The instrument
A letter of credit is a bank's promise to pay the seller on behalf of the buyer once the seller proves the goods shipped. It exists because cross-border trade has no enforceable trust. The buyer in one country doesn't trust the seller in another, and vice versa. So both sides pay banks to stand in the middle.
The buyer's bank issues the LC. The seller's bank receives it. When the seller ships, they send proof (bills of lading, certificates of origin, inspection reports, invoices) to the seller's bank, which forwards them to the buyer's bank, which checks them against the LC terms, which releases payment, which clears through correspondent banks, which finally credits the seller's account.
Every step is a paper handoff. Every step is a place documents can be lost, mismatched, rejected, or deferred for further review. Every step has fees. Every step takes days.
And critically, not every LC pays the same way. The instrument has four major variants. The variant determines whether the supplier gets cash in a week, six months, or never.
The four kinds of LC
Most people think of the letter of credit as a single instrument. It isn't. There are four major variants, and what the supplier actually receives, and when, depends entirely on which kind the buyer issues.
Sight LC. The bank pays the supplier as soon as compliant documents are accepted. Sight doesn't mean instant. Banks have five banking days under UCP 600 to examine documents at each handoff, and that examination happens at both the seller's bank and the issuing bank. Even a perfectly clean sight LC takes roughly ten calendar days from document presentation to cash. In practice, sight is the fastest variant. Suppliers prefer it. Buyers resist it because it gives them no credit period.
Usance LC (also called time LC or deferred payment LC). The bank pays on a future maturity date, typically 30, 60, 90, 120, or 180 days after sight or after the bill of lading date. The supplier ships, the documents get accepted, and then the clock starts on the usance period. A 90-day Usance LC means the supplier waits roughly three months past document acceptance to receive cash. Buyers prefer Usance because it gives them a free credit period. Suppliers tolerate it because the alternative is losing the deal.
UPAS (Usance Payable At Sight). A hybrid: the supplier is paid at sight, but the buyer doesn't repay the bank until the usance maturity. The bank funds the gap. Cleanest structure for the supplier, but expensive for the buyer because the bank charges acceptance commission plus interest. Rarely offered in practice. Buyers don't want to pay for what amounts to short-term bank financing.
Acceptance / Negotiation LC. The bank accepts a bill of exchange drawn against the LC. The accepted draft becomes a discountable instrument. The supplier can sell it to a third-party financier for immediate cash. This adds optionality but introduces another counterparty and another fee.
The variant determines everything downstream. A sight LC means cash on day 105. A 180-day Usance LC means cash on day 285, and that's only if the supplier can wait or finds someone willing to discount the receivable. Often, they can't.
The cashing trap
For Usance LCs, the supplier has three paths to cash. Each has a cost.
Path 1: Wait. Hold the LC to maturity. No discount fee. But working capital is locked for the entire usance period, which means the next order must be financed out of pocket or on a credit line. That credit line costs 6 to 12 percent APR. On a $400K order with a 90-day Usance LC, the wait itself is a $6,000 to $12,000 cost in interest the supplier carries.
Path 2: Discount the LC. Sell the future receivable to a bank for immediate cash at a discount. The rate depends on three things: the reference rate (SOFR is currently around 4.3 to 4.5 percent, EURIBOR similar), the issuing bank's credit spread, and the country corridor. Top-tier global bank in a transparent jurisdiction: SOFR plus 1 to 2 percent. Mid-tier regional bank: SOFR plus 3 to 5 percent. Weak or small bank in an opaque corridor: SOFR plus 5 to 10 percent.
Real numbers on a $400K trade. Confirmed LC from a top-tier issuer, 90-day Usance: about $7,000 in discount cost. Unconfirmed LC from a mid-tier bank, same trade: $10,000 to $15,000. A 180-day Usance from a weaker emerging-market issuer can cost $20,000 to $30,000 in discount fees alone, on top of the original LC issuance fees of 0.5 to 4 percent.
Path 3: Can't discount at all. Sometimes the LC has no buyer for the receivable. The issuing bank's credit rating is too low for any discounter's appetite. The country corridor is too risky. The LC text has soft clauses: subjective conditions like subject to buyer's inspection that make the receivable unbankable. The seller's bank has no correspondent relationship with the issuing bank.
