7. Funding and Capital Management
Every loan a bank makes has to be paid for twice: once with funding (deposits, bonds and other liabilities that provide the cash) and once with capital (the loss-absorbing equity and subordinated debt that protects those funders if the loan goes bad). This chapter explains how treasury plans and sources funding, how regulatory capital is defined, measured and planned, and how the post-2008 resolution framework makes sure that failing banks are recapitalized by their own investors rather than by taxpayers. For engineers, these topics turn into issuance registers, RWA and capital engines, MREL eligibility logic and depositor-level data that regulators expect to be accurate, reconciled and producible at very short notice.
<a id="71-funding-strategy"></a>
7.1 Funding Strategy
This subchapter covers how treasury decides where the bank's money comes from: the mix of sources, the instruments and programs used to raise wholesale funding, the role of credit ratings, the movement of funds between group entities, and the annual funding plan that ties it all together.
Funding Mix and Diversification
Why it matters. A bank's assets (mortgages, corporate loans, bonds) are mostly long-term and illiquid, while its liabilities can often be withdrawn or not rolled over at short notice. If too much funding comes from one source, one investor type, one currency or one maturity date, a single shock can remove a large share of it at once. Northern Rock (2007) depended heavily on wholesale and securitization markets that closed overnight; Silicon Valley Bank (2023) depended on a concentrated, largely uninsured base of technology-sector deposits that left in two days. Both banks were solvent on paper when the funding disappeared.
What the funding mix is. The funding mix is the composition of the bank's liabilities and equity, viewed along several dimensions at the same time:
| Dimension | Examples | Concentration risk if unbalanced |
|---|---|---|
| Source / counterparty type | Retail, SME, corporate, financial institutions, central bank | One investor group leaves together |
| Instrument | Sight deposits, term deposits, CDs, CP, repo, senior bonds, covered bonds, securitizations | One market closes (for example securitization in 2007) |
| Secured vs unsecured | Repo, covered bonds vs senior unsecured | High asset encumbrance leaves little collateral for stress |
| Tenor | Overnight to 10+ years | Large amounts maturing in the same month (refinancing cliff) |
| Currency | EUR, USD, GBP | Assets in a currency with no stable funding in that currency |
| Geography / legal entity | Home market, foreign branches, subsidiaries | Funding trapped in, or dependent on, one jurisdiction |
How diversification is measured. Treasury monitors a set of structural indicators, usually with ALCO-approved limits:
- Loan-to-deposit ratio (LDR): customer loans / customer deposits. Above 100% means part of the loan book is funded by wholesale markets.
- Wholesale funding share: wholesale funding / total funding.
- Depositor and investor concentration: share of the top 10 or top 20 counterparties in total funding.
- Maturity concentration: largest amount of wholesale funding maturing within any single month or quarter.
- Asset encumbrance ratio: encumbered assets / total assets (assets pledged in repo, covered bond pools, central bank collateral).
- Currency funding gaps: assets minus stable funding per significant currency.
Example. A bank with EUR 100bn of liabilities and equity:
Funding source EUR bn Share
---------------------------------------------------------------
Retail deposits (insured) 42 ########################################## 42%
Corporate deposits 18 ################## 18%
Senior unsecured bonds 12 ############ 12%
Covered bonds 9 ######### 9%
Repo and interbank 7 ####### 7%
Commercial paper / CDs 4 #### 4%
Subordinated debt (AT1, T2) 2 ## 2%
Equity 6 ###### 6%
---------------------------------------------------------------
Total 100
With customer loans of EUR 66bn:
- LDR = 66 / (42 + 18) = 110%. EUR 6bn of loans rely on wholesale funding.
- Wholesale share (bonds, covered bonds, repo, CP/CDs) = (12 + 9 + 7 + 4) / 94 = 34% of non-equity funding.
- If EUR 5bn of the senior bonds mature in the same quarter, that single quarter holds 42% of the senior bond stack: a refinancing cliff that the funding plan should smooth by pre-funding or issuing longer tenors.
Analogy. A diversified funding mix is like a factory with several suppliers for each critical component. One supplier going bankrupt is an inconvenience, not a shutdown. A factory with one supplier gets a better price in normal times but stops production the day that supplier fails.
Key takeaways
- The funding mix is analyzed by source, instrument, secured/unsecured, tenor, currency and legal entity.
- Concentration in any one dimension creates a single point of failure for liquidity.
- Key indicators include LDR, wholesale share, top-N concentration, maturity concentration and asset encumbrance.
- Diversification costs money in normal times; it is a deliberate trade of cost for resilience.
🧪 Practice
- Beginner: A bank has customer loans of EUR 45bn and customer deposits of EUR 50bn. Compute the LDR and explain what it says about the need for wholesale funding.
- Intermediate: Using the example above, the corporate deposits of EUR 18bn come from 40 clients, and the top 5 clients hold EUR 9bn. Which indicator captures this, and what two actions could treasury take to reduce the risk?
- Interview: "You are asked to build a funding concentration dashboard. What data would you need, and what are the hardest data problems?" Hint: counterparty hierarchies (several accounts and legal entities belonging to one economic group), consistent maturity dates for non-maturity deposits, currency conversion, linking encumbrance to specific collateral, and daily versus month-end snapshots.
Wholesale vs Retail Funding
Why it matters. Not all money behaves the same way in a crisis. A household salary account with an insured balance rarely moves because of bad news about the bank; a money market fund holding the bank's commercial paper is required by its own rules to move at the first sign of credit risk. Treasury, rating agencies and regulators therefore treat these sources very differently, and the cost difference between them is the price of that stability.
What they are.
- Retail funding: deposits from individuals and small businesses: current (sight) accounts, savings accounts, term deposits. Usually covered by deposit insurance up to a limit (see 7.3), gathered through branches and digital channels.
- Wholesale funding: funding from professional counterparties: other banks, central banks, money market funds, asset managers, insurers, large corporates. Instruments include interbank deposits, certificates of deposit (CDs), commercial paper (CP), repo, senior unsecured bonds, covered bonds, securitizations and capital instruments.
Comparison.
| Feature | Retail funding | Wholesale funding |
|---|---|---|
| Typical cost | Low interest, but high operating cost (branches, IT, staff) | Market rate plus credit spread, low operating cost |
| Stability in stress | High, especially insured operational balances | Low for short-term unsecured; higher for long-term bonds |
| Speed of raising | Slow: grows with the franchise | Fast: billions in a day when markets are open |
| Control over volume | Low: depends on customer behavior and competition | High: bank chooses size, tenor, currency |
| Price sensitivity | Moderate, increasing with digital banking | High: priced off market curves daily |
| Information sensitivity | Low (insured depositors do not monitor the bank) | High (investors react to ratings, news, spreads) |
| LCR 30-day outflow assumption | 5% stable / 10%+ less stable | 25% (operational) to 100% (unsecured from financials) |
| NSFR available stable funding factor | 95% stable / 90% less stable | 0% (financials, under 6 months) to 100% (1 year and longer) |
Secured vs unsecured wholesale funding. Secured funding (repo, covered bonds, central bank operations) is cheaper and more resilient because investors are protected by collateral, but it encumbers assets: the pledged assets are no longer available to sell or pledge in a stress, and they are removed from the pool that unsecured creditors would recover from in insolvency. High encumbrance therefore raises the cost of unsecured funding.
Example: comparing all-in cost and stress behavior. A bank needs EUR 1bn of 3-year funding. The 3-year swap rate is 2.60%.
| Option | Pricing | All-in annual cost | Assumed 30-day stress outflow |
|---|---|---|---|
| 3Y retail term deposits | 2.40% to customer + 0.60% acquisition and servicing cost | 3.00% | Low (term deposits are locked or carry a penalty) |
| 3Y senior unsecured bond | swap + 0.90% | 3.50% | 0% (no maturity within 30 days) |
| 3Y covered bond | swap + 0.35% | 2.95% | 0%, but EUR 1.15bn of mortgages encumbered (115% overcollateralization) |
| Rolling 3M CP | 3M rate + 0.20%, refinanced 12 times | Variable | Up to 100% when it matures in a stress |
The cheapest option today (rolling CP) is the least reliable; the most stable long-term options cost more or consume collateral.
Analogy. Retail deposits are like a neighborhood shop's regular customers: each spends little, but together they are steady and loyal. Wholesale funding is like a single large corporate client: big orders arrive quickly, but the client has a procurement team that will switch suppliers the moment there is a better or safer offer.
Key takeaways
- Retail funding is sticky and information-insensitive; wholesale funding is fast and scalable but flighty.
- Stability is driven by deposit insurance, operational relationships, tenor and counterparty type, not only by the label "retail" or "wholesale".
- Regulatory metrics (LCR outflow rates, NSFR stable funding factors) encode these differences.
- Secured funding is cheaper but encumbers assets and worsens the position of unsecured creditors.
🧪 Practice
- Beginner: Classify each as retail or wholesale, and secured or unsecured: a household savings account, a repo with a pension fund, a 5-year covered bond, a 3-month CD bought by a money market fund.
- Intermediate: A bank holds EUR 10bn of stable retail deposits and EUR 4bn of unsecured deposits from other financial institutions maturing within 30 days. Using LCR outflow rates of 5% and 100%, compute the 30-day stressed outflow from each and explain the difference.
- Interview: "Silicon Valley Bank's funding was mostly deposits. Why did it behave like wholesale funding?" Hint: share of uninsured balances, concentration in one industry and one venture capital network, high information sensitivity, and mobile banking making withdrawals instant.
