Commodity finance is the part of corporate banking that lends against things that move: cargoes of crude oil at sea, copper cathodes in a bonded warehouse in Rotterdam, cocoa beans waiting at a port in Côte d’Ivoire, the receivable a trader holds on a refinery that bought a cargo thirty days ago. The bank does not lend on the strength of a long-term business plan, as in leveraged or project finance (see B05 and B06). It lends on the value of a commodity it can identify, control and, if needed, sell, and it is repaid from the sale of that commodity. This is why bankers call the business “self-liquidating”: a well-structured transaction repays itself when the goods are delivered and paid for.
Inside Meridian Bank, the desk sits in Corporate Banking next to trade finance (see B11) and supply chain finance (see B12). Its hubs are Geneva, London, Singapore, Dubai, Houston and Rotterdam, where the trading houses book their deals. A European bank’s commodity desk is typically split in three:
Clients. The large independent trading houses (energy, metals and agricultural traders headquartered in Switzerland, Singapore, the Netherlands and the United Kingdom), the trading arms of oil majors and mining groups, mid-sized specialised traders, producers in emerging markets (miners, national oil companies, agricultural exporters), and independent oil and gas companies for RBL.
How the desk earns money. Commodity finance is a margin-and-fee business with a high turnover of short assets. A transactional loan may last 30 to 90 days; a borrowing-base facility is typically committed for one year and renewed every year; a PXF runs three to five years. Revenues come from (i) the drawn margin over SOFR or EURIBOR, (ii) a commitment fee on the undrawn part, (iii) an upfront or arrangement fee paid at each annual renewal, (iv) agency and security agency fees for the bank that runs the facility, and (v) ancillary flows: letters of credit and guarantees, FX, commodity and interest-rate hedging, cash management. Because assets are short and well secured, losses have historically been low in normal years, and the business is capital-light per unit of revenue. The flip side is fraud and concentration: a single fraud on warehouse receipts or duplicated invoices can wipe out years of margin, as several banks learned in Singapore in 2020 (section 8).
The team and adjacent functions. A relationship manager owns the client and the annual renewal. A structurer designs the facility: eligible assets, advance rates, sublimits, hedging and insurance requirements. A collateral manager or borrowing-base analyst, often in a middle-office unit, checks the borrowing-base certificate the client sends every week or month, reconciles it with warehouse statements and inspection reports, and blocks drawings when the base is too small. Trade operations issue and check letters of credit and shipping documents. Credit risk approves the limit and the advance rates. Compliance (sanctions, anti-money laundering, know-your-customer) is unusually present: commodities touch high-risk jurisdictions, vessels and intermediaries. Market risk validates the price assumptions; the ESG or sustainability risk team applies the bank’s sector policies (coal, oil and gas, deforestation, mining) and checks responsible-sourcing evidence. Outside the bank: inspection companies and collateral managers who verify stocks, credit insurers who insure receivables, warehouse keepers, lawyers, and the other banks of the club.
A realistic week. Monday: borrowing-base certificates arrive from twelve trader clients; the middle office flags that one copper trader’s base fell after a 4% drop in LME prices over the weekend and its headroom is now under 5%. Tuesday: a credit committee reviews the annual renewal of a USD 1bn club facility for an agricultural trader, with a new condition that soy and palm oil from high-risk origins carry EUDR-ready traceability. Wednesday: the RBL team runs its autumn redetermination for a North Sea producer with the bank’s new price deck, which for the first time includes a carbon cost on scope 1 emissions. Thursday: the sustainability risk team rejects a request to finance thermal coal cargoes for a new client, under Meridian’s coal policy. Friday: the agent bank circulates the term sheet for the renewal of the Kestrel Metals facility, the case of this module.
WHY IT MATTERS ON THE DESK — On this desk, credit risk is less about the borrower’s balance sheet than about the collateral: what it is, where it is, who controls it, what it is worth tomorrow, and whether it exists at all. The transition enters through the collateral too: which commodities the bank accepts, at what advance rate, from which supply chains, and for how long.
Commodity finance is a family of products that share one idea, lending against an identifiable commodity flow, but differ in what the bank looks at and what repays it.
| Product | Borrower | Collateral and repayment | Typical tenor |
|---|---|---|---|
| Transactional loan | Trader | One cargo; repaid by the buyer’s payment | 30 to 120 days |
| Borrowing-base facility (BBF) | Trader | Pool of inventory and receivables; base recalculated weekly or monthly | 1 year, renewed |
| Pre-export finance (PXF) and prepayment | Producer | Assigned export contracts and offtake proceeds | 2 to 5 years |
| Reserve-based lending (RBL) | Oil and gas producer | Discounted value of proven reserves; redetermined twice a year | 5 to 7 years |
A transactional loan finances one identified shipment: the trader buys a cargo of copper concentrate, the bank pays the seller (often through a letter of credit, see B11), takes security over the cargo and the documents of title (bill of lading, warehouse receipt), and is repaid when the trader’s buyer pays. The bank sees each deal, which is safe but labour-intensive.
A borrowing-base facility industrialises this. Instead of approving each cargo, the banks commit a maximum amount (the commitment) and let the trader draw up to the lower of the commitment and the borrowing base, which is the sum of eligible assets multiplied by their advance rates. Every week or month the trader sends a borrowing-base certificate listing inventory by commodity and location, receivables by debtor, hedges and cash. The agent checks it; if the base falls below the amounts drawn, the trader must repay the difference or add collateral within a few days. Large traders also have unsecured syndicated revolving credit facilities (see B03), but mid-sized traders, whose equity is thin compared with the volumes they handle, live on borrowing-base facilities.
A pre-export finance lends to a producer, a copper miner in Zambia or a cocoa cooperative in Ghana, against future exports. The producer assigns its offtake contract with a trader or end-user to the banks, and the offtaker pays into a collection account controlled by the agent. The bank’s real risk is performance: will the mine produce and ship? A prepayment is the same structure where a trader advances cash to the producer and the banks lend to the trader, with recourse to the producer’s deliveries.
A reserve-based loan lends to an oil and gas producer up to a borrowing base equal to a share of the net present value of its proven reserves, computed with a price deck and a production profile certified by an independent reservoir engineer. The base is redetermined twice a year (spring and autumn in the US market, often semi-annually in the North Sea and Africa too); if prices or reserves fall, the base falls and the borrower must repay the excess.
IN ONE SENTENCE — A borrowing base is the amount of money the bank will lend against what the borrower owns today, after a safety cushion for each asset.
INTUITION — A pawnshop lends you 60% of a watch’s resale value and 90% of a gold bar’s, because gold is easier to sell at a known price. A borrowing-base facility is an industrial pawnshop that revalues the pledged goods every week.
FORMULA
is the borrowing base; the eligible value of asset class (inventory at market price, receivables at invoice value, after removing ineligible items); its advance rate (one minus the haircut); prior-ranking claims (for example unpaid warehouse fees or a liens reserve); the commitment; the amount drawn, including letters of credit issued under the facility.
Everything interesting sits in the definitions. Eligibility decides what counts at all: inventory must be in an approved warehouse or on an approved vessel, insured, free of other liens, covered by a pledge or title-transfer arrangement valid in the jurisdiction, and usually inspected by an independent collateral manager. Receivables must be on approved debtors, not overdue by more than a set number of days, not subject to set-off, and often credit-insured. Concentration limits cap the contribution of any one debtor, location or commodity. Valuation uses a daily public price (LME for copper, nickel, aluminium and zinc; API2 or API4 indices for coal; ICE or CME for crude, cocoa, coffee, soy) minus quality and location differentials. Commodities with no liquid public price, such as cobalt hydroxide or some lithium chemicals, are valued from price-reporting agencies (Fastmarkets, Argus, Benchmark Mineral Intelligence) and receive lower advance rates.
The advance rate protects the lenders against what can happen between the last valuation and the moment they can sell the collateral after a default: the price can fall, the goods may need to be moved, sold at a discount, or found to be of lower quality. A simple model, used in credit papers to justify the number rather than to set it mechanically, is a value-at-risk over the liquidation period plus a cost of liquidation.
FORMULA
is the haircut; the standard normal quantile at confidence (2.33 at 99%); the annual volatility of the commodity price; the liquidation period in trading days; the cost of liquidation (transport, discount to a forced buyer, legal costs) as a share of value; the share of the inventory hedged by futures or fixed-price sales; the volatility of the basis that remains on hedged inventory (the difference between the physical price and the hedge). The linear mix of volatilities is a desk shortcut, not a portfolio formula.
INTERACTIVE LAB — Where an advance rate comes from
See how price volatility, liquidation period, hedging and liquidation cost set the haircut on a commodity.
| Parameter you can move | Range | Default |
|---|---|---|
| Annual price volatility | 10 to 80 % | 40 % |
| Liquidation period | 5 to 60 trading days | 25 trading days |
| Confidence level | 95 / 99 / 99.5 | 99 % |
| Liquidation cost | 0 to 15 % of value | 5 % of value |
| Share hedged | 0 to 100 % | 0 % |
| Basis volatility on hedged part | 2 to 20 % | 8 % |
| Result at default settings | Value |
|---|---|
| Haircut | 34.3 % |
| Advance rate | 65.7 % |
Try it. Set the share hedged to 100%: the advance rate jumps because only basis risk and liquidation cost remain. This is why lenders reward hedged inventory.
