IN ONE SENTENCE — Capital management is the function that measures how much of the bank’s own money each activity ties up, makes sure the total stays above what regulators require, and gives the business lines a price for using it.
Every module of this course has ended with a line called “allocated capital = 13% of RWA”. This module opens that line. Behind it sit a calculation engine that runs every month on millions of exposures, a set of regulatory texts that changed deeply on 1 January 2025, a supervisory dialogue that adds a bank-specific layer each year, and a planning process that decides whether the bank can afford next year’s growth, its dividend and its share buyback.
Where it sits. In a European universal bank such as Meridian, capital management usually reports to the Chief Financial Officer, with a strong dotted line to the Chief Risk Officer. Three teams cooperate. The RWA production team (often called “regulatory reporting” or “Pillar 1 calculation”) runs the engine, reconciles it with accounting and files the COREP templates with the supervisor every quarter. The credit risk modelling team, in Risk, owns the internal rating models (PD, and for banks allowed to, LGD and conversion factors) and defends them before the ECB’s internal model investigations. The capital planning team, in Finance, projects capital and RWA over three to five years, runs the Internal Capital Adequacy Assessment Process (ICAAP), and allocates capital budgets to business lines. A small prudential regulation team reads every new text, answers consultations, and translates rules into internal policy. Treasury issues the capital instruments (AT1, Tier 2) that complete the stack (see C15).
Clients. Capital management has no external clients. Its internal clients are every desk of this course: the corporate banker who needs a RAROC for a credit application (see B03, E03), the project finance team that needs a slotting category (see B06), the securitisation desk whose significant risk transfer trades exist to release RWA (see B16, C04), and the Board, which approves the risk appetite and the capital plan.
How it “earns” money. It does not, directly. Its value is in what it avoids and what it enables. Orders of magnitude for Meridian (case assumptions consistent with A02): RWA of EUR 118bn and CET1 capital of EUR 16.52bn, a CET1 ratio of 14.00%. One basis point of CET1 ratio is EUR 11.8m of capital. An RWA optimisation that removes EUR 1bn of RWA frees EUR 130m of allocated capital at 13%; at a 12% after-tax hurdle that capital must earn about EUR 15.6m after tax a year, so freeing it is worth roughly as much as winning a mid-sized client. Conversely, a model finding by the supervisor that adds 5% to the RWA of a portfolio can wipe out a year of business growth.
The team and adjacent functions. A capital management department in a bank of Meridian’s size counts some 40 to 80 people: data engineers, reporting specialists, quants, regulatory experts and planners. They work with Accounting (the RWA engine starts from the general ledger), Credit Risk (ratings and collateral data), Treasury (capital issuance, ALM), Investor Relations (the CET1 ratio is the first number analysts ask for), and the Sustainability team (Pillar 3 ESG disclosures and climate stress tests use the same data).
A realistic week.
WHY IT MATTERS ON THE DESK — Every price you quote in a CIB is partly a price for capital. A banker who cannot read the risk weight cannot defend a margin, cannot explain why the same borrower costs twice as much in one bank as in another, and cannot see that a regulatory reading (a conversion factor, a maturity convention) often moves the economics of a deal more than its green label.
The “product” of capital management is a ratio and a price. Learn the ratio first.
IN ONE SENTENCE — A capital ratio divides loss-absorbing own funds by risk-weighted assets, a measure of exposures adjusted for how risky regulation deems them.
INTUITION — RWA are the weight of the luggage; capital is the strength of the rope. The regulator does not care how many suitcases you carry, but how heavy they are, and it insists that the rope be at least a given fraction of that weight.
FORMULA
CET1 capital is ordinary shares plus retained earnings and reserves, minus regulatory deductions (goodwill, intangible assets, deferred tax assets that depend on future profits, shortfall of provisions against expected loss for IRB banks); is the sum over exposures of exposure at default times risk weight; covers the risk of losses on derivatives from a deterioration of counterparty credit; and are capital charges for trading and operational risk multiplied by 12.5, so that they sit on the same scale as credit.
Credit risk typically accounts for roughly 80 to 85% of a European universal bank’s RWA, operational risk for 10 to 12% and market risk and CVA for the rest. For a corporate and investment bank the credit share is what matters on the desk, and it is what this module covers.
The stack. Requirements come in layers, each with a different consequence when breached:
| Layer | Level for Meridian | What happens if breached |
|---|---|---|
| Pillar 1 minimum | CET1 4.5%, Tier 1 6%, total 8% | Licence at risk |
| Pillar 2 requirement (P2R) | 1.80% of RWA, at least 56.25% in CET1 | Same as Pillar 1: binding |
| Combined buffer | Conservation 2.5%, countercyclical 0.80%, O-SII 1.00% | Distribution restrictions (MDA) |
| Pillar 2 guidance (P2G) | 1.25% in CET1 | Supervisory dialogue, no automatic sanction |
| Management buffer | Bank’s own choice | Board’s risk appetite |
The maximum distributable amount (MDA) is the cap on dividends, AT1 coupons and variable pay that applies automatically when CET1 falls into the combined buffer. Meridian’s CET1 requirement, 9.81% of RWA, is built in A02; this module does not repeat it. Its actual ratio of 14.00% leaves 419 bp of distance to the MDA trigger.
For orders of magnitude across the euro area: the ECB’s aggregate SREP results published on 18 November 2025 put the average overall CET1 requirement and guidance for 2026 at 11.2% of RWA, broadly stable, with an average P2R of about 2.1% of RWA (total capital) and an average P2G reduced from 1.3% to 1.1%. The total capital requirement and guidance stood at 15.6%. These are averages over the significant institutions supervised by the ECB; each bank’s figure is confidential until it chooses to publish it, which listed banks do.
Countercyclical buffer. Each country sets its rate, and a bank applies the average of the rates of the countries where its credit exposures sit, weighted by own funds requirements. In France the Haut Conseil de stabilité financière kept the rate at 1% at its meeting of 9 June 2026. Meridian’s 0.80% reflects its exposures outside France.
Why 13%? Meridian allocates capital to deals at 13% of RWA: above the CET1 requirement plus guidance (about 11.1% for Meridian), below its actual ratio. It is a management number. Some banks allocate on total capital (with AT1 and Tier 2 in the denominator of cost), some on CET1 alone, some add an allocation for operational risk and for the leverage ratio. Whatever the convention, the student must ask two questions in any bank: at what ratio is capital allocated, and on which RWA: the bank’s own models, or the floor?
IN ONE SENTENCE — Exposure at default (EAD) is the amount the bank expects to have lent when the borrower defaults: the drawn balance plus a share of whatever the borrower can still draw.
is the credit conversion factor, set by Annex I of the CRR for off-balance-sheet items. CRR3 (Regulation (EU) 2024/1623, applicable since 1 January 2025) defines five buckets: 100% (full-risk items such as financial guarantees), 50% (medium risk, such as performance bonds outside the trade finance definition), 40% (other commitments, which captures committed credit lines whatever their maturity), 20% (medium-low risk, trade finance items) and 10% (unconditionally cancellable commitments, phased in from 0% until the end of 2029). The foundation IRB approach now uses the same conversion factors as the standardised approach. B03 shows what this does to a revolving facility, B11 to a guarantee line; here we use their results.
PITFALL — “It is undrawn, so it costs nothing” is the most expensive sentence in corporate banking. An undrawn RCF at 40% CCF consumes capital on 40% of its amount from the day it is signed, which is why commitment fees exist.
The standardised approach (SA) reads risk weights from tables in Part Three, Title II, Chapter 2 of the CRR. For corporates, Article 122 maps external ratings to weights; CRR3 introduced a 75% weight for the third credit quality step (BBB range), where the previous rule gave 100%. Unrated corporates are weighted 100%. Specialised lending (project, object and commodities finance) has its own article, 122a: unrated project finance is weighted 130% before operation, 100% in operation and 80% if it meets the “high-quality” conditions; object finance is weighted 100%. Commercial real estate falls under Articles 124 to 126: for income-producing real estate the weight depends on the exposure-to-value ratio (70%, 90% or 110%), and development lending takes 150%.
The SA is simple, comparable across banks and blind to everything a rating does not capture. Its weights are calibrated on average loss experience; a BBB with a fortress balance sheet and a BBB on a negative watch carry the same 75%.
IN ONE SENTENCE — Under the internal ratings-based (IRB) approach the bank supplies the probability of default (and, under the advanced variant, the loss given default and the conversion factor), and a supervisory formula turns them into the capital that covers unexpected loss at a 99.9% confidence level.
INTUITION — Expected loss is the rain you know will fall every year: you price it into the margin and provision it. Capital is the umbrella for the storm that comes once in a thousand years. The formula asks: in that storm, how much worse than average would this loan do?
FORMULA
is the capital requirement per euro of EAD; the one-year probability of default, floored at 0.05% for corporates; the loss given default (supervisory 40% for senior unsecured exposures to non-financial corporates under the foundation approach); the standard normal distribution function and its inverse; the asset correlation, which falls from 24% for the best credits to 12% for the weakest; the effective maturity in years; the maturity adjustment; the risk weight.
Term by term. The bracket is a stressed default rate minus the expected one: the formula conditions the borrower’s default on a systematic factor at its 99.9th percentile (the Vasicek one-factor model) and subtracts the average, which is supposed to be covered by margin and provisions. turns defaults into losses. The maturity term increases capital for longer loans, because a long loan has more time to be downgraded; it equals one when = 2.5 years. Multiplying by 12.5 converts a capital requirement into an RWA equivalent (12.5 is one over 8%). CRR3 removed the 1.06 scaling factor that used to multiply IRB credit RWA.
