---
title: "Liquidity and Funding Risk; Bank Treasury"
book: "Rates, Credit, XVA and Risk"
subject: quant
language: en
chapter: 24
exercises: 8
source: https://one-course.com/books/quant/6/en/chapter/24-liquidity-and-funding-risk-bank-treasury
---

# Chapter 24 — Liquidity and Funding Risk; Bank Treasury

On 8 March 2023 Silicon Valley Bank’s parent announced that it had sold 21 billion dollars of securities at an after-tax loss of 1.8 billion and would raise 2.25 billion of capital. The next day depositors withdrew more than 40 billion dollars, and the bank’s management expected more than 100 billion to follow on 10 March, about 85% of its deposits. The bank failed that day. It had invested its deposits, 94% of them uninsured, in long-dated agency mortgage securities held to maturity, with a duration of 6.2 years; when rates rose, their value fell and its depositors, a tightly connected network of technology firms and their investors, ran. Nothing in the story involves a trading desk: it is the market risk of a bank’s balance sheet and the liquidity risk of its funding, which are managed by its treasury. This chapter measures both: the regulatory liquidity ratios, the transfer prices that charge businesses for funding, the interest rate risk of the [banking book](https://one-course.com/books/quant/6/en/chapter/23-regulatory-capital-for-trading-books#def-rc-regulatory-capital-for-trading-books-books), and the behaviour of deposits.

## 24.1 Liquidity ratios

**Definition 24.1 (Liquidity coverage ratio, high-quality liquid assets).**

The *liquidity coverage ratio* (LCR) is the stock of *high-quality liquid assets* (HQLA) over the net cash outflows of a 30-day stress scenario; it must be at least 100%. HQLA are cash, central bank reserves and Level 1 securities in full, Level 2A assets (such as agency mortgage securities) after a 15% haircut, and Level 2 assets are capped at 40% of the stock; outflows apply run-off rates to each funding source, from 5% for insured stable retail deposits to 40% for non-operational corporate deposits.

**Definition 24.2 (Net stable funding ratio).**

The *net stable funding ratio* (NSFR) is available stable funding over required stable funding on a one-year horizon, weighted by how stable each liability is and how illiquid each asset is; it must be at least 100%.

**Example 24.3 (A stylised bank).**

An illustrative bank (USD billion) holds 15 of cash, 25 of five-year Treasuries, 90 of agency mortgage securities (a level-pay schedule over twelve years, duration 6.3 at purchase) and 70 of floating-rate loans, funded by 173 of deposits, 11 of term debt and 16 of equity. A year after rates rise from 1.0% to 4.5%, its HQLA at market value are 61.0: 36.6 of Level 1 (cash and Treasuries) and 24.4 of Level 2A, the MBS after the haircut and the cap. With 6% of its deposits insured and stable (run-off 5%) and 94% uninsured (40%), 30-day outflows are 65.6 and the LCR is 93%, below the minimum.

```python
def lcr(positions: list[Position], deposits: list[tuple[float, float]], market_rate: float,
        inflows: float = 0.0) -> dict:
    """HQLA at market value: Level 1 in full, Level 2A with a 15% haircut and capped at 40% of the stock;
    outflows: deposit amount x run-off; inflows capped at 75% of outflows."""
    l1 = sum(market_value(p, market_rate) for p in positions if p.hqla == "L1")
    l2 = 0.85 * sum(market_value(p, market_rate) for p in positions if p.hqla == "L2A")
    l2 = min(l2, 2.0 / 3.0 * l1)          # level 2 at most 40% of the stock: L2 <= 40/60 L1
    out = sum(a * r for a, r in deposits)
    net = out - min(inflows, 0.75 * out)
    return {"hqla": l1 + l2, "level1": l1, "level2": l2, "outflows": out, "lcr": (l1 + l2) / net}
```

***Listing 24.1.** The LCR: Level 1 and Level 2A assets with the haircut and the cap, against run-off-weighted outflows. code/firm/alm/firm_alm.py*

## 24.2 Funds-transfer pricing

**Definition 24.4 (Funds-transfer pricing).**

*Funds-transfer pricing* (FTP) charges each business that uses funding, and credits each that raises it, an internal rate: a base rate curve for the repricing tenor plus a liquidity premium for the term of the funding, so that a desk’s margin is measured after the cost of the money it uses and the treasury bears (and hedges) the interest rate and liquidity mismatch centrally.

