The IRRBB standard · RiskAtoms Academy

By the end you will be able to

1Name what the twelve SRP31 principles ask of banks and of supervisors.
2Set out how the standard defines ΔEVE and ΔNII, and the behavioural assumptions behind them.
3Compute the bank's six ΔEVE and its ΔNII, fill Template IRRBB1 and score the worst loss at 7.52% of Tier 1.
4Explain how shock sizes are calibrated, and why Basel, the EU and the UK used different sterling sizes on 31 December 2025.

Why it matters

RiskAtoms Bank's year end, 31 December 2025, falls on a seam in the rules. Basel's recalibrated shocks are to be used from 1 January 2026, and a bank with that year end discloses on them in 2026. The EU still uses the old sterling sizes. The UK switches on 1 July 2026. The same book, with the same cash flows, loses £257.2m of economic value in a parallel rise on one set of sizes and £235.4m on the other. And in that same rise its net interest income goes up by £254.7m. The standard asks for both numbers, and this lesson shows where each comes from.

Core explanation

One standard, two measures, six shocks

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.

M3.6: Interest rate risk in the banking book and ALM built the bank's banking book and measured its interest rate risk in the banking book the way an ALM team does. The rules that frame that work are in SRP31 of the Basel Framework, the standard published in April 2016 and recalibrated in July 2024. Unlike credit or market risk, IRRBB has no Pillar 1 charge. The standard works through Pillar 2 instead: principles for banks, a common way to measure and disclose, and tests that tell supervisors which banks to look at.

Twelve principles

SRP31.4, version in force from 1 January 2026
Principle	What it asks	Where
1. Identification	Identify, measure, monitor and control IRRBB; monitor and assess CSRBB	31.6–31.9
2. Governance	Board oversight of the IRRBB framework, with independent review	31.10–31.21
3. Risk appetite	Appetite and limits in both economic value and earnings	31.22–31.26
4. Measurement	Both measures, over a wide range of shock and stress scenarios	31.27–31.38
5. Assumptions	Behavioural and modelling assumptions understood, documented, tested	31.39–31.47
6. Data and models	Accurate data; independent model validation	31.48–31.61
7. Reporting	Regular reports to the board or its delegates	31.62–31.64
8. Disclosure	Disclose exposure and measurement practice (DIS70)	31.65
9. Capital	Assess IRRBB capital explicitly in the ICAAP	31.66–31.70
10. Supervisory monitoring	Supervisors collect enough information	31.71–31.73
11. Supervisory assessment	Supervisors assess exposures and management	31.74–31.81
12. Outliers	Publish outlier criteria; require action where risk is undue	31.82–31.89

Sort itBank or supervisor?

Each line is one of the twelve principles. Who does it?

Set a risk appetite for IRRBB in terms of both economic value and earnings, with limits.
Disclose the level of IRRBB exposure and how it is measured, once a year.
Assess the capital needed for IRRBB in the ICAAP.
Validate the models used to measure IRRBB independently.
Collect enough information from banks to assess their IRRBB exposures.
Publish the criteria that make a bank an outlier.
Require a bank with undue IRRBB to reduce it or hold more capital, within a set time.

Two measures, two balance sheets

Principle 4 asks for both an economic value measure and an earnings measure. L3.6.05: Economic value of equity sensitivity and L3.6.04: Net interest income sensitivity built them; the standard fixes how they are disclosed, so that banks can be compared (DIS70.4–70.5):

	ΔEVE	ΔNII
Balance sheet	Run-off	Constant
Horizon	To each cash flow's repricing date	Rolling 12 months
Equity	Excluded: no cash flows	Not an interest item
Commercial margins	May be in or out; the bank says which	In
Discounting	Risk-free zero curve (the bank discloses which)	None
Prescribed scenarios	Six (SRP31.90)	Two (SRP31.90); in the EU, parallel up and down

The bank's book was built to these rules, so its figures need no change: ΔEVE values today's positions with margins included on the SONIA curve; ΔNII rolls the book forward for a year from a base of £1,640.9m. The two measures pull in opposite directions in 4 of the six scenarios: the deposit base makes earnings rise with rates, while the receive-fixed structural hedge, part of which invests equity, makes value fall (L3.6.05: Economic value of equity sensitivity).