Three variants exist. Foundation IRB (F-IRB): own PD, supervisory LGD and CCF. Advanced IRB (A-IRB): own PD, LGD and CCF, subject to input floors. Slotting for specialised lending: no PD at all, a supervisory category instead (strong 70%, good 90%, satisfactory 115%, weak 250%, with lower weights for strong and good under 2.5 years of remaining maturity). CRR3 removed the A-IRB option for exposures to large corporates (consolidated revenues above EUR 500m) and to financial institutions: those exposures are now F-IRB at best. Meridian, as stated across the course, uses F-IRB for corporates and slotting for project finance.
INTERACTIVE LAB — The IRB risk weight as a function of PD, with the standardised weights as reference
See how the CRR3 corporate IRB formula maps PD, LGD and maturity to a risk weight, and where it crosses the standardised weights of 75% (BBB) and 100% (unrated).
| Parameter you can move | Range | Default |
|---|---|---|
| LGD | 10 to 45 % | 40 % |
| Effective maturity M | 1 to 5 years | 2.5 years |
| PD for the reading | 0.05 to 5 % | 0.2 % |
| Result at default settings | Value |
|---|---|
| IRB risk weight at this PD | 39.02 % |
| Expected loss PD x LGD | 0.080 % of EAD |
| IRB weight as a share of the BBB standardised weight (75%) | 52.0 % |
Try it. Find the PD at which the IRB weight crosses 75% with LGD 40% and M 2.5 years: just under 1%, roughly a BB+ borrower. For investment-grade borrowers, internal models halve the capital of the standardised approach.
CASE — A BBB borrower of Meridian’s grid (PD 0.20%) with the supervisory LGD of 40% and = 2.5 years carries a risk weight of 39.02%, about half of the 75% standardised weight. A BB borrower (PD 1.00%) carries 82.06%. Expected loss for the BBB is 0.08% of EAD a year.
IN ONE SENTENCE — The output floor prevents a bank’s total RWA computed with internal models from falling below a fixed percentage of the RWA it would have under the standardised approaches.
is the floor percentage, phased in under Article 465: 50% in 2025, 55% in 2026, 60% in 2027, 65% in 2028, 70% in 2029 and 72.5% from 2030; is the bank’s whole balance sheet recomputed with standardised approaches. The floor applies at the level of the bank, not deal by deal. Two transitional arrangements matter for corporate banking: until the end of 2032, IRB banks may weight unrated corporate exposures at 65% in the floor calculation (instead of 100%) if their own PD is at most 0.5%, and residential mortgages benefit from a reduced weight under conditions. The Commission’s communication on the competitiveness of the EU banking sector, published on 17 July 2026, announced proposals on precisely these two transitional arrangements, with a legislative package expected in the first quarter of 2027.
PITFALL — “The floor is 72.5% of the standardised weight, so my deal costs 72.5% × 75% = 54.4%.” Only if the bank is floor-bound. If it is not, the marginal deal costs its IRB weight; if it is, the marginal deal costs 72.5% of its standardised weight whatever its IRB weight. A bank sitting just under the floor faces a kink: it must know which side it will be on over the life of the loan.
INTERACTIVE LAB — When does the output floor bite? Meridian’s phase-in path
Project Meridian’s modelled RWA against the phased-in output floor and find the year in which the floor becomes binding, with its effect on the CET1 ratio.
| Parameter you can move | Range | Default |
|---|---|---|
| RWA with internal models (2026) | 80 to 160 EUR bn | 118 EUR bn |
| Standardised-equivalent RWA (2026) | 100 to 250 EUR bn | 165 EUR bn |
| Annual growth of modelled RWA | -2 to 8 % | 0 % |
| Annual growth of standardised-equivalent RWA | -2 to 8 % | 0 % |
| CET1 capital | 5 to 30 EUR bn | 16.52 EUR bn |
| Result at default settings | Value |
|---|---|
| Modelled RWA as a share of standardised RWA (2026) | 71.52 % |
| First year in which the floor binds (0 if never by 2030) | 2,030 |
| Floor add-on in 2030 | 1.625 EUR bn |
| CET1 ratio in 2030 with constant capital | 13.81 % |
| CET1 ratio in 2026 | 14.00 % |
Try it. Let standardised-equivalent RWA grow 2 points faster than modelled RWA (a shift towards low-PD corporates, where models save most): the floor arrives a year earlier.
CASE — Meridian’s modelled RWA are EUR 118bn and its standardised-equivalent RWA EUR 165bn (case assumption): an RWA density of 71.52% of the standardised figure. At 55% in 2026 the floor (EUR 90.75bn) is far away; at 70% in 2029 it is still below; at 72.5% in 2030 it reaches EUR 119.625bn, EUR 1.625bn above modelled RWA. With constant capital the CET1 ratio would fall from 14.00% to 13.81%. From 2030 Meridian is floor-bound at the margin: this is the planning fact that section 5.4 exploits.
The credit application of every deal shows EAD, risk weight, RWA, expected loss and RAROC, computed by a pre-deal calculator owned by capital management. After booking, the monthly engine recomputes them; differences between pre-deal and booked RWA (wrong product code, missing collateral link, wrong CCF bucket) are a recurring source of disputes that a banker who knows the mechanics catches early.
IN ONE SENTENCE — In the EU as of November 2026 there is no capital discount for green exposures and no surcharge for brown ones: sustainability enters Pillar 1 only where it changes measured risk, and it enters Pillar 2 through the supervisor’s judgement.
The CRR does not ask what a loan finances; it asks how likely the borrower is to default, how much would be lost, and how much is exposed. A green loan to a BBB borrower carries the BBB risk weight. A loan to a coal miner with the same rating carries the same. This is the principle that the green supporting factor debate (section 3.4) puts into question.
ESG nevertheless enters Pillar 1 through five doors:
| Door | Article | How ESG can move capital |
|---|---|---|
| PD | Internal models, Art. 170-180 | Transition and physical risk as rating drivers |
| LGD and collateral | Art. 208, 229 | Energy performance and physical risk in property values |
| Infrastructure supporting factor | Art. 501a | 0.75 multiplier, with an environmental assessment condition |
| Slotting criteria | Art. 153(9) RTS | ESG factors in project and object finance scoring |
| External ratings | Art. 122, 138 | Rating agencies’ treatment of ESG factors |
Article 501a multiplies the own funds requirement for credit risk by 0.75 for exposures to entities that operate or finance physical structures providing essential public services, provided a long list of conditions is met: revenues predictable and mostly from the assets, low refinancing risk, senior status, tested technology, permits obtained, and an assessment by the obligor of whether the financed assets contribute to the EU’s environmental objectives and do no significant harm. It is the closest thing in the CRR to a green factor, but it is not one: a toll road qualifies and an offshore wind farm with an unproven foundation design may not. B06 applies it to Cap Ferrat Offshore Wind. It cannot be combined with the 80% high-quality project finance weight of the standardised approach. A separate SME supporting factor (Article 501) reduces capital on small and medium-sized enterprise exposures; neither is linked to sustainability as such.
CRR3 strengthened the rules on property valuation. The value of real estate collateral is the property value under Article 229, assessed prudently and excluding expectations of price increases. Upward revaluations are capped, but the text allows an increase where modifications unequivocally increase the value, explicitly including improvements in energy performance or in the resilience of the building to physical risks. B08 uses this door: the retrofit of the Lutetia tower from EPC F to B raises the property value, which lowers the exposure-to-value ratio and the collateral-adjusted LGD. Article 208 requires that valuation and monitoring take into account environmental and climate risks. This is how a green feature can lower capital legally: through the risk parameters, documented by a valuer, not through a label.
IN ONE SENTENCE — A green supporting factor (GSF) would multiply the risk weight of qualifying green exposures by a number below one; a brown penalising factor (BPF) would multiply that of carbon-intensive exposures by a number above one.
The idea was promoted by parts of the banking industry and some policymakers from 2017, by analogy with the SME supporting factor. During the CRR3 negotiations amendments to introduce a GSF or a BPF were tabled; none was adopted. Instead the legislator mandated the EBA to assess whether a dedicated prudential treatment of exposures related to environmental or social objectives would be justified.
The EBA delivered its final Report on the role of environmental and social risks in the prudential framework on 12 October 2023. Its conclusions, which remain the reference in November 2026:
Section 5.6 puts numbers on the debate.
Pillar 2 is where supervisors deal with climate and environmental risk today. Four instruments matter.
The supervisory expectations. The ECB’s Guide on climate-related and environmental risks (November 2020) set thirteen expectations, from business strategy to stress testing. After the 2022 thematic review the ECB set staggered deadlines: materiality assessment by March 2023, full integration into governance, strategy and risk management by end-2023, and into the ICAAP and stress testing by end-2024.
The legal anchor. The Capital Requirements Directive as amended by Directive (EU) 2024/1619 (CRD6) requires banks to have robust strategies and processes to identify, measure, manage and monitor ESG risks over short, medium and long horizons (at least ten years), and to prepare plans with quantifiable targets to address the financial risks arising from the transition. It empowers supervisors to require a reduction of those risks. The EBA Guidelines on the management of ESG risks, published in January 2025, apply from 11 January 2026 to all but small and non-complex institutions, which follow by January 2027.