Debt Issuance Programs (EMTN)
Why it matters. Bond markets open and close with sentiment. A bank that needs several weeks to prepare legal documentation for each bond will miss the good days and be forced to issue on bad ones. Large banks issue dozens or hundreds of notes a year, in many currencies, sizes and structures. Writing a full prospectus for each one would be slow, expensive and error-prone.
What an EMTN program is. A Euro Medium Term Note (EMTN) program is an umbrella legal framework under which an issuer can issue many notes over time without redrafting the core documentation:
- The base prospectus describes the issuer, the risks and the general terms and conditions of all notes. In the EU it is approved by a regulator (for example in Luxembourg or Ireland) and is valid for 12 months, so it is updated annually, with supplements for material news (for example quarterly results).
- Each individual issue (a drawdown) only needs short Final Terms: amount, currency, coupon, maturity, issue price, ranking and listing.
- Programs are usually governed by English law (sometimes local law for specific instruments) and allow fixed, floating, zero-coupon and structured notes, senior preferred, senior non-preferred and subordinated ranks, and both public benchmark deals and private placements.
Related programs serve other needs:
| Program | Instruments | Typical tenor | Investors |
|---|---|---|---|
| EMTN | Senior and subordinated notes, structured notes | 1 to 30 years | Global institutional investors |
| US MTN (SEC-registered or 144A) | Notes for the US market | 1 to 30 years | US institutional investors |
| Covered bond program | Covered bonds secured on a cover pool | 3 to 15+ years | Banks, central banks, insurers |
| ECP / CD program | Euro commercial paper, certificates of deposit | Overnight to 1 year | Money market funds, corporates |
Parties involved.
| Party | Role |
|---|---|
| Issuer | The bank (or its holding company) raising funds |
| Arranger | Sets up and maintains the program documentation |
| Dealers | Banks allowed to place notes with investors; lead managers on syndicated deals |
| Fiscal / issuing and paying agent | Issues the notes into the clearing systems, pays coupons and principal |
| Trustee or fiscal agent | Represents noteholders' interests |
| Clearing systems | Euroclear and Clearstream hold and settle the notes |
| Rating agencies, lawyers, auditors | Rate the program and notes, draft documentation, provide comfort letters |
Lifecycle of a benchmark issue.
Annual program update Issue decision Execution day (T)
+-------------------+ +--------------------+ +------------------------------+
| Base prospectus |-----> | Funding plan slot, |-----> | Mandate announced, books |
| approved, ratings | | market window, | | open, guidance MS + 110 area,|
| confirmed | | dealers mandated | | books close, spread set at |
+-------------------+ +--------------------+ | MS + 85, coupon and price |
| fixed, allocation |
+---------------+--------------+
|
v
Lifecycle until maturity Settlement (T+5 typical) Final Terms signed and filed
+---------------------+ +---------------------+ +------------------------------+
| Coupons via paying |<----| Notes delivered in |<------| Swap executed to convert |
| agent, hedge swap | | Euroclear and | | fixed coupon to floating |
| accruals, buybacks, | | Clearstream, cash | | (see 6.3), booked in TMS |
| maturity repayment | | received | | |
+---------------------+ +---------------------+ +------------------------------+
Example. A bank issues a EUR 1bn 5-year senior preferred benchmark. The 5-year mid-swap (MS) rate is 2.60% and the final spread is MS + 85 bp, so the re-offer yield is 3.45%. The coupon is set at 3.375% (coupons are usually rounded down to the nearest 1/8%), so the bond is priced slightly below par:
Price = sum over t=1..5 of 3.375 / 1.0345^t + 100 / 1.0345^5
= 3.375 x 4.5215 + 84.401
= 15.260 + 84.401 = 99.661%
Gross proceeds = 1,000m x 99.661% = 996.61m
Underwriting fees (assume 0.20%) = -2.00m
Net proceeds = 994.61m
Treasury typically enters a 5-year receive-fixed / pay-floating swap at the same time, so the bank's effective funding cost becomes Euribor plus a spread, which matches the floating-rate profile of the assets it funds.
Engineering view. The treasury management system must store each issue with its legal and economic terms: ISIN, program, rank (needed for MREL, see 7.3), governing law, call dates, coupon schedule, linked hedge, and listing. Issuance data feeds the funding plan, liquidity reporting (maturity ladders), MREL eligibility, accounting (amortized cost or fair value option) and investor relations.
Analogy. An EMTN program is like a framework agreement with a supplier: the long contract is negotiated once a year, and each order afterwards is a one-page purchase order that refers to it.
Key takeaways
- An EMTN program lets a bank issue many notes quickly under one annually updated base prospectus.
- Each drawdown is documented by short Final Terms specifying amount, currency, coupon, maturity and rank.
- Benchmark bonds are priced as a spread over mid-swaps; the coupon and issue price are derived from the re-offer yield.
- Issues are usually swapped to floating, and their terms feed liquidity, MREL, accounting and reporting systems.
🧪 Practice
- Beginner: What is the difference between a base prospectus and Final Terms? Which one is updated annually?
- Intermediate: A 3-year EUR 500m note is priced at MS + 70 bp with 3-year mid-swaps at 2.50%. The coupon is set at 3.125%. Compute the re-offer yield and the issue price (annual coupons), and state whether it is above or below par.
- Interview: "Which fields of a bond issue record would you consider critical data, and which downstream systems depend on each?" Hint: rank and governing law (MREL, resolution data), maturity and call dates (liquidity ladders, MREL eligibility under one year), coupon schedule and day count (accruals, cash flows), hedge link (hedge accounting), currency (FX funding gaps).
Credit Ratings and Funding Costs
Why it matters. Wholesale investors cannot analyze every bank in depth, and many are not allowed to: their mandates often require an investment-grade rating. A bank's credit rating therefore decides who can buy its debt, how much it pays, and whether derivative counterparties and corporate clients will keep dealing with it. A downgrade can raise funding costs on every future issue and trigger immediate collateral calls on existing contracts.
What a credit rating is. A credit rating is an independent opinion on the likelihood that a borrower or a specific instrument will pay in full and on time. The three main agencies are S&P, Moody's and Fitch.
| Long-term rating (S&P / Fitch) | Moody's | Category |
|---|---|---|
| AAA | Aaa | Prime |
| AA+, AA, AA- | Aa1, Aa2, Aa3 | High grade |
| A+, A, A- | A1, A2, A3 | Upper medium grade |
| BBB+, BBB, BBB- | Baa1, Baa2, Baa3 | Lower medium grade (lowest investment grade) |
| BB+ and below | Ba1 and below | Sub-investment grade (high yield) |
Short-term ratings (A-1, P-1, F1 and so on) apply to CP and CDs and decide whether money market funds can hold them.
What drives a bank's rating.
- Standalone credit profile: asset quality, capital, profitability, funding and liquidity, business model, risk management.
- External support: parent support for subsidiaries; government support has been reduced in most ratings because resolution regimes make bail-in, not bailout, the expected outcome.
- Loss absorption below the instrument: agencies analyze how much subordinated debt and equity would absorb losses before a given instrument is hit (Moody's calls this loss-given-failure analysis).
- Sovereign and operating environment: a bank is rarely rated far above its home sovereign.
Instrument notching. Instruments of the same bank receive different ratings depending on their rank:
Covered bonds AAA (secured on a cover pool)
Deposits / senior pref. A+ (issuer rating anchor)
Senior non-preferred A- (absorbs losses before senior preferred)
Tier 2 BBB (subordinated)
AT1 BB+ (deeply subordinated, discretionary coupons)
How ratings translate into cost. The funding cost of a bond is:
Funding cost = risk-free / swap rate (same tenor) + credit spread
^
depends on rating, rank, tenor, market conditions, supply
Consequences of a downgrade.
| Channel | Effect |
|---|---|
| New issuance | Wider spreads on every future bond |
| Investor base | Funds with rating floors must sell or stop buying; short-term downgrade can shut the bank out of money market funds |
| Derivatives (CSA) | Rating-linked thresholds fall, requiring additional collateral; some contracts allow termination |
| Corporate deposits and clients | Rating-sensitive treasurers move deposits or stop using the bank as a counterparty |
| Liquidity regulation | The LCR requires banks to hold liquidity for collateral outflows from a downgrade of up to 3 notches |
Example. A bank issues EUR 5bn of 5-year senior debt each year. It is downgraded from A to BBB+ and its spread widens from 90 bp to 130 bp.
Extra cost per year of issuance = 5,000m x 0.40% = 20m per year
After 5 years (whole stack has rolled at the new spread):
outstanding at new spread = 5 x 5,000m = 25,000m
extra annual interest = 25,000m x 0.40% = 100m per year
In addition, on day 1:
CSA threshold at A = 50m, at BBB+ = 0
collateral call across 20 counterparties ~ 20 x 50m = 1,000m of liquidity
Analogy. A bank's rating works like a personal credit score. A lower score does not change the debts you already have, but every new loan costs more, some lenders refuse you, and certain contracts (such as a rental deposit) may ask for more security straight away.
Key takeaways
- Ratings decide which investors can buy a bank's debt and at what spread.
- Different instruments of the same bank are notched according to their rank in the creditor hierarchy.
- Funding cost = swap rate + credit spread; spreads widen with lower ratings and lower rank.
- A downgrade affects future funding costs and also creates immediate liquidity outflows through rating triggers.
🧪 Practice
- Beginner: Is BBB- investment grade? Is Ba1? Why does this boundary matter so much to a bank issuer?