**Example 24.5 (An FTP curve).**

With a base rate of 4.5% and an illustrative liquidity premium of $0.15\%\sqrt T$, a three-month loan is charged 4.58%, a five-year loan 4.84% and a ten-year one 4.97%; a deposit expected to stay five years is credited 4.84%. A business that funds long assets with short deposits sees the term premium in its charge instead of in the treasury’s losses.

## 24.3 Asset–liability management and rate risk in the banking book

**Definition 24.6 (Asset–liability management, IRRBB).**

*Asset–liability management* (ALM) manages the mismatch between a bank’s assets and liabilities in rates, maturities and currencies. Its central market risk is *interest-rate risk in the banking book* (IRRBB), measured in value and in earnings.

**Definition 24.7 (Economic value of equity, NII sensitivity).**

The *economic value of equity* (EVE) is the present value of the [banking book](https://one-course.com/books/quant/6/en/chapter/23-regulatory-capital-for-trading-books#def-rc-regulatory-capital-for-trading-books-books)’s asset cash flows less that of its liability cash flows (deposits slotted by behavioural models). *Net interest income sensitivity* is the change in the next year’s interest income less interest expense under a rate scenario, on a constant balance sheet.

The supervisory standard prescribes six instantaneous shocks per currency: parallel up and down, short rates up and down (largest at the short end, decaying as $e^{-t/4}$), a steepener and a flattener, sized for the US dollar at 200, 300 and 150 basis points for parallel, short and long moves ([Figure 24.1](#fig-rc-liquidity-and-funding-risk-bank-treasury-shocks); the chart of [Figure 24.2](#fig-rc-liquidity-and-funding-risk-bank-treasury-eve) marks up moves $+$ and down moves $-$). A bank whose largest EVE loss exceeds 15% of its Tier 1 capital is an outlier.

**As of September 2026 — The shock sizes.**

In July 2024 the Basel Committee recalibrated the shocks on data to December 2023, with local shock factors, a 99.9th percentile and rounding to 25 basis points, for implementation by 1 January 2026. For the US dollar the parallel and short shocks stay at 200 and 300 basis points and the long shock rises from 150 to 225. The chapter’s figures use the 2016 sizes, in force in 2022.

![The six supervisory interest rate shocks for the US dollar by tenor: parallel ±200 basis points, short-rate ±300 decaying with tenor, and the two rotations built from the short and long shocks. Data: Basel Committee IRRBB standard (2016); the chapter’s tutorial.](https://one-course.com/images/onecourse/chapters/quant-6/rc-liquidity-and-funding-risk-bank-treasury/fig-4e14891733b8.svg)

***Figure 24.1.** The six supervisory interest rate shocks for the US dollar by tenor: parallel $\pm200$ basis points, short-rate $\pm300$ decaying with tenor, and the two rotations built from the short and long shocks. Data: Basel Committee IRRBB standard (2016); the chapter’s tutorial.*

```python
def shock_curve(kind: str, sizes: dict = USD_SHOCKS) -> Callable[[float], float]:
    """Rate change at time t for one of the six supervisory scenarios."""
    def s_short(t):
        return math.exp(-t / 4.0)

    def d_short(t):
        return sizes["short"] * s_short(t)

    def d_long(t):
        return sizes["long"] * (1.0 - s_short(t))

    return {
        "parallel_up": lambda t: sizes["parallel"],
        "parallel_down": lambda t: -sizes["parallel"],
        "short_up": d_short,
        "short_down": lambda t: -d_short(t),
        "steepener": lambda t: -0.65 * abs(d_short(t)) + 0.9 * abs(d_long(t)),
        "flattener": lambda t: 0.8 * abs(d_short(t)) - 0.6 * abs(d_long(t)),
    }[kind]
```

***Listing 24.2.** The six supervisory interest rate shock scenarios. code/firm/alm/firm_alm.py*

**Example 24.8 (The stylised bank before the rise).**

At a 1% rate the bank’s EVE is 22.9. The parallel-up shock cuts it by 6.95, 43% of its Tier 1 capital of 16, nearly three times the outlier threshold; the steepener by 4.60 (6.79 with the 2024 sizes); the parallel-down shock raises it by 8.64 ([Figure 24.2](#fig-rc-liquidity-and-funding-risk-bank-treasury-eve)). The risk was measurable a year before it was realised.