The SREP outcome. Shortcomings feed the SREP score of internal governance and risk management, which can lead to qualitative measures (remediation plans) and, in principle, to a higher P2R. In its public communications the ECB has explained that climate shortcomings influenced SREP scores and, for a few banks, the P2R in earlier cycles, but it does not publish a separate “climate add-on” and the effect is small relative to the total P2R.
Enforcement. The ECB has moved from expectations to penalties. On 10 November 2025 it announced periodic penalty payments of EUR 187,650 on ABANCA for failing, for 65 days in 2024, to meet the requirement of a December 2023 decision to assess the materiality of its climate-related and environmental risks. On 13 February 2026 it announced periodic penalty payments of EUR 7,551,050 on Crédit Agricole S.A. for failing, for 75 days in 2024, to meet a similar requirement of a February 2024 decision. The amounts are small relative to the banks’ earnings; the reputational signal is not. Periodic penalty payments are capped at 5% of average daily turnover per day of infringement for up to six months under the SSM Regulation.
NOTE — The ICAAP is the bank’s own view of the capital it needs, not the regulator’s. Meridian’s ICAAP must show that climate risk was assessed for materiality, quantified in scenarios, and either covered by capital or by limits. A credible ICAAP that quantifies climate risk is the best defence against a supervisory add-on.
Article 449a of the CRR requires Pillar 3 disclosures on ESG risks. CRR3 extended them from large listed institutions to all institutions, with proportionality. On 22 June 2026 the EBA published final draft implementing technical standards amending the ESG Pillar 3 framework: fewer datapoints for large institutions (taxonomy-related templates are stopped), first-time disclosure for other institutions, and a first reference date of 31 December 2026 (31 December 2027 for small and non-complex institutions), subject to the Commission’s adoption. The Green Asset Ratio itself is disclosed under the Taxonomy Regulation’s Article 8 delegated act, not under the CRR (see A02, A04). Neither metric affects capital. Basel standards have stayed voluntary: on 13 June 2025 the Basel Committee published a voluntary framework for the disclosure of climate-related financial risks, after a consultation that had proposed a mandatory one.
Pillar 1 assumes an infinitely granular portfolio and a single systematic factor. Climate risk is the opposite: it concentrates on sectors (power, oil and gas, steel, cement, real estate) and on regions (flood zones, heat). Concentration therefore enters capital through Pillar 2 (a concentration add-on in the ICAAP, single-name and sector), through the large exposures regime (25% of Tier 1 per client group), and through limits in the risk appetite. Capital planning uses the same data in reverse: a bank that commits to a transition plan with sector targets (see B01) must check that the planned growth in renewables and grids, often long-dated and specialised lending, fits its RWA budget and its concentration limits.
A deal does not “have” a risk weight; it acquires one through a chain of decisions made by different people at different dates, and it keeps it only as long as each input stays right. The chain below follows one sustainable loan, the Borealis RCF of B03, from the first conversation to its place in the supervisor’s annual review.
INTERACTIVE LAB — The life of a deal’s capital: from pre-deal RWA to the SREP
Follow who produces which capital figure for a single loan, from origination to the supervisor’s annual review.
| # | Who | Does what | Deliverable |
|---|---|---|---|
| 1 | Relationship manager and loan structurer | Enter facility size, tenor, product type, expected utilisation and collateral in the pre-deal calculator | Indicative EAD, risk weight, RWA and RAROC |
| 2 | Credit analyst (Credit Risk) | Run the rating model, assign the obligor grade and PD, check the LGD class and any collateral eligibility | Rating and risk parameters in the credit application |
| 3 | Capital management (pre-deal desk) | Confirm the regulatory treatment: exposure class, CCF bucket, approach (F-IRB, slotting, SA), maturity convention, supporting factors | Validated capital figures and floor-bound marginal cost |
| 4 | Credit committee | Approve the deal with its RAROC, limits and conditions | Credit approval |
| 5 | Loan operations and booking | Book the facility with the correct product code, collateral links and regulatory attributes | Exposure in the core banking system |
| 6 | RWA production team | Run the monthly RWA engine, reconcile EAD with accounting balances and investigate variances | Monthly RWA report and variance analysis |
| 7 | Regulatory reporting | File COREP own funds templates and the quarterly large exposures report | COREP C 02.00, C 08.xx and large exposures returns |
| 8 | Credit risk modelling and validation | Back-test PD models annually, recalibrate if needed, report model changes to the ECB | Validation report and model change notification |
| 9 | Capital planning | Project RWA and capital over the planning horizon, including the output floor phase-in and climate scenarios | Capital plan and ICAAP |
| 10 | Sustainability reporting and Finance | Tag the exposure for Pillar 3 ESG templates, financed emissions and the Green Asset Ratio | Pillar 3 ESG disclosure and sustainability statement |
| 11 | Board risk committee | Approve the risk appetite, capital plan and ICAAP conclusions | Board resolution and ICAAP submission |
| 12 | ECB Joint Supervisory Team | Assess the ICAAP, models and climate risk management in the SREP and set P2R and P2G | SREP decision for the following year |
Step 1: pre-deal calculation. The relationship manager and the structurer enter the deal in the pricing tool. The tool is owned by capital management but sits in the front office; it applies the same rules as the monthly engine, with the rating and collateral entered by hand. The output is indicative, and every experienced banker knows that the two most common errors are made here: the wrong product type (an RCF entered as a term loan has no undrawn amount and no CCF) and a missing link to collateral (a ship loan without the mortgage recorded is unsecured, with an LGD of 40% instead of 25%).
Step 2: rating. The credit analyst runs the rating model and proposes the obligor grade. For a sustainability-linked or green deal, this is where ESG enters the risk weight legally: the rating model may include transition risk questions (carbon intensity, credibility of the transition plan, exposure to carbon pricing), and the analyst may override for a documented ESG reason. The rating is approved by the credit officer, never by the front office. Under the EBA Guidelines on ESG risk management, the credit file must also contain an ESG assessment of the client.
Step 3: regulatory treatment. For standard products the tool decides. For anything unusual (a guarantee of more than one year, a project in construction, a real estate loan that might be development lending, a structure with a public guarantor, an exposure that might qualify for the infrastructure supporting factor), the deal team should ask the pre-deal desk of capital management. The desk’s answer is a short memo: exposure class, approach, CCF, maturity, supporting factors, and whether the bank is floor-bound for the relevant period. These memos are the bank’s institutional memory of how it reads ambiguous articles.
Step 4: approval. The credit committee approves the deal on the basis of the credit application, which contains EAD, RWA, expected loss and RAROC. For large deals, a capital usage above a threshold requires the agreement of the business line’s capital budget owner.
Step 5: booking. Loan operations book the facility. The regulatory attributes (counterparty type, revenue size for the A-IRB carve-out, collateral type and value, CCF bucket, ISF eligibility flag, green flag) are entered here. An error at this stage propagates into every monthly report until someone sees it.
Step 6: monthly RWA run. The RWA engine takes the books at month end, joins exposures with ratings, collateral and counterparty data, and computes RWA exposure by exposure. The production team reconciles total EAD with the accounting balance sheet (the “finance-risk reconciliation”) and explains month-on-month changes. Large unexplained variances are escalated.
Step 7: regulatory reporting. COREP templates are filed quarterly with the supervisor: own funds and ratios (C 01.00 to C 05.00), credit risk under the SA (C 07.00) and IRB (C 08.01 to C 08.07), specialised lending slotting (C 08.06), and the output floor templates introduced with CRR3. Large exposures are reported in parallel.
Step 8: model governance. The PD model behind Borealis’s rating is validated annually by an independent unit. If the default experience departs from the calibration, the model is recalibrated, and material changes need the ECB’s prior approval. A recalibration that raises PDs by one notch for a sector raises its RWA mechanically: for a BBB portfolio, moving from 0.20% to 0.30% PD raises the F-IRB weight from about 39% to about 48%.
Step 9: capital planning and ICAAP. Capital planning projects RWA by business line under a baseline and adverse scenarios. Since 2024 Meridian’s ICAAP includes climate scenarios: transition shocks on PD for carbon-intensive sectors and physical shocks on collateral values in flood-prone areas. A02 shows the order of magnitude: doubling the PD of a EUR 14bn carbon-intensive portfolio costs about 36 bp of CET1 ratio.
Step 10: sustainability reporting. The same exposure is tagged for Pillar 3 ESG templates (sector, maturity bucket, energy performance of collateral, financed emissions) and for the Green Asset Ratio. The data come from the same booking attributes, which is why ESG data quality is now a topic for the RWA production team as much as for the sustainability team.
Steps 11 and 12: governance and supervision. The Board approves the capital plan, the risk appetite and the ICAAP. The ECB’s Joint Supervisory Team reviews them, adds its own view, and sets P2R and P2G for the following year in the SREP decision, usually notified in the last quarter.
WHO DOES WHAT — The deal team owns the inputs (product, collateral, utilisation); Credit Risk owns the rating and PD; capital management owns the regulatory treatment and the engine; Finance owns the plan and the allocation; the Board owns the appetite; the supervisor owns the final word through Pillar 2.
The case takes five deals that Meridian has already met in this course and asks the capital management questions that their own modules left in the background. Every input below comes from section 6.1 of the originating module; nothing is re-estimated.
