Hedging a Currency You Cannot Deliver: NDFs and Restricted-Currency Risk for SMEs
A mid-sized exporter wins a contract in a market where the currency does not leave the country. The invoice is denominated in a currency the buyer can pay, the supplier can receive, and nobody outside that market can settle freely. Finance does the sensible thing and asks the bank for a forward. The bank explains, politely, that a deliverable forward is not available to a company without a local entity, a local account or a permitted underlying transaction, and offers something with a different name instead.
That moment is where a surprising number of mid-market treasury losses originate. Not from a wrong view on the currency, but from signing a hedge whose mechanics were never fully explained, against an exposure that was never correctly measured.
This article covers the mechanics of non-deliverable forwards (NDFs), the three costs that sit inside the quoted rate, the basis risk that does not appear in any term sheet, and the honest question every SME should answer first: is this exposure hedgeable at all, and at what price?
Why some currencies cannot be delivered
Convertibility is a policy choice, not a property of money. Many emerging markets maintain restrictions on capital account transactions while allowing current account convertibility for genuine trade. Others impose surrender requirements, restrict who may hold balances offshore, or limit the instruments a non-resident may use.
The practical consequence for an SME is simple. You can be exposed to a currency for a year and still have no legal mechanism to take delivery of it in size, on your own balance sheet, on the date your invoice falls due. What you can usually do is agree a contract that pays you the difference between two rates in a currency you can receive.
That is an NDF. The exposure is real, the hedge is real, but the currency never moves. Only cash does.
Common NDF currencies include the Indian rupee, Chinese renminbi, Korean won, Taiwan dollar, Brazilian real, and a long tail of Asian, African and Latin American currencies. If you invoice in one of these and hold no local entity, assume you are in NDF territory until a bank tells you otherwise in writing.
The five parameters that define an NDF
An NDF is not a vague instrument, it is a contract with five defined terms. If any one of them is left unclear, you do not have a hedge, you have a dispute waiting for a rate move.
- •Notional. The amount of the restricted currency being hedged. Usually converted into a settlement currency notional for calculation purposes.
- •Contract rate. The rate agreed today, which is the rate your hedge is designed to lock in.
- •Fixing date. The date on which the reference rate is observed. Conventionally one or two business days before settlement.
- •Fixing source. The specific published rate named in the confirmation. A central bank reference rate, a page on a data provider, or a defined average. Vague references such as "the prevailing market rate" are not a fixing source.
- •Settlement currency and date. Which freely deliverable currency the difference is paid in, and when.
Settlement for a company buying the restricted currency is calculated as:
Settlement amount = Notional × (Contract rate − Fixing rate) ÷ Fixing rate
Paid in the settlement currency, typically US dollars or euro. For a company selling the restricted currency, the sign reverses. Note the division by the fixing rate: this converts the difference into a proportion of the notional, which is why NDFs are quoted and settled in round notional terms rather than in currency units.
A worked example: a UK importer with a rupee payable
An illustrative case, using round numbers.
A UK components distributor owes an Indian supplier INR 250,000,000, payable in six months. Spot USD/INR is 94.00. The company's margin is priced off that rate and it wants the number fixed.
The bank quotes a six-month NDF at 94.00, settlement in USD. Notional USD 2,659,575 (INR 250m at 94.00).
Scenario A: the rupee strengthens, fixing at 91.00.
The company buys INR in the market: USD 2,747,253. The NDF settles in its favour: 2,659,575 × (94.00 − 91.00) ÷ 91.00 = USD 87,678 received. Net USD cost: 2,659,575. Effective rate: 94.00.
Scenario B: the rupee weakens, fixing at 97.00.
The company buys INR in the market for USD 2,577,320, a saving of USD 82,255 against budget. The NDF settles against it for the same amount. Net cost: 2,659,575. Effective rate: 94.00.
Both paths land on the same number, which is the entire point. The hedge has removed the currency question from the budget, and it has also removed the chance to gain from a favourable move. Companies that expect to be congratulated for the gain in Scenario B are usually the same ones that explain away Scenario A.
