Currency is the only line in your unit economics that nobody quotes you, nobody invoices you, and nobody flags when it goes wrong.
Freight you can get a quote for. Duty you can look up. Supplier price, packaging, inspection, the prep centre — all of it arrives as a number you can put in a cell. Then the rate moves four percent between the quote and the balance payment, and the margin you approved in a spreadsheet is gone before the container sails.
I have sat in on dozens of costing reviews where FX appears as one hardcoded cell, typed in once and never revisited. That cell is usually the least examined assumption in the file — and on RMB-priced orders one of the largest variables you do not control. This is not a piece about forecasting the yuan. It is about mechanics: which risk you hold, where in the cycle, what language transfers it, and how the payment route decides the rate you end up with.
First, work out which currency pair you are actually exposed to
Most importers assume their FX problem is USD/CNY because their supplier is in China. Often it is not: the exposure depends on what you earn in against what you pay in.
You sell in USD and pay in USD. Revenue and your largest cost sit in the same currency, so a USD/CNY move does not hit your bank account directly. The exposure is indirect: it arrives as a repricing conversation at the next quotation round, or as a supplier quietly raising the USD price to rebuild the margin the yuan move took from them. Not hedgeable with a forward — you manage it with quote validity and knowing your supplier's RMB anchor.
You sell in EUR or GBP and pay in USD. Where the genuinely unmanaged exposure usually lives — European and UK sellers, and anyone on Amazon EU paying Chinese suppliers in dollars. A live cross-currency position on every open order, frequently larger in percentage terms than the CNY move people worry about, and it shows up fast because the payment date is fixed and the rate is not.
You pay in RMB. Standard for larger importers, sourcing agents and anyone with a mainland entity or a supplier who insists on local-currency settlement. Direct, visible and easy to model — and easy to get wrong, because the rate a bank quotes is not the rate in a search result.
| Your situation | Where the risk actually sits | What actually manages it |
|---|---|---|
| Sell USD, pay USD to a Chinese factory | Indirect — supplier reprices at the next quote to protect their RMB margin | Quote validity windows, an FX clause, and knowing the RMB anchor so you can recognise a repricing |
| Sell EUR/GBP, pay USD | Direct — EUR/USD or GBP/USD moves between PO and payment | A forward on the known payment amount, or invoicing in your own currency |
| Pay RMB to the factory or agent | Direct — USD/CNY across the cycle, plus the bank spread | Forward cover on the known balance, staged conversion, and a route that does not convert twice |
| Multiple markets, multiple currencies | Net exposure across all pairs, not each one separately | Net receivables against payables before paying anyone to hedge |
Half the sellers who think they need FX hedging actually need better quote hygiene: write down what you earn in, what you pay in, and when.
Why a USD quote is not the RMB price converted
Ask a Chinese factory for both prices — RMB and USD — for the same specification. The USD number is almost never the RMB number divided by the rate of the day. It is the RMB number divided by a rate the supplier considers safe for the period they expect to be paid at, plus a buffer.
That buffer is not dishonesty. It is the same logic as a builder fixing a price on a steel-exposed project: they are selling you certainty, and certainty has a price. The problem is that most buyers never ask what rate is embedded, so they cannot tell whether the buffer is 1% or 5% — or whether a later increase is a genuine cost increase or a currency adjustment in disguise.
Ask directly: “What exchange rate is built into this USD price, and how long is that rate valid for?” A supplier with a managed book answers in one sentence. One who is guessing changes the subject.
The RMB price is the real price. Everything else is a conversion with a margin of safety attached. If you only hold the USD number, you are negotiating the wrapper instead of the product — and you will lose every argument about a price increase because you have nothing to compare against.
The exposure is the whole cycle, not the wire date
Most cost models treat FX as a single event: the day the balance goes out. An order actually carries currency exposure across four to six months, and each stage behaves differently.
| Stage | Typical length | What is fixed | Where the currency risk sits |
|---|---|---|---|
| Quote issued | 30–60 days validity | Nothing — an offer | Supplier's buffer covers the window; when it expires, so does the price |
| Sample and spec approval | 2–5 weeks | Specification, not price | Silent drift: the approved spec gets priced at today's rate, not the sampled one |
| PO signed, deposit paid | Day zero | Unit price and deposit | Only the deposit portion is economically locked |
| Production | 30–60 days | Unit price, contractually | Supplier absorbs or reprices at renewal — unless the contract lets them back out |
| Balance payment | Before shipment | Amount due | The live moment: the rate that applies to the largest payment |
| Ocean transit | 25–45 days | FOB cost is fixed | None on goods cost — but duty converts at the customs rate, not yours |
| Payment terms (net 30/60) | If you have them | Nothing until settlement | An open position for the whole term — terms quietly become a currency bet |
Two rows matter most. Customs conversion: duty is computed by the importing authority using its own published rate on a fixed cycle — not the rate your bank charged, not the rate on the invoice. On a high-duty product, a move between invoice date and entry date changes landed cost with no invoice to show for it. Payment terms: net 30 or net 60 feels like a concession, but it is an open currency position for that period, held by whoever owes the money — if the currency you owe appreciates while you hold the balance, the terms cost you part of what they gave you.
The arithmetic that makes this real
The numbers below are illustrative; substitute your own cost structure.
Take a product selling at 100 in your reporting currency. Cost of goods is 50 and RMB-denominated, so it moves with the rate. Freight, duty, fees and marketing are another 25, fixed in your currency. Gross profit is 25 — a 25% gross margin.
The yuan appreciates 4% between the quote you costed and the balance payment. Cost of goods goes 50 to 52. Gross profit goes 25 to 23.