Question. Meridian’s Group Treasurer and the head of capital management prepare a note for the Asset and Liability Committee (ALCO): “What does our sustainable lending cost in capital, under each approach we might be asked to apply, and which regulatory choices move that cost most?” The note must cover five recently approved deals, chosen because they span the main exposure classes of a CIB.
Data (taken from the originating modules).
| Deal | Exposure class | Meridian’s amount | PD / LGD | Approach used in its module |
|---|---|---|---|---|
| B03 Borealis, SL RCF | Corporate, off-balance | EUR 75m commitment, 20% drawn | 0.20% / 40% | F-IRB, M 2.5 |
| B06 Cap Ferrat, offshore wind | Specialised lending, project finance | EUR 150m | 0.50% / 40% | Slotting strong, ISF |
| B08 Lutetia, office retrofit | IPRE | EUR 75m | 1.00% / 20% | F-IRB, collateral LGD |
| B09 Nordhavn, methanol vessels | Object finance, guaranteed | EUR 32m | 1.00% / 25% | F-IRB, physical collateral |
| B11 Ventera, green guarantee line | Corporate, trade contingent | EUR 240m outstanding | 0.20% / 40% | F-IRB, 20% CCF |
Additional data, also from the originating modules: Borealis is rated BBB by agencies; Cap Ferrat is in operation, in the slotting category “strong”, and meets the conditions of the infrastructure supporting factor; Lutetia’s stabilised exposure-to-value is 53.1%; Nordhavn’s vessels are financed at a 60% loan-to-value with a ten-year tenor and a weighted average life of 7.5 years; Ventera is a BBB contractor whose guarantees have original tenors of 4 months (bid bonds, EUR 10m), 18 months (advance payment guarantees, EUR 60m), 30 months (performance bonds, EUR 120m) and 24 months (warranty bonds, EUR 50m). Meridian allocates 13% of RWA.
Reading. These five deals carry four different green or sustainable labels (sustainability-linked, green project, green building, transition shipping, green trade). The point of the exercise is to see that the capital calculation does not look at any of the labels, and to find what it does look at.
Question. How much exposure does each deal carry in regulatory terms?
Calculation. Two deals are off-balance-sheet in part. For Borealis, the drawn 20% of EUR 75m (EUR 15m) counts in full and the undrawn EUR 60m at the 40% CCF of committed facilities: EUR 24m, so EAD is EUR 39m. For Ventera, all EUR 240m of guarantees outstanding are trade-related contingent items in bucket 4 under the reading that Meridian’s policy applies (B11, section 6.1): 20% CCF, so EAD is EUR 48m. Cap Ferrat (EUR 150m), Lutetia (EUR 75m, the stabilised phase) and Nordhavn (EUR 32m) are drawn loans with EAD equal to the amount outstanding.
Result. Total EAD is EUR 344m for EUR 572m of commitments and guarantees (75 + 150 + 75 + 32 + 240).
PITFALL — The CCF is where small words move big numbers. If Meridian applied the 50% bucket to Ventera’s guarantees with an original tenor above one year (EUR 230m of the 240m), EAD would rise to EUR 117m, almost two and a half times more (B11, section 5.7). The green label on the line changes nothing; the reading of “generally of less than one year” in Article 4(1)(80) changes a lot.
Question. What would these exposures cost a bank without internal models, and what does Meridian need to know about it even though it has models?
Calculation. Standardised weights by deal:
Result. Standardised RWA are EUR 29.25m, EUR 112.5m, EUR 52.5m, EUR 32m and EUR 36m, a total of EUR 262.25m, and standardised capital at 13% is EUR 34.09m.
Reading. Meridian needs this number for two reasons. It is the basis of the output floor; and it is what a competitor without models (a smaller bank, a credit fund benchmarking its returns against banks) sees. If unrated, Borealis and Ventera would be weighted 100% under the SA, but an IRB bank may use 65% in its floor calculation until the end of 2032 for unrated corporates with an internal PD of at most 0.5%.
Question. What do the deals cost under the approaches Meridian actually uses, and what changes when Meridian becomes floor-bound in 2030?
Calculation. Applying the IRB formula of section 2.4 with Meridian’s convention = 2.5 years:
INTERACTIVE LAB — Capital of five sustainable deals: standardised, IRB and the output floor
Recompute EAD, RWA and capital for the five deals of the case under Meridian’s internal approaches and the standardised approach, and switch the regulatory choices that move them: maturity convention, slotting or F-IRB for the project, supporting factor, rating status, output floor level.
| Parameter you can move | Range | Default |
|---|---|---|
| F-IRB maturity convention (Art. 162(1)) | Fixed 2.5 years / Calculated (Art. 162(2)) | Fixed 2.5 years |
| Approach for the offshore wind project | Slotting strong / F-IRB (PD 0.50%) | Slotting strong |
| Infrastructure supporting factor on the project | yes / no | yes |
| Standardised treatment of Borealis and Ventera | Rated BBB (75%) / Unrated, floor transitional (65%) / Unrated (100%) | Rated BBB (75%) |
| Output floor percentage | 55 / 60 / 65 / 70 / 72.5 | 72.5 % |
| Utilisation of the Borealis RCF | 0 to 100 % | 20 % |
| Hypothetical green factor on all five (1 = none) | 0.5 to 1.25 x | 1 x |
| Result at default settings | Value |
|---|---|
| Effective maturity of the guarantee book if calculated | 1.125 years |
| Total EAD | 344.00 EUR m |
| RWA Borealis (internal) | 15.22 EUR m |
| RWA Cap Ferrat (internal) | 78.75 EUR m |
| RWA Lutetia (internal) | 30.77 EUR m |
| RWA Nordhavn (internal) | 16.41 EUR m |
| RWA Ventera (internal) | 18.73 EUR m |
| Total RWA, internal approaches | 159.88 EUR m |
| Total RWA, standardised approach | 262.25 EUR m |
| Floor percentage of standardised RWA | 190.13 EUR m |
| Internal RWA as a share of standardised RWA | 60.96 % |
| Capital at 13%, internal approaches | 20.78 EUR m |
| Capital at 13%, standardised | 34.09 EUR m |
| Marginal capital at 13% for a floor-bound bank | 24.72 EUR m |
| Regulatory expected loss (internal parameters) | 0.5996 EUR m |
| Capital effect of calculated maturity | 2.44 EUR m |
| Capital effect of losing the supporting factor | 3.41 EUR m |
| Capital effect of a 0.75 green factor | -5.20 EUR m |
Try it. Switch the maturity convention to ‘Calculated’: the four long-dated green loans gain capital and the guarantee book loses some. Then set the floor to 55% (the 2026 level) and to 72.5% (2030) and compare the floor-bound capital with the internal figure.
Result. Under Meridian’s approaches the five deals consume EUR 159.88m of RWA and EUR 20.78m of capital, an RWA density of 60.96% of the standardised figure. The regulatory expected loss, the sum of PD × LGD × EAD, is EUR 0.5996m a year: under 0.2% of EAD. The floor at 72.5% of standardised RWA is EUR 190.13m, so a floor-bound Meridian would count EUR 24.72m of capital for these deals at the margin.
Reading, deal by deal. The bar chart is more instructive than the totals.
CASE — For the ALCO note: five sustainable deals, EUR 344m of EAD, EUR 20.78m of capital today; EUR 24.72m once the floor binds in 2030 under current rules. The difference, EUR 3.93m of capital, costs about EUR 0.63m of pre-tax earnings a year at Meridian’s 12% after-tax hurdle (EUR 3.93m × 12% / 75%). None of these numbers depends on the deals being green.
The last figure needs its own test, which the next block provides along with the remaining marginal-cost readings of the case.
INTERACTIVE LAB — The marginal cost of the output floor for one deal
Compare the capital and the pre-tax earnings needed to meet the hurdle for one exposure, under its internal risk weight and as a floor-bound marginal exposure.
| Parameter you can move | Range | Default |
|---|---|---|
| Exposure at default | 1 to 500 EUR m | 344 EUR m |
| Internal risk weight (average) | 10 to 150 % | 46.4765 % |
| Standardised risk weight (average) | 20 to 150 % | 76.2355 % |
| Floor percentage | 55 / 60 / 65 / 70 / 72.5 | 72.5 % |
| RAROC hurdle after tax | 8 to 16 % | 12 % |
| Result at default settings | Value |
|---|---|
| Capital under internal weight (13%) | 20.78 EUR m |
| Capital if floor-bound (13%) | 24.72 EUR m |
| Extra capital when floor-bound | 3.93 EUR m |
| Extra pre-tax earnings to meet the hurdle | 0.629 EUR m |
| Same, in bp of EAD | 18.3 bp |
| Floor-bound risk weight | 55.27 % |
Try it. Set the floor to 55%, the 2026 level: the ‘extra’ capital is negative, which means the floor is irrelevant for a deal priced today but booked in a bank that is not floor-bound. The question for a seven-year loan is which side of the floor the bank will be on in 2030.
Reading. The average internal weight of the five deals is 46.48% (EUR 159.88m over EUR 344m) and the average standardised weight 76.24% (EUR 262.25m over EUR 344m). For a floor-bound bank, the five deals need EUR 0.63m more pre-tax earnings a year, 18.3 bp of EAD, to earn the same 12% hurdle. For Borealis alone (EAD EUR 39m, weights 39.02% and 75%) the floor-bound weight is 54.38% and the extra earnings needed are about EUR 0.125m a year, 31.9 bp of EAD: more than ten times the largest annual revenue loss from the sustainability ratchet computed in B03 (EUR 10,800). On investment-grade RCFs it is the floor, not the green feature everyone negotiates about, that decides the economics. This is why European banks lobbied for the transitional 65% weight for unrated investment-grade corporates and why the Commission’s July 2026 communication targets it.
