L/C Letter of Credit: What It Is and Why It Matters When Importing from China

An L/C, or letter of credit, is a bank’s written promise to pay your supplier once the supplier hands over a specific stack of documents that matches the credit exactly. Your money does not move because the goods arrived, or because they were the right goods. It moves because the paperwork matched. That distinction sounds like a technicality until the first time a shipment sits in a warehouse while a bank refuses to release funds over a date that is three days out of line.

What an L/C Actually Is, and Who Is On the Hook

Strip away the jargon and a documentary credit swaps one promise for another. Instead of trusting a supplier you have never met to ship after you pay, you get your bank to promise the supplier’s bank that payment will happen. Under the ICC’s Uniform Customs and Practice for Documentary Credits — UCP 600, the rulebook nearly every credit is issued under — that promise is irrevocable and definite, and it is owed against a complying presentation rather than against the goods themselves.

Four parties usually appear. You are the applicant, the one who asks the bank to issue. Your bank is the issuing bank and carries the payment obligation. The supplier is the beneficiary, the party entitled to be paid. A bank in the supplier’s country advises the credit, and sometimes adds its own separate promise on top, at which point it becomes the confirming bank. That confirmation exists for one reason: the seller does not trust your bank or your country enough to rely on it alone.

The part importers consistently underestimate is how narrow the bank’s job is. The bank decides, on the basis of the documents alone, whether they appear on their face to be compliant. Nobody at the bank inspects your cartons, calls the factory, or forms a view on whether the goods are any good. A perfect presentation covering defective goods gets paid. A flawless shipment with a typo on the invoice can be refused. Once you internalise that split, most L/C behaviour that looks irrational starts making sense.

Not everyone is allowed to be on either end

Banks apply eligibility rules at both ends of the credit, and this is where China-origin orders start to diverge from the textbook. Bank of China’s published terms for issuing a credit require the applicant to hold a business licence proving its legal operation and scope of business, and to hold the qualification to engage in import and export trade.

The mirror image applies on the beneficiary side. A party that cannot legally export in its own name is a party that will struggle to sit on a credit as beneficiary — which is precisely the situation of most sellers inside a wholesale market.

The Rules That Decide Whether You Get Paid

Three deadlines and one standard govern almost every dispute. The standard is strict compliance: as the law firm HFW puts it in its client guide, UCP 600 imposes a doctrine of strict compliance and there is no de minimis rule. Trivial-looking errors are still errors. Banks are not being obstructive when they refuse over a misspelling; they are following the only rule they are permitted to apply.

On timing, a bank has a maximum of five banking days following the day of presentation to determine whether documents comply. That ceiling is itself a reform — UCP 600 cut the examination window down from the seven days allowed under the old UCP 500, which matters to a seller waiting on cash. Separately, documents must reach the bank within the time limits stated in the credit and within 21 calendar days after the shipment date. Miss that window and the presentation is stale, no matter how perfect the paperwork.

Consistency across the file is the trap that catches most first-timers. Data in one document need not be identical to data in another, but it must not conflict with it or with the credit. Meanwhile the goods description on the commercial invoice must correspond with the description in the credit, while other documents may describe the goods in general terms. So the invoice is held to a tighter standard than the packing list — a asymmetry that surprises people who assume all documents are checked the same way.

How often does this go wrong? The ICC Banking Commission estimates that between 65% and 80% of documents presented under documentary credits worldwide are refused on first presentation. The ICC also lists what is actually going wrong, and it is mundane: timing issues around expiry, shipment and presentation periods, plus conflicting data between documents, missing documents, and goods descriptions that do not match the credit.

There has been no meaningful improvement over time — when the ICC last surveyed the question, 58.9% of respondents reported no change in refusal rates at all, against 26.7% seeing a decrease and 12.3% an increase.

Refusal is not the same as non-payment, and it helps to keep the two apart. When a bank refuses, it must give a single notice stating that it is refusing and listing each discrepancy it is refusing on. From there the documents are usually corrected and re-presented, or you as applicant waive the discrepancies and instruct payment anyway. What you have lost is time and fees, not usually the money — a distinction that matters when your supplier is panicking about a refusal notice.

