Government debt is catching up with Europe’s banks

Reuters’ closing report for 7 October put European bank shares 3.3% lower, compared with a 1% decline in the STOXX 600. The sell-off brought several balance-sheet questions into focus: what happens to existing bonds, how quickly funding becomes more expensive, what hedges actually protect, and whether borrowers can withstand the shock. Banks still have substantial buffers. The changing composition of those buffers helps explain why government debt is troubling shareholders again. ²

Documentary investigation, information checked through 8 October 2026. Available banking disclosures mainly describe end-2025 and June 2026. Illustrative calculations are identified separately from market observations and reported bank data.

There is an apparent puzzle at the heart of the story. Higher interest rates should allow banks to charge more for lending. Why sell their shares when bond yields rise? The answer lies on both sides of the balance sheet. Banks earn interest, but they also fund their assets, hedge their exposures and provide collateral. A rate increase changes several prices on different schedules. Its overall effect depends on the contracts already in place and on what caused money to become more expensive.

Reuters’ intraday report on 7 October recorded a 3.5% fall in the banking index, with Société Générale, Deutsche Bank, UniCredit and Intesa Sanpaolo each down more than 4%. The closing report put the sector’s decline at 3.3%. These observations belong to different moments. They establish a sharp share-price reaction, without measuring the losses recognised in banks’ accounts that day. ^{1, 2}

Understanding the episode requires following the shock through the system. It begins in energy and government bond markets, passes through portfolios often treated as conservative, reaches banks’ own funding and can eventually affect a mortgage offer or a company’s ability to service its debt. One particularly revealing change has occurred in an unglamorous part of bank treasury operations: liquidity buffers now contain more sovereign exposure.

When higher rates become an unwelcome development

A bond yield is the return an investor requires at the current market price, assuming the payments used in the calculation arrive as scheduled. A central bank policy rate is a different price, used to steer short-term monetary conditions. Bond yields incorporate expectations about future policy, inflation, compensation for committing money over time, and credit and liquidity premiums associated with a particular issuer. ^{5, 20}

When yields rise alongside stronger economic activity, banks may enjoy greater demand for credit and healthier borrowers. An energy-driven increase brings a less comfortable combination. More expensive oil takes spending power from households, squeezes some businesses and can keep inflation elevated. Expected banking revenues become less reliable just as part of the funding base becomes more costly. Higher lending rates then have to compensate for a wider range of risks.

Monetary conditions had already tightened before the trading session in question. On 10 September, the ECB raised its policy rates by 25 basis points. A basis point is one hundredth of a percentage point. That decision and the market repricing on 7 October are separate events, with different transmission schedules. Treating them as a single interest-rate shock would obscure which contracts actually reset and when. ¹⁸

The level of a government bond yield also needs to be distinguished from its spread over another issuer. A similar rise in French and German yields points towards a common component. If the French yield rises faster, its spread over the Bund widens. The spread captures the relative price investors demand; it cannot by itself be converted into a default probability. Differences in market liquidity and demand for safe assets can also move the comparison. ⁵

The difficult combination for a bank shareholder is falling values on existing bonds, more expensive funding and a weaker outlook for borrowers. Investors may lower their expectations for future earnings while requiring a higher return to hold bank equity. The share price can adjust before the quarter’s accounts are prepared. That interpretation fits the concerns reported on 7 October. Determining how much of the decline each channel caused would require a separate empirical analysis, beyond the evidence in a market news report.