- Intermediate: A bank's CSAs have the following thresholds: 100m at A-, 25m at BBB+, 0 at BBB. It currently posts no collateral with 12 counterparties, each with an exposure above 100m. Compute the collateral call for a 1-notch and a 2-notch downgrade from A-.
- Interview: "How would you design a system to calculate the liquidity impact of a rating downgrade?" Hint: a contract-level inventory of rating triggers (CSA thresholds, independent amounts, termination events, rating-linked deposit clauses), the current exposure per netting set, a rating scenario per agency, and aggregation by legal entity and currency.
Intragroup Funding
Why it matters. Banking groups are not single companies: they are made of a parent (often a holding company), operating banks, foreign subsidiaries, branches and special-purpose entities. Capital and liquidity are regulated per legal entity, and host regulators want to protect local depositors. As a result, cash that is surplus in one entity cannot always be used where it is needed (it may be trapped), and the way funds move inside the group must be documented, priced at arm's length and recorded in both entities.
Funding models.
| Model | How it works | Advantages | Disadvantages |
|---|---|---|---|
| Centralized | Parent or main operating bank raises wholesale funding and lends it to subsidiaries | One strong name in the market, economies of scale, better pricing | Subsidiaries depend on parent; host regulators may resist |
| Decentralized (self-funded subsidiaries) | Each subsidiary funds itself locally, mainly with local deposits | Local resilience, aligns with multiple-point-of-entry resolution, fewer intragroup exposures | Higher cost, smaller issues, duplicated treasury capability |
| Hybrid | Local deposits plus parent support for long-term or foreign-currency needs | Balances cost and resilience | More complex governance and pricing |
Intragroup instruments.
- Intragroup loans and deposits (term or revolving).
- Cash pooling between entities in the same country and currency.
- Intragroup repo and collateral transfers.
- Capital injections and subordinated loans (including internal MREL, see 7.3).
- Guarantees and liquidity facilities.
Constraints.
| Constraint | Effect on intragroup funding |
|---|---|
| Large exposure limits (25% of Tier 1 capital per counterparty group) | Limit how much a subsidiary can lend to its parent, unless an exemption applies |
| Local liquidity requirements (LCR, NSFR per entity) | Subsidiary must keep its own buffer; surplus may not be counted elsewhere |
| Ring-fencing (for example UK retail ring-fenced banks) | Restricts funding flows between the ring-fenced bank and the investment bank |
| Transfer pricing (tax, OECD arm's length principle) | Intragroup rates must match what independent parties would charge |
| Resolution planning | Funding should be structured so that losses can be passed up to the resolution entity |
Example.
+-------------------------------+
Investors -------> | HoldCo (resolution entity) | issues 3Y senior at swap + 0.90%
(bonds) +---------------+---------------+
| internal loan at swap + 0.95%
| (arm's length: HoldCo cost + 5 bp)
v
+-------------------------------+
| OpCo Bank (EUR) | -----> lends to subsidiary
+---------------+---------------+ in USD via FX swap
|
v
+-------------------------------+
| Subsidiary (USD) | standalone cost would be
| local deposits + group loan | USD swap + 1.50%
+-------------------------------+
The subsidiary saves 55 bp by borrowing from the group, but the group's exposure to it rises, and the subsidiary's regulator may require that a minimum share of its funding be local or long-term.
Engineering view. Every intragroup deal must be booked in both entities with a shared identifier, the same economic terms and opposite signs. Consolidated reports eliminate these positions, so mismatches (one side booked late, different FX rates, different maturity dates) create consolidation breaks. Intercompany reconciliation is a standard daily or monthly control.
Analogy. Intragroup funding is like family members lending to each other: it is cheaper and faster than going to a bank, but each household is still audited separately, may be told to keep its own emergency savings, and the tax office expects the interest rate to be a fair one.
Key takeaways
- Liquidity and capital are regulated per legal entity, so group surplus is not automatically available where needed.
- Groups choose between centralized, decentralized and hybrid funding models.
- Intragroup transactions are limited by large exposure rules, local requirements, ring-fencing, transfer pricing and resolution planning.
- Both legs of each intragroup deal must match and reconcile, because they are eliminated on consolidation.
🧪 Practice
- Beginner: Give two reasons why a regulator might stop a subsidiary from lending its surplus cash to the parent.
- Intermediate: A parent issues at swap + 0.80% and lends to a subsidiary. The subsidiary's standalone funding cost is swap + 1.40%. Propose an arm's length internal rate and explain how it could be justified to a tax authority.
- Interview: "Intercompany balances do not net to zero on consolidation. How would you investigate?" Hint: match by shared deal ID first, then compare booking dates, FX rates used for revaluation, accrual conventions and cut-off times between entities in different time zones.
Funding Plans and Budgeting
Why it matters. If treasury only reacts to funding needs as they arise, it is forced to issue whatever the market offers on that day, possibly during a market closure or just after bad news. Funding needs can be predicted from the business plan and the maturity profile, so banks plan them one to three years ahead. Supervisors require banks to submit funding plans and check that they are realistic.
What a funding plan is. A funding plan projects the balance sheet over the planning horizon (usually 3 years) and specifies how the resulting funding requirement will be covered: by instrument, currency, tenor and quarter, subject to constraints.
How it is built.
- Balance sheet projection: loan growth, deposit growth, securities portfolio and retained earnings from the business plan.
- Maturities: wholesale debt, covered bonds, central bank funding and capital instruments that mature or become callable.
- Regulatory needs: liquidity buffer size (LCR), stable funding (NSFR), MREL and capital instruments (see 7.2 and 7.3).
- Funding requirement: the gap between uses and sources.
- Allocation: split the requirement into instruments within constraints (investor capacity, maturity concentration limits, encumbrance limits, rating targets).
- Cost budget: projected spreads give the funding cost, which feeds the NII budget and the FTP liquidity curve (see 6.4).
- Stress and contingency: test the plan under adverse scenarios and link it to the contingency funding plan.
- Execution tracking: actual vs plan, updated as markets and the balance sheet change.
Example: year-1 funding requirement.
Uses of funds EUR bn
------------------------------------------------------
Loan growth 8.0
Wholesale debt maturing 6.0
Increase in liquidity buffer 1.0
------------------------------------------------------
Total uses 15.0
Sources from the business EUR bn
------------------------------------------------------
Deposit growth 4.0
Retained earnings 0.5
------------------------------------------------------
Total internal sources 4.5
Wholesale funding requirement = 15.0 - 4.5 = 10.5
Allocation, taking MREL and capital needs into account:
| Instrument | EUR bn | Assumed spread over swaps | Reason |
|---|---|---|---|
| Covered bonds | 4.0 | 35 bp | Cheapest long-term funding, limited by cover pool size |
| Senior preferred | 3.0 | 85 bp | Diversification, investor base |
| Senior non-preferred | 2.5 | 120 bp | Needed to meet MREL subordination |
| Tier 2 | 0.5 | 200 bp | Replace a Tier 2 bond losing capital recognition |
| AT1 | 0.5 | 400 bp | Fill the Tier 1 requirement |
| Total | 10.5 | ~95 bp weighted |
Weighted spread = (4.0 x 35 + 3.0 x 85 + 2.5 x 120 + 0.5 x 200 + 0.5 x 400) / 10.5
= (140 + 255 + 300 + 100 + 200) / 10.5
= 995 / 10.5 = 94.8 bp
Annual spread cost = 10,500m x 0.948% = ~99.5m per year (on top of swap rates)
Note how the 1.0bn of capital instruments (Tier 2 and AT1) contribute 300 of the 995 bp-billions: capital is by far the most expensive part of the liability stack.
Analogy. A funding plan is like a household that knows its mortgage fixed rate expires in 18 months. It does not wait until the last week to look for a new loan; it compares offers early, avoids refinancing during a holiday when banks are closed, and keeps some savings in case rates spike.
Key takeaways
- The funding plan turns the business plan and maturity profile into a wholesale funding requirement.
- The requirement is allocated across instruments, currencies and tenors within liquidity, MREL, capital and encumbrance constraints.
- Planned spreads drive the funding cost budget, NII projections and FTP.
- The plan is stress tested, tracked against actual issuance and revised as conditions change.
🧪 Practice
- Beginner: List four inputs needed to compute a bank's wholesale funding requirement for next year.
- Intermediate: In the example, senior non-preferred spreads widen from 120 bp to 160 bp before the bank issues. Recompute the weighted spread and the annual spread cost. Then suggest a reallocation that keeps MREL compliance and explain its trade-offs.
- Interview: "Halfway through the year, the bank has issued 30% of its plan, and spreads have widened by 50 bp. What would you recommend, and what data would you look at?" Hint: liquidity runway (survival horizon), upcoming maturities, LCR and NSFR headroom, MREL deadlines, the cost of waiting versus issuing now, alternative instruments (covered bonds, secured funding), and slowing asset growth.
<a id="72-capital-fundamentals"></a>
7.2 Capital Fundamentals
This subchapter explains what bank capital is, how regulators define and measure it against risk-weighted assets and total exposure, which buffers sit above the minimum, and how banks plan capital over several years.
Regulatory vs Economic Capital
Why it matters. Banks fund themselves mostly with other people's money; typically 90 to 95% of the balance sheet is debt and deposits. If loans go bad, someone has to absorb the loss. Capital is the layer of funding that absorbs losses first, so that depositors and senior creditors are repaid in full. A common misconception is that capital is a pile of cash held in a vault: it is not. Capital sits on the liability side of the balance sheet. It is the part of the funding that does not have to be repaid and can be written down without the bank defaulting. Liquidity (Chapter 5) is about having cash to pay obligations today; capital is about remaining solvent when losses arrive.