Six shocks, and where their sizes come from

L3.6.06: The standard shock scenarios built the six scenarios from three sizes per currency: parallel, short and long. The sizes are not chosen by judgement. SRP98 sets a method, and the July 2024 recalibration rewrote it to run on each currency's own history (calibration):

ExploreFrom rate history to a shock size (SRP98.57–98.62)

Click a box to see what happens there.

1. Data

Daily rates for nine tenor buckets per currency, from 3 January 2000 to 29 December 2023 (98.57). The 2016 calibration stopped in 2015.

Basel SRP31.90 Table 2 (2016 and July 2024 versions); EU Delegated Regulation 2024/856, Annex Part A; PRA Rulebook, ICAA 9.11
Sterling sizes, bp	Parallel	Short	Long
Basel, 2016 standard (to 31 Dec 2025)	250	300	150
Basel, SRP31.90 from 1 Jan 2026	275	425	250
EU, Delegated Regulation 2024/856 Annex	250	300	150
UK, ICAA 9.11 to 30 Jun 2026	250	300	150
UK, ICAA 9.11 from 1 Jul 2026	275	425	250

So on 31 December 2025 three rulebooks gave two answers as at 3 Nov 2026. Basel asks banks to use its new sizes from 1 January 2026, and a bank with a 31 December year end discloses them in 2026 on its 31 December 2025 data (the July 2024 standard, introduction). The EU's supervisory outlier test uses its own Annex, still on the 2016 sterling sizes; we found no amendment adopting the 2024 sizes. The UK changed rule 9.11 of its ICAA Part on 1 July 2026 (PRA LIAF02/25, October 2025). The bank's worst ΔEVE is £257.23m on Basel's sizes Basel and £235.42m on the EU's EU, which were also the UK's at the year end UK.

Behaviour: deposits and prepayment

Principle 5 is where most of the judgement sits. The standard names the key assumptions: how long non-maturity deposits stay and how far their rates follow the market, how fast fixed-rate loans prepay, and how term depositors redeem early (SRP31.39–31.41). Banks must document them, test their sensitivity and review them at least once a year (31.45–31.47). L3.6.09: Non-maturity deposits set the bank's deposit model and L3.6.08: Optionality in loans its prepayment model. Measured against the caps the standardised framework sets (SRP31.112, Table 4), the bank's choices sit inside:

Deposit	£m	Core share	Core cap	Average maturity	Maturity cap
Personal current accounts	22,000	81%	90%	2.5y	5.0y
Instant-access savings	10,000	30%	70%	1.5y	4.5y
Operational corporate deposits	5,000	50%	50%	1.5y	4.0y
Corporate call accounts	3,000	10%	50%	1.0y	4.0y

One cap binds: operational deposits would be 60% core on the bank's own history, and the wholesale cap holds them at 50%. Fixed-rate mortgages prepay at 8% a year in the base case, 6.4% when rates rise and 9.6% when they fall: the 0.8 and 1.2 multipliers of SRP31.119. The EU goes further in its guidelines: modelled NMDs may not have a weighted average repricing date beyond five years, and deposits from financial customers, other than operational deposits, may not be modelled behaviourally at all (EBA/GL/2022/14, paras 110–111).

Disclosure

Principle 8 makes the numbers public, once a year (DIS70). Table IRRBBA describes how the bank measures and manages IRRBB, including the average and longest repricing maturity it gives its NMDs. Template IRRBB1 reports ΔEVE and ΔNII under each scenario, for this year and last, with the maximum and Tier 1. The worked example fills it in for the bank.

The formal version

For scenario $i$ , with $CF(t_k)$ the rate-sensitive cash flows of the banking book at time $t_k$ in years and $R(t)$ the continuously compounded risk-free zero rate (M-A05: The number e and continuous growth), the two measures and the test are (DIS70.4–70.5; SRP31.82):

\begin{aligned} \text{EVE}_i &= \sum_{k} CF_i(t_k)\, e^{-R_i(t_k)\,t_k}, \qquad R_i(t) = \max\big(R_0(t) + \Delta R_i(t),\ \text{floor}\big) \\ \Delta\text{EVE}_i &= \text{EVE}_i - \text{EVE}_0 \\ \Delta\text{NII}_i &= \text{NII}^{\,12\text{m}}_i - \text{NII}^{\,12\text{m}}_0 \quad \text{(constant balance sheet)} \\ \text{Loss} &= \max_{i=1,\dots,6}\ \max\big(0,\ -\Delta\text{EVE}_i\big), \qquad \text{outlier if } \text{Loss} > 15\% \times \text{Tier 1} \end{aligned}