With an annual volatility of 40%, a liquidation period of 25 trading days, 99% confidence and a 5% liquidation cost, the model gives a haircut of 34.3% and an advance rate of 65.7%. This is the profile of an unhedged battery-metal intermediate such as cobalt hydroxide, and it is why the case uses 65% for cobalt. Fully hedged LME copper cathodes in an approved warehouse, with 6% basis volatility, a 20-day liquidation period and 3% cost, give 93.1%; credit committees then round down to 90% to cover operational risk. The model explains the ranking of advance rates (hedged exchange-traded metals highest, illiquid unhedged intermediates lowest), while the final numbers come from policy grids and negotiation.
PITFALL — Hedging removes price risk from the collateral but creates a liquidity need for the trader. If prices rise, the short futures position loses money and the trader must pay variation margin to its broker in cash, while the physical inventory gain is only realised when the goods are sold. Many borrowing-base facilities therefore include a margin-call sublimit or accept the positive mark-to-market of hedges as an eligible asset. A trader that is fully hedged can still run out of cash.
A borrowing-base facility is priced like a revolving credit facility (see B03): a drawn margin over SOFR (most commodities are traded in US dollars) or EURIBOR, a commitment fee on the undrawn part (typically 30 to 40% of the margin), an upfront fee paid at each annual renewal, and an agency fee for the bank that runs the base. Letters of credit issued under the facility pay an LC commission. The margin reflects the trader’s standalone credit, the strength of control over the collateral, the mix of commodities and the competition among the 10 to 20 banks that are active in a given hub.
For the trader the economic question is the cost of carry. Holding inventory costs storage, insurance and financing; the trader earns it back from the contango in the forward curve, from arbitrage between locations and qualities, and from blending. When the bank cuts an advance rate by ten points, the trader must fund that slice with equity, which is scarce and expensive. This is why advance rates, more than the margin, are the real battleground of a renewal.
The risks of commodity finance are specific and well catalogued.
A borrowing-base facility is documented under English or New York law, often starting from LMA or LSTA templates for the facility agreement, completed by a security package: pledges or assignments over inventory, assignments of receivables and insurance, account charges over collection accounts, and tripartite agreements with warehouse keepers and collateral managers. The borrowing-base schedule to the agreement defines eligible assets, advance rates, concentration limits and reporting frequency. A collateral management agreement binds the inspection company that controls the goods. RBLs add a reserves report from an independent engineer and the banks’ price deck. PXFs add the assigned offtake contract and a notice and acknowledgement from the offtaker.
Commodity finance sits in an awkward place in the transition. It finances the fossil fuels the economy is trying to leave (oil, gas, coal), the metals it needs to electrify (copper, nickel, cobalt, lithium, graphite) and the agricultural commodities that drive most tropical deforestation (soy, palm oil, cattle, cocoa, coffee, rubber, wood). The same desk may do all three in one week. The sustainable layer therefore does not appear mainly as a “green” product. It appears in four places:
What does not change: the bank still lends against a liquid, controlled, correctly valued commodity. A battery metal is not safer collateral because it serves the transition; often it is the opposite.
Most European banks restrict new upstream oil and gas financing (new fields, Arctic, oil sands) while still financing traders’ flows and, selectively, producers. Policies differ and change often: on the desk, the operative document is the bank’s own sector policy. The voluntary net-zero alliances that once framed these commitments have weakened: members of the Net-Zero Banking Alliance voted in October 2025 to cease its operations, after the departure of most large US, Canadian, Japanese and several UK banks. In the EU, supervisory expectations and CRD6 transition plans now carry the weight (see A02 and E12).
For RBL, the transition enters the price deck and the cost base. A bank’s price deck is the set of forward prices (usually below the forward curve, sometimes with a long-term flat price) that its engineering team approves twice a year. Two climate adjustments are now discussed in credit committees:
The EU Methane Regulation (Regulation (EU) 2024/1787) adds a third channel for gas and oil sold into the EU: importers will have to show monitoring, reporting and verification equivalent to EU rules, then methane-intensity values, then compliance with maximum intensity values. On 14 October 2026 the Commission proposed to postpone these importer obligations (MRV equivalence from 1 January 2027 to 31 May 2029; intensity reporting from August 2028 to August 2029; maximum intensity from August 2030 to August 2031); in late October and early November 2026 the Parliament and the Council adopted positions, and the trilogue was pending at the time of writing.
FORMULA
is production in year (million boe), declining at a constant rate; the deck price in USD per boe; the annual drift of the deck (zero for a flat deck, negative in a transition scenario); the operating cost per boe; the scope 1 and 2 intensity in kgCO2e per boe; the carbon price in USD per tonne; the discount rate (10% by RBL convention, “NPV10”); the advance ratio applied to the NPV. Taxes and capital expenditure are ignored to keep the mechanics visible; in practice they matter a lot.
INTERACTIVE LAB — RBL borrowing base under a carbon cost and a transition price deck
Compare the effect of a carbon price on scope 1 and 2 emissions with the effect of a falling price deck on an RBL borrowing base.
| Parameter you can move | Range | Default |
|---|---|---|
| Production in year 1 | 2 to 20 mmboe | 8 mmboe |
| Annual decline rate | 5 to 25 % | 12 % |
| Reserve life | 5 to 20 years | 12 years |
| Deck price, year 1 | 40 to 100 USD/boe | 70 USD/boe |
| Annual drift of the deck | -8 to 2 % | 0 % |
| Operating cost | 5 to 45 USD/boe | 25 USD/boe |
| Scope 1 and 2 intensity | 5 to 60 kgCO2e/boe | 25 kgCO2e/boe |
| Carbon price | 0 to 200 USD/t | 80 USD/t |
| Discount rate | 8 to 15 % | 10 % |
| Advance ratio on NPV | 40 to 70 % | 60 % |
| Result at default settings | Value |
|---|---|
| NPV of reserves | 1,456.2 USD m |
| Borrowing base | 873.7 USD m |
| Reduction of NPV due to carbon cost | 4.4 % |
Try it. Set the carbon price to 0 and the drift to -4%: the transition price deck cuts the base far more than a carbon price of 80 USD/t on operational emissions.
With the default field (8 mmboe in year 1, 12% decline, 12 years, USD 70 per boe, USD 25 operating cost), the borrowing base is USD 914.3m without carbon cost. A carbon price of USD 80 per tonne on 25 kgCO2e per boe adds USD 2 per boe of cost and reduces the base to USD 873.7m, a 4.4% cut. A transition deck in which prices fall 4% a year, without any carbon cost, reduces the base to USD 753.2m, a cut of about 18%. The lesson is one of proportion: for an RBL, the carbon cost of the producer’s own emissions matters less than the demand-side scenario embedded in the price deck, because scope 3 (the combustion of the oil sold) is where most of the emissions are, and it reaches the producer through prices, not through a carbon bill.
Thermal coal is the commodity on which European bank policies converge most. The usual architecture is: no new clients with a significant share of revenue from thermal coal; existing clients financed only if they have a phase-out plan; complete exit by around 2030 in the EU and OECD countries and by around 2040 elsewhere; dedicated rules for traders (a cap on the share of coal in their traded volumes or in the borrowing base, declining to zero). Metallurgical (coking) coal, used to make steel, is often treated separately because it has fewer substitutes today.
Two regulatory hooks make coal exposure visible even without a ban. First, the EU Paris-aligned Benchmark rules (Commission Delegated Regulation (EU) 2020/1818, Article 12) exclude companies that derive 1% or more of their revenues from the exploration, mining, extraction, distribution or refining of hard coal and lignite; a coal trader falls in this category. Second, the EBA Pillar 3 ESG template on banking-book transition risk (template 1, see E06) requires banks to show their exposures to companies excluded from Paris-aligned benchmarks. A trader with 20% of revenue from coal will therefore appear in that column of Meridian’s disclosure until it exits.
Copper, nickel, cobalt, lithium and graphite supply chains are concentrated in a few countries, with known human-rights and environmental risks: artisanal mining and child labour in the cobalt belt of the Democratic Republic of Congo, tailings and water use in copper and lithium, high-carbon nickel processing in Indonesia. They are also exposed to state intervention: the DRC suspended cobalt exports in February 2025, extended the suspension in June, and replaced it from 16 October 2025 with annual export quotas (96,600 tonnes for each of 2026 and 2027), which moved cobalt prices and the value of traders’ inventories.
The reference standard is the OECD Due Diligence Guidance for Responsible Supply Chains of Minerals from Conflict-Affected and High-Risk Areas (third edition, 2016). It sets a five-step framework (management systems; identify and assess risks; design and implement a response; independent third-party audit at identified points of the supply chain; annual public reporting) and, in its Annex II, a model supply-chain policy listing the risks to avoid or mitigate (worst forms of child labour, forced labour, support to armed groups, bribery, money laundering). Industry programmes put it into practice: the Responsible Minerals Initiative’s smelter assessments, the LME’s requirement that every listed brand implement the OECD guidance and report under an audit or red-flag track, and site certifications such as the Copper Mark or IRMA.
Regulation is catching up, more slowly than announced:
For the commodity banker, the practical translation is an eligibility test: inventory and receivables of high-risk minerals count in the borrowing base only if the trader shows an OECD-aligned due-diligence system and valid audit evidence for the relevant smelters or mine sites. When the evidence lapses, the asset drops out of the base. Section 5.3 puts a number on this.
The EU Deforestation Regulation (Regulation (EU) 2023/1115, EUDR) prohibits placing on the EU market, or exporting from it, cattle, cocoa, coffee, oil palm, rubber, soya and wood, and products derived from them, unless they are deforestation-free (produced on land not deforested after 31 December 2020), legal in the country of production, and covered by a due-diligence statement filed in the EU information system. The statement must include the geolocation of every plot of land where the commodity was produced. Countries are benchmarked as low, standard or high risk, which sets the intensity of checks.