A different kind of liquidity cushion

The ECB’s May 2026 Financial Stability Review describes banks replacing some central bank reserves with sovereign bonds, predominantly issued in the euro area. The overall liquidity cushion can remain large while becoming more sensitive to market prices. The same review points to declining domestic sovereign concentration in most countries, a development that reduces one potential source of vulnerability. Both changes matter. ⁴

The EBA dashboard released on 25 September provides a more recent observation for its EU/EEA sample: sovereign bond holdings within high-quality liquid assets increased by 8.7% in the first half of 2026, while cash balances declined. This extends the evidence of a changing liquidity-buffer composition. It measures neither portfolios’ net sensitivity nor losses from the October trading session. ²⁴

The substitution has an understandable commercial purpose. A central bank reserve balance can settle a payment immediately. A government bond pays a coupon, can be sold and can be pledged as collateral. Some sovereign securities qualify among the most liquid assets under prudential rules. Yet their market value falls when the yield demanded by buyers rises. They can serve a liquidity purpose while carrying duration and valuation risk. ^{4, 8}

Within its EU/EEA sample, the EBA reports that on-balance-sheet sovereign exposures, comprising loans and debt securities, increased from approximately €3.65 trillion at the end of 2024 to €4.18 trillion at the end of 2025. Relative to Common Equity Tier 1 capital, the exposure rose from about 206% to 232%. CET1 is the highest-quality core of regulatory capital. This comparison divides an exposure by a capital stock; it is neither a loss estimate nor an immediate recapitalisation requirement. ³

“Sovereign exposures” is broader than a portfolio of publicly traded French bonds. Multiplying the aggregate by the price decline of a single ten-year security would produce a misleading sector loss estimate. Maturities, issuers, currencies and hedges vary. Accounting classifications differ too. The growing stock identifies a more important transmission channel, while leaving the net sensitivity to be established.

Treasury teams have to combine ready access to money, income generation and control of risk. Holding only very short-term reserves avoids much of the market-price volatility of long bonds, but leaves income exposed to the next short-rate adjustment. Buying longer securities can lock in contractual income while making their market value more sensitive. Each choice stabilises one dimension and creates a commitment in another.

This is why the size of a liquidity buffer is only the beginning of the analysis. Its contents, the amount already pledged and the stability of the funding it supports determine how much protection it provides when conditions change. Two similarly sized cushions can perform differently under the same market move.

How a yield change reprices a bond

For annual payments, the pricing formula is P = Σ(C / (1 + y)^t) + N / (1 + y)^n: C is the coupon, N the redemption amount, n the number of years and y the yield expressed as a fraction. The example below uses C = 3 and N = 100.

Consider an illustrative bond bought for €100, paying a fixed annual coupon of €3 and repaying its €100 principal in ten years. At a required yield of 3%, it is worth €100. If the required yield immediately rises to 4.5%, the same future payments are worth approximately €88.13. That is an 11.87% price decline. The calculation discounts every coupon and the principal repayment, excluding tax, default, fees and hedging.

Nothing about the contractual payment schedule has changed. A new buyer wants a better return and can achieve it with the old coupon only by paying a lower price. The original holder can still receive the scheduled payments if the issuer honours its obligations. Selling immediately, however, crystallises the difference between the purchase price and what the market will now pay.

Time magnifies the effect. With the same coupon and yield change, a two-year bond is worth about €97.19, while a twenty-year bond is worth €80.49. Distant payments are more affected by an increase in the rate used to discount them. Bond investors summarise this price sensitivity using duration. Maturity is one influence on duration, alongside the timing and size of payments; a ten-year maturity does not automatically mean a duration of ten years.

The example explains how a government can remain fully capable of paying its interest while its creditors incur a mark-to-market loss. It also explains why the government’s marginal borrowing cost differs from the average cost of its outstanding debt. Existing fixed coupons survive. New issuance and refinancing gradually bring higher market rates into the interest bill.

The bank holding the bond faces the same timing problem. Its contractual income may be unchanged, but the asset exchanges for less cash. The ability to wait for repayment therefore has economic value. That ability depends on stable funding, foreseeable cash needs and hedging arrangements, as well as the borrower’s creditworthiness.

A price fall should consequently prompt two different questions: what income is still due, and how much cash can be raised against the asset today? Focusing on only one gives an incomplete account of the risk.

