add-on is the rate move itself.

Spark AtomSVB and the hidden part

SVB's losses sat mostly in bonds held to maturity, at amortised cost, so they did not show in its capital (L3.6.12: Credit spread risk in the banking book and the SVB case). The EBA guidelines, published nearly five months before SVB failed, already excluded nothing from CSRBB because of its accounting. Across Module 3.6's whole spread stress, including fixed-rate corporate loans, the bank's economic loss is £148.4m, 4.34% of Tier 1; only £82.7m of it would show in CET1.

Key takeaways

The standardised framework slots cash flows into 19 buckets, caps the bank's deposit assumptions, scales its prepayment and redemption rates by fixed multipliers, and adds automatic options at +25% volatility.
Non-core deposits are overnight; core deposits are capped at 90%, 70% or 50% of the balance with average maturities of 5, 4.5 or 4 years.
On the bank: £229.24m from the slotted flows plus £40.77m for the sold tracker cap gives £270.01m, 7.90% of Tier 1, against Module 3.6's £257.23m.
CSRBB is the spread above risk-free, for credit and liquidity: Basel asks banks to monitor it; the EBA guidelines set the scope and leave the method to the bank; no accounting category is excluded.
Illustrative CSRBB on the buffer bonds: £94.7m, 2.77% of Tier 1, and the LCR falls from 157.4% to 156.6%.

Reaction

Check what stuck

Question 1 · objective 3

The bank's standardised measure is £270.01m. Its Tier 1 is £3,419.94m. What is the measure as a percentage of Tier 1?

Question 2 · objective 2

Under parallel up the slotted flows lose £229.24m. The bank's sold tracker cap rises in value by £40.77m. What is the standardised ΔEVE for parallel up, loss positive, in £m?

£m

Question 3 · objective 4

In the illustrative CSRBB stress, the buffer bonds lose £94.72m. As a percentage of Tier 1?

Question 4 · objective 1

Under the standardised framework, a current-account balance the bank's model treats as 95% stable. How much can be slotted beyond overnight?

AAt most 90%, with an average maturity of no more than 5 yearsB95%, the model's figureCAt most 70%, with an average maturity of no more than 4.5 yearsDNone: all non-maturity deposits are overnight

Question 5 · objective 5

In the EU's standardised approach, how are a pension fund's call deposits treated?

AAll non-core: slotted overnightBUp to 50% core, average maturity up to 4 yearsCUp to 70% coreDAs fixed-rate funding to their contractual date

Glossary

Standardised framework: Standardised framework: the Basel method for computing ΔEVE with prescribed buckets, caps and multipliers instead of the bank's own models (SRP31.94–31.129). A supervisor can mandate it; a bank can choose it.
Time bucket: Time bucket: one of 19 repricing intervals, from overnight to over 20 years, each with a midpoint at which its net cash flow is discounted.
Core deposits: Core deposits: the part of a stable non-maturity deposit balance that is unlikely to reprice even when rates move; it may be slotted beyond overnight, up to a cap.
CPR: CPR (conditional prepayment rate): the yearly share of a loan book repaid early. The standard scales a base CPR by 0.8 when rates rise and 1.2 when they fall.
TDRR: TDRR (term deposit redemption ratio): the share of fixed-term deposits withdrawn early. The standard scales a base ratio by 1.2 when rates rise and 0.8 when they fall.
KAO: KAO: the add-on for automatic interest rate options: the change in their value under each scenario with implied volatility 25% higher; sold options add to the loss.
Bachelier model: Bachelier (normal) model: an option model in which the underlying rate is normally distributed, so volatility is in basis points rather than as a percentage of the rate.
CSRBB: CSRBB (credit spread risk in the banking book): the risk from changes in the market price of credit and liquidity risk, above the risk-free rate, that IRRBB and default risk do not explain (SRP31.3; EBA/GL/2022/14).
CS01: CS01: the value lost when a credit spread widens by one basis point.

Deep Dives & Origins

Deep Dives

L3.6.08: Optionality in loans: the bank's own prepayment and option models, which the framework replaces with multipliers. Intermediate.
L4.9.01: Pillar 2 and the SREP: Pillar 2, where IRRBB and CSRBB capital is set. Intermediate.

Origins

Basel Committee on Banking Supervision, SRP31: Interest rate risk in the banking book, paras 31.3, 31.9 and 31.94–31.129, version in force from 1 January 2026, bis.org.
Basel Committee on Banking Supervision, Interest rate risk in the banking book (d368), April 2016, paras 120–132 (prepayment, term deposits, automatic options and the calculation with aggregation).
Commission Delegated Regulation (EU) 2024/857 (standardised and simplified standardised approaches), OJ L, 24 April 2024, Arts 8, 9, 10, 13, 17, 18, 23 and 25.
European Banking Authority, Guidelines on IRRBB and CSRBB (EBA/GL/2022/14), 20 October 2022, paras 8, 110–111, 120–125, 144, 154 and 156.
Prudential Regulation Authority, PRA Rulebook: Internal Capital Adequacy Assessment Part, rules 9.1A–9.1C and 9.13–9.43, as at 3 November 2026; SS31/15, October 2025 near-final version, paras 2.9A and 2.12B; PS15/26, May 2026, para 5.10.

Module complete

You have finished M4.5: Liquidity and IRRBB regulation. Try the Lab Test for M4.5: Liquidity and IRRBB regulation →

Watch this lesson

Video version

Prefer to watch? The same lesson as a 14-minute video will sit here, made from this page's script.

Previous · L4.5.06

Supervisory outlier tests

Next · Lab Test

M4.5: Liquidity and IRRBB regulation