This document-examination machinery is the same one your bill of lading has to survive, which is why the transport document is the single most refused item in the file.

Why a Yiwu Market Booth Cannot Present Your L/C Documents

Here is the part no general L/C explainer will tell you, and it is the single most useful thing to understand before you ask a Yiwu supplier for L/C terms. The booth you are buying from is, in the overwhelming majority of cases, structurally incapable of being your beneficiary. Not unwilling — incapable.

Goods bought inside the Yiwu market complex typically leave China under customs supervision code 1039, the market procurement trade mode, governed by the General Administration of Customs announcement revising the market procurement supervision measures (Announcement No. 221 of 2019, building on the original pilot under Announcement No. 54 of July 2014).

The mode exists because the old general trade path could not cope with a buyer taking 40 SKUs from 30 different sellers. Under 1039 a registered market operator files one consolidated export declaration at the place of purchase, covering goods gathered from many separate booths, and the value of a single customs declaration is capped at USD 150,000. The model started in Yiwu and has since been replicated to 39 pilot markets across China.

Now line that up against what a credit demands. The credit names one beneficiary, and that beneficiary presents a commercial invoice in its own name whose goods description corresponds with the credit, together with a transport document showing the full set of originals in the form the credit stipulates. A booth operating under 1039 has none of that machinery. It is not the declarant on the export — the market operator is.

Under 1039 exports receive VAT exemption without any export VAT refund, so the booth has no reason to run the VAT-invoice paper trail a general trade exporter maintains. The booth typically hands you a pro forma or a hand-written receipt, and that is not a document a bank will examine under UCP 600 Article 18.

So when a buyer emails eleven booths asking them to accept an L/C, the replies are refusals or silence, and the buyer concludes that Yiwu suppliers are unprofessional. They are not. They are correctly declining an instrument they cannot perform under.

What a sourcing agent does at this point is not persuasion — it is restructuring. The credit has to be re-pointed at a single party holding genuine import-export qualification, who can be named as beneficiary, issue one commercial invoice covering the consolidated shipment, and present a full document set. That party then settles with the individual booths domestically, in RMB, on booth-appropriate terms. The booths never touch the credit.

This has a second-order consequence people miss. Because one invoice now covers goods from many booths, the goods description on that invoice has to be drafted so it corresponds with the credit while still covering a genuinely mixed carton mix. Ask for a credit that specifies eleven separate product descriptions in detail and you have engineered a discrepancy into the file before the goods even ship.

The workable pattern is a credit whose description is broad enough to be truthfully satisfied by a consolidated invoice, with the SKU-level detail living in the packing list, where the looser “must not conflict” standard applies rather than the invoice’s “must correspond” standard.

How the Process Runs, and Which Documents the Credit Will Demand

The transaction has a fixed shape, and knowing where you sit in it tells you when you still have leverage and when you have none. HFW’s client guide sets out the structure of a typical documentary credit in ten steps, and it is worth walking through them in order because the buyer’s real decision points are clustered at the very start.

1. Seller and buyer agree the sales contract, with payment to be made by L/C.
2. The buyer requests the issuing bank to issue the credit. This is the bank’s own irrevocable undertaking to pay the beneficiary on compliance with the conditions, and the buyer agrees to indemnify the bank and gives a pledge over the documents.
3. The credit is issued and sent to the seller’s local bank, the advising bank.
4. The advising bank examines the credit and informs the seller, adding its own undertaking if it is also confirming.
5. The seller ships the goods.
6. The seller presents documents under the credit to the advising or nominated bank.
7. Documents are checked. If they are in order, payment is made and the documents are forwarded to the issuing bank.
8. The issuing bank checks the documents and reimburses the confirming bank.
9. Documents are released against payment from the buyer, or other arrangements.
10. The buyer uses the documents to obtain possession of the goods.

Notice that everything the buyer controls happens in steps one and two. After the credit is issued, your influence over the file is limited to paying for amendments or waiving discrepancies. That is why the checklist further down this page is worth more than any amount of chasing once the goods are moving — and why, on a consolidated Yiwu order, the beneficiary question has to be settled in step one rather than discovered at step six when the booths cannot produce a presentable invoice.