Where the loss appears in the accounts

The €88.13 figure is a market valuation. Its accounting consequences depend on the bond’s classification. For debt instruments, IFRS 9 distinguishes amortised cost, fair value through other comprehensive income and fair value through profit or loss. Classification follows the business model and the contractual cash-flow characteristics. A bank cannot simply switch categories to avoid an inconvenient loss when the market turns. ^{6, 7}

At amortised cost, the carrying amount follows contractual cash flows and the effective interest method, with allowances for expected credit losses. A market-yield change alone does not produce the same accounting revaluation as a fair-value portfolio. Selling the asset can crystallise a loss, while deterioration in repayment prospects can require impairment. ^{6, 21}

For debt at fair value through other comprehensive income, or FVOCI, the relevant valuation movement passes through a separate component of comprehensive income and equity. Interest income and expected credit losses follow their own recognition rules. Accumulated amounts are recycled on disposal under the applicable rules for debt instruments. Prudential effects also depend on hedges, tax and regulatory adjustments, so a gross portfolio markdown is not a direct measure of the change in the CET1 ratio. ^{6, 7}

At fair value through profit or loss, or FVPL, valuation movements enter earnings. A hedged portfolio may have offsetting movements on derivatives. The positions and their accounting treatment need to be considered together rather than presenting the decline in a single asset line as the group’s final loss. ^{6, 19}

The EBA reports that 59.5% of sovereign exposures at end-2025 were held at amortised cost. Another 17.1% were measured through profit or loss and 20.2% through OCI. These categories, as cited in the report, do not sum precisely to 100%; the figures in this investigation do not allocate the residual or silently normalise the percentages. ³

Accounting classification supplies important information while leaving the liquidity question intact. A bond with a lower market price raises less cash even if its carrying amount changes little. Conversely, a recognised fair-value loss can coexist with funding stable enough to avoid a forced sale. The distinction becomes useful when assets, liabilities and their maturities are examined as a single financial structure.

That structure also determines whether an apparently patient investment strategy is genuinely affordable. Intending to hold a bond to maturity is different from having the resources to do so through a period of stress. The relevant evidence lies in funding contracts, available collateral and cash-flow assumptions, not simply in a portfolio label.

Collateral is assessed at today’s price

Banks can borrow against bonds. The lender applies a haircut to the asset’s valuation to establish a prudent collateral amount. The Eurosystem’s framework uses daily valuation and allows additional collateral or cash to be required when the haircut-adjusted value becomes insufficient. This is one route by which a market-price movement turns into a liquidity requirement. ^{8, 9}

Return to the hypothetical bond and apply an unchanged 5% haircut, chosen solely for illustration. At a price of €100, it supports €95 of funding. At €88.13, it supports approximately €83.72. Maintaining the initial borrowing requires about €11.28 of additional haircut-adjusted collateral value, subject to the terms of the transaction. The lender has not increased the haircut; the lower market price alone creates the shortfall.

This arithmetic does not describe an identified margin call on 7 October. It shows what a sustained market disturbance would make important: assets already pledged, spare collateral, financing agreements and safety margins. A large liquid portfolio can provide valuable protection. Its effectiveness depends in part on how much remains available rather than already supporting another obligation.

Derivatives introduce another set of cash flows. A hedge can gain value as rates rise and offset a bond markdown. Depending on the direction of the positions and the agreements involved, collateral exchanges can either supply or consume liquidity. Assuming that every rate increase necessarily creates adverse margin calls for banks would ignore the purpose and direction of their hedges. ^{19, 20}

The funding bill arrives in instalments

Banks obtain funding from depositors and financial markets. Bond issuance is priced using a reference curve and compensation for the risk of the bank and the particular instrument. A covered bond, a senior unsecured obligation and a subordinated security expose investors to different risks. Sovereign stress can influence these funding conditions through banks’ asset holdings, ratings, economic environment and the credibility of public support. ^{5, 22}

A French bank’s borrowing rate is therefore not mechanically the OAT yield plus a fixed number. Swap rates are also used as benchmarks, and each issuer has its own funding structure. The bank can nevertheless face a higher reference curve and a wider credit spread at the same time. Sovereign yields become a warning about its financing environment without supplying a universal pricing formula.