Its application has been postponed twice. After Regulation (EU) 2024/3234 and then Regulation (EU) 2025/2650 (in force since 26 December 2025), it applies from 30 December 2026 for large and medium operators and traders and from 30 June 2027 for micro and small operators. The Commission’s simplification review of 4 May 2026 confirmed the December 2026 date, and a delegated regulation of September 2026 adjusted the product scope. Financial institutions are not obliged parties: a bank that finances cocoa or soy does not file due-diligence statements. But its collateral does fall under the regulation: a cargo of soy beans without geolocation data cannot be sold into the EU, so its realisable value, and therefore its eligibility in a borrowing base, depends on the trader’s traceability. Agricultural facilities are therefore adding EUDR compliance as an eligibility criterion for EU-bound goods (see A05 for the nature dimension).
Large traders have used sustainability-linked margins on their facilities for several years, under the Sustainability-Linked Loan Principles of the LMA, APLMA and LSTA (revised in March 2025, see B03 and A04). The SLLP require KPIs that are material to the borrower’s core business, targets that go beyond business as usual, annual verification and transparent reporting. For a trader, credible KPIs are emissions of its own assets (ships, terminals, refineries) or the share of volumes from certified sites. Weak KPIs are those already required by the bank’s own policy (coal exit) or by law (EUDR compliance): rewarding a borrower with a lower margin for not breaking the bank’s policy or the law fails the ambition test.
WHO DOES WHAT — At a renewal, the sustainable layer is negotiated in three documents: the borrowing-base schedule (eligibility and sublimits), the information undertakings (audit evidence, traceability data, coal volumes) and the margin grid (sustainability-linked adjustment). The first two move far more money than the third.
The lifecycle below follows the annual renewal of a club borrowing-base facility for a metals trader, with Meridian Bank as mandated lead arranger, facility agent and security agent. A new facility follows the same path with a longer due diligence; an RBL replaces steps 6 and 10 with the reserves report and the semi-annual redetermination.
INTERACTIVE LAB — Annual renewal of a borrowing-base facility
Follow a borrowing-base facility from the renewal request to daily monitoring, with the transition checks at each step.
| # | Who | Does what | Deliverable |
|---|---|---|---|
| 1 | Relationship manager | Receives the trader’s renewal request and new business plan (commodity mix, volumes, coal exit path) | Renewal request and call report |
| 2 | Structurer | Drafts the indicative term sheet: commitment, advance-rate grid, eligibility, sublimits, margin grid | Indicative term sheet |
| 3 | Sustainability risk team | Screens the client against coal, mining and deforestation policies; reviews responsible-sourcing system and audit evidence | ESG due-diligence note with conditions |
| 4 | Compliance (KYC and sanctions) | Refreshes KYC, checks origins, vessels, counterparties and high-risk jurisdictions | KYC refresh and sanctions clearance |
| 5 | Credit analyst | Analyses audited accounts, trading limits, hedging policy, stress of the borrowing base; proposes rating and LGD | Credit application |
| 6 | Collateral manager and inspection company | Field visit to warehouses, stock count, review of warehouse agreements and title documents | Collateral audit report |
| 7 | Credit committee | Approves limit, hold, advance rates and conditions, including coal sublimit and eligibility tests | Credit approval with conditions |
| 8 | Agent and club banks | Circulate term sheet, agree allocations, negotiate facility agreement and borrowing-base schedule with counsel | Signed amendment and restatement agreement |
| 9 | Legal counsel and security agent | Perfect pledges and assignments, tripartite warehouse agreements, insurance endorsements | Conditions precedent satisfied |
| 10 | Borrowing-base analyst (agent middle office) | Checks each weekly certificate against prices, inspection reports, aged receivables and audit evidence | Approved borrowing base and availability notice |
| 11 | Agent and trader treasury | Processes drawings and LC requests within availability; triggers mandatory prepayment on any deficiency | Drawdown notices, deficiency notices |
| 12 | Sustainability-linked verifier and agent | Annual KPI verification and margin adjustment; report on coal volumes and sourcing audits | Verification report and margin certificate |
| 13 | Finance and risk reporting | Book RWA, expected credit loss stage, financed emissions and Pillar 3 ESG templates | Regulatory and sustainability reporting entries |
Request and term sheet (steps 1 and 2). Three months before the facility’s anniversary, the trader’s treasurer sends the relationship manager the plan for the coming year: expected volumes by commodity, warehouse locations, debtor book, hedging policy. The structurer turns it into a term sheet. The borrowing-base schedule is negotiated line by line: each commodity gets an advance rate, a concentration limit and eligibility criteria. For a trader in transition, the term sheet now includes a coal sublimit schedule and responsible-sourcing conditions.
ESG and compliance screening (steps 3 and 4). The sustainability risk team applies the bank’s policies: is the client within the coal policy thresholds, and does its exit path meet the policy dates? Does it have an OECD-aligned due-diligence system for minerals, with audits of its main smelters and mines? For agricultural traders, is it ready for EUDR? The output is a short note with conditions, which goes to the credit committee with the credit application. Compliance refreshes KYC and screens origins, vessels and counterparties against sanctions lists; commodity desks have a high share of alerts because of trade routes through sanctioned or high-risk countries.
Credit and collateral (steps 5 to 7). The credit analyst works on two levels: the trader as a company (equity, profitability, risk management, trading limits, value-at-risk of its book) and the collateral (stress of the borrowing base to price shocks, dilution of receivables, fraud controls). The collateral audit by an independent inspection company checks that the goods exist, are where they should be, of the right quality and free of third-party claims. The credit committee approves the limit, Meridian’s final hold and the conditions.
Documentation and closing (steps 8 and 9). For a renewal, the parties sign an amendment and restatement of the existing facility agreement. The security package must be refreshed for new warehouses and new debtors. Conditions precedent include legal opinions, insurance certificates naming the security agent as loss payee, and warehouse attornment letters.
Monitoring (steps 10 and 11). The weekly or monthly certificate is the heartbeat of the facility. The analyst reprices inventory at the latest LME, index or price-reporting-agency prices, deducts ineligible items (aged receivables, goods without valid inspection, minerals without valid audit evidence) and publishes availability to the trader and the banks. If drawings exceed the base, the trader must prepay within the cure period (two to five business days is common). Deficiencies that are not cured are an event of default.
Annual verification and reporting (steps 12 and 13). If the facility has a sustainability-linked margin, an external verifier checks KPI performance once a year and the agent applies the margin adjustment. Finance and risk book the exposure, its RWA, its IFRS 9 stage and its contribution to financed emissions and Pillar 3 ESG templates.
PITFALL — The borrowing-base analyst is usually junior and checks dozens of certificates a week. Most commodity frauds were “visible” in the certificates (receivables on the same debtor in two banks, stock in a tank that another bank also financed) but nobody compared across banks. Electronic registries of warehouse receipts and cross-bank checks of invoices are the industry’s answer since 2020; they work only if the analyst has time and authority to block a drawing.
Question. Who is the borrower, what does it ask for, and why is this renewal different from the previous ones?
Data (case assumptions). Kestrel Metals Trading SA is a fictional Geneva-based trader founded in 2009. It started in thermal coal (South African and Colombian coal sold to European and Asian utilities) and moved into base metals ten years ago. In 2025 it had revenues of USD 6.0bn, a tangible net worth of USD 450m and financial debt of USD 750m, of which about USD 480m on average under its borrowing-base facility and the rest under bilateral transactional lines. It sells about 5 Mt of thermal coal a year. Meridian rates Kestrel BB (one-year PD 1.00%), a typical grade for a mid-sized trader with thin equity relative to its volumes.
The existing facility is a USD 600m committed borrowing-base facility, one year, renewed each December, provided by a club of four banks with USD 150m each. Meridian is mandated lead arranger, facility agent and security agent. Pricing: SOFR plus 175 bp drawn, a commitment fee of 35% of the margin, an upfront fee of 50 bp of the commitment at each renewal, an agency fee of USD 0.2m a year paid to Meridian. SOFR is assumed at 3.50% for the case.
In its renewal request of September 2026, Kestrel announces a strategic shift: exit thermal coal by the end of 2029 and grow in battery metals (cobalt hydroxide from the DRC, lithium carbonate from South America and China) alongside its copper and nickel business. It asks the club to keep the commitment at USD 600m, to raise advance rates on battery metals, and to add a sustainability-linked margin “to recognise the transition”.
The credit analyst’s first pass on Kestrel as a company (case assumptions) gives the following picture. Gross margin is thin, about 2.5% of revenue, as for most traders: Kestrel earns on volumes, logistics and arbitrage, not on price direction, and its board limits open price positions to a value-at-risk of USD 8m a day. Net profit was USD 55m in 2025. Leverage looks high on paper (USD 750m of debt against USD 450m of equity) but most of the debt is matched by inventory and receivables that turn over in 30 to 60 days; the analyst therefore reads leverage net of readily marketable inventories, a standard adjustment for traders. Ownership is private, by the founding family and management, so there is no listed equity to absorb a shock quickly. The main weaknesses are concentration (three debtors account for 40% of receivables) and the coal book, which earns a quarter of gross profit although it is less than a tenth of revenue.
Reading. Meridian’s coal policy forbids financing thermal coal for traders beyond 2029, so a coal exit plan is a condition for renewing at all, not a favour to reward. The real questions for the club are: how much can Kestrel borrow against its new inventory mix, how risky is that inventory, what responsible-sourcing conditions make the battery metals acceptable collateral, and what does the bank earn for the risk.