The credit itself decides which documents are required, and every credit is drafted differently — but the commercial invoice and the transport document are the two that do the heavy lifting, and HFW notes that specific requirements also apply to other documents such as insurance documents and certificates. Each additional document the credit calls for is another surface on which a discrepancy can appear, which is the practical argument for keeping the list as short as the deal allows.

The US International Trade Administration puts the reason plainly: the required documents are detailed and prone to errors and discrepancies, and to avoid payment delays and extra fees they should be prepared by trained professionals.

Three rules govern the file regardless of what is on the list. Original documents are required under UCP 600 Article 17, so a scanned copy is not a presentation. Transport documents and insurance documents must be dated. And the transport document must be a full set of originals, with a charterparty bill of lading ruled out — while whether it is made out “to order” or to a straight named consignee has to match what the credit says, because a mismatch there is a listed discrepancy.

On a consolidated Yiwu order there is a fourth question worth asking in step one. For any certificate the credit demands, confirm early that the party named as beneficiary can actually obtain one covering goods it bought domestically from eleven separate sellers. A credit that calls for a document nobody in the chain can produce is a refusal that was written into the terms before the goods existed.

What an L/C Actually Costs You

Quoted L/C pricing is usually given as a vague percentage, which hides how the charge actually behaves. Commission is typically charged per quarter of the credit’s validity, not once. Take a real published schedule: HBL’s Schedule of Bank Charges for January to June 2026 sets import L/C opening commission at 0.40% per quarter for the first quarter and 0.25% per sub-quarter after that, on annual volumes up to Rs. 25 million, with a floor of Rs. 2,500 per credit.

Fee levels vary widely between banks and countries, so treat those figures as one bank’s published tariff rather than a global rate — but the per-quarter structure is the norm, and it means a credit left open for six months costs materially more than the headline suggests.

The charges that actually hurt are the event-driven ones. The same 2026 schedule prices a discrepancy in import L/C documents at US$ 75, plus US$ 20 correspondence charges. Amendments run Rs. 1,400 flat per transaction plus Rs. 1,000 in SWIFT charges, and if the amendment increases the amount or extends the shipment period, commission is recharged on top.

Transferring an export credit to another beneficiary carries its own flat fee — Rs. 2,000 in that schedule. None of this is exotic. It is the routine cost of a file that needed three corrections.

Run the arithmetic against a typical Yiwu order and the case often collapses on its own. On a US$ 30,000 mixed consignment, opening commission at those rates plus SWIFT and courier is a few hundred dollars before anything goes wrong — and given the 65-80% first-presentation refusal rate, something usually does. Add two discrepancy charges and an amendment, and you are paying real money for an instrument that still does not inspect your goods.

On small mixed orders the protection you actually want is an inspection before the balance is paid, which an L/C does not give you at any price.

The Types You Will Be Offered, and Which Ones Matter

A Chinese issuing bank will quote you a menu. ICBC’s published import L/C product lists transferable, back-to-back, sight payment, deferred payment, negotiation, and usance forms. Most of those distinctions collapse into two questions that actually affect you: when does the money leave, and can the credit be split?

On timing, a sight credit pays when compliant documents are presented; a usance or deferred credit pays at a defined later date, which is effectively supplier-funded credit and is priced accordingly. On splitting, the transferable credit is the one that gets pitched at multi-supplier buyers, and it deserves scepticism. Under UCP 600 Article 38 a credit is only transferable if it expressly says so, and a transferred credit cannot be transferred onward again to a subsequent beneficiary — one hop only. It also carries its own transfer fee per the schedule above.

People hear “transferable” and imagine it solves the Yiwu multi-booth problem. It generally does not, and the reason goes back to the 1039 structure. Transfer moves part of the credit to a second beneficiary who must then present documents in its own name. A market booth that is not the export declarant and issues no bankable commercial invoice is no more able to perform as a second beneficiary than as the first.

Transfer is a useful tool when the ultimate suppliers are themselves qualified exporters — a real factory with its own export licence, for instance. It is not a workaround for market-booth structure, and treating it as one just adds fees to the same dead end.