Expected vs unexpected loss. Lending always produces some losses. Banks distinguish two parts of the loss distribution:
Probability
|
| ***
| * **
|* **
|* ***
|* ****
|* *****
|* *******
|* ***********_______________
+-----+-----------------------------------+---------------> Annual loss
| |
EL 99.9% quantile
|<--->|<--------------------------------->|<---- tail ---->
Expected Unexpected loss (UL) Losses beyond capital:
loss: priced covered by CAPITAL bank fails (0.1% of years)
in and covered
by provisions
- Expected loss (EL) is the average loss: a cost of doing business, covered by pricing and loan loss provisions.
- Unexpected loss (UL) is the loss above the average that occurs in bad years. Capital exists to cover it up to a chosen confidence level.
Two ways to decide how much capital is enough.
| Aspect | Regulatory capital | Economic capital |
|---|---|---|
| Defined by | Regulators (Basel framework, CRR/CRD in the EU) | The bank's own models |
| Purpose | Minimum legal requirement, comparability across banks | Internal steering: risk appetite, pricing, capital allocation |
| Risks covered | Pillar 1: credit, market, operational, CVA | All material risks, including IRRBB, concentration, business, reputational |
| Confidence level | Implicit (the IRB formula uses 99.9% over 1 year) | Chosen by the bank, often linked to a target rating (for example 99.95% for AA) |
| Diversification | Limited recognition | Modeled correlations between risks and portfolios |
| Measurement | Rules, formulas, approved models, floors | Internal models (credit portfolio models, VaR, scenario analysis) |
| Binding? | Yes: breaches lead to supervisory action | No, but used in ICAAP (see Capital Planning and ICAAP) |
Example. A portfolio of 1,000 corporate loans of EUR 1m each, each with a probability of default (PD) of 1% and a loss given default (LGD) of 45%:
EL = PD x LGD x EAD = 1% x 45% x 1,000m = 4.5m (covered by pricing/provisions)
Bank's credit portfolio model:
99.9% one-year loss quantile = 30.0m
Economic capital = 30.0m - 4.5m = 25.5m
Regulatory (standardized, unrated corporates at 100% RW):
RWA = 1,000m x 100% = 1,000m
Minimum total capital = 8% x 1,000m = 80.0m
The two figures differ because the standardized approach uses one simple risk weight for all unrated corporates, while the internal model reflects this portfolio's actual PD, diversification and granularity. Banks use economic capital to compute risk-adjusted return on capital (RAROC):
RAROC = (revenues - costs - expected loss) / economic capital
Analogy. Regulatory capital is like the legal minimum tire tread depth: everyone must meet it and it is easy to check. Economic capital is the tread depth a careful driver chooses for the specific route, weather and load. The driver must obey the law, but plans the trip on their own assessment.
Key takeaways
- Capital is loss-absorbing funding on the liability side, not a cash reserve.
- Expected losses are covered by pricing and provisions; capital covers unexpected losses.
- Regulatory capital is rule-based and binding; economic capital is the bank's own estimate and covers all material risks.
- Economic capital feeds risk-adjusted performance measures such as RAROC.
🧪 Practice
- Beginner: Explain in one sentence each the difference between capital and liquidity, and between expected and unexpected loss.
- Intermediate: A loan portfolio earns revenue of EUR 60m, costs EUR 20m, has an expected loss of EUR 10m and consumes EUR 200m of economic capital. Compute RAROC. If the bank's hurdle rate is 12%, does the portfolio create value?
- Interview: "Why would a bank's economic capital be higher than its regulatory capital for some risks and lower for others?" Hint: risks outside Pillar 1 (IRRBB, concentration, business risk), diversification benefits, conservative regulatory risk weights for low-risk granular portfolios, and floors.
CET1, AT1 and Tier 2
Why it matters. Not every source of loss-absorbing funding works equally well. Ordinary shares absorb losses immediately and never have to be repaid; a subordinated bond absorbs losses only once the bank is failing, and pays a coupon that cannot be stopped without a default. Regulators therefore divide capital into tiers, based on how reliably each instrument absorbs losses.
Going concern vs gone concern.
- Going-concern capital absorbs losses while the bank keeps operating: CET1 and AT1 (together, Tier 1).
- Gone-concern capital absorbs losses once the bank has failed or is in resolution: Tier 2 (and, beyond regulatory capital, bail-inable debt; see 7.3).
The tiers.
| Feature | CET1 (Common Equity Tier 1) | AT1 (Additional Tier 1) | Tier 2 |
|---|---|---|---|
| Main instruments | Ordinary shares, share premium, retained earnings, accumulated other comprehensive income (AOCI), other reserves | Perpetual deeply subordinated bonds ("CoCos") | Subordinated bonds and loans |
| Maturity | Perpetual | Perpetual; issuer call allowed after at least 5 years with supervisory approval, no incentive to redeem | At least 5 years original maturity |
| Distributions | Fully discretionary dividends | Fully discretionary, non-cumulative coupons (a cancelled coupon is lost forever) | Contractual coupons; missing one is a default |
| Loss absorption | First loss, automatically | Write-down or conversion to shares when CET1 falls below a trigger (at least 5.125%) and at the point of non-viability | At the point of non-viability or in insolvency |
| Rank in insolvency | Last | Above CET1, below Tier 2 | Above AT1, below senior debt |
| Pillar 1 minimum (% of RWA) | 4.5% | Tier 1 total 6% (so AT1 can fill up to 1.5%) | Total capital 8% (so Tier 2 can fill up to 2%) |
Regulatory deductions from CET1. Accounting equity is not all usable capital. Items that would have little or no value in a crisis, or that would count twice, are deducted. The main ones are:
- Goodwill and other intangible assets.
- Deferred tax assets that rely on future profitability.
- For IRB banks, any shortfall of provisions compared with expected loss.
- Significant investments in other financial institutions above thresholds.
- Defined-benefit pension fund assets.
- Prudential filters: the cash flow hedge reserve and gains or losses on own liabilities due to the bank's own credit standing are removed.
Tier 2 amortization. In its last 5 years before maturity, a Tier 2 instrument's recognized amount declines straight-line (on a daily basis), because capital that is about to be repaid offers less protection.
Example. A bank with RWA of EUR 50bn:
CET1 calculation EUR bn
-------------------------------------------------------------
Share capital and share premium 2.0
Retained earnings 6.0
Accumulated other comprehensive income -0.3
less: goodwill -0.8
less: other intangibles (software) -0.4
less: deferred tax assets relying on future profits -0.2
-------------------------------------------------------------
CET1 6.3
AT1: perpetual notes, 7% CET1 trigger 1.0
Tier 1 7.3
Tier 2: bond A, 1.5bn, 8 years remaining -> 100% 1.5
bond B, 0.5bn, 3 years remaining -> 3/5 x 0.5 0.3
Tier 2 1.8
Total capital 9.1
CET1 ratio = 6.3 / 50 = 12.6%
Tier 1 ratio = 7.3 / 50 = 14.6%
Total capital ratio = 9.1 / 50 = 18.2%
AT1 in practice. In March 2023, about CHF 16bn of Credit Suisse AT1 notes were written down to zero as part of the rescue by UBS, while shareholders received UBS shares. The contractual terms of those notes allowed a write-down upon government support. This went against the usual expectation that AT1 ranks above equity, and it shows why investors read AT1 terms and conditions closely.
Analogy. The capital tiers are like layers of protection in a car crash. The crumple zone (CET1) absorbs the impact first and is designed to be destroyed. The airbag (AT1) deploys when the impact passes a threshold. The seatbelt (Tier 2) is what is left when everything else has given way.
Key takeaways
- CET1 is the highest-quality capital: permanent, first to absorb losses, fully discretionary distributions.
- AT1 is perpetual with discretionary coupons and a trigger-based write-down or conversion; Tier 2 is dated subordinated debt that absorbs losses at the point of non-viability.
- Pillar 1 minimums are 4.5% CET1, 6% Tier 1 and 8% total capital of RWA.
- Accounting equity becomes CET1 only after regulatory deductions and filters; Tier 2 amortizes in its last 5 years.
🧪 Practice
- Beginner: Rank CET1, AT1, Tier 2 and senior bonds from first to last in absorbing losses.
- Intermediate: A Tier 2 bond of EUR 800m has 2 years and 6 months left to maturity. What amount counts as Tier 2 capital today? What happens to the total capital ratio of the example bank if this bond replaces bond B?
- Interview: "Why does an AT1 bond pay a much higher coupon than a Tier 2 bond from the same issuer?" Hint: perpetual maturity with call risk, discretionary non-cumulative coupons, a going-concern trigger that can be hit before failure, and deeper subordination.
Risk-Weighted Assets (RWA)
Why it matters. A EUR 100m loan to a AAA-rated government and a EUR 100m loan to a start-up do not carry the same risk, so they should not need the same capital. Risk-weighted assets convert each exposure into a risk-adjusted amount, so that capital ratios reflect what the bank has actually lent to, not just how much.
What RWA is.
Capital ratio = Capital / RWA
RWA = sum over exposures of ( exposure amount x risk weight )
+ 12.5 x capital requirements for market, operational and CVA risk
The factor 12.5 is 1 / 8%: it converts a capital requirement back into an RWA-equivalent, so that all risks can be added together.
