Trade Loan

An exporter signs a NZ$2.0m contract with an offshore buyer on sixty-day terms. Between signing and getting paid sit five months: two months of production to fund, a month on the water, then the buyer's credit period. For those five months the business needs money that exists precisely for this trade — not more, not longer. A term loan is the wrong shape; an overdraft is the wrong size; a revolving line may not exist or may be needed for everything else. The trade loan is built for exactly this: finance advanced against one identifiable transaction, with tenor matched to the trade cycle and repayment coming from the trade's own proceeds.

The product's essence is specificity and self-liquidation. Repayment does not rest on the borrower's general trading, or a promise of instalments — it rests on one transaction completing and paying. That is the cleanest repayment logic in lending, and its exact weakness: a term loan survives one failed deal, a revolver absorbs one bad month, but when a trade loan's trade fails, the design fails with it. So a desk reading a trade loan cannot stop at the client — it must read the trade: the buyer, the contract, the goods, the route the money takes home.

The trade itself — named parties, goods, amount, payment terms, shipping basis, tied one-to-one to the facility. The advance — commonly around 70-85% of trade value, never the whole of it, so the client keeps skin in the game and the buffer absorbs currency, disputes and freight. Tenor matched to the cycle — production plus transit plus the credit period, typically anywhere from one to six months; longer than the cycle and the bank is carrying idle time risk, shorter and the product stops making sense. A single-rail rate — no commitment fee; short tenor; priced between a revolver (the risk is more definable) and pure LC discounting (there is no issuing bank standing behind it).

Two features are the product's fingerprint. The drawdown trigger: funds release against evidence the trade exists and is moving — purchase order, contract, invoice, shipping documents — so every dollar drawn is anchored to something inspectable. The repayment trigger: the buyer's payment routes to a collection account the bank controls, and the loan clears from it automatically. That account is not paperwork — it is the physical implementation of self-liquidation, and most of the ways this product fails in practice begin with money arriving somewhere else. Security wraps around both: an assignment of the trade receivable first, a general security agreement behind it.

The structure is simpler than a revolver — one rate, no commitment fee — but short tenors make percentages misleading, so the honest arithmetic runs in dollars against the trade. A constructed illustration, every figure invented for teaching:

Two percent of the trade sounds small until it meets the client's margin. If the exporter earns 25% gross on this contract — NZ$500,000 — the financing consumes roughly 8% of the gross profit. Acceptable. If the margin is 8%, financing consumes a quarter of it, and the honest question belongs on the table before the facility does: is the client trading for themselves, or working the trade to pay the bank? Thin-margin trade plus full-cost finance is a business model with the client's name on it and someone else's economics.

Repayment runs down a three-layer waterfall, and the layers weaken as you descend. Primary — the design: the buyer pays, into the collection account, and the loan clears. Everything is priced on this path, and it stands on three checkable conditions: the payment terms are in the contract, the account control is real, and the buyer has been told where to pay. Secondary — the client: if the trade fails, can the client's other cash flows cover this principal in the month it matures? Usually weaker than it sounds — nobody plans a standalone repayment for a loan designed to repay itself. Tertiary — enforcement: collect from the buyer directly under the assignment, or enforce general security. Cross-border, slow, and never the basis of a day-one decision. A desk that finds itself relying on the third layer at approval is not pricing a trade loan; it is writing a distressed asset with extra steps.

Because repayment is one transaction, the risk reading goes granular. On the buyer: five years of relationship history is not the same fact as “will pay this invoice on time” — the desk wants the last twelve months of actual payment behaviour, disputes included. On the trade's reality: documents are verified, not collected — related-party buyers, off-market pricing, unusual routes and too-generous terms are the classic fingerprints of a trade constructed to raise finance rather than to sell goods. On jurisdiction: a willing buyer behind capital controls is an unpaid loan with good intentions, and an assignment that is unenforceable in the buyer's country is comfort, not security. And on the portfolio: ten trade loans that are each individually sound can still be one buyer, one commodity, or one country ten times over — the single-trade logic must not be allowed to hide the aggregate.

The stacked-loan drift. A client doing continuous small trades gets a trade loan for each, and the portfolio quietly recreates invoice finance without its monitoring. The product boundary is worth holding: identifiable single trades take trade loans; continuous flow takes a revolving facility or invoice finance.

History mistaken for credit. “We have dealt with this buyer five years” answers a relationship question, not a payment question. The file wants recent payment performance — days-to-pay against terms, disputes, partial payments — because long relationships fail on the trade nobody checked.

The leaked collection account. The account is set up, and then the client tells the buyer to pay the usual account — often innocently. The proceeds land in general funds, get absorbed, and the self-liquidating loan is suddenly unsecured. Before drawdown, two confirmations: the buyer has acknowledged the payment instruction, and the invoice actually shows the collection account.

The rolling extension. Maturity passes, the buyer “needs another month,” the loan extends — three times, silently, while the real news (the buyer defaulted months ago) waits. A missed trade-loan maturity is an escalation event on the day it happens, because every extension is the self-liquidation assumption failing in public. Two extensions is a limit; the third conversation is restructuring.

The step change waved through. A client whose largest-ever order was NZ$1.0m asks to finance NZ$5.0m for the same buyer. Scale is not linear: production capacity, supply chain, and the buyer's own reasons for a five-fold order all need re-underwriting, and the structural answer is usually staged drawdowns against milestones, not one cheque against enthusiasm.

The closest published file is Issue 02 — an exporter with USD receipts, NZD costs, and a multi-currency request the senior read structures around what is evidenced, which is the same discipline this product runs on: lend against the trade you can see. The sibling product where the buyer's bank stands behind payment is LC discounting; the arithmetic for sizing any working capital gap is the cash conversion cycle; and the borrower's-side view of how a bank weighs all this is How Banks Read You.