Confirmation is the one option worth paying for in the right circumstances. Adding a confirming bank gives the seller a second, separate undertaking from a bank in its own jurisdiction. If a Chinese supplier is nervous about your bank or your country’s transfer risk and that nervousness is blocking the deal, confirmation is the lever that unblocks it. If your bank is well known internationally and the supplier is comfortable, paying for confirmation is money spent on reassurance nobody needed.

Is an L/C Right for Your Order?

The honest answer for most readers arriving at this page from a Yiwu sourcing question is no — and that is worth saying plainly, because a lot of trade content sells the L/C as the mature, professional choice regardless of order shape. The US International Trade Administration frames it more carefully, recommending letters of credit for higher-risk situations, where the importer’s credit is unacceptable or unavailable, where the trading relationship is new or less established, or where extended payment terms are being requested.

It also notes flatly that the instrument is labour-intensive and relatively expensive, and that the required documents are detailed and prone to errors and discrepancies.

That last exclusion deserves emphasis because it is the one buyers rationalise away. An L/C protects you against not being paid against documents. It does nothing about the goods. If the thing keeping you awake is whether the cartons contain what you ordered at the quality you agreed, the instrument that addresses it is a pre-shipment inspection with the balance held back until you have the report — not a bank credit.

Buyers who conflate the two end up paying L/C fees for quality protection they never bought, then discovering the gap after the container lands.

Worth keeping in perspective: as a payment instrument the L/C works. Reporting on the ICC Trade Register 2025, which draws on more than 47 million trade finance and export finance transactions with exposures above USD 23 trillion contributed by 21 global banks, puts default rates across major trade finance products including letters of credit below 0.3% overall.

The ICC does not publish the split for import versus export credits in its public summary, so anyone quoting you a precise import-L/C default percentage is filling in a number the source withholds. The instrument is low-risk. The question is whether your order shape justifies its cost and friction.

What to Check Before You Let the Credit Be Issued

Once a credit is issued, every change costs a fee and burns days. The leverage is all in the draft. Work through the terms in this order before your bank transmits anything, because each item below maps to one of the discrepancy categories the ICC actually records.

• Beneficiary name and legal capacity: confirm the named beneficiary holds import-export qualification and is the party that will appear as exporter on the declaration. If those are two different entities, stop and fix it now.
• Goods description: keep it broad enough that a consolidated commercial invoice can correspond with it truthfully. Push SKU-level detail into the packing list.
• Latest shipment date with real slack: timing issues are the ICC’s most common discrepancy category. Build in more buffer than the supplier’s optimistic estimate.
• Presentation period: the default is 21 calendar days after shipment. If documents have to travel between several parties first, negotiate a longer window into the credit rather than hoping.
• Transport document form: to order versus straight consignee is a listed discrepancy when the credit and the document disagree. Decide who is meant to control the goods, then state it once, consistently.
• Full set of originals: confirm how many originals the credit requires and that the carrier will actually issue that many.
• Documents nobody can produce: strike any certificate the beneficiary cannot realistically obtain. A credit calling for a document that does not exist is a guaranteed refusal.
• Partial shipment and transhipment permissions: mixed consolidated cargo frequently needs both allowed. Prohibiting them by default creates discrepancies later.
• Who pays which charges: settle discrepancy fees, amendment fees and confirmation costs in the credit terms, before there is a dispute about them.

The single highest-value habit is having the beneficiary review the draft credit before it is transmitted. It costs nothing, takes a day, and catches the terms the seller knows it cannot meet.

Rules and bank practice differ by country and by product, so treat this as a working checklist rather than legal advice, and confirm the specifics with your bank or a trade finance specialist for your own transaction. The Incoterm you agreed also has to line up with the documents the credit demands — an EXW sale cannot produce a bill of lading in the seller’s name, and credits are routinely drafted as if it could.

What We Check on a Consolidated Document Set, and Where We Stop

Because a credit is examined on paper, the useful work happens before the paper is created. The documented procedure our sourcing service runs on a consolidated Yiwu order lines up against the discrepancy categories the ICC actually records, and it is worth stating plainly — including its limits, because an agent who claims to solve the bank’s job is overselling.

Goods from separate booths are received into a 3,000 sqm warehouse with 30 days of free storage, which is what buys the timing slack a credit needs: the latest shipment date has to clear the slowest booth, not the average one, and storage that costs nothing for the first month is what makes waiting for booth eleven a scheduling decision rather than a financial one.

