Non-maturity deposits · RiskAtoms Academy

By the end you will be able to

1Explain the two behaviours of a deposit with no maturity: how long the balance stays and how much of a rate change the bank passes on (the deposit beta).
2Split a deposit into stable and volatile, then core and non-core, under the Basel caps.
3Slot core deposits into repricing buckets and check their average maturity against the caps.
4Show on the bank's book how these assumptions drive NII, EVE and the outlier test.

Why it matters

A current account has a contractual maturity of one day. Read literally, RiskAtoms Bank owes £40bn that could all leave tomorrow and reprice tomorrow. Nobody runs a bank on that reading: most of the money has sat there for years, at a rate that barely moves. Between the contract and the behaviour lies a model, and the model decides the answer. Treat every one of those deposits as overnight and the bank's worst fall in economic value is £1,503m, 42.9% of Tier 1, nearly three times the outlier line. Model them as the bank does and it is £257m, 7.35%. Same balance sheet, same deposits.

Core explanation

A deposit has two clocks

Illustrative dataRiskAtoms Bank plc's banking book

The same £67bn balance sheet as Module 3.5, on one made-up reporting date, in pounds, with how each line reprices, its customer rate, a SONIA curve with Bank Rate at 4.00%, a £19.2bn macro hedge on the fixed-rate mortgages and a £16bn structural hedge. Tier 1 capital is taken as £3.5bn. The figures are round illustrative values, not any bank's data. Every lesson in Module 3.6 uses this book, so the figures agree.

This lesson uses the book's four non-maturity deposits, their stable shares (from deposit history, illustrative), betas and slotting. Nothing is added.

A non-maturity deposit (NMD) has no end date and an administered rate: the bank decides when and how far to move it. To measure its interest rate risk, ALM needs two estimates. The first is volume: how much of the balance stays, and for how long. The second is price: how much of a move in Bank Rate the bank passes on to the customer, the deposit beta. A low beta means the deposit is cheap funding whose cost barely moves, which is what made it valuable to the bank in L3.6.04: Net interest income sensitivity.

Stable and volatile

The first cut is by behaviour over time. History shows that some of the balance swings with salaries, bills and seasons, and some sits there through all of it. The part that stays with a high degree of likelihood is the stable balance. Personal current accounts are the stickiest: 90% stable in the book. Corporate call accounts, moved by treasurers watching rates, are 50%.

Core and non-core

Stable is not the same as fixed-rate. A savings balance can stay for years and still reprice every time Bank Rate moves: it stays because the bank keeps paying a market rate. For interest rate risk the useful part is the stable balance that also does not reprice, the core deposits. The book estimates it with a simple rule: core share = stable share × (1 − beta). A beta of 0.6 on instant-access savings means most of the stable money behaves like floating funding, so only 30% is core. The rest is non-core and counted as overnight.

The Basel standardised framework then caps the core share by category, because a bank's own estimate is exactly the kind of number that drifts in its favour. Operational corporate deposits would be 60% core by the rule, but wholesale deposits are capped at 50%, so the book uses 50%.

Deposit	Balance	Rate	Stable	Beta	Core share	Core	Average maturity
Personal current accounts	£22.0bn	0.50%	90%	0.10	81%	£17.82bn	2.5y (cap 5.0)
Instant-access savings	£10.0bn	3.00%	75%	0.60	30%	£3.00bn	1.5y (cap 4.5)
Operational corporate deposits	£5.0bn	1.00%	80%	0.25	50% (cap 50%)	£2.50bn	1.5y (cap 4.0)
Corporate call accounts	£3.0bn	3.50%	50%	0.80	10%	£0.30bn	1.0y (cap 4.0)
Total	£40bn				59.1%	£23.62bn

Behavioural maturity and slotting

Core balances run off slowly. Slotting places them in the repricing buckets of L3.6.03: Repricing gap analysis as if they were a ladder of fixed-rate deposits. The book spreads each core balance in equal monthly cohorts: current accounts over 5 years (an average maturity of 2.5 years, against a cap of 5), savings and operational deposits over three, call accounts over two. Non-core sits in the overnight bucket. This is the behavioural view L3.5.04: Cash flow ladders and maturity gaps used for liquidity; here it sets when the deposits reprice, not when they leave.