Question. Several modules of this course used = 2.5 years in the F-IRB formula; some described it as fixed by regulation, others as a bank choice requiring permission, and others as something banks now have to calculate. Which is right under CRR3, and what difference does it make for these five deals?
The rule. Article 162(1) of the CRR, as replaced by point (88)(a) of Article 1 of Regulation (EU) 2024/1623, now reads in substance: for exposures for which an institution has not received permission to use own estimates of LGD, the maturity value shall be applied consistently and either set at 2.5 years (0.5 years for securities financing transactions) or, alternatively, calculated in accordance with paragraph 2. Paragraph 2 (as amended by point (88)(b)) applies to institutions using own LGD estimates, which must calculate .
Three consequences follow. First, a fixed 2.5 years remains lawful for F-IRB under CRR3; the claim that CRR3 forces F-IRB banks to compute effective maturity is wrong for the EU, even though the Basel framework allows national supervisors to require it. Second, the choice is now framed as the institution’s own, subject to consistency; the former text instead gave competent authorities the decision, as part of the IRB permission, whether the institution should calculate . Third, “consistently” excludes cherry-picking: a bank cannot use calculated maturity for its short trade book and the fixed 2.5 years for its long project loans. Banks on the advanced approach (own LGD) must calculate .
FORMULA
is the cash flow (principal, interest and fees) contractually payable by the obligor in period ; is the time in years; the ratio is a cash-flow-weighted average life. For instruments without a schedule, is the maximum remaining time (in years) that the obligor is permitted to take to discharge its obligations, which for a committed revolving facility means its final maturity. The floor of one year does not apply to certain short-term exposures listed in Article 162(3), notably short-term self-liquidating trade transactions, which can take their actual maturity; the cap of five years always applies.
INTUITION — The fixed 2.5 years is like an insurer charging every driver the premium of a five-year-old car. Calculating charges the old car more and the new car less. Which one you prefer depends on what you drive.
Calculation. With calculated maturity:
INTERACTIVE LAB — Effective maturity: what Article 162 does to long green loans and short trade instruments
Plot the F-IRB risk weight against effective maturity for the parameter sets of the case and compare it with the fixed 2.5-year convention.
| Parameter you can move | Range | Default |
|---|---|---|
| Parameter set | BBB unsecured (B03, B11) / Project PD 0.50% (B06) / IPRE LGD 20% (B08) / Vessels LGD 25% (B09) | BBB unsecured (B03, B11) |
| Effective maturity M | 1 to 5 years | 5 years |
| Result at default settings | Value |
|---|---|
| Risk weight at this M | 59.56 % |
| Risk weight at M = 2.5 years | 39.02 % |
| Change against the 2.5-year convention | 52.7 % |
| Maturity adjustment b | 0.21064 |
Try it. Compare the BBB curve with the IPRE curve: the maturity effect is proportionally larger for low PDs, because a good borrower has more room to be downgraded over five years.
Result. Total internal RWA rise from EUR 159.88m to EUR 178.69m, capital from EUR 20.78m to EUR 23.23m, and the density from 60.96% to 68.14% of standardised RWA. Deal by deal: Borealis +52.7%, Lutetia and Nordhavn +34.4%, Ventera −29.0%.
Reading. The fixed 2.5 years is a subsidy to long-dated lending and a tax on short-dated lending. Sustainable finance is dominated by long-dated assets: project finance, real estate retrofits, vessels, grids. A bank that moved to calculated maturity would see the capital cost of its green book rise by a third, while its trade and short-term book would get cheaper. This is a regulatory choice with a larger effect on green lending than any supporting factor so far discussed, and it is rarely mentioned in the green capital debate. It also explains why F-IRB banks in the EU overwhelmingly keep the fixed value, and why the 0.25-year variant of B11 for letters of credit is only available to a bank that calculates maturity for its whole F-IRB book.
PITFALL — Do not confuse the maturity in the formula with the remaining maturity used in slotting. For slotting, a remaining maturity under 2.5 years lowers the weight of “strong” from 70% to 50% and of “good” from 90% to 70%, whatever the F-IRB convention.
Question. A trade association asks Meridian to support a call for a 0.75 green supporting factor on taxonomy-aligned exposures, and a 1.25 brown penalising factor on fossil-fuel exposures. What would they do to Meridian’s capital, and what would a 0.75 factor claim about risk?
Data (case assumptions). Meridian’s taxonomy-aligned lending carries EUR 10bn of RWA. Its most carbon-intensive corporate portfolio is the EUR 14bn book of A02, weighted at 82.06%: EUR 11.49bn of RWA. CET1 capital EUR 16.52bn, RWA EUR 118bn.
Calculation. A factor that cuts the risk weight of a BBB loan (PD 0.20%, LGD 40%, = 2.5) by 25% must be compared with the PD that would produce the same weight through the formula. We solve for such that .
FORMULA
is the supporting factor (below one) or penalising factor (above one); the probability of default that gives the same risk weight as the factor; the relative fall in default probability that the factor implicitly assumes, holding LGD and maturity unchanged.
INTERACTIVE LAB — What a green supporting factor assumes about default risk, and what it does to the CET1 ratio
Translate a supporting or penalising factor into the equivalent change in PD, and compute the effect on Meridian’s CET1 ratio of applying it to the green and brown books.
| Parameter you can move | Range | Default |
|---|---|---|
| PD of the green borrower | 0.05 to 4 % | 0.2 % |
| LGD | 10 to 45 % | 40 % |
| Green supporting factor | 0.5 to 1 x | 0.75 x |
| RWA of the green book | 0 to 40 EUR bn | 10 EUR bn |
| Brown penalising factor | 1 to 1.5 x | 1.25 x |
| RWA of the brown book | 0 to 40 EUR bn | 11.49 EUR bn |
| Total RWA | 50 to 200 EUR bn | 118 EUR bn |
| CET1 capital | 5 to 30 EUR bn | 16.52 EUR bn |
| Result at default settings | Value |
|---|---|
| Risk weight before the factor | 39.02 % |
| PD giving the same weight as the factor | 0.1198 % |
| Implied fall in PD | 40.1 % |
| RWA released by the GSF | 2.50 EUR bn |
| CET1 ratio with the GSF | 14.30 % |
| CET1 ratio with the BPF | 13.67 % |
| CET1 ratio with both | 13.96 % |
Try it. Set the PD to 1.00%: a 0.75 factor now assumes that green borrowers default half as often. Then set the factor to 0.9: the implied PD gap falls to about 17%, closer to what some empirical studies find for energy-efficient mortgages.
Result. For a BBB borrower the 0.75 factor is equivalent to cutting the PD from 0.20% to 0.1198%, a fall of 40.1%; for a BB borrower (PD 1.00%) the equivalent PD is 0.4941%, a fall of 50.6%. Applied to Meridian’s EUR 10bn green RWA, the factor releases EUR 2.5bn of RWA and lifts the CET1 ratio from 14.00% to 14.30%. The brown factor on the EUR 11.49bn carbon-intensive book lowers it to 13.67%; both together give 13.96%.
Reading. Three points for the ALCO note.
The brown penalising factor has a different logic, sometimes called the “double materiality” or prudential-plus-policy argument: it would make banks pay for the systemic risk of financing emissions, even if individual borrowers are not riskier today. Supervisors have resisted it for the same reason: capital requirements are meant to cover risk to the bank, not to steer credit, and using them to steer credit risks undermining their credibility. The honest summary: both factors are policy instruments dressed as prudential ones; Pillar 1 has so far refused them, and the work has moved to Pillar 2 and to better risk parameters.
WHY IT MATTERS ON THE DESK — On the desk, you will hear “green assets should be cheaper in capital”. The answer that shows you understand the issue is a number: a 25% capital discount is a claim that the borrower defaults 40% less often, and nobody has shown that.
Question. Suppose the ECB, in its next SREP, concludes that Meridian’s integration of climate risk into the ICAAP remains insufficient and raises its P2R by 25 bp. What does that cost, and how should it be allocated to deals?
Calculation. The P2R applies to all RWA. At least 56.25% must be met with CET1 and 75% with Tier 1. Capital management must decide whether to absorb the add-on in the management buffer or to raise the allocation rate charged to businesses.
INTERACTIVE LAB — The cost of a climate-related Pillar 2 add-on
Compute the extra CET1 and total capital needed after a P2R increase, the fall in MDA distance, and the effect on the capital allocated to the five deals if the add-on is passed on to the business lines.
| Parameter you can move | Range | Default |
|---|---|---|
| P2R increase | 0 to 100 bp | 25 bp |
| Total RWA | 50 to 200 EUR bn | 118 EUR bn |
| Allocation rate before the add-on | 10 to 16 % | 13 % |
| RWA of the five deals | 50 to 300 EUR m | 159.88 EUR m |
| RAROC hurdle after tax | 8 to 16 % | 12 % |
| Result at default settings | Value |
|---|---|
| Extra CET1 required | 165.9 EUR m |
| Extra total capital required | 295.0 EUR m |
| Fall in the distance to MDA | 14.06 bp |
| Allocation rate if passed on (CET1 part) | 13.1406 % |
| Extra capital allocated to the five deals | 0.2248 EUR m |
| Bank-wide pre-tax earnings to remunerate the extra CET1 | 26.55 EUR m |
Try it. Compare the EUR 0.22m of extra capital on the five deals with the EUR 3.93m of the output floor (section 5.4) and the EUR 2.45m of calculated maturity (section 5.5): the climate add-on is the smallest of the three regulatory effects on these deals.