![Change in the stylised bank’s economic value of equity under the six supervisory shocks, at a 1% rate. The dashed line is the outlier threshold, a loss of 15% of Tier 1 (2.4); the parallel-up and steepener scenarios cross it. Data: the chapter’s tutorial.](https://one-course.com/images/onecourse/chapters/quant-6/rc-liquidity-and-funding-risk-bank-treasury/fig-cf5ecaddc7a8.svg)

***Figure 24.2.** Change in the stylised bank’s [economic value of equity](#def-rc-liquidity-and-funding-risk-bank-treasury-eve) under the six supervisory shocks, at a 1% rate. The dashed line is the outlier threshold, a loss of 15% of Tier 1 (2.4); the parallel-up and steepener scenarios cross it. Data: the chapter’s tutorial.*

## 24.4 Deposits: the behavioural model

**Definition 24.9 (Non-maturity deposit, deposit beta).**

A *non-maturity deposit* can be withdrawn at any time and pays a rate the bank sets; a behavioural model splits it into a core part, expected to stay and slotted over several years, and a volatile part. Its *deposit beta* is the share of a change in market rates passed on to the deposit rate.

Deposit models turn a legally overnight liability into a long synthetic one: they justify investing deposits in long assets. If the beta is low, rising rates widen the margin and NII rises; if depositors demand a high beta, or leave, the margin reverses.

**Example 24.10 (Two betas).**

The stylised bank earns NII of 4.27 a year at a 1% rate. With a [deposit beta](#def-rc-liquidity-and-funding-risk-bank-treasury-nmd) of 0.35, a 200 basis point rise adds 0.49 a year; with a beta of 0.80, the same rise costs 1.07 ([Figure 24.3](#fig-rc-liquidity-and-funding-risk-bank-treasury-nii)). The fixed-rate securities earn the same in both cases; the difference is what the depositors take.

![One-year net interest income sensitivity of the stylised bank to a parallel rise in rates, for two deposit betas. The sign of the bank’s earnings exposure depends on how depositors behave. Data: the chapter’s tutorial.](https://one-course.com/images/onecourse/chapters/quant-6/rc-liquidity-and-funding-risk-bank-treasury/fig-eb0ad18ec715.svg)

***Figure 24.3.** One-year [net interest income sensitivity](#def-rc-liquidity-and-funding-risk-bank-treasury-eve) of the stylised bank to a parallel rise in rates, for two [deposit betas](#def-rc-liquidity-and-funding-risk-bank-treasury-nmd). The sign of the bank’s earnings exposure depends on how depositors behave. Data: the chapter’s tutorial.*

## 24.5 The 2023 regional-bank failures

**Definition 24.11 (Held-to-maturity portfolio).**

A *held-to-maturity portfolio* holds securities the bank intends and is able to keep until maturity, carried at amortised cost: their market losses do not appear in earnings or in regulatory capital. Selling some can force the whole portfolio to be marked.

**Definition 24.12 (Bank run).**

A *bank run* is a withdrawal of deposits faster than the bank can meet from its liquid assets, driven by depositors’ fear that others will withdraw first: individually rational, collectively self-fulfilling.

**Example 24.13 (The stylised bank a year later).**

After the 350 basis point rise, the MBS book held to maturity has lost 19.2% of its value, USD 17.2 billion, 108% of the bank’s equity; the Treasuries have lost 3.9. On the day, the bank can pay out its cash and sell its Treasuries at market: 36.1, 20.9% of its deposits ([Figure 24.4](#fig-rc-liquidity-and-funding-risk-bank-treasury-liquidity)). Anything more requires pledging or selling the MBS, which realises the loss.

![The stylised bank’s liquid resources a year after the rise, at market values, against its deposits. Only cash and Treasuries can be turned into payments on the day of a run: 20.9% of deposits. Data: the chapter’s tutorial.](https://one-course.com/images/onecourse/chapters/quant-6/rc-liquidity-and-funding-risk-bank-treasury/fig-5e20e7e28dcc.svg)

***Figure 24.4.** The stylised bank’s liquid resources a year after the rise, at market values, against its deposits. Only cash and Treasuries can be turned into payments on the day of a run: 20.9% of deposits. Data: the chapter’s tutorial.*