Sort itWhich Basel category is it?

The caps depend on who holds the deposit and how the account is used. Put each account in its category.

A personal current account with a salary paid in every month
A pensioner's account that pays no interest and is rarely used
An instant-access savings account paying 3.0%
An online bonus saver, interest-bearing, used once or twice a year
A manufacturer's operational account that pays its suppliers
A partnership of solicitors' office account

What the assumptions do to the bank

The two clocks show up in the two lenses. In NII the beta matters: when Bank Rate rises, a deposit costs more only by beta × the rise, after the bank's 1-month lag. In EVE the slotting matters: a core balance slotted over years is a long fixed-rate liability, worth less to its holders when rates rise, which is a gain to the bank.

Deposit	Core	Year-1 NII, +100bp	ΔEVE, parallel up	ΔEVE, parallel down
Personal current accounts	£17.82bn	−£20.2m	+£1,031.3m	−£1,128.2m
Instant-access savings	£3.00bn	−£55.0m	+£115.4m	−£121.8m
Operational corporate deposits	£2.50bn	−£11.5m	+£93.8m	−£99.0m
Corporate call accounts	£0.30bn	−£22.0m	+£8.1m	−£8.4m
Total	£23.62bn	−£108.6m	+£1,248.6m

The deposits gain £1,248.6m of value in parallel up, most of it the current accounts. That gain is what offsets the fall in value of the fixed-rate mortgages and bonds. It is also why the bank invests £12bn of the core current accounts in the structural hedge of L3.6.10: Hedging the banking book: the hedge turns that behavioural maturity into fixed-rate income.

Three ways deposit assumptions go wrong

The formal version

For deposit category $i$ with balance $D_i$ , stable share $s_i$ , beta $\beta_i$ , Basel cap $c_i$ and core slotted in $n = 12Y_i$ equal monthly cohorts over $Y_i$ years:

\begin{aligned} C_i &= \min\!\big(s_i(1-\beta_i),\; c_i\big)\, D_i, \qquad N_i = D_i - C_i \\ \bar m_i &= \frac{1}{n}\sum_{k=1}^{n} \frac{k - \tfrac12}{12} = \frac{1}{12n}\cdot\frac{n^2}{2} = \frac{n}{24} = \frac{Y_i}{2} \;\le\; \bar m_i^{\text{cap}} \\ \Delta \text{NII}_i &= -D_i\,\big(\max(0,\, r_i + \beta_i\,\Delta r) - r_i\big)\,(1 - \ell) \\ \Delta \text{EVE}_i &= -\sum_{k} \text{CF}_{ik}\Big(e^{-(z(t_k) + \Delta z(t_k))\,t_k} - e^{-z(t_k)\,t_k}\Big) \end{aligned}

$C_i$ is core and $N_i$ non-core (overnight); $\bar m_i$ the average maturity of the slotting; $\ell$ the lag before the bank moves its rate (1 month); $\text{CF}_{ik}$ the deposit's slotted cash flows of interest and principal, discounted on the SONIA zero curve $z$ today and after the shock $\Delta z$ (L3.6.05: Economic value of equity sensitivity). The minus sign in front makes a fall in the liability's value a gain to the bank. For the current accounts:

\begin{aligned} C &= \min(0.9 \times 0.9,\; 0.9) \times 22{,}000 = \text{£}17{,}820\text{m} \\ \bar m &= 5/2 = 2.5 \text{ years} \le 5 \\ \Delta\text{NII}_{+100} &= -22{,}000 \times 0.1 \times 1\% \times \tfrac{11}{12} = -\text{£}20.17\text{m} \end{aligned}

Try it

Pick a deposit and change its three assumptions. The rest of the book stays as it is, so the outlier test shows what one modelling choice does to the whole bank.