On the goods themselves, pre-shipment inspection is performed at 100% produced and 80% packed against ANSI/ASQ Z1.4 (ISO 2859-1) sampling, with critical defects at zero allowed, major at AQL 2.5 and minor at AQL 4.0, and the PDF report — hi-res photos, video tests and measurement data — is delivered within 24 hours. Standalone inspection runs at a flat daily rate, published as $199/man-day. Cargo insurance is quoted at 0.3%.

That 24-hour report turnaround is the number that matters most to an L/C timetable. The presentation clock starts at the shipped-on-board date and runs 21 calendar days, so an inspection result that lands the next day still leaves room to correct a carton count or a marking before the invoice and packing list are finalised. An inspection that reports a week later does not.

Where we stop is equally important. An agent can align the commercial invoice with the credit’s goods description, keep the SKU detail in the packing list, confirm the transport document’s consignee form matches what the credit stipulates, and make sure the full set of originals exists.

An agent cannot make a bank accept a discrepant file, cannot waive a discrepancy on your behalf — only the applicant can do that — and cannot turn a booth into a qualified beneficiary. Nor does any inspection regime give you the bank-side protection an L/C withholds: the credit still pays on matching paper covering defective goods. Anyone promising otherwise is describing a service that does not exist under UCP 600.

A Worked Example: One Credit, Eleven Booths

Take an illustrative case. A buyer assembles a US$ 62,000 order: kitchen storage from four booths, seasonal decorations from three, hardware and tools from another four. Eleven sellers, one 40ft container, and a bank at home that has offered a sight credit. Here is how the structure has to be built for the credit to survive examination.

First, the beneficiary. The eleven booths cannot be named — none is the export declarant. One qualified exporter is named as sole beneficiary, and it is that entity’s name that must appear on the commercial invoice and be consistent across the file. Second, the value check: at US$ 62,000 the order sits comfortably under the USD 150,000 single-declaration cap, so it can move as one 1039 declaration rather than being split, which would have meant multiple document sets against one credit and a much harder presentation.

Third, the description. The credit is drafted to describe the goods in terms broad enough that one invoice covering all eleven booths’ cartons corresponds with it — “assorted household storage, seasonal decorative items and hand tools” rather than eleven itemised lines. The SKU breakdown, carton counts and marks live in the packing list, where the standard is only that data must not conflict.

Fourth, the timeline, which is where these deals fail. Goods trickle in from eleven booths over roughly two weeks. They are checked and consolidated, then the container loads, and the shipped-on-board date on the bill of lading is what starts the 21-day presentation clock. The bank then has up to five banking days to examine.

So the latest shipment date in the credit needs to sit well beyond the slowest booth’s realistic delivery, not the average — one late seller moves the loading date for the whole container, and a missed latest-shipment date is a discrepancy that no amount of careful invoicing repairs.

Fifth, the honest verdict on this file. It works, but the buyer is paying opening commission per quarter, carrying real refusal risk on a document set assembled from eleven sources, and still has no bank-side protection on quality. For an order of this shape, a deposit against a signed order plus the balance released after a satisfactory pre-shipment inspection covers the actual risk more cheaply.

Where the credit genuinely earns its place is when the buyer’s own lender or corporate policy requires documentary settlement — then the structure above is how you make it survive contact with a wholesale market. Either way, the mechanics of consolidating goods from separate suppliers into one shipment is the part that determines whether the document set holds together.

Conclusion

A letter of credit is a documentary instrument, and it behaves exactly like one: it pays on matching paper, refuses on mismatched paper, and takes no view on your goods. For a single large order from a qualified factory it is a sound way to bridge trust. For a mixed Yiwu consignment drawn from booths exporting under market procurement mode, the instrument fits so badly that the real work is restructuring who sits on the credit at all.

If you are weighing an L/C against a deposit-and-balance arrangement for a Yiwu order, work out first whether your actual exposure is payment risk or product risk, since only one of them is something a bank credit addresses. From there, the terms in the checklist above are worth walking through with whoever will be named as beneficiary before anything is transmitted.

Questions fréquemment posées