Components and approaches.
| Risk type | Typical share of RWA | Approaches |
|---|---|---|
| Credit risk (incl. counterparty) | 75 to 85% | Standardized (risk weights by exposure class, external rating, loan-to-value); internal ratings-based (IRB): foundation (bank estimates PD) or advanced (bank estimates PD, LGD, EAD) |
| Market risk | 2 to 10% | Standardized or internal models under the Fundamental Review of the Trading Book (FRTB) |
| Operational risk | 8 to 15% | Standardized approach based on a business indicator (Basel III final) |
| CVA risk | 1 to 3% | Basic or standardized CVA approach |
Selected standardized credit risk weights (Basel III final).
| Exposure | Risk weight |
|---|---|
| Sovereign rated AAA to AA- | 0% |
| Corporate rated A+ to A- | 50% |
| Unrated corporate | 100% |
| Regulatory retail (small, granular) | 75% |
| Residential mortgage, loan-to-value 60 to 80% | 30% |
| Residential mortgage, loan-to-value above 100% | 70% |
| Past-due exposures (unsecured part) | 100 to 150% |
Off-balance sheet items (undrawn commitments, guarantees) are first converted into exposure equivalents with credit conversion factors (CCF), then risk-weighted.
IRB in one formula. Under IRB, the capital requirement K per exposure comes from a single-factor portfolio model at 99.9% confidence:
K = LGD x [ N( ( G(PD) + sqrt(R) x G(0.999) ) / sqrt(1 - R) ) - PD ] x maturity adjustment
RWA = K x 12.5 x EAD
N = standard normal CDF, G = its inverse, R = asset correlation (set by regulation)
IRB banks can have much lower RWA for safe portfolios, which led to concerns about model variability. Basel III final therefore introduced an output floor: total RWA cannot be lower than 72.5% of the RWA calculated with standardized approaches. In the EU it is phased in from 50% in 2025 to 72.5% in 2030.
Example.
Exposure EUR bn RW / method RWA EUR bn
------------------------------------------------------------------------------------
Residential mortgages (LTV 60-80%) 10.0 30% 3.00
Unrated corporate loans 5.0 100% 5.00
Corporate loans rated A 2.0 50% 1.00
Government bonds (AA) 4.0 0% 0.00
Undrawn commitments 2.0 x CCF 40% 0.8 100% 0.80
------------------------------------------------------------------------------------
Credit risk RWA 9.80
Market risk: capital requirement 80m x 12.5 1.00
Operational risk: capital requirement 120m x 12.5 1.50
------------------------------------------------------------------------------------
Total RWA 12.30
RWA density = RWA / total assets = 12.3 / 21.0 = 58.6%
Moving EUR 1bn from unrated corporate loans to mortgages at 30% saves EUR 0.7bn of RWA. This is why business mix decisions are also capital decisions.
Engineering view. An RWA engine calculates risk weights at exposure level for millions of positions. It depends on counterparty classification, ratings, collateral allocation, CCFs, netting sets and model parameters. Every number must be traceable to source data and reconciled to the general ledger, and the results feed COREP regulatory reporting (see Chapter 9) and capital planning.
Analogy. RWA works like an insurer's view of a portfolio of policies. Two portfolios with the same total insured value can need very different reserves: one insures brick houses in safe areas, the other wooden houses in a wildfire zone. Risk weights are the bank's "risk zone" multipliers.
Key takeaways
- RWA converts exposures into risk-adjusted amounts; capital ratios divide capital by RWA.
- Credit risk dominates RWA; market, operational and CVA risks are added via 12.5 x capital requirement.
- The standardized approach uses prescribed risk weights; IRB uses the bank's own PD, LGD and EAD in a regulatory formula.
- The output floor limits how far IRB RWA can fall below standardized RWA.
🧪 Practice
- Beginner: Compute the RWA of a EUR 200m unrated corporate loan and a EUR 200m AAA government bond under the standardized approach.
- Intermediate: A bank's IRB RWA is EUR 40bn and its standardized RWA would be EUR 60bn. Compute the floored RWA with a 72.5% output floor, and the impact on a CET1 ratio of 13% (computed on IRB RWA).
- Interview: "RWA jumped 5% overnight, but the business says nothing changed. How would you investigate?" Hint: compare exposure-level results between the two days; look for changes in ratings, counterparty classification, collateral or netting data, model parameter releases, FX movements, and missing data defaulting to the highest risk weight.
Leverage Ratio
Why it matters. Risk weights can be wrong. Before 2008, AAA-rated structured securities carried very low risk weights and caused large losses; internal models can underestimate risk; and banks have an incentive to optimize their way to low RWA. Some banks that failed in 2008 reported strong risk-based ratios while holding equity of only 2 to 3% of their total assets. The leverage ratio is a simple, non-risk-based backstop that limits how big the balance sheet can be relative to capital, however safe the assets appear.
What it is.
Leverage ratio = Tier 1 capital / Total exposure measure >= 3% (Basel minimum)
- G-SIBs must hold an additional leverage ratio buffer equal to 50% of their G-SIB risk-based buffer.
- Some jurisdictions set different calibrations (for example the UK applies 3.25% for large banks and excludes central bank reserves; US banks face a supplementary leverage ratio with enhanced requirements for the largest groups).
The exposure measure. It is broader than accounting total assets:
| Component | Treatment |
|---|---|
| On-balance sheet assets | Accounting value, minus assets deducted from Tier 1 |
| Derivatives | Replacement cost plus potential future exposure under SA-CCR, multiplied by 1.4; limited recognition of collateral |
| Securities financing (repo, securities lending) | Gross receivables with limited netting, plus a counterparty credit risk add-on |
| Off-balance sheet items | Notional x credit conversion factor, with a floor of 10% |
Which constraint binds? A bank must meet both the risk-based and the leverage requirement. Which one binds depends on its RWA density (RWA / exposure measure):
Leverage binds when 3% x Exposure > Tier 1 requirement % x RWA
i.e. when RWA / Exposure < 3% / Tier 1 requirement %
Example: Tier 1 requirement of 11% of RWA
break-even RWA density = 3% / 11% = 27%
Banks with mortgage-heavy books, large government bond portfolios or large repo businesses have low RWA density and are often leverage-constrained.
Example. The example bank from the previous topics has Tier 1 capital of EUR 7.3bn and RWA of EUR 50bn.
Exposure measure EUR bn
---------------------------------------------
On-balance sheet assets (after deductions) 180
Derivatives (SA-CCR, x 1.4) 15
Securities financing transactions 20
Off-balance sheet (after CCFs) 10
---------------------------------------------
Total exposure measure 225
Leverage ratio = 7.3 / 225 = 3.24% (minimum 3.0%, headroom 0.24% = 0.54bn of Tier 1)
RWA density = 50 / 225 = 22% (below the 27% break-even: leverage is binding)
If treasury grows the repo book by EUR 10bn of low-risk, well-collateralized trades, RWA barely moves, but the exposure measure rises to 235bn and the leverage ratio falls to 3.11%.
Treasury implications. The leverage ratio makes balance sheet size itself costly. It discourages low-margin, high-volume activities such as repo matched books and holding excess central bank reserves, and in the past it encouraged "window dressing": shrinking the balance sheet just before quarter-end reporting dates. Basel disclosure standards now require banks to publish leverage ratios based on daily averages of securities financing transactions to expose this behavior.
Analogy. RWA-based ratios are like loading rules that depend on what each truck carries; the leverage ratio is the weight limit on the bridge, which applies regardless of the cargo.
Key takeaways
- The leverage ratio is Tier 1 capital divided by a broad, non-risk-weighted exposure measure, with a 3% Basel minimum.
- It is a backstop against model risk and the underestimation of risk weights.
- The exposure measure includes derivatives, securities financing and off-balance sheet items.
- Low-RWA-density banks are typically leverage-constrained, which makes balance sheet size itself a scarce resource.
🧪 Practice
- Beginner: A bank has Tier 1 capital of EUR 4bn and an exposure measure of EUR 120bn. Compute the leverage ratio and the headroom above 3%.
- Intermediate: For the example bank, how much additional exposure could treasury add before the leverage ratio falls to 3.0%? What does that imply for the pricing of repo trades?
- Interview: "Why would a bank with a 15% CET1 ratio still be capital-constrained?" Hint: low RWA density makes the leverage ratio binding; also consider MREL leverage-based requirements and the output floor.
Capital Buffers
Why it matters. A minimum requirement that cannot be breached is a cliff edge: if a bank's capital falls below it, the regulator must intervene, which can itself trigger a crisis of confidence. Capital buffers create a usable zone above the minimum. Banks build them up in good times and can use them in a stress, but while they are inside the buffer zone, their distributions (dividends, AT1 coupons, bonuses) are restricted so that earnings go back into capital.
The capital stack.
% of RWA (CET1)
^
| +-------------------------------------------+
| | Management buffer (bank's own choice) | no regulatory consequence
| +-------------------------------------------+
| | Pillar 2 Guidance (P2G) | supervisory expectation, not binding,
| +-------------------------------------------+ not part of the MDA calculation
| | Systemic buffers (G-SIB / O-SII / SyRB) | \
| | Countercyclical buffer (CCyB, 0 to 2.5%+) | > Combined buffer requirement (CBR)
| | Capital conservation buffer (2.5%) | / breach => MDA restrictions
| +-------------------------------------------+
| | Pillar 2 Requirement (P2R, bank-specific) | \ breach => regulatory intervention,
| | Pillar 1 minimum (4.5% CET1) | / possible withdrawal of licence
| +-------------------------------------------+
+-------------------------------------------------->
| Buffer | Size | Purpose |
|---|---|---|
| Capital conservation buffer (CCB) | 2.5% | General buffer to absorb losses in stress |
| Countercyclical buffer (CCyB) | 0 to 2.5% (can be higher), weighted by the location of the bank's credit exposures | Built up during credit booms, released in downturns (many authorities released it in 2020) |
| G-SIB / O-SII buffer | G-SIB 1 to 3.5%; O-SII set nationally | Extra loss absorbency for systemically important banks |
| Systemic risk buffer (SyRB) | Set nationally | Structural or sectoral systemic risks (for example residential real estate) |
Pillar 2 Requirement (P2R) is set by the supervisor after its annual review (SREP in the EU). At least 56.25% of P2R must be met with CET1 and at least 75% with Tier 1.