**As of September 2026 — March 2023.**

According to the Federal Reserve’s review of April 2023, Silicon Valley Bank lost over 40 billion dollars of deposits on 9 March 2023 and expected over 100 billion more on 10 March, roughly 85% of its deposit base; 94% of its deposits had been uninsured at the end of 2022, and its [held-to-maturity portfolio](#def-rc-liquidity-and-funding-risk-bank-treasury-htm) had a duration of 6.2 years. On 12 March the Federal Reserve created the Bank Term Funding Program, lending for up to a year against eligible securities valued at par, and Silicon Valley Bank and Signature Bank were resolved with all depositors protected. At the end of 2022 Silicon Valley Bank had reported over 15 billion dollars of unrealised losses on its held-to-maturity securities, 89% of its common equity tier 1 capital (US Government Accountability Office). First Republic Bank, with 229.1 billion dollars of assets, was closed on 1 May 2023 and its deposits and most assets sold to JPMorgan Chase; the FDIC estimated the cost to its insurance fund at about 13 billion dollars.

## 24.6 Tutorial: a balance sheet under shocks

**Goal.** Compute the stylised bank’s EVE and NII sensitivities under the supervisory shocks, its LCR, and the outflow its liquid assets can meet. **End state:** the numbers of Examples [24.3](#ex-rc-liquidity-and-funding-risk-bank-treasury-bank), [24.8](#ex-rc-liquidity-and-funding-risk-bank-treasury-eve), [24.10](#ex-rc-liquidity-and-funding-risk-bank-treasury-beta) and [24.13](#ex-rc-liquidity-and-funding-risk-bank-treasury-run) and the four charts.

1. **Balance sheet** : `positions()` , `deposits(beta)` , `bank(rate)` .
2. **Value and earnings** : `eve_table()` , `nii_table()` .
3. **Liquidity** : `lcr_now()` , `liquid_same_day()` , `htm_loss()` .
4. **Charts** : `fig_rc_alm.py` .

**What to change next.** Hedge the MBS with a pay-fixed swap of the same duration and rerun the EVE shocks; halve the core share of deposits and see the EVE change sign.

## 24.7 Build: the ALM engine

**Purpose.** The firm’s banking-book risk: EVE and NII under supervisory and internal scenarios, liquidity ratios, transfer prices, and run capacity; the balance-sheet side of the risk engine.

**Interface.** `shock_curve(kind)`, `SCENARIOS`; `Position`, `Deposits`, `BalanceSheet` with `pv`, `delta_eve`, `nii`; `market_value`; `lcr`; `ftp_rate`; `run_capacity`; `bond_price_change`, `duration`; `eve_ratio`.

**Rules.** Flat base curve with scenario shifts; annual cash flows; core deposits slotted in equal annual amounts; LCR with the Level 2 cap and the Level 2A haircut; outlier test at 15% of Tier 1.

**Acceptance tests.** `code/firm/alm/tests/`: shock shapes; bond price change against duration; a matched book has no EVE sensitivity; the Level 2 cap; the FTP identity; run capacity.

**Stretch.** A full yield curve with basis and optionality in loans and deposits; NSFR factors; dynamic balance sheets; intraday liquidity.

Sources and further reading

- Board of Governors of the Federal Reserve System, *Review of the Federal Reserve’s Supervision and Regulation of Silicon Valley Bank* , April 2023.
- Federal Reserve press release of 12 March 2023 on the Bank Term Funding Program.
- Basel Committee on Banking Supervision, *The Liquidity Coverage Ratio and liquidity risk monitoring tools* (2013); *The net stable funding ratio* (2014); *Interest rate risk in the banking book* (2016).

## 24.8 Exercises

**Exercise 24.1 ★.**

A bank has 30 of Level 1 assets and 40 of Level 2A assets before haircuts. What is its HQLA?

**Solution of Exercise 24.1.**

Level 2A after the haircut is $0.85\times40 = 34$, capped at $\tfrac23\times30 = 20$ (Level 2 at most 40% of the stock); HQLA $= 30+20 = 50$.

**Exercise 24.2 ★.**

Estimate the price fall of a portfolio with duration 6.2 when yields rise 350 basis points, and say why the exact figure differs.

**Solution of Exercise 24.2.**

$-6.2\times3.5\% = -21.7\%$ to first order; the exact fall is smaller (19.2% for the chapter’s MBS) because of positive convexity, and because the duration itself falls as yields rise.

**Exercise 24.3 ★.**

What is the short-rate shock at five years in US dollars?

**Solution of Exercise 24.3.**

$300\times e^{-5/4} = 86$ basis points.