$CF_i$ carries the scenario's behaviour: prepayment at $\gamma_i \times$ the base rate, with $\gamma_i$ = 0.8 or 1.2.
$\Delta R_i(t)$ is the scenario shape of L3.6.06: The standard shock scenarios built from the currency's sizes; the floor is the supervisor's, no higher than zero (SRP31.93). The book uses 0%, which binds only in short down, where the overnight rate would reach −0.33%.

The size, step by step. Writing the six steps of SRP98 as one line, for scenario type $s$ (parallel, short or long) in currency $c$ , with $B_s$ its tenor buckets and $h$ six months:

\begin{aligned} \Delta\bar r_{s,c}(t) &= \frac{1}{\lvert B_s\rvert} \sum_{j \in B_s} \big( r_{j,c}(t) - r_{j,c}(t - h) \big) \\ q_{s,c} &= \text{99.9th percentile of } \lvert \Delta\bar r_{s,c}(t) \rvert, \ t \in [2000, 2023] \\ \bar R_{s,c} &= \text{round}_{25\text{bp}}\Big( \min\big(\text{cap}_s,\ \max(100\text{bp},\ q_{s,c})\big) \Big) \end{aligned}

The formula is our compact writing of the steps as SRP98.57–98.62 word them, read in the July 2024 standard; the caps are 400bp, 500bp and 300bp.

How the measure moves with the size. Differentiate EVE with respect to a parallel shift $\Delta$ (M-B01: Derivatives): each discount factor $e^{-(R+\Delta)t}$ brings down $-t$ , so

\begin{aligned} \frac{d\,\text{EVE}}{d\Delta}\Big|_{\Delta = 0} &= -\sum_k t_k\, CF(t_k)\, e^{-R_0(t_k)\,t_k} = -9{,}258 \\ \text{per bp:}\quad \text{DV01} &= 9{,}258 \times 0.0001 = \text{£}0.926\text{m} \\ \text{secant, 250 to 275bp:}\quad \frac{-257.23 - (-235.42)}{25} &= -0.8725\ \text{£m per bp} \end{aligned}

The slope between the EU and Basel sizes (M-A10: Slopes and rates of change) is smaller than the tangent at today's curve, mainly because each further basis point takes a little less value off a fixed cash flow (L3.6.05: Economic value of equity sensitivity). So the 25bp difference in size explains the gap between the two regimes almost exactly: 25 × 0.8725 = £21.81m, against the DV01 estimate of £23.15m.

Try it

Start at Basel's 275bp and step down to the EU's 250bp: economic value loses about £21.8m less. Then go to 200bp, the size the PRA uses as a benchmark for smaller firms. Watch the two pairs of lines: economic value falls when rates rise, earnings fall when they drop, and the earnings loss in parallel down grows faster than the gain in parallel up.

Try itHow big is the shock?

Parallel shock size, up and down275bp

Economic value (run-off)

ΔEVE, parallel up−£257.2m

ΔEVE, parallel down+£251.7m

Worst fall ÷ Tier 1 (test: 15%)7.52%

Earnings (one year, constant)

ΔNII, parallel up+£254.7m

ΔNII, parallel down−£263.3m

Worst fall ÷ Tier 1 (EU line: 5%)7.70%

ΔEVE, upΔEVE, downΔNII, upΔNII, down

From 250bp to 275bp, parallel up costs a further −£21.8m: about £0.873m for each extra basis point, the slope of the solid blue line here. It flattens slowly as the size grows.

Worked example

RiskAtoms Bank's IRRBB1 at 31 December 2025

£ millions, sterling only, Basel sizes 275/425/250bp with the book's 0% floor.