Result. A 25 bp P2R add-on requires EUR 165.9m of extra CET1 (EUR 295m of total capital), narrows the MDA distance by 14.06 bp, and if passed on through the allocation rate raises it from 13% to 13.1406%, adding EUR 0.2248m of capital to the five deals. Meridian must earn about EUR 26.55m more before tax each year to remunerate that CET1 at its hurdle.
Reading. The add-on is real money for the bank (EUR 26.55m a year is roughly the RAROC contribution of a mid-sized corporate coverage team) but small for any single deal. Its main effect is on the bank’s behaviour: it falls on everyone, green or brown, so it rewards the bank only for fixing its risk management, which is its purpose. That is why supervisors prefer it to Pillar 1 factors. Since 2025 the ECB has also used periodic penalty payments, which hit the profit and loss account, not capital (section 3.5).
Concentration. The same SREP would look at concentration. Cap Ferrat’s EUR 150m is a small fraction of Meridian’s Tier 1 capital, far from the 25% large exposure limit; but Helios Renewables, Cap Ferrat’s lead sponsor, is also a bond issuer and a client of several desks (see C01). The sector view matters more. Offshore wind is a small number of projects with similar technologies, a few turbine manufacturers and the same weather systems: a Pillar 2 concentration analysis groups them, and the ICAAP may set aside capital for a common-mode failure (a design defect across a turbine model, a cable failure pattern) that Pillar 1, with its single systematic factor, ignores. The irony of a successful green strategy is that it creates concentrations that Pillar 1 does not see.
Question. The numbers are on the table. What does the ALCO actually decide, and who acts on each decision?
Decision 1: the pricing basis for long-dated deals. The committee confirms that every new deal maturing after 2029 is priced on the higher of its internal RWA and 72.5% of its standardised RWA. For the five deals already approved nothing changes in the contract; what changes is their contribution to the business lines’ capital budgets from 2030, and therefore the bonus pool. The head of corporate banking objects that Borealis and Ventera will look less profitable through no fault of the desk. Capital management answers that the cost is real for the bank and that the transitional 65% for unrated investment-grade corporates, and any outcome of the Commission’s 2027 package, will be reflected as soon as they are law, not before. Owner: Finance (pricing tool parameters), with Capital Management.
Decision 2: the maturity convention. The committee keeps the fixed 2.5 years for F-IRB, documents the choice as required by the consistency condition of Article 162(1), and asks the model risk function to run the calculated-maturity variant every quarter as a sensitivity in the ICAAP. The reason is not only the cost of the five deals (EUR 2.45m of capital). Moving to calculated maturity would require reliable contractual cash-flow data for every exposure, would penalise the long-dated sustainable assets the bank’s transition plan wants to grow, and would reward only the short-dated trade book. The committee notes that the Basel framework lets supervisors impose calculation and that this could change; the sensitivity keeps the bank ready. Owner: Capital Management, with Credit Risk Modelling.
Decision 3: the infrastructure supporting factor. The factor saves EUR 3.41m of capital on Cap Ferrat alone, more than any other item except the floor. The committee asks Credit Risk to re-check each condition of Article 501a annually, including the obligor’s environmental assessment, and asks Internal Audit to sample the eligibility files of the project finance book. A factor claimed without its file is a supervisory finding waiting to happen. Owner: Credit Risk, project finance desk.
Decision 4: the position on the green supporting factor. The trade association’s request is answered in two lines: Meridian does not support a factor without evidence of a risk differential, and it supports work on ESG risk drivers in PD and LGD models, on collateral valuation and on harmonised data. The committee notes that a factor would lift the CET1 ratio from 14.00% to 14.30% (stage 5), a tempting figure, and that a bank that relies on regulatory relief rather than on risk measurement is exposed if the relief is withdrawn. Owner: Prudential Regulation, Public Affairs.
Decision 5: Pillar 2 readiness. The committee asks the ICAAP team to present, before the next SREP, a climate scenario that changes at least one decision: a sector limit, a pricing adjustment or a capital buffer. The purpose is to show the Joint Supervisory Team that climate risk is measured and used, which is the best defence against an add-on of the kind computed in stage 6 and against periodic penalty payments. Owner: Risk, Capital Planning.
Reading. None of the five decisions turns on the label of a deal. Each turns on a regulatory reading, a data requirement or a governance step. This is the real content of “inside the bank” for sustainable finance: the label is decided by the deal team and the sustainability specialists; the capital is decided by people who read articles, maintain data and defend models, and the two worlds must talk to each other for the bank’s green ambitions to be financeable.
| Figure | Value | Where |
|---|---|---|
| Total EAD of the five deals | EUR 344m | 5.2 |
| Borealis EAD; Ventera EAD | EUR 39m; EUR 48m | 5.2 |
| Standardised RWA; capital at 13% | EUR 262.25m; EUR 34.09m | 5.3, 5.4 |
| Internal RWA (M 2.5); capital | EUR 159.88m; EUR 20.78m | 5.4 |
| Density of internal over standardised RWA | 60.96% | 5.4 |
| Floor at 72.5%; capital if floor-bound | EUR 190.13m; EUR 24.72m | 5.4 |
| Extra pre-tax earnings when floor-bound | EUR 0.63m, 18.3 bp of EAD | 5.4 |
| Borealis floor-bound weight; extra earnings | 54.38%; EUR 0.125m | 5.4 |
| Regulatory expected loss | EUR 0.5996m | 5.4 |
| Ventera effective maturity if calculated | 1.125 years | 5.5 |
| Internal RWA with calculated maturity; capital | EUR 178.69m; EUR 23.23m | 5.5 |
| Borealis RW at M 5; Ventera RW at M 1.125 | 59.56%; 27.72% | 5.5 |
| PD equivalent to a 0.75 GSF (BBB) | 0.1198%, a 40.1% cut | 5.6 |
| Meridian CET1 with GSF; with BPF; both | 14.30%; 13.67%; 13.96% | 5.6 |
| 25 bp P2R add-on: CET1 needed; MDA effect | EUR 165.9m; 14.06 bp | 5.7 |
| Add-on allocated to the five deals | EUR 0.2248m | 5.7 |
| Meridian floor binding year; add-on in 2030 | 2030; EUR 1.625bn | 2.5 |
The ranking that the ALCO note should end with. On these five deals, the regulatory choices that move capital are, in decreasing order: the output floor (+EUR 3.93m of capital when it binds), the maturity convention (+EUR 2.45m if calculated), the infrastructure supporting factor on Cap Ferrat (−EUR 3.41m, already in the base case; see B06), a hypothetical 0.75 green factor (−EUR 5.20m on internal RWA, which no rule grants), and a 25 bp climate P2R add-on (+EUR 0.22m). The green labels themselves move nothing.
Section 5 computed RWA deal by deal. Inside the bank, three further conventions decide what a business line actually pays for those RWA.
Which RWA. Meridian’s pricing tool shows both the internal and the standardised RWA of each deal. Until the floor binds, capital is allocated on internal RWA. Capital planning has decided (case assumption) to price new deals with a final maturity after 2029 on the higher of the internal RWA and 72.5% of the standardised RWA, because those deals will live in a floor-bound bank for most of their life. Some banks instead allocate the bank-level floor add-on to business lines pro rata to their standardised RWA, which spreads the cost across all deals; others allocate it to the portfolios that “cause” it (the low-PD corporates and mortgages). Each method sends a different signal to the desks, and capital management has to explain to the Board which behaviour it wants.
Which capital. Meridian allocates 13% of RWA, a CET1-equivalent figure. The Tier 1 and Tier 2 instruments that also count towards requirements are cheaper than equity; their cost is charged through funds transfer pricing, not through the RAROC denominator.
Which other constraints. Capital is not the only regulatory resource. The leverage ratio (Tier 1 over a non-risk-weighted exposure measure, minimum 3% plus any add-on) binds for banks with low-risk balance sheets; off-balance items enter it with the standardised CCFs subject to a 10% floor. For Meridian, with an RWA density of around 40% of its leverage exposure, the risk-weighted ratio binds first, so the leverage ratio is monitored rather than allocated. MREL, the resolution requirement, adds a need for bail-inable debt above capital; Treasury charges its cost through FTP (see C15, E04).
The RAROC lines for Ventera’s guarantee line come from B11 (section 5.7) and are not re-estimated: revenues EUR 1.435m (commissions EUR 1.405m plus issuance fees EUR 0.03m), operating costs 40% of revenues, a contingent liquidity charge of EUR 48,000 (the 40 bp HQLA carry on a 5% outflow assumption applied to the EUR 240m outstanding), expected loss EUR 38,400 (0.20% × 40% × EUR 48m). Tax 25%. What changes here is the RWA line.
FORMULA
is revenues; the operating costs; the contingent liquidity charge from Treasury; the expected loss; the 25% tax rate; the risk-weighted assets under the treatment considered.