The maturity schedule prevents the new price being applied instantly to the whole debt stock. Imagine a bank with €10 billion of fixed-rate bonds outstanding. It refinances one quarter of that amount at a rate one percentage point higher. The €2.5 billion renewed generates an additional €25 million of annual interest expense at a full-year run rate. In this example, the remaining €7.5 billion keeps its contractual coupon until its own maturity.

The €25 million figure is an annual rate after refinancing, rather than necessarily the expense recognised in the first financial year. That depends on the issuance date. The example excludes changes in balance-sheet size, hedging, tax and deposit pricing. Its purpose is to show why the same funding stock can be manageable or troublesome depending on the distribution of maturities.

Prefunding changes that exposure. In its 30 July fixed-income presentation, Deutsche Bank reported €9 billion of issuance completed by 17 July against an annual programme of €10 billion to €15 billion. Part of the year’s funding had consequently been raised before October’s turbulence. The figure neither prices the next issue nor covers every future requirement. It demonstrates why the date attached to a funding plan matters. ¹²

In its update published on 30 September, the bank put year-to-date issuance of funding instruments at approximately €13 billion. This more recent figure complements the July snapshot; it cannot establish the remaining funding need or the cost of the next issue. ²⁵

Deposits add another moving part. Stable customers can reduce reliance on market funding, but higher rates on savings accounts or term deposits compress margins. New loans and reinvestments can earn more at the same time. Net interest income depends on the relative speed of these adjustments and on the volume of lending customers can actually sustain.

There is consequently no single moment at which “higher rates” arrive on a bank’s income statement. Some liabilities reset promptly, others at maturity. Some loans float, others retain an old coupon. Hedging adds another schedule. A careful assessment follows those schedules instead of multiplying a headline rate change by the entire balance sheet.

A hedge protects a particular exposure

Asset-liability management tracks interest rate risk in the banking book, or IRRBB. Net interest income, or NII, is interest income less interest expense over a period. Economic value of equity, or EVE, measures the net present value of cash flows from assets, liabilities and off-balance-sheet positions under the model’s assumptions. The two measures answer questions on different horizons. ²⁰

Hedges reshape exposure, subject to limits associated with the reference curve and customer behaviour. A swap-linked hedge may cushion a broad rate move without fully neutralising a premium specific to one sovereign. A deposit legally withdrawable on demand may behave like stable funding in practice, but the assumed stability needs to be tested rather than taken for granted. ^{19, 20}

Deutsche Bank describes investing and hedging certain rate-insensitive deposits through a rolling profile typically extending over ten years. This spreads the repricing of income through time. It is a concrete example of treasury management that looks quite different from a single bet on the next ECB decision. ¹¹

The bank’s June 2026 prudential disclosures show the modelled consequences under standardised scenarios. In the parallel upward-shift scenario, Deutsche Bank reports a €6,694 million decline in economic value and a €63 million decline in one-year net interest income. In the parallel downward-shift scenario, the measures are a €1,647 million increase in economic value and a €682 million decline in net interest income. ¹⁰

Those amounts must be interpreted separately. Economic value looks across future cash flows; the income measure covers a year. The regulatory calculations use specific conventions, including a static balance sheet and constant exchange rates for the reported income sensitivity. They are neither accounting losses from 7 October nor estimates confined to sovereign bonds. They also do not reproduce the illustrative 150-basis-point yield increase used earlier. ^{10, 23}

The useful lesson is about exposure rather than league tables. Short-term interest income can be relatively stable while longer-term economic value remains sensitive. Banks reporting strong income can still have very different maturity ladders, derivatives and concentrations. Their risk disclosures must be read alongside the assumptions, especially when the actual disturbance combines rates, credit spreads and a changing economic outlook.