Cost one: forward points, which are interest rates in disguise
An NDF rate is not a forecast. It is spot plus or minus the interest rate differential between the two currencies over the hedge period. Where the restricted currency carries a higher implied yield than the settlement currency, the forward sits at a discount to spot, and a company buying that currency forward pays more than today's spot for the privilege of deferring the payment.
For several Asian and Latin American currencies, that discount is structural rather than cyclical. A treasurer looking at an NDF level 4% above spot is not being overcharged by the bank. They are being shown the cost of not holding the currency for six months, and the cost is a real economic item that belongs in the margin calculation from day one.
This is also why quoting an NDF hedge into a fixed-price contract requires more care than quoting a G10 forward. The differential moves, and a tender priced at one differential can be loss-making at another before any spot move occurs.
Cost two: the onshore and offshore spread
Restricted currencies frequently trade in two places at once: the domestic onshore market, where the central bank manages liquidity and access, and an offshore market where the currency is freely traded but never delivered. The two rates are linked by arbitrage that is imperfect precisely because capital does not move freely.
That gap is the onshore/offshore basis. It widens when domestic liquidity tightens, when capital controls are enforced, or when sentiment diverges from the official fixing. An NDF is priced off the offshore leg. If your real transaction happens in the onshore market, you have hedged a rate you will not transact at, and the difference is your cost of being wrong about the basis.
Where a company has onshore access through a local entity or a permitted transaction, an onshore deliverable forward is often cheaper and better matched. The NDF is the fallback for companies operating from outside, not the default for everyone with the exposure.
Cost three: access, and the regulatory price of it
The third cost is not in the rate at all. It is in the conditions attached to being allowed to hedge, and those conditions have been moving.
India provides the clearest recent example. In October 2026 the Reserve Bank of India tightened the framework around rupee derivatives: the threshold for entering foreign exchange derivative transactions without establishing an underlying exposure was cut from USD 100 million to USD 5 million equivalent across all authorised dealers, users were barred from rebooking a cancelled contract (rollover at maturity remains permitted), and a cash Foreign Exchange Risk Reserve of 20% of the rupee equivalent notional was introduced for contracts above USD 2 million notional where the dealer purchases foreign currency against the rupee. A follow-up circular restricted non-deliverable derivative contracts involving the rupee offered to non-resident users to hedging purposes only.
China moved the other way on one specific cost: in February 2026 the People's Bank of China cut the foreign exchange risk reserve ratio on forward foreign exchange sales from 20% to zero, effective 2 March 2026, materially reducing the cost of buying dollars forward onshore.
Two lessons follow. First, the price of a restricted-currency hedge is partly a policy variable, and policy can change between the date you price a contract and the date you hedge it. Second, a company whose hedging depends on a facility that regulators can withdraw is not running a policy, it is running an assumption.
Add the ordinary operational frictions on top: documentation, credit lines or collateral support annexes for larger notional, minimum ticket sizes that put the instrument out of reach for small exposures, and settlement mechanics that require someone to actually monitor a fixing date.
Where the basis risk hides
Three mismatches account for most NDF disappointments in mid-market companies.
The fixing mismatch. You hedged the invoice date. The customer paid eleven days late. The fixing happened on the contract date, the spot transaction happened on the payment date, and the hedge protected the wrong day. Fix by hedging to a conservative expected settlement date and accepting a small roll, not by assuming invoices are paid on time.
The source mismatch. The confirmation names a central bank reference rate. Your bank's spot rate on the day you actually convert is a commercial rate with a spread. The two rarely coincide. Read the fixing source before signing, and model the expected gap.
The exposure mismatch. You hedged gross receivables while paying local suppliers in the same currency. That is a hedge against a cash flow that partly offsets itself, and it costs a spread twice. Net first, always. Systematic netting typically removes 25% to 40% of gross exposure at no cost, and it is the cheapest improvement available before an NDF is even discussed.
When the NDF is unavailable, expensive or simply wrong
For many SMEs the correct conclusion is that the instrument is not the answer. The alternatives are not compromises; several are structurally better.
Invoice in a deliverable currency. The cleanest hedge is often a commercial decision made at contract stage. Shifting invoicing to USD or euro transfers the currency risk to a counterparty better placed to carry it, at the cost of a price concession that is usually smaller than the hedge cost plus the basis risk.