Question. What borrowing base does Kestrel’s balance sheet support at renewal, and how much headroom does it leave?
Data. Kestrel’s November 2026 certificate and the advance-rate grid agreed in the term sheet (case assumptions):
| Asset | Eligible value (USD m) | Hedged share | Advance rate |
|---|---|---|---|
| Copper cathodes, LME brands, approved warehouses | 220 | 100% | 90% |
| Nickel cathodes and briquettes, LME brands | 60 | 90% | 85% |
| Cobalt hydroxide, DRC origin, audited supply | 50 | 0% | 65% |
| Lithium carbonate, battery grade | 40 | 30% | 60% |
| Thermal coal, ARA and Richards Bay stocks | 120 | 80% | 75% |
| Credit-insured receivables | 150 | n.a. | 90% |
| Other receivables on approved debtors | 80 | n.a. | 80% |
The advance rates follow the logic of section 2.3: hedged LME metals highest, unhedged battery-metal intermediates lowest, receivables high when insured. Lithium gets the lowest rate despite partial hedging because its futures markets are thin and its price fell by more than 80% between late 2022 and 2024, a reminder of how fast these markets move.
Calculation. Each line contributes its value times its advance rate: copper 198.0, nickel 51.0, cobalt 32.5, lithium 24.0, coal 90.0, insured receivables 135.0, other receivables 64.0, a total of USD 594.5m. The commitment is 600, so availability is the borrowing base, 594.5. With average drawings of USD 480m, headroom is USD 114.5m. The average advance rate on inventory is 395.5 / 490 = 80.7%.
INTERACTIVE LAB — Kestrel’s borrowing base at renewal
Compute the borrowing base, availability and headroom from inventory, receivables and advance rates, and see the effect of a failed responsible-sourcing test.
| Parameter you can move | Range | Default |
|---|---|---|
| Copper cathodes | 0 to 400 USD m | 220 USD m |
| Nickel | 0 to 200 USD m | 60 USD m |
| Cobalt hydroxide | 0 to 200 USD m | 50 USD m |
| Lithium carbonate | 0 to 200 USD m | 40 USD m |
| Thermal coal | 0 to 200 USD m | 120 USD m |
| Insured receivables | 0 to 300 USD m | 150 USD m |
| Other receivables | 0 to 200 USD m | 80 USD m |
| Cobalt audit evidence valid (1 yes, 0 no) | 1 / 0 | 1 |
| Amount drawn | 0 to 600 USD m | 480 USD m |
| Result at default settings | Value |
|---|---|
| Borrowing base | 594.5 USD m |
| Availability (lower of commitment and base) | 594.5 USD m |
| Headroom | 114.5 USD m |
| Average advance rate on inventory | 80.7 % |
Try it. Switch the cobalt audit evidence to 0: USD 32.5m of availability disappears overnight, without any change in prices.
Reading. At first sight Kestrel is comfortable. But this headroom is an average. Traders draw most when they build stock and when prices are high; Kestrel’s peak drawings in 2026 reached USD 560m, leaving USD 34.5m of headroom at peak (simulator of section 5.4 with no price shock). Section 5.4 shows what a price shock does at that point.
PITFALL — A borrowing base is not a credit limit that the bank grants once a year; it is a number that moves every week with prices, stock levels, receivable ageing and eligibility. A trader can be in default without any loss in its business, simply because a certificate shows a base below drawings and it has no cash to cure the gap.
Question. How should the club make the battery metals acceptable, and what happens to the base if the sourcing evidence fails?
Data. Kestrel’s cobalt hydroxide comes from two industrial mines in the DRC copper-cobalt belt and one processing intermediary that blends material from several sites. Its lithium carbonate comes from a South American producer and a Chinese converter. The ESG due-diligence note (step 3 of section 4) concludes that Kestrel has an OECD-aligned policy (based on Annex II) and a supply-chain team, but that the intermediary has no current third-party audit and buys some material from sites near artisanal mining zones.
Design. The term sheet adds three conditions to the borrowing-base schedule:
Calculation. In the simulator of section 5.2, setting the cobalt evidence to invalid removes 50 × 65% = USD 32.5m from the base, which falls from 594.5 to USD 562.0m. Headroom at average drawings falls from 114.5 to USD 82.0m, and at peak drawings of 560 from 34.5 to just USD 2.0m.
Reading. The eligibility test is where responsible sourcing has teeth. It costs Kestrel nothing if its suppliers are audited, and USD 32.5m of liquidity if they are not. This incentive is far stronger than any margin step-down (section 5.6). It also protects the banks: cobalt that fails an audit becomes harder to sell to battery makers subject to the EU Batteries Regulation from August 2027 and to buyers that require LME-compliant material, so its liquidation value is genuinely lower. The test is not a moral add-on; it reflects a real collateral risk.
WHY IT MATTERS ON THE DESK — The same logic applies to agricultural traders with EUDR: soy or palm oil without geolocation cannot be sold into the EU from 30 December 2026, so it is worth less as collateral for EU-bound flows. Eligibility criteria convert sustainability regulation into a number in the borrowing base.
Question. At peak drawings, what happens to the base if metal and coal prices fall sharply, and does hedging protect the banks?
Data. Peak drawings USD 560m. A stress scenario of a uniform 25% fall in all commodity prices (case assumption, close to the moves seen in nickel, cobalt and lithium in single quarters over 2022 to 2025). Hedged shares from the table of section 5.2. Receivables are already invoiced and do not move with prices. Hedged inventory keeps its value because the gain on the hedge offsets the loss on the physical goods.
Calculation. The fall reduces the value of the unhedged part of each inventory line by 25%. The unhedged parts are: copper 0, nickel 6, cobalt 50, lithium 28, coal 24, a total of USD 108m. The loss of base is the sum of each loss times its advance rate: nickel 6 × 25% × 85% = 1.275, cobalt 50 × 25% × 65% = 8.125, lithium 28 × 25% × 60% = 4.2, coal 24 × 25% × 75% = 4.5, a total of USD 18.1m. The base falls from 594.5 to USD 576.4m, still above peak drawings of 560: no deficiency.
If the cobalt evidence has also lapsed (section 5.3), cobalt drops out entirely and the stressed base is USD 552.0m, below drawings: Kestrel must prepay a deficiency of USD 8.0m within the cure period. If Kestrel had not hedged at all, the same 25% shock would cut the base by USD 98.9m to USD 495.6m, a deficiency of USD 64.4m.
FORMULA
is the price shock (25% in the case); the hedged share of inventory line ; its value before the shock; receivables, which do not move with prices; the amount drawn.
INTERACTIVE LAB — Price shock, hedging and the borrowing-base deficiency
Stress Kestrel’s borrowing base for a fall in commodity prices and measure the mandatory prepayment at peak drawings.
| Parameter you can move | Range | Default |
|---|---|---|
| Fall in commodity prices | 0 to 50 % | 25 % |
| Amount drawn | 300 to 600 USD m | 560 USD m |
| Cobalt audit evidence valid (1 yes, 0 no) | 1 / 0 | 1 |
| Hedging as share of policy hedges | 0 to 100 % | 100 % |
| Result at default settings | Value |
|---|---|
| Stressed borrowing base | 576.4 USD m |
| Deficiency to prepay | 0.0 USD m |
| Fall in the base | 18.1 USD m |
| Headroom (negative means deficiency) | 16.4 USD m |
Try it. Set hedging to 0% and move the shock: the deficiency appears at a price fall of about 9%. Then restore hedging and set the cobalt evidence to 0.
What the agent does on the day. Suppose the stress happens in March 2027 with the cobalt evidence lapsed. On Monday morning the borrowing-base analyst reprices the certificate, finds a base of USD 552.0m against USD 560m drawn, and sends a deficiency notice to Kestrel and the club. Kestrel has three business days to cure. Its options, in order of preference: add USD 8.0m of new insured receivables from shipments invoiced over the weekend, repay from the cash it holds at another bank, or sell part of its unhedged lithium. The relationship manager calls the treasurer the same morning; the sustainability risk team checks whether the intermediary’s audit can be renewed within the 60-day remediation period, which would restore the cobalt line and lift the stressed base back to USD 576.4m. The credit analyst updates the watch-list assessment: a cured deficiency is a warning, not a default, but two deficiencies in a quarter would trigger a review of the advance-rate grid.
Reading. Three lessons. First, hedging is the banks’ best friend: Kestrel’s policy hedges cut the loss of base from USD 98.9m to USD 18.1m in a 25% shock. Second, the transition increases price risk in the base: the unhedged USD 78m of cobalt and lithium produces most of the stressed loss although these metals are a small part of inventory. Third, risks combine: a sourcing problem and a price fall together turn a comfortable facility into a deficiency. The credit paper should therefore stress eligibility and prices jointly, not one at a time.
WHO DOES WHAT — When a deficiency notice goes out, the trader’s treasurer calls the agent within the hour. The usual cures are to repay from cash, sell inventory, add new receivables to the certificate or ask for a temporary waiver. The agent cannot grant a waiver alone: it needs the majority lenders, typically two thirds of commitments, which in a club of four means at least three banks.
Question. How should the facility enforce Meridian’s coal policy, and what does Kestrel’s transition do to its borrowing capacity over four years?