A hedge is best judged against the risk it was designed to manage. Measuring it solely against the latest market narrative can make a sensible protection look ineffective, or conceal an exposure that the chosen instrument never covered.

Oil also reaches the loan file

Sovereign debt accounts for only part of a bank’s assets. An energy shock affects financed businesses as well. A transport company or manufacturer that cannot pass through higher input costs has less cash available to service its borrowing. Households may reduce other spending as their energy bill rises. Depending on the shock’s duration, loan demand can weaken while the risk on existing credit increases.

Charging a higher rate on a new loan produces a benefit only if the income is collected. Additional interest can be absorbed by more defaults, impairment or lower volumes. Expected credit losses are intended to recognise deterioration before the final default. The consequences emerge over time and vary across consumer credit, smaller businesses, commercial property and large corporate portfolios. ^{6, 16}

French housing finance illustrates the importance of contractual terms. Banque de France data published on 6 October show that 99.5% of new housing credit in August was fixed-rate, on a definition that includes renegotiations and excludes bridge loans. This is a share of the month’s loan production, calculated from non-seasonally adjusted data. It is not a measure of the entire mortgage stock. ¹⁴

For an existing fixed-rate borrower, a market-yield increase does not automatically change the scheduled instalment. The pressure shifts towards new buyers, potential refinancing and the bank’s margin as its resources renew. The borrower remains exposed through income and other spending. Job loss or a larger energy bill can undermine affordability even when the mortgage payment itself is unchanged.

The average rate on new housing loans excluding renegotiations reached 3.32% in August, following 3.27% in June and 3.30% in July. These rates exclude fees and insurance. They show gradual transmission already visible before October, rather than measuring the impact of the trading session being investigated. ¹⁴

Higher rates can also support reported earnings. Société Générale announced 14.9% year-on-year growth in second-quarter net interest income for its French Retail Banking, Private Banking and Insurance businesses. The issuer’s disclosure concerns that specific segment, rather than the entire group. It is a useful reminder that assets repricing over time can boost income, without supporting an extrapolation to every activity or future quarter. ¹³

The central issue is the balance of these effects. Fixed contracts redistribute the timing of the shock; they do not remove it from the financial system. Some borrowers gain protection, some prospective borrowers face reduced purchasing power, and banks manage the resulting mismatch through their funding and hedges.

Banks and governments enter the same feedback loop

Transmission becomes more concerning when it feeds back on itself. Sovereign stress can damage bank asset values and funding conditions. Banks that tighten lending can weaken activity. Lower tax receipts and higher public expenditure can then make government finances more difficult. The cycle can start with repricing and precautionary decisions, before a government has missed any payment. ⁵

Its strength depends on concentrations and alternatives. A bank may diversify its securities, hedge rates or find a different source of funding. It remains connected to its customers’ economy and its financing markets. The reduction in domestic sovereign concentration noted by the ECB is therefore a useful buffer, without making banks independent of their surroundings. ^{4, 5}

The ECB also has an instrument designed to counter certain unwarranted, disorderly market dynamics that threaten monetary-policy transmission: the Transmission Protection Instrument, or TPI. Activation is a Governing Council decision, with criteria concerning economic policy and debt sustainability. The framework distinguishes market dysfunction from persistent tensions driven by fundamentals. It supplies no automatic spread threshold that guarantees the value of bank bond portfolios. ¹⁷

That distinction matters. An available intervention tool can reduce some disorderly scenarios while leaving the economic consequences of a lasting fiscal problem intact. Bank investors have to consider the capacity to intervene, its conditions and each institution’s exposures together. A bad trading day is insufficient evidence for an inevitable banking crisis; equally, central bank instruments do not amount to unconditional protection from every sovereign repricing.