In these cases, the supplier has $400K of paper they cannot turn into money. They've shipped goods, fronted production cost, paid factory deposits, and now hold an instrument that pays sometime in the next six months, if at all. They wait, or they walk.
The 1860s lock-in
Letters of credit were designed in the 1860s when transatlantic shipping took weeks and a single buyer–seller relationship might be the only trade contact a country had. The design worked: a neutral bank stood as the guarantor.
160 years later the design is unchanged. The world has email, the world has containers, the world has Stripe. International trade still runs on the same paper-based instrument. The reason it persists is that the incumbents profit from it. Banks like it because every step is a billable event. Regulators like it because there's an established audit trail. Insurers like it because they've underwritten the risk model for a century.
Buyers and sellers are the only parties who lose, and they're the only parties without a seat at the table when the rules are written.
The timeline
Pick a real order: $400,000 of nitrile gloves from a Malaysian factory to a US distributor. The standard LC timeline runs like this.
Days 0–21. LC negotiation, application, bank credit review, drafting, issuance via SWIFT MT700, advising bank verification, seller review, amendments. Most LCs need 2 to 3 amendment rounds: wrong port code, missing inspection clause, dated payment terms, undefined currency. Each amendment cycles back through both banks. In practice, a clean issued LC takes 14 to 21 calendar days; complex deals or first-time relationships can run 30 days or more before production can begin.
Days 21–65. Production and shipping. The factory takes 30 days for a typical production run and another 10 to 14 days for ocean transit on Asia-to-US routes. The supplier has now committed capital, materials, and labor against an LC they haven't yet been paid against.
Days 65–70. Goods arrive at the destination port. Seller compiles the LC presentation: bills of lading, certificates of origin, inspection reports, packing list, commercial invoice, insurance certificate. 17 documents typical, each requiring signatures, dates, and reference numbers that must match the LC text exactly.
Days 70–77. Seller's bank examines the documents under UCP 600 Article 14, 5 banking days maximum.
Days 77–91. If they passed, documents are forwarded to the issuing bank, which gets its own 5 banking days. Banks reject 60 to 70 percent of presentations on first pass for documentary discrepancies. A missing signature, a mismatched port, a date off by a day.
Days 91–105. Discrepancy waiver from the buyer's bank if needed. More fees. Documents re-presented. Payment finally released through correspondent banks. Seller's account credited. That's the sight LC.
For Usance LCs, add the usance period. 30-day Usance: cash by day 135. 90-day Usance: cash by day 195. 180-day Usance: cash by day 285. By the time a 180-day Usance LC settles, nine months have passed since the buyer first applied for the LC. The supplier financed that entire period.
The full bill
Direct LC fees run 0.5 to 4 percent of the trade value. On a $400K order at 2 percent, that's $8,000. Paid by the buyer formally, but priced into the cost of goods, so the supplier pays it too. Plus discrepancy fees ($200 to $500 each), amendment fees ($150 to $500 each, with most LCs needing 2 to 3), and SWIFT and correspondent bank charges.
If the LC is Usance and the supplier wants cash early, discounting adds another layer. A 90-day Usance from a top-tier issuer costs roughly 1.75 percent of face value to discount, $7,000 on $400K. From a mid-tier or weaker issuer, the same trade costs $10,000 to $15,000. A 180-day Usance from an emerging-market bank can cost $20,000 to $30,000 in discount fees alone.
Indirect costs are larger. 105 days of working capital tied up on a sight LC, or 195 to 285 days on a Usance LC, means the supplier needs a credit line to keep operating. That line costs 6 to 12 percent APR. On a $400K order with a 195-day Usance, that's $13,000 to $26,000 in interest the supplier carries while waiting for the LC to clear.
Add it up. On a single $400K trade with a Usance LC, the supplier can pay $30,000 to $50,000 in combined LC fees, discount costs, working capital interest, and discrepancy charges. 8 to 12 percent of the trade value.