Maximum Distributable Amount (MDA). If CET1 falls into the combined buffer, the bank may only distribute a fraction of its profits, depending on which quartile of the buffer it is in:
| Position within the combined buffer | Share of profits that may be distributed |
|---|---|
| Top quartile | 60% |
| Second quartile | 40% |
| Third quartile | 20% |
| Bottom quartile | 0% |
Example. Using the example bank (CET1 EUR 6.3bn, RWA EUR 50bn, CET1 ratio 12.6%):
Pillar 1 CET1 4.500%
P2R 2.0% x 56.25% (CET1 part) 1.125%
Capital conservation buffer 2.500%
Countercyclical buffer 1.000%
O-SII buffer 1.000%
---------------------------------------------------
MDA threshold (CET1) 10.125%
P2G 1.500%
---------------------------------------------------
Supervisory expectation 11.625%
MDA distance = 12.6% - 10.125% = 2.475% x 50bn = EUR 1.24bn of CET1 (before considering AT1/T2 shortfalls)
Combined buffer = 2.5 + 1.0 + 1.0 = 4.5%, spanning 5.625% to 10.125%; each quartile = 1.125%
If CET1 fell to 9.5%: top quartile (9.0% to 10.125%) -> at most 60% of profits may be distributed
Why investors watch the MDA distance. AT1 coupons count as distributions. A bank close to its MDA threshold may be forced to cancel AT1 coupons, so the MDA distance is a key input to AT1 pricing.
Analogy. Buffers work like the reserve section of a car's fuel tank. The warning light comes on well before the tank is empty; you can still drive, but you are expected to stop spending on extras and refuel.
Key takeaways
- Buffers sit above Pillar 1 and P2R and can be used in stress, but using them restricts distributions.
- The combined buffer requirement includes the conservation, countercyclical and systemic buffers.
- The MDA threshold (Pillar 1 + P2R + CBR) is the trigger for automatic distribution limits; P2G is a non-binding expectation above it.
- The MDA distance is closely watched by management, supervisors and AT1 investors.
🧪 Practice
- Beginner: List the components of the combined buffer requirement and the size of the capital conservation buffer.
- Intermediate: A bank has P2R of 2.5%, CCyB of 0.5%, O-SII buffer of 1.5% and RWA of EUR 80bn. Compute its MDA threshold in CET1 terms and the CET1 amount needed to stay 2% above it.
- Interview: "Why do regulators release the countercyclical buffer in a crisis instead of keeping banks safer?" Hint: procyclicality: banks protecting ratios by cutting lending deepen the downturn; releasing the buffer lets them absorb losses and keep lending.
Capital Planning and ICAAP
Why it matters. Capital ratios change every month, driven by profits and losses, dividends, balance sheet growth, regulatory changes and market movements. Raising new equity takes time, dilutes shareholders and is hardest exactly when it is needed most. Banks therefore project their capital position several years ahead, under normal and stressed conditions, and plan actions in advance.
The capital plan. The plan projects capital resources and requirements over 3 to 5 years:
- Capital resources: CET1 movements (profits, dividends, buybacks, OCI, deductions), AT1 and Tier 2 issuance, calls and amortization.
- Capital requirements: RWA growth from the business plan, model and methodology changes, regulatory changes (for example the output floor phase-in), FX effects.
- Targets: ratios above the MDA threshold plus P2G plus a management buffer, leverage ratio and MREL (see 7.3).
- Capital actions: dividend policy, buybacks, AT1/Tier 2 issuance, RWA optimization (significant risk transfer securitizations, portfolio sales), equity issuance as a last resort.
The capital walk. Changes in the CET1 ratio are explained step by step:
Opening CET1 ratio
+ net profit
- dividends and buybacks
+/- OCI and changes in deductions
- RWA growth (business)
- RWA regulatory and model changes
+/- FX
= Closing CET1 ratio
ICAAP. The Internal Capital Adequacy Assessment Process is the Pillar 2 process in which a bank assesses, documents and defends its own view of the capital it needs. Under the ECB's guide it has two perspectives:
| Perspective | Question | Measures |
|---|---|---|
| Normative | Can the bank meet all regulatory and supervisory capital requirements over the next 3+ years, in baseline and adverse scenarios? | Regulatory ratios, MDA distance, leverage ratio, MREL |
| Economic | Does the bank hold enough internal capital to cover all material risks, measured with its own methods? | Economic capital vs internal capital (see Regulatory vs Economic Capital) |
The ICAAP, together with the ILAAP for liquidity, is an input to the supervisor's annual review (SREP), which sets the P2R and P2G. Supervisory stress tests (EBA/ECB in the EU, the Federal Reserve's stress tests in the US, which set the stress capital buffer) test capital plans against common adverse scenarios.
Example. The example bank projects one year ahead:
| Line | Baseline (EUR bn) | Adverse (EUR bn) |
|---|---|---|
| Opening CET1 | 6.30 | 6.30 |
| Net profit / (loss) | +0.90 | -0.60 |
| Dividends (50% payout; suspended in adverse) | -0.45 | 0.00 |
| Share buyback | -0.20 | 0.00 |
| OCI (bond portfolio revaluation) | -0.05 | -0.30 |
| Closing CET1 | 6.50 | 5.40 |
| Opening RWA | 50.0 | 50.0 |
| Business growth (+5%) / credit migration (+8%) | +2.5 | +4.0 |
| Output floor phase-in | +1.5 | +1.5 |
| Closing RWA | 54.0 | 55.5 |
| CET1 ratio | 12.04% | 9.73% |
| MDA threshold | 10.125% | 10.125% |
In the baseline, the bank stays 1.9 percentage points above its MDA threshold. In the adverse scenario, it falls 0.4 points below, so distributions would be restricted. The plan must document management actions that would restore the ratio: for example cancelling the buyback, RWA reduction through a significant risk transfer, slower lending growth. Issuing AT1 would not help, because the shortfall is in CET1.
Engineering view. Capital planning combines balance sheet projections, P&L forecasts, RWA calculators and scenario generators. The key requirements are consistency across models (the same scenario must drive credit losses, NII and RWA), versioned assumptions, reproducible runs and a clear audit trail from each number in the ICAAP report back to its inputs.
Analogy. Capital planning is like a pilot's flight plan with fuel calculations: the plan covers the expected route, an alternate airport and holding time in bad weather, and defines in advance at what fuel level the pilot must divert.
Key takeaways
- Capital plans project capital resources and requirements over 3 to 5 years under baseline and adverse scenarios.
- The capital walk explains every movement in the CET1 ratio from opening to closing.
- ICAAP is the bank's own assessment of capital adequacy, with normative and economic perspectives, and feeds the SREP.
- Plans must include credible management actions for stress, recognizing which capital tier a shortfall affects.
🧪 Practice
- Beginner: Name three actions a bank can take to increase its CET1 ratio, at least one acting on the numerator and one on the denominator.
- Intermediate: In the adverse scenario above, how much RWA would the bank need to remove to restore its CET1 ratio to the 10.125% MDA threshold, assuming CET1 stays at 5.40bn?
- Interview: "How would you design a capital planning platform so that the same scenario gives consistent results for credit losses, NII and RWA?" Hint: a central scenario repository with versioned macro variables, shared reference data and balance sheet projection, model orchestration with explicit dependencies, run IDs linking all outputs, and reconciliation of starting positions to reported figures.
<a id="73-resolution-and-loss-absorption"></a>
7.3 Resolution and Loss Absorption
This subchapter explains the post-2008 framework that allows a failing bank to be recapitalized by its own investors rather than by taxpayers: bail-in, the minimum amounts of loss-absorbing debt banks must hold, the plans prepared in advance for recovery and resolution, and the deposit insurance schemes that protect small depositors.
Bail-In Mechanics
Why it matters. In 2008 and 2009, governments injected hundreds of billions into failing banks, because letting large banks enter ordinary insolvency would have frozen payments, deposits and markets. This created a "too big to fail" problem: creditors of large banks expected to be rescued, so they lent cheaply and did not discipline risk-taking. Resolution regimes (the FSB Key Attributes of 2011, the EU Bank Recovery and Resolution Directive (BRRD) and Single Resolution Mechanism, Title II of the US Dodd-Frank Act, the UK Banking Act 2009) give authorities tools to restructure a failing bank while keeping its critical functions running. The central tool is bail-in.
What bail-in is. Bail-in is the power of the resolution authority to write down liabilities, or convert them into equity, so that the bank's own shareholders and creditors absorb losses and recapitalize it. The bank keeps operating; its owners change.