Substantial buffers, measured before the shock

The ECB’s latest supervisory statistics, released on 15 September, describe significant institutions under its direct supervision in the second quarter of 2026. Their aggregate CET1 ratio was 16.00%, the liquidity coverage ratio was 154.91%, and the non-performing loan ratio was 2.17%, using the measure excluding central bank cash balances and other demand deposits. These are starting conditions, reported with a lag, rather than a real-time account of October. ¹⁵

CET1 compares loss-absorbing capital with risk-weighted assets. A ratio of 16% does not mean that 16% of the unweighted balance sheet is available as cash. The liquidity coverage ratio compares liquid assets with net outflows under a thirty-day stress scenario. A high aggregate is valuable protection, but it does not describe how the assets are distributed across banks or how stable their market value will be. ^{15, 26}

Even a seemingly better asset-quality ratio requires checking both numerator and denominator. Between March and June, the amount of non-performing loans increased by 2.1%, while the loan denominator grew by 2.6%. The ratio edged down from 2.18% to 2.17%. The arithmetic explains how the proportion of a problem can fall while its absolute amount rises. ¹⁵

Stress tests provide another reference point. The 2025 EU-wide exercise covered 64 banks representing roughly three quarters of EU banking assets. Aggregate CET1 remained above 12% under its adverse scenario. This is evidence of loss-absorbing capacity within a hypothetical framework, whose starting balance sheets, shocks and conventions differ from October 2026. ¹⁶

Capital and liquidity levels therefore put limits on what can sensibly be inferred from an equity-index decline. They do not eliminate the need to inspect maturity schedules and concentrations. A strong aggregate may conceal a weaker institution. A bank can also remain solvent while its shares lose value because expected distributions or future profitability look less attractive.

What the next disclosures need to establish

Turning market concern into a banking diagnosis will require comparable observations. Interest-rate sensitivities must be matched with sovereign holdings, hedges and their dates. Completed bond issues need to be distinguished from announced funding programmes and secondary-market yields. Wider credit spreads, deposits moving between institutions and a concentrated refinancing schedule are different pieces of evidence. Their interaction matters more than any isolated headline.

On the lending side, weaker borrowers, migration towards riskier credit categories and impairment charges will reveal the economic transmission. Lower new lending can reflect reduced demand, tighter supply or both. Separating the effects of oil, interest rates and fiscal conditions will also be necessary: they can affect the same customers without sharing the same initial cause.

This investigation did not identify a disclosure that assigns a precise accounting loss to the 7 October session alone, or establishes a general increase in collateral haircuts that day. The available evidence documents a larger sovereign exposure, several transmission channels and substantial remaining buffers. The size and distribution of any net losses remain open questions.

The underlying economic connection is already visible. Public borrowing contributes to the pricing environment in which banks operate: the return on their securities, their capacity to pledge collateral, the cost of their resources and the health of their borrowers. When several of those conditions deteriorate together, the expected benefit from higher rates becomes harder to retain. The share-price decline is a reason to examine that interaction, including its delays and protections.

Scope and method

The European equity index, the EBA’s EU/EEA banking sample and ECB-supervised significant institutions cover different populations. They answer different questions here and their balance sheets are not combined. ECB quarterly comparisons use the single 15 September 2026 release, including its revisions. Corporate sources describe their own businesses; forecasts and simulations are identified accordingly. The bond-pricing, collateral and refinancing examples can be reproduced from the inputs and annual-discounting formula stated in the article. Yields, maturities, the haircut and the refinancing share are illustrative assumptions. The investigation makes no investment recommendation.

Reading balance sheets and collateral

The guide to bank health explains capital, liquidity and loan quality. Our investigation of the ECB’s collateral rules follows the eligibility of pledged assets. To distinguish a price change from a default and trace who absorbs losses, read Sovereign default: who pays?.