Design. Rather than excluding coal immediately, which would cut USD 90m of base and force a fire sale of Kestrel’s coal business, the club agrees a coal sublimit: the contribution of thermal coal to the borrowing base may not exceed a share of the total base that declines each year, 20% in 2027, 12% in 2028, 6% in 2029 and 0% from 2030. A share of the base, rather than a fixed amount, is used so that the cap tightens automatically if the rest of the business shrinks.
FORMULA
is the contribution of coal to the base (value times advance rate); the sublimit share for the year; the contribution of all other assets. The cap is computed on the other assets because coal itself is part of the base.
Data. Kestrel’s plan (case assumptions): coal inventory falls by USD 40m a year, from 120 at renewal to zero at the 2030 renewal, and is replaced value for value by battery metals (half cobalt, half lithium), whose average advance rate is 62.5% instead of coal’s 75%. Copper, nickel, the existing battery metals and receivables stay as in section 5.2, so the other assets contribute USD 504.5m before any replacement.
Calculation. In 2027 coal contributes 90 and the cap is 20/80 × 504.5 = 126.1, so it does not bind: the base is 594.5. In 2028, coal is 80 (contribution 60), the replacement metals add 40 × 62.5% = 25, other assets contribute 529.5, the cap is 12/88 × 529.5 = 72.2: no binding, base USD 589.5m. In 2029 the base is USD 584.5m and in 2030, with no coal left, USD 579.5m. If instead Kestrel kept its coal at 120, the sublimit would bind from 2028 and the base would fall to 573.3, then 536.7, then USD 504.5m in 2030.
INTERACTIVE LAB — Coal sublimit and borrowing base, 2027 to 2030
Compare the borrowing base when the trader delivers its coal exit plan with the base when it keeps its coal and the sublimit binds.
| Parameter you can move | Range | Default |
|---|---|---|
| Coal inventory reduction per year | 0 to 60 USD m | 40 USD m |
| Advance rate on replacement battery metals | 50 to 90 % | 62.5 % |
| Contribution of other assets | 300 to 700 USD m | 504.5 USD m |
| Result at default settings | Value |
|---|---|
| Base 2028, plan delivered | 589.5 USD m |
| Base 2029, plan delivered | 584.5 USD m |
| Base 2030, plan delivered | 579.5 USD m |
| Base 2028, coal kept at 120 | 573.3 USD m |
| Base 2029, coal kept at 120 | 536.7 USD m |
| Base 2030, coal kept at 120 | 504.5 USD m |
| Coal share of the base in 2027 | 15.1 % |
| Coal cap 2028, plan delivered | 72.2 USD m |
| Base lost 2027 to 2030, plan delivered | 15.0 USD m |
| Base lost 2027 to 2030, coal kept | 90.0 USD m |
Try it. Raise the advance rate on battery metals to 75%: the transition becomes neutral for the base. Kestrel’s request for higher battery-metal advance rates is really a request to make its transition liquidity-neutral.
Reading. Two findings, one expected and one less so. The expected one: the sublimit is a backstop. If Kestrel delivers its plan the cap never binds; if it does not, the cap removes coal from the base mechanically and costs Kestrel USD 90.0m of capacity by 2030. The less expected one: delivering the plan also costs Kestrel USD 15.0m of base (594.5 to 579.5), because battery metals receive lower advance rates than hedged coal. The transition, at constant inventory, makes Kestrel’s balance sheet less financeable. This is the economic reason behind its request for higher battery-metal advance rates; the simulator shows that 75% would make the shift neutral. The club’s answer depends on whether hedging tools for cobalt and lithium deepen and on Kestrel’s hedging policy, not on the green credentials of the metals.
PITFALL — Granting higher advance rates to battery metals “because they are green” is a classic error. It mixes a sustainability preference with a collateral valuation. If the bank wants to support the transition, it can do so through the margin, through tenor or through a separate dedicated facility, but the haircut must reflect liquidation risk, otherwise the bank’s loss in a stress falls on the assets it most wanted to support.
Question. What does the facility cost Kestrel, and how much is the sustainability-linked margin worth compared with the eligibility and advance-rate decisions?
Data. Pricing from section 5.1: SOFR 3.50%, margin 175 bp, commitment fee 35% of the margin, upfront fee 50 bp, agency fee USD 0.2m. Average drawings on the whole facility USD 480m. The club accepts a two-way sustainability-linked adjustment of up to 5 bp, split equally between two KPIs that are not already required by the bank’s policy or the law:
| KPI | Baseline 2025 | 2027 | 2028 | 2029 |
|---|---|---|---|---|
| Share of copper, nickel and cobalt volumes from sites with Copper Mark, IRMA or equivalent site certification | 25% | 40% | 55% | 70% |
| Carbon intensity of chartered shipping, gCO2 per tonne-km, change from 2025 | 0% | -5% | -10% | -15% |
Each KPI met lowers the margin by 2.5 bp; each missed raises it by 2.5 bp. The coal exit is deliberately not a KPI: it is already a condition of the facility. The cobalt audit is not a KPI either: it is an eligibility criterion.
Calculation. Kestrel pays interest of 480 × (3.50% + 1.75%) = USD 25.2m, a commitment fee of 120 × 35% × 1.75% = USD 0.735m on the undrawn USD 120m, an upfront fee of 600 × 0.50% = USD 3.0m and the agency fee of USD 0.2m: a total of USD 29.135m, or 6.07% of average drawings. If both KPIs are met, the margin falls by 5 bp and Kestrel saves 480 × 0.05% + 120 × 35% × 0.05% = USD 0.261m a year. By contrast, losing the cobalt eligibility removes USD 32.5m of availability, which Kestrel must replace with equity; at a 15% cost of equity against a 5.25% cost of debt, this costs USD 3.17m a year, about 12.1 times the value of the ratchet.
INTERACTIVE LAB — What the facility costs Kestrel, and what the ratchet is worth
Compute the annual cost of the borrowing-base facility for the trader and compare the sustainability-linked margin with the cost of lost availability.
| Parameter you can move | Range | Default |
|---|---|---|
| SOFR | 1 to 6 % | 3.5 % |
| Drawn margin | 100 to 300 bp | 175 bp |
| Sustainability-linked adjustment | -5 / -2.5 / 0 / 2.5 / 5 | 0 bp |
| Average drawings | 200 to 600 USD m | 480 USD m |
| Upfront fee | 0 to 100 bp | 50 bp |
| Availability lost to ineligibility | 0 to 100 USD m | 32.5 USD m |
| Kestrel’s cost of equity | 8 to 25 % | 15 % |
| Result at default settings | Value |
|---|---|
| Annual cost of the facility | 29.135 USD m |
| All-in cost on average drawings | 6.07 % |
| Effect of the ratchet on the annual cost | 0.000 USD m |
| Annual cost of replacing lost availability with equity | 3.17 USD m |
| Lost-availability cost as a multiple of a full 5 bp ratchet | 12.1 x |
Try it. Set the adjustment to -5 bp, then set lost availability to zero: compare the two outputs and you see which lever matters to the trader’s treasurer.
Reading. The sustainability-linked margin is real but small; the eligibility criteria and the advance rates are what move the trader’s economics. This matches what the Helios case showed for the greenium in bonds (see C01): the price signal exists, it is not the main reason the client changes behaviour. The ratchet is still useful: it creates an annual, verified reporting line on KPIs that matter, which feeds the bank’s own transition monitoring.
Question. How many tonnes of CO2 does Meridian report for its exposure to Kestrel, and how does the transition plan change the figure?
Data. Under the PCAF standard for business loans to unlisted companies, the attribution factor is the outstanding amount divided by the borrower’s total equity plus debt. Meridian’s average outstanding is USD 120m (25% of USD 480m). Kestrel’s equity plus debt is 450 + 750 = USD 1,200m. Kestrel’s own scope 1 and 2 emissions (offices, warehouses it operates) are about 25 ktCO2e a year (case assumption). Its material scope 3 is category 11, use of sold products: the combustion of the coal it sells, 5 Mt a year at an emission factor of 2.4 tCO2 per tonne (case assumption close to IPCC default values for bituminous coal). Metals sold have no combustion emissions; their upstream emissions (category 1) are left out here to keep the example readable. Coal volumes follow the inventory path of section 5.5.
FORMULA
is financed emissions; the bank’s outstanding; the borrower’s total equity and debt; the borrower’s scope 1 and 2 emissions; its relevant scope 3; the coal volume sold (Mt); the emission factor (tCO2 per tonne).
Calculation. The attribution factor is 120 / 1,200 = 10%. Scope 3 category 11 is 5 × 2.4 = 12 MtCO2, so Kestrel’s total is 12,025 ktCO2e and Meridian’s share is 1,202.5 ktCO2e a year, or about 10.0 ktCO2e per USD m lent. As the coal volumes fall, the attributed figure falls to 802.5 in 2028, 402.5 in 2029 and 2.5 ktCO2e in 2030.
INTERACTIVE LAB — Financed emissions attributed to Meridian through Kestrel
Compute PCAF financed emissions of a loan to a coal and metals trader and follow them along the coal exit path.
| Parameter you can move | Range | Default |
|---|---|---|
| Coal sold in 2027 | 0 to 15 Mt | 5 Mt |
| Emission factor of coal | 1.8 to 2.8 tCO2/t | 2.4 tCO2/t |
| Meridian outstanding | 50 to 300 USD m | 120 USD m |
| Kestrel equity plus debt | 600 to 3000 USD m | 1200 USD m |
| Kestrel scope 1 and 2 | 0 to 200 ktCO2e | 25 ktCO2e |
| Coal inventory reduction per year | 0 to 60 USD m | 40 USD m |
| Result at default settings | Value |
|---|---|
| Attribution factor | 10.0 % |
| Financed emissions 2027 | 1,202.5 ktCO2e |
| Financed emissions per USD m lent, 2027 | 10.02 ktCO2e |
| Financed emissions 2028 | 802.5 ktCO2e |
| Financed emissions 2029 | 402.5 ktCO2e |
| Financed emissions 2030 | 2.5 ktCO2e |
Try it. Double Kestrel’s equity plus debt: the attributed emissions halve although nothing changed in the coal sold. Attribution factors move with the borrower’s balance sheet.