**Exercise 24.4 ★★.**

Why does the stylised bank gain EVE when rates fall, and what does that say about its position?

**Solution of Exercise 24.4.**

Its assets are longer than its liabilities (long fixed-rate securities against deposits slotted over five years at most): it is long duration, like a receiver of fixed in a swap. It gains when rates fall and loses when they rise.

**Exercise 24.5 ★★.**

Why is a high share of uninsured deposits a liquidity risk and not only a credit issue for depositors?

**Solution of Exercise 24.5.**

Uninsured depositors lose money if the bank fails, so any doubt about its solvency gives each of them a reason to leave first; when they are few, large and connected, they leave together. Their run-off is fast and correlated, which is a liquidity risk to the bank whatever its capital.

**Exercise 24.6 ★★.**

How does an FTP curve change the incentive of a business that lends long?

**Solution of Exercise 24.6.**

The business is charged the term funding rate of its loans (4.84% for five years in the example), so a long loan earns only its credit and service margin; the term and liquidity premium goes to the treasury, which funds and hedges it. Without FTP the business would book the curve’s slope as profit.

**Exercise 24.7 ★★★.**

*Coding.* Compute the one-year NII change for a 350 basis point rise with [deposit betas](#def-rc-liquidity-and-funding-risk-bank-treasury-nmd) of 0.35 and 0.80.

**Solution of Exercise 24.7.**

$+0.86$ a year with a beta of 0.35, $-1.87$ with 0.80.

**Exercise 24.8 ★★★.**

*Find the flaw.* “Our securities are held to maturity and guaranteed by government agencies; rate rises cannot hurt us because we will never sell them.”

**Solution of Exercise 24.8.**

Holding to maturity avoids realising the loss only if the bank never needs to sell or pledge the securities; their economic value has fallen all the same, the bank earns below-market yields on them for years, and a deposit outflow can force a sale. Government guarantees remove credit risk, not interest rate or liquidity risk.

## 24.9 Problem: March 2023

**Problem 24.1.**

Weekend problem — a run on a solvent-looking bank

The stylised bank of this chapter is a year into the rate rise. Its HTM portfolio is carried at cost; its uninsured depositors are connected and follow the same news.

**Part I — The balance sheet.**

1. Give the unrealised loss on the MBS and its share of equity.
2. Give the loss on the Treasuries.
3. What would the bank’s equity be if both were marked?
4. Why did the supervisory EVE test already flag the risk at a 1% rate?
5. What hedge would have removed most of it, and at what cost to earnings?

**Part II — Liquidity.**

6. Give the LCR and why it is below 100%.
7. Give the outflow the bank can meet on the day.
8. Compare it with the first day of SVB’s run as a share of deposits.
9. What sources of funding could the bank use on day two?
10. How did the Bank Term Funding Program change the bank’s options?

**Part III — Behaviour.**

11. Why is a 94% uninsured deposit base different from a retail base of the same size?
12. How does a [deposit beta](#def-rc-liquidity-and-funding-risk-bank-treasury-nmd) rising with rates change NII and EVE?
13. What does announcing a securities sale and a capital raise signal to depositors?
14. Why can a run happen to a bank that is solvent at amortised cost?
15. Which ratio would have measured the concentration of depositors?

**Part IV — Judgement.**

16. What should the treasury have done in 2021?
17. Should held-to-maturity losses count in capital?
18. How should deposit models be validated?
19. State the *named result* : the held-to-maturity unrealised loss as a share of capital and the one-day outflow at which the [bank runs](#def-rc-liquidity-and-funding-risk-bank-treasury-run) out of liquid assets.
20. In one sentence: what killed the bank, market risk or liquidity risk?