The creditor hierarchy. Losses are allocated bottom-up, in the reverse order of insolvency ranking:
Last to absorb losses
+---------------------------------------------------------+
| Excluded: covered (insured) deposits, secured liabilities| not bailed in
| (covered bonds up to collateral value), client assets, | (DGS takes the place of
| liabilities to banks < 7 days, salaries, taxes | covered depositors)
+---------------------------------------------------------+
| Preferred deposits: uninsured part of individual and |
| SME deposits |
+---------------------------------------------------------+
| Senior preferred debt, other deposits, derivatives | bail-inable
+---------------------------------------------------------+
| Senior non-preferred debt | bail-inable
+---------------------------------------------------------+
| Tier 2 | write-down/conversion
+---------------------------------------------------------+
| AT1 | write-down/conversion
+---------------------------------------------------------+
| CET1 (shareholders) | wiped out first
+---------------------------------------------------------+
First to absorb losses
Key principles.
- Point of non-viability (PONV): resolution starts when the bank is failing or likely to fail, there is no private alternative, and resolution is in the public interest. Otherwise, the bank enters normal insolvency.
- Write-down and conversion of capital instruments: at PONV, CET1, AT1 and Tier 2 are written down or converted before or together with resolution.
- No creditor worse off (NCWO): no creditor may receive less than in normal insolvency. Creditors who do can be compensated.
- Valuation: an independent valuation estimates the losses and the amount of conversion needed.
- Resolution weekend: decisions are typically taken between market close on Friday and opening on Monday, so that clients and markets see a functioning bank on Monday.
- Other tools may be used alongside bail-in: sale of business, bridge institution, asset separation (bad bank).
Example. A bank with EUR 100bn of assets suffers EUR 8bn of losses. Its post-resolution RWA is EUR 40bn, and the authority wants a 10% CET1 ratio after resolution.
Before Step 1: absorb Step 2: recapitalize After
(EUR bn) 8bn of losses to 10% x 40bn = 4bn (EUR bn)
-------------------------------------------------------------------------------------
Assets 100 -8 92
-------------------------------------------------------------------------------------
Covered deposits 50 50
Other deposits 20 20
Covered bonds (secured) 8 8
Senior preferred 10 10
Senior non-preferred 4 -4 converted to equity 0
Tier 2 2 -2 written down 0
AT1 1 -1 written down 0
CET1 (old shareholders) 5 -5 wiped out 0
New CET1 (former SNP) - +4 4
-------------------------------------------------------------------------------------
Total liabilities+equity 100 92
Old shareholders and capital instrument holders lose everything; senior non-preferred holders become the new shareholders; depositors and senior preferred creditors are untouched. No public money is used.
Real cases. In June 2017, Banco Popular was the first bank resolved by the EU's Single Resolution Board: its equity, AT1 and Tier 2 were written down or converted, and it was sold to Santander for EUR 1 over a weekend. In 2023, Credit Suisse's AT1 notes were written down in a government-supported takeover rather than a formal bail-in, which showed that contractual terms also matter.
Engineering view. Bail-in can only be executed if the authority knows, within hours, exactly what liabilities exist. Banks must be able to produce a full liability data set: every instrument and contract with its rank, amount, maturity, governing law, whether it contains a contractual recognition of bail-in (needed for liabilities under non-EU law), secured status and counterparty. They must also provide data for the valuation and be able to execute the bail-in in their own systems (for example cancelling and reissuing securities through central securities depositories).
Analogy. Bail-in is a queue for losses. Shareholders stand at the front, then AT1, Tier 2 and senior non-preferred holders. Insured depositors are not in the queue at all. Each group pays until the bill is settled, and the next group in line pays nothing.
Key takeaways
- Bail-in recapitalizes a failing bank by writing down or converting its own liabilities, not with public money.
- Losses follow the reverse creditor hierarchy; covered deposits and secured liabilities are excluded.
- No creditor may be worse off than in normal insolvency (NCWO).
- Executing bail-in requires complete, accurate liability data that can be produced within hours.
🧪 Practice
- Beginner: In the example, which creditors lose money, and which become the new owners of the bank?
- Intermediate: Rework the example with losses of EUR 10bn instead of EUR 8bn. Which instruments are fully written down, how much senior non-preferred is used for loss absorption, and how much senior preferred must be converted to reach a 10% CET1 ratio on EUR 40bn of RWA?
- Interview: "A regulator asks for a full liability data template by Sunday noon for a resolution weekend. What would make this hard in a typical bank?" Hint: liabilities spread across many systems (deposits, securities issuance, derivatives, trade payables), missing rank or governing law fields, intragroup positions, accrued interest, and the need for a consistent cut-off time across systems.
MREL and TLAC
Why it matters. Bail-in only works if, at the point of failure, the bank has enough liabilities that can be bailed in safely. If the only remaining liabilities are deposits and short-term operational payables, bailing them in would trigger a run and disrupt the economy, and the authority might have to choose between a bailout and chaos. Regulators therefore require banks to hold a minimum amount of loss-absorbing equity and debt before any trouble starts.
Two frameworks, one idea.
| Feature | TLAC (Total Loss-Absorbing Capacity) | MREL (Minimum Requirement for own funds and Eligible Liabilities) |
|---|---|---|
| Set by | FSB global standard (2015) | EU BRRD; set bank by bank by the resolution authority (for example the SRB) |
| Applies to | G-SIBs only | All EU banks (requirement may equal own funds for banks expected to be liquidated) |
| Calibration | At least 18% of RWA and 6.75% of leverage exposure (since 2022) | Bank-specific: loss absorption + recapitalization amounts, plus a market confidence charge |
| Subordination | Generally required | Required for G-SIIs and large banks (at least 13.5% of RWA / 5% of leverage exposure under Pillar 1), and as set by the authority |
| Buffers | Combined buffer requirement sits on top of the RWA requirement | Combined buffer requirement sits on top of MREL as % of RWA; breach triggers MREL-MDA restrictions |
How MREL is calibrated (simplified).
MREL (% of RWA) = Loss absorption amount (Pillar 1 + P2R)
+ Recapitalization amount (Pillar 1 + P2R after resolution)
+ Market confidence charge (combined buffer excl. CCyB, minus 1.25%)
Combined buffer requirement must be met in addition, with CET1 not used for MREL.
Eligibility. An instrument counts towards MREL/TLAC only if it:
- Has at least 1 year of remaining maturity.
- Is fully paid up and not funded, directly or indirectly, by the bank itself.
- Is unsecured and not guaranteed by the bank or a related entity.
- Has no set-off rights or acceleration rights for the holder, except in insolvency.
- Is not a structured note whose return is linked to derivatives.
- Includes contractual recognition of bail-in if governed by non-EU law (for EU banks).
How subordination is achieved.
| Type | How | Common in |
|---|---|---|
| Structural | A holding company issues ordinary senior debt and lends the proceeds to operating subsidiaries, often as internal MREL; HoldCo creditors rank below all OpCo creditors | US, UK, Switzerland |
| Contractual | Debt contractually ranks below other senior debt: senior non-preferred (introduced in France in 2016, harmonized across the EU in 2017) | Most EU banks |
| Statutory | The law ranks certain liabilities (for example deposits) above senior debt | Germany (earlier approach), deposit preference regimes |
Internal MREL / TLAC. In groups, material subsidiaries must hold internal loss-absorbing instruments issued to the resolution entity (for TLAC, 75 to 90% of the external requirement), so that losses in a subsidiary can be passed up to the parent without the subsidiary entering resolution itself.
Example. Using the example bank (RWA EUR 50bn, P1 8%, P2R 2%, combined buffer 4.5% of which CCyB 1.0%):
Loss absorption amount 8% + 2% = 10.00%
Recapitalization amount 8% + 2% = 10.00%
Market confidence charge (4.5% - 1.0%) - 1.25% = 2.25%
------------------------------------------------------------------
MREL 22.25%
+ combined buffer requirement (stacked on top) 4.50%
------------------------------------------------------------------
Effective requirement 26.75% x 50bn = 13.375bn
Available (EUR bn):
Own funds (CET1 6.3 + AT1 1.0 + Tier 2 1.8) 9.1
Senior non-preferred, > 1 year remaining 3.0
Senior preferred, > 1 year remaining, eligible 2.5
------------------------------------------------------------------
Total MREL-eligible 14.6 (29.2% of RWA)
Headroom = 14.6 - 13.375 = 1.225bn
Now suppose EUR 1.5bn of senior preferred bonds pass into their final year: they stop counting on that day, and headroom falls to -0.275bn. MREL eligibility creates a maturity cliff one year before contractual maturity, which is why the funding plan (see 7.1) refinances MREL instruments well ahead of maturity.
Analogy. MREL and TLAC are like a requirement for every ship to carry enough lifeboats for all passengers before leaving port. Building lifeboats in the middle of a storm is not an option.
Key takeaways
- TLAC (G-SIBs, global) and MREL (EU, bank-specific) require minimum amounts of bail-inable capital and debt.
- MREL is calibrated as loss absorption plus recapitalization plus market confidence, with buffers stacked on top.
- Only instruments with at least 1 year remaining maturity and meeting strict conditions are eligible; subordination is achieved structurally, contractually or by statute.
- Eligibility roll-off one year before maturity must be built into funding plans and monitoring systems.
🧪 Practice
- Beginner: Which of these count towards MREL: a 5-year senior non-preferred bond issued last year, a senior bond maturing in 9 months, a secured covered bond, an AT1 note?
- Intermediate: Recompute the example bank's MREL requirement if its P2R rises to 3% and its O-SII buffer to 1.5%. Does it still comply with EUR 14.6bn of eligible instruments?
- Interview: "Design the data model and daily process for MREL eligibility monitoring." Hint: instrument master with rank, governing law, bail-in recognition clause, call and maturity dates, holder information (to exclude self-held or group-held positions); daily eligibility flags, a 12-month forward roll-off projection and alerts when headroom drops below a threshold.