Four percent on one input became two points of gross margin — 8% of gross profit. On a 15% margin product with the same cost structure, the same move takes more than a quarter of gross profit. A 4% move in a quarter is not an extreme event; it is an ordinary one.
The mistake is not the move — it is discovering it late. If the rate is a variable in the model, a 4% move shrinks a number before you place the order and you can decide whether the deal still clears your hurdle. If it is a hardcoded cell, the same move shows up as a worse-than-expected month — with the container already on the water.
What an exchange-rate clause should actually say
Buyers try to solve this with “prices are subject to exchange rate fluctuation.” That sentence transfers nothing: it does not say from what, to what, or who pays. A clause both sides can live with has four working parts.
- A named base rate and its source. The rate the price was built on, from a specific published source on a specific date, so both parties can look it up and get the same number.
- A neutral band. Moves inside the band change nothing. This is what makes the clause acceptable to a supplier: they are not exposed to every tick, only to moves that hurt.
- Shared adjustment outside the band. Beyond it, the difference is split — commonly 50/50, sometimes with a cap on your exposure. Shared pain is why this gets signed instead of argued about.
- A reset trigger and a floor/ceiling. Recalculation at each PO or at a fixed review date, with an absolute cap in either direction so neither side carries a tail event.
The argument that works is not “the yuan moved, so I want a discount.” It is: “We both want a price that does not need renegotiating every quarter. Here is a mechanism that protects you if the currency goes against you and me if it goes against me. Inside a three percent band, neither of us does anything.” And if you have the clause, apply it when the move favours the supplier — a one-way clause is not a clause, it is a threat, and it gets priced into your next quote.
The payment route decides the rate you actually get
Two importers can pay the same invoice on the same day and get different economics: the route decides how many times the money converts and at what spread.
Watch the spread, not the headline. The rate on a bank's contract note is not the published mid-market rate; the difference is their margin, and on small transfers it is often the largest single FX cost in the transaction. Ask for it in basis points and compare two providers on the same day for the same amount.
Do not convert twice. If you hold EUR, convert to USD to pay a supplier, and the supplier converts USD back to RMB, two spreads were paid to move money that only needed to move once. Where the supplier or agent can settle in the currency you already hold, one leg disappears.
Know which yuan you are dealing with. There are two prices: the onshore rate, traded inside the mainland within a managed band around a daily fixing published by the central bank, and the offshore rate, which trades freely and can sit at a premium or discount. Which applies depends on who receives the money and where — a Hong Kong trading entity usually means offshore, a mainland factory paid in RMB means onshore with its own documentation.
Intermediary charges are almost always deducted from the amount received, so agree up front who bears them. A supplier short-paid by forty dollars on a five-figure invoice will mention it — at the worst possible moment. For the payment methods themselves and the fraud patterns around them, see paying Chinese suppliers safely.
Tools that do not require a treasury department
1. Price the buffer in deliberately. The cheapest hedge is a margin that survives a bad quarter. If a 4% adverse move breaks the deal, the deal was too thin — that is a pricing decision. Model at a rate slightly worse than spot and treat anything better as unplanned upside.
2. Net your natural exposure first. Receivables and payables in the same currency are already a partial hedge; only cover the net remainder. Sellers with US revenue and USD supplier payments are more naturally hedged than they realise — the gap is usually the EUR/GBP leg.
3. Use a forward for known amounts. For a balance payment whose size and date you know, a forward locks the rate today. Banks and several payment providers offer them to small businesses, usually against a credit line or margin. Hedge the payment you have a PO for, not the volume you hope to do next year.
4. Stage conversions instead of timing them. Converting roughly equal amounts on a schedule removes the single worst outcome — converting the year's requirement on the worst day of the year.
5. Choose the invoice currency on purpose. Quoting in your selling currency puts the risk on the supplier, who prices it in. Quoting in RMB gives transparency and precise hedging, at the cost of carrying the risk. Neither is wrong; drifting into one without deciding is.
And four things to avoid. Do not speculate — hedging more than your committed orders is a currency position financed by working capital. Do not over-hedge and cancel — a forward you cannot deliver against must be closed out, a cash cost with no goods attached. Do not switch suppliers over a currency move — requalifying a factory costs months, and a supplier who reprices transparently with a stated rate beats a cheaper one you have never audited.
The checklist to put in the PO file
- Record the currency of the agreement; if it is not RMB, record the RMB anchor price and the rate embedded in the USD figure.
- State quote validity in days, with an expiry date.
- Write the FX clause with a named rate source, a band, a sharing rule and a cap.
- Model landed cost at a rate worse than spot, and note the assumption next to the cell.
- Convert duty at the importing authority's published rate, not your bank's.
- Write the currency you earn in versus the currency you pay in on one line. That is your exposure.
- Price the FX spread as a cost line and compare providers annually.
- For any balance large enough to notice, decide in advance: forward, staged conversion, or accept the exposure on purpose.
- Review the FX assumption quarterly alongside freight and duty — the three lines that move without an invoice.
The Bottom Line
Currency risk on China sourcing is not a forecasting problem. It is a bookkeeping problem most importers have not booked. The exposure is created the day you accept a quote in a currency you do not earn, it runs the length of the order cycle, and it is settled by whichever mechanism — clause, forward, buffer or luck — is in place when the balance falls due.
Fix the cheap things first: know your RMB anchor, know which pair you are exposed to, put a real clause in the agreement, and stop treating the rate as a constant. Those four moves cost nothing and cover most of the risk. What remains is what forwards are for.
Cost the order the way it will actually land
The free mini-guide walks through the full landed-cost build-up — freight, duty, currency, fees and the assumptions most cost sheets leave out — plus the document set and inspection checkpoints that decide whether a container arrives profitable.
Browse the China sourcing guides