Data quality. Under the PCAF scoring, emissions reported by the borrower and verified by a third party score 1, unverified reported emissions score 2, and emissions computed from physical activity data (here, tonnes of coal sold times an emission factor) score 3. Kestrel publishes no emissions report, so Meridian’s sustainability reporting team builds the scope 3 figure from the coal volumes Kestrel reports under the facility’s information undertakings, a score of 3 for category 11. The coal volume reporting that the club required to police the sublimit (section 5.5) thus doubles as the data source for the bank’s financed emissions: one undertaking, two uses, which is how sustainability data should be designed into a facility.
Reading. The figure is enormous relative to the loan: about 10 ktCO2e per USD m, many times what a typical corporate loan carries, and it comes almost entirely from scope 3. Three cautions. First, the same tonnes are counted by the utility that burns the coal (its scope 1) and by its lenders, so financed emissions add up to more than the world’s emissions; they are a measure of exposure, not of responsibility. Second, a trader’s scope 3 is volatile: a few large cargoes more or less change the figure by hundreds of kilotonnes. Third, the exit plan reduces Meridian’s figure to almost nothing by 2030, but the coal may well be traded by someone else: the bank’s portfolio decarbonises faster than the real economy. This is the central debate on portfolio-level targets (see E07).
Question. What does the credit committee approve, and on what arguments?
The debate. The credit paper is presented in November 2026 by the relationship manager and the credit analyst, with the ESG due-diligence note attached. The committee’s chief risk officer raises three points. First, the trader’s request for higher battery-metal advance rates: the simulator of section 5.5 shows it would make the transition neutral for Kestrel’s liquidity, but section 5.4 shows that cobalt and lithium already produce most of the stressed loss. Second, the RAROC: the deal clears the hurdle today (section 6.2) but falls below it in 2030 once coal collateral has been replaced by assets Meridian does not recognise for capital. Third, the reputational risk of financing cobalt from the DRC copper belt, where a single press report on child labour at a supplier site could force the bank out of the relationship.
The decision. The committee approves the renewal at USD 600m with Meridian’s hold unchanged at USD 150m, subject to the following terms:
Reading. The final terms put each sustainability concern in the instrument that can carry it: policy in the sublimit, collateral risk in eligibility and advance rates, client ambition in the margin, the bank’s own profitability in a repricing clause. A committee that tried to make one instrument (usually the margin) do all four would get a weak result on each.
| Item | Value | Source |
|---|---|---|
| Borrowing base at renewal | USD 594.5m | 5.2 |
| Headroom at average drawings of USD 480m | USD 114.5m | 5.2 |
| Average advance rate on inventory | 80.7% | 5.2 |
| Base if cobalt evidence lapses | USD 562.0m | 5.3 |
| Stressed base, 25% price fall, hedged | USD 576.4m | 5.4 |
| Deficiency, stress plus cobalt ineligible, drawings 560 | USD 8.0m | 5.4 |
| Deficiency, stress, no hedging | USD 64.4m | 5.4 |
| Coal share of the base in 2027 | 15.1% | 5.5 |
| Base in 2030, plan delivered / coal kept | USD 579.5m / 504.5m | 5.5 |
| Annual cost of the facility for Kestrel | USD 29.135m (6.07%) | 5.6 |
| Value of a 5 bp ratchet per year | USD 0.261m | 5.6 |
| Cost of lost cobalt availability per year | USD 3.17m | 5.6 |
| Attribution factor | 10.0% | 5.7 |
| Financed emissions 2027 / 2030 | 1,202.5 / 2.5 ktCO2e | 5.7 |
| Meridian RAROC (section 6) | 12.05% | 6.2 |
Exposure class. The first question is whether Kestrel’s facility is a corporate exposure or a commodities finance specialised-lending exposure. The Basel definition (CRE30, transposed in Article 147(8) of the CRR) reserves the specialised-lending treatment for structured short-term lending against exchange-traded commodities where the borrower has no other activities and no independent capacity to repay, so that the structure, not the borrower, carries the credit. Kestrel is a diversified trader with its own equity, other lenders and many activities: the facility is a corporate exposure, rated on Kestrel’s own PD, with the commodities recognised as collateral. Slotting would apply to a single-purpose vehicle financing one stock of oil or metal.
Approach. Under CRR3, banks may no longer use the advanced IRB approach (own LGD and CCF estimates) for corporates with consolidated revenues above EUR 500m. Kestrel’s USD 6.0bn of revenue puts it in that group, so Meridian uses F-IRB: own PD, regulatory LGD, regulatory credit conversion factors.
Exposure at default. The facility is committed and not unconditionally cancellable, so the undrawn part receives a 40% conversion factor. Meridian’s share at average use: drawn 120, undrawn 30, EAD = 120 + 40% × 30 = USD 132m.
LGD with collateral. F-IRB recognises receivables and “other physical collateral” if legal and operational conditions are met (Article 199 CRR): for physical collateral, notably liquid markets for disposal and well-established, publicly available market prices, frequent revaluation and enforceable security. Meridian’s credit risk control recognises LME-traded copper and nickel and index-priced coal as eligible physical collateral, but not cobalt hydroxide or lithium carbonate, whose prices come from price-reporting agencies rather than liquid exchange markets (a conservative case assumption; practice differs between banks). The CRR3 F-IRB formula, taken from Basel CRE32, blends unsecured and secured LGDs:
FORMULA
is the exposure value; the unsecured LGD (40% for senior claims on non-financial corporates); the value of collateral of type allocated to the exposure; its haircut (40% for receivables and for other physical collateral); the secured LGD (20% for receivables, 25% for other physical collateral); the secured part.
Meridian’s 25% share of eligible collateral is 25% × (220 + 60 + 120) = USD 100m of physical collateral and 25% × 230 = USD 57.5m of receivables. After haircuts, the secured parts are 60.0 and 34.5; the unsecured part is 132 − 94.5 = 37.5. Hence = (40% × 37.5 + 25% × 60 + 20% × 34.5) / 132 = 27.95%. With PD 1.00% and M = 2.5 years, the risk weight is 57.35% and the RWA USD 75.7m, against 82.06% (RWA 108.3) without collateral recognition.
INTERACTIVE LAB — RWA of Meridian’s share under four treatments
Compute the F-IRB LGD with physical and receivable collateral and compare RWA under F-IRB, standardised, slotting and the output floor.
| Parameter you can move | Range | Default |
|---|---|---|
| PD (one year) | 0.1 to 5 % | 1.0 % |
| Eligible physical collateral, whole facility | 0 to 600 USD m | 400 USD m |
| Eligible receivables, whole facility | 0 to 400 USD m | 230 USD m |
| Maturity M | 2.5 / 1 | 2.5 years |
| Utilisation | 0 to 100 % | 80 % |
| Result at default settings | Value |
|---|---|
| EAD (Meridian share) | 132.0 USD m |
| F-IRB LGD with collateral | 27.95 % |
| F-IRB risk weight | 57.35 % |
| RWA, F-IRB with collateral | 75.7 USD m |
| F-IRB risk weight, unsecured | 82.06 % |
| RWA, F-IRB unsecured | 108.3 USD m |
| RWA floor, 72.5% of standardised | 95.7 USD m |
Try it. Set M to 1 year: if the bank’s F-IRB treatment uses the facility’s contractual maturity, the risk weight falls sharply. Then set physical collateral to 280 to see the 2030 mix without coal.
Comparison. Under the standardised approach, Kestrel is an unrated corporate at 100%, and physical collateral is not recognised: RWA USD 132m. The output floor, phased in from 50% in 2025 to 72.5% in 2030, compares the bank’s total IRB RWA with 72.5% of its standardised RWA, here USD 95.7m for this exposure. Because the floor applies at bank level, a single deal does not “pay” the floor, but commodity books with low IRB weights are among those that make it bind; some banks now allocate a share of the floor to such desks. The EU transitional relief that lets unrated corporates count at 65% in the floor calculation is reserved for investment-grade-equivalent borrowers, so it does not help a BB trader. Slotting weights (70% for “good” with residual maturity under 2.5 years, 115% for “satisfactory”) are shown for comparison only.
Maturity. Meridian’s F-IRB convention is M fixed at 2.5 years. Under Article 162(1) CRR as amended by Regulation (EU) 2024/1623, a bank without own-LGD permission applies M consistently, either fixed at 2.5 years (0.5 years for securities financing transactions) or calculated under Article 162(2) for its whole book, with no mixing. Where M is calculated, a revolving facility is measured on its maximum contractual termination date, which for a one-year renewable facility gives about one year, and the risk weight drops from 57.35% to about 45.5%; but the same choice would raise the weight of every exposure longer than 2.5 years. The point matters for commodity desks, whose short facilities are one of the few areas where the maturity input changes capital materially.