**Solution of Problem 24.1.**

**1.** USD 17.2 billion, 108% of equity. **2.** USD 3.9 billion. **3.** $16-17.2-3.9 = -5.1$: negative. **4.** A 200 basis point rise already cut EVE by 6.95, 43% of Tier 1, far above the 15% outlier threshold: the duration mismatch was visible before any loss. **5.** Pay-fixed swaps of matching duration: EVE immune to parallel moves, but paying fixed at market rates on the notional reduced the carry of the securities to about the floating rate. **6.** 93%: the uninsured deposits carry a 40% run-off, and the MBS count only after the haircut and within the Level 2 cap. **7.** USD 36.1 billion, 20.9% of deposits. **8.** SVB lost over 40 billion dollars on 9 March, about a quarter of the roughly 165 billion implied by its management’s figures (140 billion being 85%); the stylised bank could not meet that. **9.** Repo of the Treasuries and MBS, the discount window against collateral at market value less haircuts, Federal Home Loan Bank advances; each takes time and marks the securities. **10.** It lent against securities at par, not market value, for up to a year: the bank could borrow on its MBS without realising the 19% loss. **11.** Retail deposits are insured, many and dispersed, and move slowly; uninsured corporate deposits are large, concentrated and informed, and move together. **12.** NII falls as deposit costs rise while fixed assets do not reprice; EVE falls further because the deposits are worth less as a cheap funding source. **13.** That the bank had to sell at a loss and needed capital: a signal of weakness that invited the run. **14.** Because the run concerns liquidity at market value: the depositors can be paid only from assets whose market value is below their book value, and each depositor fears being last. **15.** None of the standard ratios: concentration and connectedness of depositors need their own metrics (share of uninsured, largest depositors, sector concentration). **16.** Hedge the duration of the securities, keep them shorter, price deposits with a realistic beta, and hold liquidity for the uninsured base. **17.** At least in supervisory measures and [stress tests](https://one-course.com/books/quant/6/en/chapter/22-stress-testing-and-scenarios#def-rc-stress-testing-and-scenarios-stress): capital that ignores them overstates the loss-absorbing capacity in a run. **18.** Against history by cohort and rate cycle, against peers, with stress scenarios of rapid outflows, and by reviewing the concentration of the base. **19.** Named result: *March 2023*: the stylised bank’s HTM unrealised loss is 108% of its equity, and its liquid assets run out after a one-day outflow of 20.9% of deposits (USD 36.1 billion). **20.** Both: market risk made the bank insolvent at market value, and liquidity risk turned that into failure in a day.

## 24.10 Interview questions

**Interview question 24.1 ★ bank, risk.**

Define the LCR and the NSFR. What does each protect against?

**Solution of Interview question 24.1.**

LCR: HQLA against 30 days of stressed net outflows, a short-term survival buffer. NSFR: stable funding against the stable funding needs of assets over a year, a limit on maturity transformation.

*What the interviewer is looking for: the two horizons and what each constrains.*

**Interview question 24.2 ★★ bank.**

What is [funds-transfer pricing](#def-rc-liquidity-and-funding-risk-bank-treasury-ftp) and why do banks need it?

**Solution of Interview question 24.2.**

An internal price for funds by tenor (base curve plus liquidity premium) charged to asset businesses and credited to liability businesses; it measures margins net of funding, centralises rate and liquidity risk in treasury, and prices liquidity use.

*What the interviewer is looking for: the mechanism and the incentives.*

**Interview question 24.3 ★★ risk, researcher.**

EVE and NII sensitivity can have opposite signs. Explain and give an example.

**Solution of Interview question 24.3.**

EVE values all future cash flows; NII measures the next year. A bank with long fixed assets and sticky deposits loses EVE when rates rise (the assets fall) but gains NII for a while (floating assets reprice, deposits do not).

*What the interviewer is looking for: horizon difference and an example.*

**Interview question 24.4 ★★ researcher.**

How would you model [non-maturity deposits](#def-rc-liquidity-and-funding-risk-bank-treasury-nmd)?

**Solution of Interview question 24.4.**

Balances split into core and volatile, with decay rates estimated from history by segment; deposit rates as a function of market rates (beta, lags, floors); stress overlays for outflows; validation by cycle, including rate rises.

*What the interviewer is looking for: balances, rates and validation.*

**Interview question 24.5 ★★★ bank, risk.**

What went wrong at Silicon Valley Bank, in risk-management terms?

**Solution of Interview question 24.5.**

Long-duration securities funded by concentrated, uninsured deposits, unhedged rate risk, reliance on held-to-maturity accounting, weak liquidity stress testing, and a public announcement that triggered a coordinated run.

*What the interviewer is looking for: duration, funding concentration and governance.*

**Interview question 24.6 ★★★ trader, bank.**

How would you hedge a bank’s EVE exposure, and what does the hedge do to earnings and accounting?

**Solution of Interview question 24.6.**

Pay-fixed swaps against the securities or receive-floating on the deposits’ modelled duration; this reduces NII when curves are upward sloping, and needs hedge accounting to avoid volatility in reported earnings.

*What the interviewer is looking for: instrument, carry cost and accounting.*