Try itModel one deposit and see what it does to the bank

Deposit

Stable share90%

Deposit beta0.10

Core slotted over5.0 years

Core: slotted out over time£17.82bn
Non-core: treated as overnight£4.18bn
Personal current accounts, 0.50% today£22.00bn

Core still slottedMost the Basel caps allow (90% core, 5.0-year average)

Core share (cap 90%)81.0%

Average maturity (cap 5.0y)2.50 years

Core balance£17.82bn

Non-core, overnight£4.18bn

Year-1 NII, Bank Rate +100bp−£20.2m

Its ΔEVE, parallel up+£1,031.3m

Its ΔEVE, parallel down−£1,128.2m

Bank's worst ΔEVE−£257.2m (Parallel up)

Share of Tier 1 (outlier above 15%)7.35%

Room to the threshold£268m

Core share = stable share × (1 − beta), capped. Slot the current accounts over three years instead of five and the bank fails the outlier test in parallel up: the shorter the deposits look, the less they offset the fixed mortgages. Slot them over ten years (an average of five, still inside the cap) and the worst shock flips to parallel down, and the bank fails again from the other side. The assumption, not the deposits, moves the number.

Worked example

Modelling £22bn of current accounts

From one balance to a slotted ladder

Step 1 of 6

1
Stable
Ten years of balances show 90% of the £22bn stays: £19.8bn stable, £2.2bn volatile.

What ALCO should ask

The deposit model moves the outlier test from 7.3% to 17.9% with one change of slotting. ALCO should see the result under the alternatives, not just the chosen one, and the model should be validated against how the deposits behaved when rates actually moved.

Regulatory lens

Basel. The 2016 IRRBB standard's standardised framework (now in the Basel Framework, SRP31) asks banks to separate NMDs by the nature of the deposit and depositor, split each category into stable and non-stable parts using observed volume changes over the past ten years, and identify core deposits as the stable part unlikely to reprice even when rates change a lot. Non-core deposits go in the overnight bucket. Core deposits are capped at 90% of retail transactional deposits with an average maturity of at most 5 years, 70% and 4.5 years for retail non-transactional, and 50% and 4 years for wholesale.

Europe. The EBA's 2022 guidelines on IRRBB and CSRBB cap the weighted average repricing date of modelled retail and non-financial wholesale deposits at five years, per currency, and say deposits from financial customers should not be modelled behaviourally, except operational deposits. The rules in detail come in M4.5: Liquidity and IRRBB regulation; how deposit models are built and tested comes in M6.7: Liquidity, ALM and behavioural models.

Spark Atoms

Spark Atom28 pence in the pound

In July 2023 the UK Financial Conduct Authority published a 14-point action plan on cash savings. It found that nine of the largest providers had passed on 28% of Bank Rate rises to easy-access savers between January 2022 and May 2023, against 51% for notice and fixed-term deposits. In this lesson's language, an easy-access beta of about 0.28. The regulator's concern was fair value to savers; ALM's concern is that a beta set partly by public pressure can move fast.

Spark AtomTen years of nothing happening

Basel asks for ten years of volume history to find the stable balance. A bank that measured it in 2021 looked back over a decade in which Bank Rate never rose above 0.75%. Depositors had no reason to move money for a better rate, so the history showed balances that never left. It said little about what would happen when Bank Rate reached 5%.

Key takeaways

A non-maturity deposit has two behaviours to model: how long the balance stays (volume) and how much of a rate change the bank passes on (the beta).
Core deposits are stable and unlikely to reprice; the book uses stable share × (1 − beta), capped by Basel at 90%, 70% and 50%. RiskAtoms Bank's core is £23.62bn of £40bn.
Non-core deposits are overnight; core is slotted into repricing buckets with an average maturity inside the caps (5, 4.5 and 4 years).
The beta drives NII; the slotting drives EVE. Slotted, the deposits gain £1,249m of value in parallel up, offsetting the fixed-rate assets.
The assumptions decide the answer: all overnight, the worst ΔEVE is 42.9% of Tier 1; slotted as the book does, 7.35%.