Assumptions (Meridian conventions plus case data): FTP on drawn USD amounts of 21 bp, made of the 6 bp liquidity premium for one-year funding and a 15 bp cost of raising USD through FX swaps (case assumption); HQLA carry of 40 bp on a liquidity buffer equal to 10% of the undrawn commitment (the LCR outflow rate on committed credit facilities to non-financial corporates); cost-income 40% on lending revenues (net margin, commitment fee, upfront fee) and 60% on agency and ancillary revenues; ancillary revenues (FX, commodity hedges, letters of credit) of USD 0.6m a year with USD 2.5 of RWA per euro of revenue (applied to dollars here); EL = PD × LGD* × EAD.
| Line | Calculation | USD m |
|---|---|---|
| Net interest margin | 120 × (175 − 21) bp | 1.848 |
| Commitment fee | 30 × 35% × 175 bp | 0.184 |
| Upfront fee | 150 × 50 bp | 0.750 |
| HQLA carry | − 30 × 10% × 40 bp | −0.012 |
| Agency and ancillary fees | 0.2 + 0.6 | 0.800 |
| Revenues | sum | 3.570 |
| Operating costs | 40% × 2.770 + 60% × 0.800 | 1.588 |
| Expected loss | 1.00% × 27.95% × 132 | 0.369 |
| RWA | 75.70 + 2.5 × 0.6 | 77.20 |
| Allocated capital | 13% × 77.20 | 10.04 |
| RAROC | (3.570 − 1.588 − 0.369) × 75% / 10.04 | 12.05% |
INTERACTIVE LAB — RAROC of Meridian’s USD 150m share
Work out Meridian’s RAROC line by line and see the effect of the sustainability-linked margin, utilisation and the collateral mix after the coal exit.
| Parameter you can move | Range | Default |
|---|---|---|
| Drawn margin | 100 to 300 bp | 175 bp |
| Sustainability-linked adjustment | -5 / -2.5 / 0 / 2.5 / 5 | 0 bp |
| Utilisation | 20 to 100 % | 80 % |
| Upfront fee | 0 to 100 bp | 50 bp |
| Ancillary revenue | 0 to 2 USD m | 0.6 USD m |
| Eligible physical collateral, whole facility | 0 to 600 USD m | 400 USD m |
| PD | 0.1 to 5 % | 1.0 % |
| Maturity M | 2.5 / 1 | 2.5 years |
| Result at default settings | Value |
|---|---|
| Net interest margin | 1.848 USD m |
| Commitment fee | 0.184 USD m |
| Revenues | 3.570 USD m |
| Operating costs | 1.588 USD m |
| Expected loss | 0.369 USD m |
| RWA | 77.20 USD m |
| Allocated capital | 10.04 USD m |
| RAROC | 12.05 % |
Try it. Set the adjustment to -5 bp and physical collateral to 280: both transition effects together push the deal well below the hurdle.
Reading. The deal clears the 12% hurdle by a hair (12.05%), and only thanks to the agency and ancillary revenues that Meridian earns as agent and relationship bank; a participant bank without them would be below. The sustainable features move the RAROC in three directions:
A desk facing this outcome has honest options: reprice battery-metal tranches, require more hedging so that the collateral is more liquid, work with credit risk control to recognise exchange-traded lithium or cobalt contracts once they are liquid enough, or accept a below-hurdle hold for strategic reasons with an explicit decision of the credit committee.
A one-year revolving facility is cheap to fund in term premium but uses the bank’s dollar liquidity. Meridian, a euro bank, raises dollars partly through FX swaps, and the cost of the cross-currency basis is part of the FTP charge assumed above. Under the LCR, the undrawn commitment to a non-financial corporate carries a 10% outflow; under the NSFR, drawn loans to non-financial corporates with residual maturity under one year require 50% stable funding and undrawn committed facilities 5%. Self-liquidating trade transactions such as letters of credit receive lower factors, one reason banks like to push traders’ usage into LCs rather than cash drawings.
For Meridian, the drawn amounts are financial assets at amortised cost: the facility is held to collect, and its cash flows (SOFR plus a margin) are solely payments of principal and interest. The sustainability-linked adjustment is a contingent feature; the IFRS 9 amendments published in May 2024 and applicable from 1 January 2026 clarify that such features can pass the SPPI test when the resulting cash flows are not significantly different from those of an otherwise identical loan without the feature, which is clearly the case for a 5 bp two-way ratchet (see E05). The upfront fee is part of the effective interest rate of the expected drawings and is spread over the commitment period; the agency fee is service revenue recognised over time under IFRS 15. Expected credit losses are 12-month in stage 1; a borrowing-base deficiency that is not cured, a waiver request or a rating downgrade are typical triggers of a significant increase in credit risk (stage 2). The LGD used for ECL should reflect the collateral and its price sensitivity: in a downturn scenario, commodity prices and default probabilities move together, a wrong-way risk to build into the forward-looking scenarios.
| Risk | What can happen | Control | Owner |
|---|---|---|---|
| Collateral fraud | Duplicate pledges, fake receipts or invoices | Independent inspection, title registries, debtor verification | Agent middle office, collateral manager |
| Price risk on unhedged stock | Base falls below drawings | Advance-rate grid, hedging covenant, weekly revaluation | Structurer, market risk |
| Thin markets for battery metals | Forced sale far below index | Lower advance rates, concentration limits, PRA price checks | Credit risk |
| Sourcing failure | Collateral unsellable, reputational damage | OECD-aligned eligibility test, audit evidence, remediation period | Sustainability risk, agent |
| EUDR non-compliance | EU-bound goods blocked | Traceability as eligibility criterion for EU flows | Sustainability risk, credit |
| Coal policy breach | Facility finances coal beyond policy dates | Declining sublimit, coal volume reporting, annual policy review | Relationship manager, sustainability risk |
| Sanctions and money laundering | Prohibited origins, vessels, parties | KYC refresh, vessel and origin screening | Compliance |
| Weak sustainability-linked KPIs | Margin reward for business as usual | SLLP ambition test, external verification, no KPI that duplicates policy or law | Sustainability coordinator, credit |
| Misleading reporting | Battery-metal financing presented as green | Labelling policy; GAR and PCAF rules applied as written | Finance, sustainability reporting |
| Wrong-way risk | Prices fall when default probability rises | Joint stress of prices and PD, downturn LGD | Credit risk, risk methodology |
Red-flag checklist for a commodity facility with transition features
Fraud and the 2020 reset. In April 2020 Hin Leong Trading, a Singapore oil trader, collapsed owing about USD 3.5bn to 23 banks; court-appointed managers reported fabricated documents on a large scale. A Bloomberg tally published in July 2020 put potential losses for banks from commodity-trading frauds and failures over the previous six years at about USD 9.3bn (Hin Leong, Agritrade, Zenrock, Hontop and others). Some banks reduced or closed commodity finance activities in Asia, while most tightened controls (Daily Maverick/Bloomberg, 9 July 2020). In February 2023 Trafigura announced a USD 577m charge after discovering that cargoes of nickel it had bought did not contain the expected metal, a reminder that even large traders can be victims of collateral fraud.
Price shocks. In March 2022 the LME suspended nickel trading and cancelled trades after an unprecedented price spike, and battery-grade lithium carbonate prices fell by more than 80% between late 2022 and 2024. Cobalt shows how policy moves prices: the DRC export suspension from February 2025 and the quotas from 16 October 2025 (96,600 tonnes a year for 2026 and 2027) reversed a fall to multi-year lows.
Default experience. The ICC Trade Register 2025 (data to 2024, 26 trade and export finance banks) confirms low default rates for short-term trade products and warns that the CRR3 output floor raises capital charges mainly for balance-sheet products such as short-term import and export loans, while letters of credit are less affected. The Trade Register does not report commodity finance separately.
Regulatory calendar (as of November 2026).
| Rule | Status | Date |
|---|---|---|
| EUDR (Reg. 2023/1115, amended by 2025/2650) | Applies to large and medium operators and traders | 30 December 2026 |
| EUDR, micro and small operators | Applies | 30 June 2027 |
| Batteries Regulation due diligence | Postponed | 18 August 2027 |
| CSDDD after Omnibus I (Dir. 2026/470) | Applies to in-scope companies | 26 July 2029 |
| Methane Regulation importer MRV equivalence | Commission proposal of 14 October 2026 to move it to 31 May 2029; trilogue pending | 1 January 2027 in the current text |
| CRR3 output floor | Phase-in | 50% in 2025 to 72.5% in 2030 |
Climate alliances. The Net-Zero Banking Alliance ceased operations after a members’ vote in October 2025. Coal and oil and gas policies now rest on banks’ own commitments, supervisory expectations and, in the EU, CRD6 transition plans.
Research. An IMF working paper (Dong, Hengge, Valencia and Varghese, 2025) finds that lenders charge carbon-intensive US borrowers higher loan spreads, and that borrowers signalling environmental commitments get discounts that shrink as carbon intensity rises. The evidence fits the case: pricing signals exist, but they are modest relative to the effect of eligibility and access to credit.
Open debates.
INTERACTIVE LAB — Commodity finance in transition: check your understanding
Test the mechanics of borrowing bases, RBL, collateral eligibility, capital and financed emissions.
Q1. A trader has a commitment of USD 600m and a borrowing base of USD 560m. It has drawn USD 500m. What is its headroom?
Answer: B. Availability is the lower of the commitment and the base, 560; headroom is 560 minus 500.
Q2. Why do hedged LME copper cathodes get a higher advance rate than unhedged cobalt hydroxide?
Answer: B. The haircut covers price moves over the liquidation period; hedging and liquid markets shrink that risk.
Q3. In an RBL, what usually reduces the borrowing base most in a transition scenario?
Answer: C. Scope 3 effects reach the producer through prices; the case shows an 18% cut from the deck versus 4.4% from carbon cost.