Across global trade ($32 trillion in 2024), letters of credit consume roughly $2 trillion in cycle costs, fees, and tied-up working capital every year. Most of it is invisible because nobody itemises it. It's just the cost of doing business.
The $30 million lesson
EDMA Group has been a glove distributor for years. We sourced from manufacturers in Asia (primarily China) and sold to distributors and direct buyers in the US and Europe. The standard payment structure was an LC from the buyer to us, against which we'd pay the manufacturer separately or attempt a back-to-back instrument.
In 18 months we lost over $30 million in revenue to letters of credit. Not LCs that were fraudulent. Not LCs that were unfunded. The kind of loss had two shapes.
The Usance trap. A typical deal: a US client wanted to issue an LC against a shipment from a Chinese factory. The LC structure required negotiation between the US issuing bank and the supplier's Chinese bank. The supplier wanted payment certainty before producing. The two banks didn't have a working correspondent relationship. The procedure was negotiated between the banks for 30 days before the LC could even be issued. Production didn't start during that month.
Even after issuance, the supplier in China couldn't cash it at survivable terms. If the LC had been structured as transferable, the transfer process introduced its own complexity and cost: additional fees, additional verification, additional time. If it wasn't transferable, the supplier had no direct claim on the LC at all; we had to wait to receive payment from the US client first, then pay the supplier separately, which added another 60 to 90 days on top of everything else. And critically, the US and Chinese banks weren't friendly to each other, meaning no Chinese bank would advance funds against the LC at a reasonable rate. The supplier's options were to wait six months for actual cash, or to discount at 8 to 10 percent APR. Many manufacturers wouldn't accept either option, and the deal would die before it shipped.
The discrepancy trap. Sight LCs that broke because of documentary discrepancies. A port code on a B/L didn't match the LC, a certificate of origin was dated one day before the shipment instead of one day after, a packing list had a typo on the carton count. Each broken LC meant the seller didn't get paid on time. Each delay cascaded. Production lots couldn't be funded. Shipments couldn't be insured. The next order couldn't be placed.
We lost manufacturer relationships because of payment delays caused by LCs that broke after we'd already shipped. We lost client relationships because of orders that died in 30 days of bank-to-bank negotiation. The cumulative cost across roughly 18 months was over $30 million in revenue that should have been ours.
The protocol thesis didn't start in a whitepaper. It started in our own AR ledger. We needed an alternative because the existing system was actively destroying our business. Read the full EDMA case study →
Root cause: The US issuing bank and the Chinese supplier's bank had no working correspondent relationship. The supplier wanted payment certainty before producing; the buyer wanted Usance terms.
Compounding cost: Even after the LC was issued, no Chinese bank would advance funds against it at a survivable rate. Supplier's options narrowed to: wait six months for cash, or discount at 8–10% APR. Many manufacturers wouldn't accept either.
Outcome: Deals died before they shipped. Lost manufacturer relationships through repeated payment delays.
Cascade: Each broken LC meant the seller didn't get paid on time. Each delay pushed downstream. Production lots couldn't be funded. Shipments couldn't be insured. The next order couldn't be placed.
Compounding cost: Lost manufacturer relationships from payment delays. Lost client relationships from orders that died in 30 days of bank-to-bank negotiation. Revenue evaporated across the AR ledger.
Outcome: $30M+ in revenue that should have been ours, gone to documentary friction.
The opportunity
Three things have to be true at once for an LC alternative to work.
The buyer and seller need shared operational ground: one source of truth for what was ordered, manufactured, inspected, shipped, delivered. That's the operating-system layer. TradeOS.
Financiers need to see verified deals and price risk on real operational data, not on a bank's underwriting opinion from 1958. That's the matching layer. Global Trade Marketplace.
Payment needs to release on verified events, not on paper presentations. The bills of lading and inspection certs become cryptographic proofs anchored to the milestone they describe. That's the rail layer, Settlement with EDSD stablecoin, EMT milestone tokens, and PoV consensus on the EDMA L2.
We didn't design this to disrupt LCs. We designed it because LCs were costing us $30 million we couldn't afford to lose. The disruption is what happens when something better exists.