From the six shocks to the disclosure

Step 1 of 5

1

Run the six scenarios

Scenario	ΔEVE, Basel sizes	ΔEVE, EU/UK 2025 sizes	ΔNII, year 1
Parallel up	−£257.23m	−£235.42m	+£254.69m
Parallel down	+£251.72m	+£227.12m	−£263.27m
Steepener	−£79.59m	−£40.40m	−£191.60m
Flattener	+£10.22m	−£2.72m	+£253.74m
Short rates up	−£119.41m	−£84.76m	+£337.46m
Short rates down	+£99.78m	+£72.31m	−£307.42m

Base NII for the year ahead is £1,640.87m. Economic value falls in the three scenarios that raise rates where the book is long fixed, and NII falls in the three that cut short rates.

Regulatory lens

The standard is SRP31 of the Basel Framework, with the calibration method in SRP98 and disclosure in DIS70; the recalibrated shocks should be implemented by 1 January 2026. Statements on jurisdictions are as at 3 November 2026 as at 3 Nov 2026.

In the UK

The PRA writes the standard into chapter 9 of its Internal Capital Adequacy Assessment Part: systems for interest rate risk and for credit spread risk in non-trading activities (9.1, 9.1A), the scenarios (9.7) and the sterling sizes (9.11), which moved to 275/425/250bp on 1 July 2026 after consultation LIAC01/25 (July 2025) and finalisation LIAF02/25 (October 2025). Supervisory Statement SS31/15, in the version effective from 1 January 2027, asks firms to calculate the EVE impact at least quarterly (2.9B) and to report on IRRBB to the management body at least quarterly (2.10A). UK disclosure uses Table UK IRRBBA and Template UK IRRBB1 (Disclosure (CRR) Part, Article 17a). The tests are in L4.5.06: Supervisory outlier tests.

In the EU

Article 84 of the Capital Requirements Directive requires systems for IRRBB and CSRBB, and Article 98(5) sets the supervisory outlier tests. Commission Delegated Regulation (EU) 2024/856, applying since 14 May 2024, fixes the scenarios, the common assumptions (for example, CET1 and perpetual own funds without a call date are excluded from EVE, and NII is measured over one year on a constant balance sheet) and the sizes in its Annex. The EBA guidelines on IRRBB and CSRBB (EBA/GL/2022/14) apply from 30 June 2023, and from 31 December 2023 for CSRBB. Disclosure follows Article 448 of the CRR, in the EU IRRBBA and IRRBB1 formats.

In the US

We found no IRRBB outlier test in the US agencies' rules. Their 2010 interagency advisory (Federal Reserve SR 10-1) expects banks to consider both earnings and economic value, warns that parallel shocks of ±200bp may not be enough and asks for larger moves (for example up and down 300bp and 400bp), and says earnings exposure is best projected over at least two years. The SVB case of L3.6.12: Credit spread risk in the banking book and the SVB case is where that approach was tested.

Spark Atoms

Spark AtomFrom 20% of total capital to 15% of Tier 1

The 2016 standard tightened the outlier threshold to 15% of Tier 1, from 20% of total capital (Basel Committee press release, 21 April 2016). The EU's Capital Requirements Directive, as adopted in 2013, flagged a bank whose economic value fell by more than 20% of its own funds after a 200bp change in rates. The bank would have passed that old test with ease: its ±200bp loss is £190.9m.

Spark AtomSterling's short shock grew by 125bp

In the 2024 recalibration the sterling short shock rose from 300bp to 425bp: 125bp, as much as the Hong Kong dollar (250 to 375bp) and the Swedish krona (300 to 425bp), and more than the euro (250 to 350bp), the Swiss franc (150 to 250bp) or the US dollar, whose short size stayed at 300bp. The dollar's parallel size stayed at 200bp too.

Key takeaways

SRP31 has twelve principles: nine for banks (identify, govern, set appetite, measure, model, validate, report, disclose, capitalise) and three for supervisors (monitor, assess, act on outliers).
ΔEVE uses a run-off balance sheet with equity excluded; ΔNII uses a constant balance sheet over 12 months. Six scenarios apply to EVE, two to NII.
Shock sizes come from each currency's six-month rate changes, 2000–2023: the 99.9th percentile, floored at 100bp, capped at 400/500/300bp, rounded to 25bp.
Sterling: Basel 275/425/250bp from 2026; the EU still 250/300/150bp; the UK switched on 1 July 2026. The bank's worst ΔEVE is £257.2m (7.52% of Tier 1) or £235.4m (6.88%).
Template IRRBB1 discloses ΔEVE and ΔNII by scenario, the maximum and Tier 1, once a year; Table IRRBBA explains the method, including NMD maturities.