Q4. A coal sublimit caps coal at 10% of the base and other assets contribute USD 450m. What is the maximum coal contribution?
Answer: B. The cap is s/(1-s) times the other assets: 0.10/0.90 x 450 = 50.
Q5. Which statement about the EUDR is correct as of November 2026?
Answer: B. After Regulation 2025/2650, application is 30 December 2026 for large and medium operators; banks are not obliged parties.
Q6. Which minerals does the EU Conflict Minerals Regulation 2017/821 cover?
Answer: B. It covers 3TG; cobalt, lithium, nickel and graphite are covered by the Batteries Regulation due-diligence rules.
Q7. Why is Kestrel’s facility treated as a corporate exposure rather than commodities-finance specialised lending?
Answer: B. Specialised lending applies where the borrower has no independent capacity and the structure carries the credit.
Q8. Under CRR3 F-IRB, what secured LGD applies to eligible other physical collateral?
Answer: C. LGDS is 25% for other physical collateral, 20% for receivables and real estate, after a 40% haircut.
Q9. Why can the coal exit reduce Meridian’s RAROC on Kestrel?
Answer: B. Less eligible collateral raises F-IRB LGD and RWA, as section 6.2 shows (12.05% to 11.06%).
Q10. Which sustainability-linked KPI would fail the ambition test for Kestrel?
Answer: C. Rewarding compliance with the bank’s own policy is business as usual, not ambition.
Q11. A bank lends USD 50m to a trader with equity plus debt of USD 500m whose scope 3 category 11 is 4 MtCO2. Financed emissions (ignoring scope 1 and 2) are:
Answer: B. Attribution factor 10% times 4,000 ktCO2 = 400 ktCO2.
Q12. What does the Taxonomy contribute to the GAR for a borrowing-base facility to a non-EU metals trader?
Answer: C. Commodity trading is not a Taxonomy activity and the trader publishes no Taxonomy KPIs.
Q13. What happens when a weekly certificate shows a base below drawings?
Answer: B. A borrowing-base deficiency triggers a mandatory prepayment; failure to cure is an event of default.
Q14. Which IFRS 9 change, applicable from 1 January 2026, matters for sustainability-linked margins?
Answer: B. The May 2024 amendments clarify how contingent ESG features are assessed under the SPPI test.
A trader pledges aluminium worth USD 100m (fully hedged, advance rate 85%), cocoa worth USD 80m (50% hedged, advance rate 70%) and insured receivables of USD 120m (advance rate 85%). The commitment is USD 250m and drawings are USD 220m. (a) Compute the base and headroom. (b) Compute the base after a 20% fall in cocoa prices. (c) USD 30m of the cocoa is EU-bound and lacks geolocation data after 30 December 2026, and the facility excludes it: compute the base before any price shock.
Solution. (a) Base = 100 × 85% + 80 × 70% + 120 × 85% = 85 + 56 + 102 = USD 243m; availability = min(250, 243) = 243; headroom = 243 − 220 = USD 23m. (b) The unhedged cocoa is 40; its value falls by 8; the base falls by 8 × 70% = 5.6 to USD 237.4m; headroom USD 17.4m. Aluminium does not move because it is fully hedged; receivables are invoiced. (c) Excluding USD 30m of cocoa removes 30 × 70% = 21; base = USD 222m and headroom only USD 2m. The EUDR exclusion is almost four times as damaging as the price shock.
Other assets contribute USD 400m to a trader’s base and coal contributes USD 70m. The sublimit for the year is 12% of the base. (a) Is the sublimit binding? (b) What is the base?
Solution. (a) The maximum coal contribution is 12/88 × 400 = USD 54.55m. Coal contributes 70, so the cap binds. (b) Base = 400 + 54.55 = USD 454.55m, instead of 470 without the sublimit: the sublimit costs USD 15.45m of capacity. Check: 54.55 / 454.55 = 12.0%.
A participant bank holds USD 100m of a borrowing-base facility, 70% drawn on average. Margin 200 bp, FTP 20 bp, commitment fee 35% of the margin, upfront fee 40 bp, HQLA carry 40 bp on 10% of the undrawn, no agency or ancillary revenue, cost-income 40%, PD 1.00%, LGD 30%, risk weight 60% (given), CCF 40%, capital 13% of RWA, tax 25%. (a) Compute RAROC. (b) What margin would reach 12%?
Solution. (a) EAD = 70 + 40% × 30 = 82; RWA = 60% × 82 = 49.2; capital = 6.396. Revenues: net margin 70 × 1.80% = 1.26; commitment fee 30 × 35% × 2.00% = 0.21; upfront fee 0.40; HQLA carry −30 × 10% × 0.40% = −0.012; total 1.858. Costs 40% = 0.743. EL = 1% × 30% × 82 = 0.246. RAROC = (1.858 − 0.743 − 0.246) × 75% / 6.396 = 10.19%. (b) For 12%, after-tax profit must be 0.12 × 6.396 = 0.7675, pre-tax 1.0234, so revenues must be (1.0234 + 0.246) / 0.6 = 2.1156. Revenues are 80.5 × m + 0.248 (70 drawn plus 35% of 30 undrawn, net of FTP 0.14 and plus fees 0.388), so m = 1.8676 / 80.5 = 2.32%, a margin of about 232 bp. The participant, without agency and ancillary revenue, needs 32 bp more than the margin it gets: the usual economics of club participations, which are held for relationship reasons.
1. Walk me through a borrowing-base facility. A borrowing-base facility is a committed revolving facility, usually one year, where the borrower can draw up to the lower of the commitment and a borrowing base. The base is the sum of eligible inventory and receivables multiplied by advance rates that reflect price volatility, liquidity and control. The borrower reports a certificate weekly or monthly and the agent recalculates availability. If the base falls below drawings, the borrower must prepay or add collateral within a short cure period. The bank’s risk is mainly the existence, value and control of the collateral, not the borrower’s long-term plan.
2. How do you set an advance rate? I start from the liquidation risk: how far can the price move between the last valuation and the sale after a default, and what does it cost to sell. A value-at-risk over a liquidation period plus a liquidation cost gives a first number; hedging, exchange listing and storage location adjust it. Then the policy grid and the peer banks’ rates set the final level. Hedged LME metals get around 90%, illiquid unhedged intermediates 60 to 65%. I would not raise a rate because a commodity is “green”.
3. How does climate change affect an RBL? Through three channels: a carbon cost on the producer’s own emissions, a transition price deck that lowers future oil and gas prices, and regulation such as the EU Methane Regulation for gas sold into Europe. In most fields the price deck matters far more than the carbon cost on scope 1 and 2, because combustion emissions reach the producer through demand and prices. Redeterminations twice a year transmit these changes quickly into the base. Banks also face policy limits on new upstream projects that shape which RBLs they renew.
4. How would you treat a coal-exposed trader that wants to stay a client? I would check the bank’s coal policy first: thresholds, phase-out dates and conditions for existing clients. If the client has a credible exit plan, I would put it into the documentation with a declining coal sublimit in the borrowing base, coal volume reporting and a clear end date. I would not reward the exit with a margin step-down, because it is a condition of the facility. I would also flag the effects on Pillar 3 disclosures and financed emissions, and track whether coal flows simply move to an affiliate.
5. What is the role of responsible-sourcing due diligence in a metals facility? For cobalt, lithium, nickel and other minerals from high-risk areas, the OECD guidance sets the reference: a five-step due-diligence framework and the Annex II risks. In a borrowing-base facility I would turn it into an eligibility criterion: minerals count in the base only with valid third-party audit evidence for their suppliers. This protects the bank because non-compliant material is harder to sell to buyers subject to the Batteries Regulation or LME rules. It also gives the trader a strong incentive, measured in lost liquidity rather than in basis points.
6. Is a loan to a battery-metals trader green? Not by default. Trading is not a Taxonomy activity, so it does not count in the Green Asset Ratio, and a general working-capital facility has no use-of-proceeds framework. It can be a sustainability-linked loan if it has ambitious, material KPIs verified annually. Calling it green because of the metals would be a greenwashing risk, especially if sourcing risks are not managed. The honest label is a transition-relevant facility with responsible-sourcing conditions.
7. How do you compute the capital on a secured commodity facility under CRR3? First I decide the exposure class: corporate if the trader has independent capacity, specialised lending if the structure alone carries the credit. For a large corporate, CRR3 requires F-IRB, so I take the own PD, compute EAD with a 40% conversion factor on the undrawn commitment, and blend LGDs: 40% unsecured, 25% for eligible physical collateral and 20% for receivables after 40% haircuts. Physical collateral must meet CRR conditions, notably liquid markets and public prices. Then the risk weight comes from the IRB formula, and the output floor at 72.5% of standardised RWA constrains the bank overall.
8. A client’s KPI-linked margin was met, but its financed emissions rose. How do you explain it to a manager? The KPIs measure specific things, for example certified sourcing or shipping intensity, while financed emissions depend on the borrower’s total scope 1, 2 and relevant scope 3 and on the attribution factor. If the client sold more coal, or if our outstanding rose relative to its equity plus debt, financed emissions rise regardless of the KPIs. It may also be a data effect: a new emission factor or better data quality. I would decompose the change into volume, attribution and data effects, and check whether the KPIs remain material; if coal volumes drive the figure, the facility’s coal sublimit, not the ratchet, is the lever.
Regulation (EUR-Lex)
Basel and supervisory texts
Standards and market bodies
Regulatory status updates used for dates
Market events
Academic and theory