INTERACTIVE LAB — Ventera’s green guarantee line: RAROC under four capital treatments
Recompute the RAROC of the B11 guarantee line when only the capital treatment changes, and compare the effect of the regulatory treatment with the effect of the green pricing discount and of a hypothetical green supporting factor.
| Parameter you can move | Range | Default |
|---|---|---|
| Capital treatment | F-IRB, M 2.5 years / F-IRB, calculated M / Standardised / Floor-bound (72.5% of SA) | F-IRB, M 2.5 years |
| Green discount on the eligible part | 0 to 15 bp | 0 bp |
| Green-eligible outstanding (non-gas) | 0 to 240 EUR m | 200 EUR m |
| Hypothetical green factor on RWA | 0.5 to 1 x | 1 x |
| Result at default settings | Value |
|---|---|
| RWA | 18.73 EUR m |
| Allocated capital (13%) | 2.435 EUR m |
| Revenues after discount | 1.435 EUR m |
| Profit after tax | 0.5809 EUR m |
| RAROC | 23.86 % |
| Revenues needed for a 12% RAROC | 0.793 EUR m |
Try it. Apply a 5 bp green discount and then switch the treatment to ‘Floor-bound’: the discount costs under 2 points of RAROC, the floor about 7. Then set a 0.75 green factor with the standardised treatment and see how a policy factor would rescue a deal that capital rules make marginal.
Line by line, in EUR m, at default values (F-IRB, = 2.5):
| Line | Amount | Note |
|---|---|---|
| Revenues | 1.435 | Commissions and fees (B11) |
| Operating costs (40%) | −0.574 | Lending cost-income ratio |
| Contingent liquidity charge | −0.048 | Treasury charge (B11, section 6.3) |
| Expected loss | −0.0384 | 0.20% × 40% × 48 |
| Profit before tax | 0.7746 | |
| Profit after tax (25%) | 0.581 | |
| RWA; capital at 13% | 18.73; 2.435 | 39.02% × EUR 48m |
| RAROC | 23.86% | Hurdle 12% |
Result. The same guarantee line earns 23.86% under Meridian’s F-IRB convention, 33.59% if Meridian calculated effective maturity, 17.12% if Meridian were floor-bound, and 12.41% under the standardised approach, barely above the hurdle (revenues needed for 12% would be EUR 1.392m instead of EUR 0.793m under F-IRB). A 5 bp green discount on the EUR 200m eligible part reduces revenues to EUR 1.335m and the RAROC to 22.01%. A hypothetical 0.75 green factor would raise it to 31.81%.
How the sustainable feature changes it. The green label changes one line, revenues, by the discount Meridian chooses to grant (none in the base case), and adds work that the RAROC does not show (eligibility checks on each guarantee, reporting). It does not change RWA, expected loss, liquidity or tax. The regulatory treatment changes the RAROC by between 7 and 11 points in either direction; the green discount by less than 2. A standardised bank competing with Meridian for Ventera needs revenues 75% higher than Meridian’s to earn the same hurdle; this is why the trade guarantee business concentrates in IRB banks and why the output floor, by narrowing that gap, matters for the competitive map of green trade finance.
Capital is also a source of funding: each euro of CET1 allocated to a deal funds a euro of assets without interest cost. Meridian’s FTP framework, like most, credits business lines with a “capital benefit” equal to allocated capital times a risk-free rate; this course ignores it by convention (it would add about 2.50% × EUR 2.435m ≈ EUR 0.06m a year to Ventera’s revenues before tax). The point matters for capital planning: when rates are high, the capital benefit partly offsets the cost of capital, and capital-heavy businesses look less expensive than in a zero-rate world.
Liquidity regulation runs in parallel with capital and uses similar conversion ideas. Committed facilities generate LCR outflows (10% for undrawn credit lines to non-financial corporates, 30% for liquidity facilities), and trade contingent items a low single-digit percentage set nationally; the NSFR requires stable funding for a share of off-balance-sheet commitments. For the five deals, Treasury’s charges were computed in their modules: the RCF’s undrawn EUR 60m generates an HQLA carry (B03), the guarantee line EUR 48,000 a year (B11), the long-dated term loans a liquidity premium of 6 bp per year of tenor on their funding (B06, B08, B09). Capital management and ALM must look at both together: a green loan that is cheap in capital because it is collateralised may be expensive in liquidity because it is long (see E04).
IRB capital covers unexpected loss; expected loss is supposed to be covered by provisions. The CRR enforces the link: for IRB exposures, if accounting provisions (IFRS 9 expected credit loss allowances) are lower than the regulatory expected loss, the shortfall is deducted from CET1; if they are higher, the excess can be added to Tier 2 capital up to 0.6% of IRB credit RWA. For the five deals, regulatory expected loss is EUR 0.5996m. IFRS 9 stage 1 provisions use a 12-month expected loss with point-in-time PDs and economic LGDs, often lower than the regulatory floor values (the supervisory LGD of 40% is conservative for senior secured corporates). A shortfall of a few hundred thousand euros on these deals would reduce CET1 euro for euro. Climate risk enters IFRS 9 through forward-looking scenarios and management overlays; the ECB has repeatedly criticised banks for not reflecting climate risk in ECL (see E05). An overlay that raises ECL on carbon-intensive sectors reduces CET1 through the profit and loss account but can also reduce an IRB shortfall: the net capital effect depends on which side of the regulatory expected loss the bank sits.
Pillar 3 ESG. All five exposures appear in the Pillar 3 ESG templates of Article 449a: by sector (NACE), by maturity bucket and, for Lutetia, by the energy performance of the collateral. Under the EBA’s June 2026 final draft standards, large institutions stop publishing the taxonomy-related templates, so the banking book taxonomy alignment ratio (BTAR) template disappears from Pillar 3, subject to adoption by the Commission.
Green Asset Ratio. The GAR, published under the Taxonomy Regulation’s Article 8 delegated act, counts taxonomy-aligned exposures in the numerator. Of the five deals, Cap Ferrat (renewable electricity generation, if the do-no-significant-harm and minimum safeguards are documented) and possibly Lutetia’s retrofit (if the renovation criteria are met) can enter it; Borealis’s sustainability-linked RCF is general-purpose and does not, nor does the guarantee line by default; Nordhavn’s dual-fuel vessels depend on the criteria for maritime transport and on the fuel actually used. The capital calculation and the GAR therefore look at entirely different attributes of the same exposure.
Financed emissions (PCAF). Each deal carries attributed emissions: Meridian’s share of the client’s or project’s emissions in proportion to the loan over enterprise value or project cost. For Cap Ferrat, attributed emissions are close to zero in operation; for Nordhavn, they depend on the methanol’s origin; for Ventera, the guarantee is off-balance and the PCAF standard has no attribution method for guarantees comparable to the one for loans. No capital rule uses these numbers; the ICAAP’s transition scenarios and the CRD6 transition plan do.
NOTE — Capital, GAR and financed emissions are three views of the same book: risk, taxonomy alignment and climate impact. A deal can be good on one and poor on another. Capital management’s role is to make sure that the bank’s targets on the second and third do not quietly assume capital that the first does not provide.
Capital figures are produced by a long chain of data and judgement. The risks below are those that capital management and its auditors look for, with the control and the function that owns it.
| Risk | Control | Owner |
|---|---|---|
| Wrong CCF bucket or product code | Product-to-bucket mapping, pre-deal desk memo for non-standard items | Capital management |
| Collateral not linked or not eligible | Collateral eligibility checklist, monthly unlinked-collateral report | Credit risk, operations |
| Stale property or vessel values | Revaluation frequency rules, valuation independence | Valuation unit |
| PD model miscalibration | Annual back-testing, independent validation | Model validation |
| Inconsistent maturity convention | Policy on Article 162(1), system-wide parameter | Capital management |
| ISF claimed without all conditions | Eligibility file per exposure, second review | Credit risk, capital management |
| Floor effect ignored in pricing | Pricing tool shows internal and floor-bound RWA | Finance |
| Climate risk not in ICAAP | Materiality assessment, scenario analysis, Board sign-off | Risk, capital planning |
| ESG overrides used to cut capital | Override log, limits on upward overrides, audit sampling | Credit risk, internal audit |
| Pillar 3 ESG data errors | Data lineage, reconciliation with booking | Sustainability reporting |
| Green flag used as capital argument | Policy: no capital relief from labels; challenge in committee | Capital management |
The greenwashing risk in capital management is subtle. Nobody can legally cut a risk weight because a loan is called green, but there are softer routes, and a capital manager should recognise them.
Red-flag checklist.
PITFALL — The opposite error exists too: a credit officer who refuses any ESG input in the rating “because there is no green supporting factor” ignores that the EBA guidelines expect ESG risk drivers to be considered in credit risk assessment. Risk-based inputs are legitimate; purpose-based relief is not.
Regulation in force (November 2026).
Supervisory data.
Green supporting factor.
Disclosure.
Orders of magnitude. For a BBB corporate loan: F-IRB weight about 39% (M 2.5), 60% (M 5), standardised 75%, floor-bound 54.4%. For project finance in operation: slotting strong 70%, 52.5% with the supporting factor. For IPRE at 53% exposure-to-value: standardised 70%. For object finance: standardised 100%.