Reaction

Check what stuck

Glossary

IRRBB: Interest rate risk in the banking book (IRRBB): the risk to a bank's capital and earnings from rate moves that affect its banking book positions (SRP31.1).
SRP31: SRP31: the chapter of the Basel Framework that holds the IRRBB standard, first published in April 2016 and recalibrated in July 2024.
Economic value of equity (EVE): Economic value of equity (EVE): the present value of the banking book's rate-sensitive cash flows, assets less liabilities plus hedges, with equity itself left out.
Net interest income (NII): Net interest income (NII): interest earned less interest paid. The standard's earnings measure is the change in NII over the next twelve months.
Run-off balance sheet: Run-off balance sheet: today's positions only, each running to its repricing date with nothing new written. Used for ΔEVE (DIS70.4).
Constant balance sheet: Constant balance sheet: whatever matures or reprices is replaced by the same business, so the balance sheet keeps its size and mix. Used for ΔNII (DIS70.5).
Shock calibration: Calibration: the method in SRP98.57–98.63 that turns each currency's history of six-month rate changes into a parallel, short and long shock size.
Non-maturity deposit (NMD): Non-maturity deposit (NMD): a deposit the customer can withdraw at any time, such as a current account. Banks model how long it really stays and how far its rate follows the market.
Template IRRBB1: Template IRRBB1: the annual fixed-format disclosure of ΔEVE and ΔNII under each prescribed scenario, for this year and last, with the maximum and Tier 1 (DIS70).
Table IRRBBA: Table IRRBBA: the annual qualitative disclosure of how the bank defines, measures and manages IRRBB, with the average and longest repricing maturity of its NMDs (DIS70).
ICAAP: ICAAP: the bank's own assessment of the capital it needs for all its risks, including IRRBB, which Pillar 1 does not charge (Principle 9).

Deep Dives & Origins

Deep Dives

L4.5.06: Supervisory outlier tests: the outlier tests of Basel, the EU and the UK on the same figures. Intermediate.
L4.5.07: The standardised IRRBB framework and CSRBB: the standardised framework and credit spread risk. Intermediate.
M6.6: Liquidity, ALM and behavioural models: how deposit and prepayment models are built and tested. Advanced.

Origins

Basel Committee on Banking Supervision, SRP31: Interest rate risk in the banking book, version in force from 1 January 2026 (published 16 July 2024) and version effective 15 December 2019, bis.org.
Basel Committee on Banking Supervision, Recalibration of shocks for interest rate risk in the banking book (d578), July 2024, with the revised SRP98.57–98.63, bis.org.
Basel Committee on Banking Supervision, Interest rate risk in the banking book (d368), April 2016, and press release of 21 April 2016, bis.org; Principles for the Management and Supervision of Interest Rate Risk, July 2004, para 84.
Basel Committee on Banking Supervision, DIS70: Interest rate risk in the banking book, version effective 15 December 2019, bis.org; and the Swiss SIF copy.
Commission Delegated Regulation (EU) 2024/856 of 1 December 2023 (supervisory shock scenarios, common modelling and parametric assumptions, large decline), OJ L, 24 April 2024.
European Banking Authority, Guidelines on IRRBB and CSRBB (EBA/GL/2022/14), 20 October 2022.
Commission Implementing Regulation (EU) 2022/631 of 13 April 2022 (IRRBB disclosure templates).
Prudential Regulation Authority, PRA Rulebook: Internal Capital Adequacy Assessment Part, chapter 9, as at 3 November 2026; Low impact amendments LIAC01/25 (July 2025) and LIAF02/25 (October 2025); SS31/15, October 2025 near-final version effective 1 January 2027; PRA Rulebook: Disclosure (CRR) Part, Article 17a.
Board of Governors of the Federal Reserve System and others, Interagency Advisory on Interest Rate Risk Management (SR 10-1), 11 January 2010.

Watch this lesson

Video version

Prefer to watch? The same lesson as a 12-minute video will sit here, made from this page's script.