Reaction

Check what stuck

Question 1 · objective 2

Retail instant-access savings of £10bn: 80% is stable and the deposit beta is 0.4. With core share = stable share × (1 − beta), capped at 70%, how much is core, in £m?

£m

Question 2 · objective 2

Wholesale (corporate) deposits: 90% stable, beta 0.2. What core share does the standardised framework allow, in %?

%

Question 3 · objective 3

A bank slots its core retail current accounts in equal monthly cohorts over six years. What is their average maturity, and does it fit the Basel cap?

ASix years, above the five-year capBThree years, inside the five-year capCThree years, above the 2.5-year capDSix years, inside the cap because current accounts are transactional

Question 4 · objective 1

£5bn of savings with a beta of 0.3. Bank Rate rises 200bp and the bank moves its rate a month later. By how much does year-1 net interest income change, in £m? (Use a minus sign for a fall.)

£m

Question 5 · objective 4

Why does treating every non-maturity deposit as overnight make RiskAtoms Bank look far riskier in the parallel up shock?

ABecause overnight deposits pay more interestBBecause it removes the deposits' value gain when rates rise, which offsets the value lost on fixed-rate mortgages and bondsCBecause the Basel caps forbid overnight depositsDBecause NII falls when deposits are overnight

Glossary

Non-maturity deposit: Non-maturity deposit (NMD): a deposit the customer can withdraw at any time, with a rate the bank sets and can change, such as a current account or an instant-access savings account.
Administered rate: Administered rate: a rate the bank sets itself and changes when it chooses, usually some time after the market moves, rather than one tied to an index by contract.
Stable balance: Stable balance: the part of a deposit category that history shows stays with the bank with a high degree of likelihood. Basel asks for ten years of volume history.
Deposit beta: Deposit beta: the share of a change in the market rate that the bank passes on to the deposit rate. A beta of 0.1 means a 1 percentage point rise in Bank Rate raises the deposit rate by 0.1 points.
Core deposits: Core deposits: the stable balances that are also unlikely to reprice, even when rates move a lot. They behave like long, fixed-rate funding.
Non-core deposits: Non-core deposits: everything that is not core. The standardised framework treats them as overnight.
Slotting: Slotting: placing a deposit's balances into repricing time buckets according to how it behaves, not its contractual maturity of one day.
Transactional account: Transactional account: a retail account with regular transactions, such as a salary paid in, or one that pays no interest. Other retail accounts are non-transactional.
Zero floor: Zero floor: a customer deposit rate cannot go below zero, so in a large rate fall the bank cannot cut it by the full beta.

Deep Dives & Origins

Deep Dives

Basel Committee on Banking Supervision, Interest rate risk in the banking book, Standards, April 2016, paragraphs 110 to 115 (non-maturity deposits) and Table 2. Intermediate.
European Banking Authority, Guidelines on IRRBB and CSRBB (EBA/GL/2022/14), October 2022, section 4.3.5 on IRRBB measurement assumptions, paragraphs 110 to 112 on deposits. Advanced.

Origins

Basel Committee on Banking Supervision, Interest rate risk in the banking book, Standards (d368), Bank for International Settlements, April 2016.
Basel Committee on Banking Supervision, Basel Framework, SRP31 Interest rate risk in the banking book, version in force from 1 January 2026.
European Banking Authority, Guidelines on IRRBB and CSRBB, EBA/GL/2022/14, 20 October 2022, paragraphs 110 to 112.
Financial Conduct Authority, FCA sets out 14-point action plan on cash savings, press release, 31 July 2023.
Board of Governors of the Federal Reserve System, Review of the Federal Reserve's Supervision and Regulation of Silicon Valley Bank, April 2023.

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.