What changed recently. The removal of the 1.06 scaling factor and the cut in F-IRB LGD from 45% to 40% lowered IRB capital for corporates in 2025; the 40% CCF for committed lines in F-IRB (75% before) lowered it further for RCFs. The output floor works in the other direction and its effect grows every year until 2030. Taken together, CRR3 has made investment-grade corporate lending cheaper in the short run for IRB banks and more expensive in the long run for floor-bound ones.
Open debates.
INTERACTIVE LAB — Capital, RWA and the green debate: check your understanding
Test the mechanics of CRR3 capital and the arguments of the green supporting factor debate.
Q1. Under CRR3, what CCF applies to an undrawn committed RCF under F-IRB?
Answer: B. F-IRB now uses the standardised CCFs; committed lines are in the 40% bucket.
Q2. What does Article 162(1) as amended by Regulation (EU) 2024/1623 say about M for an F-IRB bank?
Answer: C. The amended text gives the two options, applied consistently. Own-LGD banks must calculate.
Q3. A BBB corporate with PD 0.20% and LGD 40% has an F-IRB weight of about 39% at M = 2.5. What happens at M = 5?
Answer: B. The maturity adjustment raises the weight to 59.56%.
Q4. Which output floor percentage applies in 2026?
Answer: C. Phase-in: 50% in 2025, 55% in 2026, 60% in 2027, up to 72.5% in 2030.
Q5. For which borrowers does the output floor bite hardest?
Answer: B. Models save most on low-PD exposures, so the gap with the standardised weight is largest there.
Q6. What did the EBA’s October 2023 report conclude on a green supporting factor?
Answer: C. The EBA cited design and calibration problems and insufficient evidence of risk differentials.
Q7. A 0.75 factor on a BBB loan (PD 0.20%, LGD 40%) is equivalent to a PD of about:
Answer: C. Solving the IRB formula gives 0.1198%, a 40% fall.
Q8. The infrastructure supporting factor of Article 501a is best described as:
Answer: B. It is risk-motivated, includes an environmental assessment condition, and is not limited to green assets.
Q9. Can the ISF and the 80% high-quality project finance weight be combined under the SA?
Answer: B. The two reliefs are not cumulative.
Q10. What share of P2R must at least be met with CET1?
Answer: C. 56.25% in CET1 and 75% in Tier 1.
Q11. For an IRB bank, if IFRS 9 provisions are below the regulatory expected loss:
Answer: B. Shortfalls are deducted from CET1; excesses may count in Tier 2 within a limit.
Q12. Which instrument has the ECB used to sanction late climate materiality assessments?
Answer: B. ABANCA (2025) and Credit Agricole (2026) received periodic penalty payments.
Q13. Under the CRR3 SA, a corporate in the third credit quality step (BBB) is weighted:
Answer: C. CRR3 introduced 75% for that step; 65% is the transitional floor weight for unrated investment-grade corporates.
Q14. Why does calculated maturity raise the capital of green lending more than of trade finance?
Answer: B. The maturity adjustment increases with M, and sustainable assets are typically long-dated.
Meridian commits EUR 120m in a four-year RCF to a BB borrower (PD 1.00%), unrated by agencies. Average utilisation 30%. F-IRB with LGD 40% and M 2.5 years. Compute EAD, F-IRB RWA and capital, standardised RWA, and the floor-bound RWA at 72.5%. Is the floor relevant for this deal?
Solution. Drawn EUR 36m; undrawn EUR 84m × 40% = EUR 33.6m; EAD EUR 69.6m. F-IRB weight at PD 1.00%: 82.06%; RWA EUR 69.6m × 82.06% = EUR 57.11m; capital at 13% = EUR 7.42m. Standardised: unrated corporate, 100%: EUR 69.6m. Floor-bound: 72.5% × EUR 69.6m = EUR 50.46m, below the internal RWA. The floor does not bite on this deal even in a floor-bound bank, because the internal weight (82.06%) is already above 72.5% of the standardised weight. The transitional 65% for unrated corporates does not apply either, since the PD exceeds 0.5%. Lesson: the floor is a problem for investment-grade lending, not for sub-investment grade.
Meridian lends EUR 50m to an A-rated corporate (PD 0.07%) in a seven-year bullet green loan. F-IRB, LGD 40%. Compute RWA under the fixed 2.5-year convention and with calculated maturity. Compare with the standardised weight (A rating: 50%) and the floor at 72.5%.
Solution. The maturity adjustment is . At = 2.5 the weight is 21.36%, RWA EUR 10.68m. With calculated maturity, a seven-year bullet has a cash-flow-weighted life above five years, so = 5: weight 35.61%, RWA EUR 17.80m, two thirds more. Standardised RWA: EUR 25m; floor at 72.5%: EUR 18.125m. With calculated maturity the internal RWA almost reach the floor-bound level, so for this long, high-quality loan the maturity convention matters more than the floor. With fixed maturity, a floor-bound bank would count EUR 18.125m instead of EUR 10.68m.
A EUR 100m green term loan to a BBB borrower (PD 0.20%, LGD 40%, M 2.5) earns EUR 0.9m a year of net margin over FTP. Costs 40% of revenues; expected loss PD × LGD × EAD; capital 13% of RWA; tax 25%. Compute the RAROC, the RAROC with a hypothetical 0.8 green supporting factor, and the revenue needed for a 12% RAROC with and without the floor at 72.5% of a 75% standardised weight.
Solution. RWA = 39.02% × EUR 100m = EUR 39.02m; capital EUR 5.07m. Expected loss 0.2% × 40% × 100 = EUR 0.08m. Profit after tax = (0.9 − 0.36 − 0.08) × 0.75 = EUR 0.345m. RAROC = 0.345 / 5.07 = 6.80%, below the hurdle. With a 0.8 factor, capital falls to EUR 4.06m and the RAROC rises to 8.50%: still below the hurdle. For 12%: (0.6 × Rev − 0.08) × 0.75 = 12% × 5.07, so Rev = EUR 1.486m, 149 bp. Floor-bound (RWA 72.5% × 75% × 100 = EUR 54.375m, capital EUR 7.07m): Rev = EUR 2.018m. A generous green factor does not rescue an underpriced deal; pricing does.
1. Why does an undrawn credit line consume capital? Because regulation assumes that borrowers draw their lines as they approach default. The credit conversion factor converts the undrawn amount into an exposure at default: 40% for committed facilities under CRR3, in both the standardised and the foundation IRB approaches. A EUR 100m undrawn line therefore counts as EUR 40m of exposure. That is why commitment fees exist and why banks dislike large undrawn back-up lines at low prices.
2. Explain the output floor in two minutes. The output floor limits how far internal models can reduce RWA: total RWA cannot fall below a percentage of what the standardised approaches would give, 72.5% when fully phased in by 2030 (55% in 2026). It applies at bank level. For a bank that is floor-bound, the marginal cost of a new deal is the floor percentage of its standardised weight, not its internal weight. The effect is strongest on low-PD exposures, where models save most, such as investment-grade corporates and prime mortgages.
3. Does CRR3 require F-IRB banks to calculate effective maturity? No. Article 162(1), as replaced by Regulation (EU) 2024/1623, says that for exposures without own LGD estimates the maturity is applied consistently and either set at 2.5 years (0.5 for securities financing transactions) or calculated under paragraph 2. Banks using own LGD estimates must calculate it. The choice matters: for a BBB loan, moving from 2.5 to 5 years raises the risk weight from about 39% to about 60%, and for long-dated green assets the calculated option is materially more expensive.
4. Should there be a green supporting factor? I would start with what it means: a 0.75 factor on a BBB loan is equivalent to assuming its PD falls by about 40%. The evidence for such a gap is weak and mixed, and the EBA declined to recommend a factor in 2023. A factor would also release capital on deals that already pass the hurdle and would be neutralised for floor-bound banks unless applied to the standardised approach. Risk-based channels, such as ESG drivers in PD models and energy performance in collateral values, are more defensible.
5. How does climate risk affect a bank’s capital today? Through three channels. In Pillar 1, indirectly: transition and physical risks can raise PDs and lower collateral values, which raises RWA. In Pillar 2, directly: the ICAAP must quantify climate risk, and SREP findings can lead to qualitative measures or a higher P2R. And through enforcement: the ECB has imposed periodic penalty payments on banks that missed deadlines for climate materiality assessments.
6. What is the infrastructure supporting factor and is it green? It is a 0.75 multiplier on the capital requirement for exposures to qualifying infrastructure, under Article 501a. It requires predictable revenues, senior ranking, tested technology and an environmental assessment by the obligor, among other conditions. It is not a green factor: a toll road can qualify and an innovative renewable project may not. It cannot be combined with the 80% high-quality project finance weight in the standardised approach.
7. A banker says the green loan should have a lower RAROC hurdle. What do you answer? That the hurdle is the cost of equity and does not depend on the label unless the bank decides, as a strategic choice, to subsidise green lending. If it does, it should do so transparently, for instance through a capped allowance or an FTP incentive, and measure who pays. The capital itself is driven by the risk weight, which the label does not change. The real levers are the PD, collateral, tenor and pricing.
8. How would you explain an RWA increase of EUR 600m in one month to the CFO? I would decompose it into volume (new and repaid exposures), quality (rating migrations and default flows), parameters and methodology (model changes, policy changes), foreign exchange, and data corrections. For each driver I would give the amount and the top exposures. Then I would distinguish recurring effects from one-offs and say whether the change affects the floor position or the capital plan.
Regulation (EU primary sources).
Supervisory and regulatory publications.
Academic and reference works.