Sources and documents

1. S01 · Reuters · 2026-10-07European bank shares fall as bond yields surge, spreads widen. Intraday observation on 7 October; distinct from the closing report.
2. S02 · Reuters / AOL Canada · 2026-10-07Europe stocks retreat after three-day rise as rising yields, oil dent sentiment. Final update at 16:42 UTC; closing report for 7 October. Reuters republications are not independent confirmations.
3. S03 · European Banking Authority · 2026-06Risk Assessment Report, June 2026. End-2025 EU/EEA sample; on-balance-sheet sovereign exposures and accounting categories.
4. S04 · European Central Bank · 2026-05Financial Stability Review, May 2026, section 3.4 and Chart 3.11
5. S05 · Bank for International Settlements / CGFS · 2011-07-11The impact of sovereign credit risk on bank funding conditions, CGFS Papers No 43
6. S06 · IFRS Foundation · Accessed 2026-10-08IFRS 9 Financial Instruments
7. S07 · HM Revenue & Customs · Accessed 2026-10-08CFM21840: IFRS 9, classification of financial assets, tests
8. S08 · European Central Bank · Accessed 2026-10-08Risk mitigation in Eurosystem collateral operations
9. S09 · European Central Bank · Accessed 2026-10-08Risk control: haircuts and variation margins
10. S10 · Deutsche Bank · As of 2026-06-30 · accessed 2026-10-08Pillar 3 Report as of June 30, 2026, EU IRRBB1, printed page 117. EU IRRBB1 table, printed page 117; regulatory scenarios as of 30 June.
11. S11 · Deutsche Bank · 2026-07-29Interim Report as of June 30, 2026, NII methodology. Net interest income management and deposit hedging methodology.
12. S12 · Deutsche Bank · 2026-07-30Q2 2026 Fixed Income Conference Call, slide 14. Slide 14; issuance completed by 17 July and the announced annual programme.
13. S13 · Société Générale · 2026-07-30Second quarter and first half 2026 results. Net interest income for French Retail Banking, Private Banking and Insurance.
14. S14 · Banque de France · 2026-10-06Crédits aux particuliers, août 2026. August 2026 lending; rates exclude renegotiations, fixed-rate share includes them and excludes bridge loans.
15. S15 · European Central Bank Banking Supervision · 2026-09-15Supervisory banking statistics for significant institutions, second quarter 2026. Directly supervised significant institutions, Q2 2026; quarterly comparisons include revisions.
16. S16 · European Banking Authority · 2025-08-012025 EU-wide Stress Test: Results
17. S17 · European Central Bank · 2022-07-21The Transmission Protection Instrument
18. S18 · European Central Bank · 2026-09-10 · correction 2026-09-11Monetary policy decisions, 10 September 2026
19. S19 · European Central Bank · 2022-05Interest rate risk exposures and hedging of euro area banks’ banking books
20. S20 · Basel Committee / BIS · 2024-07-16 · effective 2026-01-01Application guidance on interest rate risk in the banking book
21. S21 · European Central Bank Banking Supervision · 2023-07-28Overall amount of unrealised losses in euro area banks’ bond portfolios contained
22. S22 · European Central Bank · 2025-05Financial Stability Review, May 2025, bank funding section 3.3
23. S23 · BIS Financial Stability Institute · 2017-06-24IRRBB: Pillar 2 standardised framework, Executive Summary
24. S24 · European Banking Authority · 2026-09-25EU/EEA banks display strength amid a challenging risk environment. High-quality liquid assets in the first half of 2026, EU/EEA sample; separate from ECB aggregates.
25. S25 · Deutsche Bank · 2026-09-30Key updates communicated during Q3 2026. Page 4: approximately €13 billion of funding instruments issued year to date, as of 30 September.
26. S26 · Basel Committee / BIS · 2022-12-08LCR20: High-quality liquid assets and net cash outflows. Sections 20.1, 20.4 and 20.5: liquid assets and net outflows over thirty days. Current version accessed on 8 October 2026.