Build a natural hedge. Local costs against local revenue, or local sourcing against local sales, reduce net exposure before any financial instrument is used. Where a market is large enough to justify local operations, this is the lowest-cost risk reduction available.
Borrow in the restricted currency. A local-currency loan matched against local receivables creates a liability in the same currency as the asset. Where onshore borrowing is permitted, this is often cheaper than the NDF alternative and needs no derivative documentation.
Use a pass-through clause. For long contracts, a pricing clause that adjusts with the reference rate beyond a defined band shares the risk explicitly instead of hedging it at a spread. Institutional customers resist this, but it is negotiable more often than finance teams assume.
Shorten the exposure. Faster collection, milestone billing, or receivables finance against the specific invoice reduces the window in which the rate can move. Halving the exposure period halves the risk without a single trade.
Provision and accept. For exposures below roughly 3% of EBITDA, an explicit budget provision and a stated policy of no hedging can be entirely rational. The cost of hedging small exposures often exceeds the expected loss, and the operational overhead is real.
A decision framework
The table is a starting point for a policy discussion, not a formula to apply mechanically.
| Situation | Preferred approach | Why |
|---|---|---|
| Exposure under 3% of EBITDA, short cycle | Accept and provision | Hedge cost and admin exceed expected loss |
| Exposure material, onshore access available | Local entity, deliverable forward | Cheaper, better matched, fewer basis issues |
| Exposure material, no onshore access, active NDF market | NDF ladder against net exposure | Only practical instrument, but price the basis |
| Exposure material, NDF expensive or restricted | Reprice in hard currency, or pass-through clause | Moves risk to the party who can carry it |
| Long-dated, uncertain, tender-stage exposure | Option, or hedge on award only | Never hedge a forecast you may not win |
| Local borrowing permitted | Local-currency debt against receivables | Natural match, no derivative documentation |
The pre-trade checklist
Before the first NDF, six items should be settled in writing, not in conversation.
- •Confirm the fixing source and the fallback. What happens if the reference rate is not published on the fixing date, for instance because of a market holiday or a disruption? Confirmations normally contain a fallback; read it.
- •Confirm the netting set and the credit terms. An uncollateralised NDF is a credit exposure to your bank, and theirs to you. Understand whether margin can be called, and in what circumstances.
- •Confirm the accounting treatment. Under IFRS 9, cash flow hedge accounting requires documentation at inception, a defined hedged item, and a test of economic relationship. Retrofitting documentation after the trade is not permitted, and without it, settlement gains and losses land in profit and loss rather than other comprehensive income, which creates exactly the earnings volatility the hedge was meant to remove.
- •Confirm who monitors the fixing. A named person, a diarised date, and a reconciliation to the spot transaction executed on the same day. This is the single most common operational failure.
- •Confirm the tax treatment. Settlement receipts and payments are usually taxed as trading items or as part of the underlying, depending on jurisdiction and documentation. Check before the first settlement, not after.
- •Confirm the exit. If the exposure disappears, what does it cost to close the hedge? An NDF has no delivery, so exit is a cash settlement at market. Know that number in advance.
The point of all of this
Restricted-currency exposure is not a niche problem for large corporates. Any SME that sells into a market with capital controls, or buys from one, carries it. The instrument available to hedge that exposure is a cash-settled contract that never delivers the currency, priced off an offshore market that may not match the rate at which the company actually transacts, under a regulatory framework that can change the cost of the hedge between pricing and execution.
None of that makes NDFs unusable. It makes them a tool that has to be understood before it is signed, sized against net exposure after natural offsets, priced with forward points and basis included in the contract margin, and monitored by a named person against a documented fixing. Companies that do those four things get a hedge that behaves. Companies that treat an NDF as a forward with a different name tend to discover the difference in the settlement, at a rate they did not choose, on a date they did not plan for.
The first step is not a trade. It is a one-page exposure map: which currencies, how much, for how long, and whether each one can be delivered to your own balance sheet. Most companies find that half their apparent exposure nets away, a quarter is hedgeable onshore, and only a small remainder genuinely needs an offshore instrument. That map, not the hedge ratio, is what makes the policy defensible.