Recovery and Resolution Plans
Why it matters. Banks fail fast: Silicon Valley Bank went from a public announcement to closure in about two days. There is no time to design a strategy during the crisis. Recovery and resolution planning forces banks and authorities to agree in advance what the bank would do to save itself, and what the authority would do if that fails, and to remove the practical obstacles to both.
Two plans, two owners.
| Aspect | Recovery plan | Resolution plan |
|---|---|---|
| Prepared by | The bank (approved by its board, assessed by the supervisor) | The resolution authority, with data and input from the bank |
| Situation | Bank under severe stress but still viable | Bank failing or likely to fail |
| Goal | Restore viability with the bank's own measures, without public support | Maintain critical functions, protect depositors and public funds, avoid contagion |
| Content | Indicators, thresholds, recovery options, scenarios, governance, communication | Preferred resolution strategy, critical functions, resolvability assessment, removal of impediments |
| Update | At least annually | At least annually |
In the US, large banks submit resolution plans ("living wills") themselves under Title I of the Dodd-Frank Act, showing how they could be resolved under the Bankruptcy Code.
Recovery plan components.
- Recovery indicators across capital, liquidity, profitability, asset quality and market-based measures (share price, CDS spreads), each with an early warning threshold and a recovery threshold.
- Recovery options: issue capital, suspend dividends, sell subsidiaries or portfolios, deleverage, liability management exercises, access central bank facilities, with an estimate of their impact, timing and feasibility.
- Scenarios: idiosyncratic, market-wide and combined, used to test that the options would be sufficient (overall recovery capacity).
- Governance and escalation: who decides, how fast, and how the market and supervisors are informed.
Resolution plan components.
- Strategy: single point of entry (bail-in at the top holding company, losses passed up from subsidiaries) or multiple point of entry (separate resolution of regional subgroups); for small banks, normal insolvency with deposit insurance payout.
- Critical functions: deposits, payments, lending, clearing access that must continue.
- Resolvability requirements for the bank: MREL, operational continuity in resolution (shared services and IT must keep working), continued access to financial market infrastructures, resolution-resilient contracts (no termination solely because of resolution), funding in resolution (estimating needs and identifying collateral for central bank facilities), valuation readiness and management information systems that can produce data quickly.
From business as usual to resolution.
Business as usual --> Early warning --> Recovery --> Early intervention --> FOLTF --> Resolution
| | | | | |
indicators within indicator breaches bank activates supervisor can "failing authority
appetite EW threshold: recovery plan require actions, or likely takes control:
escalation, options replace management to fail" bail-in, sale,
heightened declared bridge bank
monitoring
|<- resolution
weekend ->|
Example: recovery indicator dashboard.
| Indicator | Current | Early warning | Recovery threshold | Status |
|---|---|---|---|---|
| CET1 ratio | 12.6% | 11.5% | 10.5% | Green |
| Leverage ratio | 3.24% | 3.40% | 3.15% | Amber (below early warning) |
| LCR | 145% | 120% | 105% | Green |
| NSFR | 118% | 110% | 103% | Green |
| Return on assets (rolling 4 quarters) | 0.45% | 0.20% | -0.20% | Green |
| 5Y senior CDS spread | 95 bp | 150 bp | 250 bp | Green |
The leverage ratio has crossed its early warning threshold (consistent with the leverage example in 7.2). This triggers escalation to ALCO and the board and closer monitoring, but not yet recovery options.
Analogy. A recovery plan is the fire safety plan that the building's occupants follow themselves: extinguishers, evacuation routes and fire wardens. A resolution plan is the fire department's plan for the same building: where to park the trucks, where the water connections are, and which walls can be broken. Both are drawn up before any fire.
Key takeaways
- Recovery plans are owned by the bank and describe how it would restore viability using its own measures.
- Resolution plans are owned by the authority and describe how a failing bank would be resolved while protecting critical functions.
- Recovery indicators have early warning and recovery thresholds that drive escalation.
- Resolvability depends on operational and data readiness: MREL, continuity of services and FMI access, resilient contracts, funding and valuation data.
🧪 Practice
- Beginner: Who writes the recovery plan, and who writes the resolution plan? Give one example of a recovery option.
- Intermediate: For the dashboard above, the bank's CDS spread jumps to 180 bp and the LCR drops to 115%. Which statuses change, and what escalation would you expect?
- Interview: "What does 'operational continuity in resolution' mean for a treasury system hosted by a group service company?" Hint: service contracts that survive resolution, no termination clauses triggered by resolution, pre-funded service costs, documented service catalogues, separable data and access, and the ability to keep payments and FMI connectivity running for the resolved entity.
Deposit Insurance Schemes
Why it matters. Individual depositors cannot analyze a bank's balance sheet, and they know that if the bank runs out of cash, those who withdraw first get paid in full. That makes bank runs self-fulfilling: even a rumor can cause one. Deposit insurance removes the incentive to run for most depositors by guaranteeing their deposits up to a limit, and it protects small savers who could not be expected to monitor their bank.
What a deposit guarantee scheme (DGS) is. A DGS repays eligible depositors up to a coverage limit if their bank fails. It is funded by contributions from banks, not by depositors.
| Feature | EU (DGS Directive) | United States (FDIC) | United Kingdom (FSCS) |
|---|---|---|---|
| Coverage limit | EUR 100,000 per depositor per bank | USD 250,000 per depositor, per insured bank, per ownership category | GBP 120,000 per depositor per bank licence (since December 2025, previously GBP 85,000) |
| Temporary high balances | Higher protection for limited periods after specific life events (for example a house sale), set nationally | Not applicable | Up to GBP 1.4m for 6 months after specific life events |
| Payout time target | 7 working days | Usually by the next business day | Within 7 days |
| Funding | Ex-ante fund, target 0.8% of covered deposits, risk-based contributions | Deposit Insurance Fund, minimum reserve ratio of 1.35% of insured deposits, risk-based assessments | Annual levies on firms |
| Role beyond payout | Can contribute to resolution, up to what a payout would have cost | Also the resolution authority for insured banks; uses least-cost resolution | Can contribute to resolution |
Eligibility. Coverage usually applies to individuals, businesses and charities. Deposits of banks, investment firms, insurers, pension funds and (in most cases) public authorities are excluded. Coverage is per depositor per bank licence, so two brands operating under one banking licence share a single limit.
Deposit insurance and resolution. Covered deposits are excluded from bail-in. If a bank is resolved, the DGS takes the place of covered depositors in the creditor hierarchy, with a super-preferred ranking. A strong resolution framework and deposit insurance work together: the first protects the system, the second protects small savers.
Moral hazard. Insured depositors have no reason to choose a safer bank, which weakens market discipline. This is one reason for coverage limits, risk-based contributions and the requirement that large depositors and investors are bail-inable. When Silicon Valley Bank failed in March 2023, more than 90% of its deposits exceeded the insurance limit, and the US authorities invoked a systemic risk exception to protect all depositors, an exceptional decision that renewed the debate about coverage design.
Example: aggregation for one depositor (EU, EUR 100,000 limit).
Depositor: Anna, at Bank X (two brands, one licence)
---------------------------------------------------------------
Current account (brand A) 30,000
Savings account (brand A) 60,000
Joint account with partner: 80,000 / 2 owners 40,000
Term deposit (brand B, same licence) 20,000
---------------------------------------------------------------
Total eligible balance 150,000
Covered 100,000
Uncovered (preferred deposit, ranks above senior) 50,000
The partner's share of the joint account (EUR 40,000) is aggregated with the partner's own accounts, not with Anna's.
Engineering view: the single customer view (SCV). To pay out within days, banks must be able to produce a file listing each depositor's aggregated eligible balance, covered amount and contact details. In the UK, the SCV file must be producible within 24 hours. Building it requires:
- A unique depositor identity across all products, systems and brands (deduplication of customers).
- Correct ownership splits for joint and trust accounts.
- Eligibility flags (excluding financial institutions and other excluded depositors).
- Accrued interest up to the failure date and currency conversion at the reference rate.
- Flags for accounts that must not be paid automatically (sanctions, legal disputes, dormant accounts).
- Regular testing, because the file is used only in a crisis but must work the first time.
Analogy. Deposit insurance works like a promise to everyone in a queue outside a ticket office that there are enough seats for them all. Once people believe it, they stop pushing to the front, and the queue stays calm.
Key takeaways
- Deposit insurance prevents runs by guaranteeing eligible deposits up to a limit per depositor per bank licence.
- Coverage limits are EUR 100,000 in the EU, USD 250,000 per ownership category in the US and GBP 120,000 in the UK.
- Covered deposits are excluded from bail-in; the DGS takes their place in the creditor hierarchy.
- Fast payouts depend on accurate single customer view data, which banks must be able to produce at very short notice.
🧪 Practice
- Beginner: A depositor has EUR 70,000 at Bank A and EUR 70,000 at Bank B (separate licences). How much is covered in the EU? How much if A and B were two brands of the same licence?
- Intermediate: In the US, a couple has USD 400,000 in a joint account and each partner has USD 200,000 in a single account at the same bank. Compute the insured and uninsured amounts, explaining the ownership categories.
- Interview: "Design a single customer view file generator that must run within 24 hours of a failure." Hint: customer master and identity matching across systems, ownership rules engine, eligibility and exclusion flags, interest accrual to a cut-off time, reconciliation of totals to the general ledger, and regular dry runs with data quality metrics.
<a id="8-market-and-counterparty-risk-in-treasury"></a>