The SwapEazi reference: method and governance

What every SwapEazi figure is measured against, how it is built, how precise it is, what is excluded, and how it is kept independent.

1. What the SwapEazi reference is

The SwapEazi reference is an independent measure of the fair mid-market rate between two currencies at a point in time. Every SwapEazi figure is measured against it: the cost of a bank payment, an FX deal, a card transaction, a stablecoin swap on a DEX, or the conversion in and out of a tokenised fund.

Its purpose is to answer one question the same way for everyone: how far a conversion was from the fair rate, with an honest statement of how precisely that can be known. It is not a tradable price, it is not a quote, and no one is bound to deal at it.

2. How a cost is calculated

Cost is reconstructed from what was sent and what landed. The provider's quote is not needed, and is not used: it describes the price the provider chose to show.

• Effective rate = amount landed ÷ amount sent.
• All-in cost (%) = (1 − effective rate ÷ mid-market rate) × 100.

The all-in cost captures the exchange spread, visible fees, correspondent deductions and receiving-bank charges in one number. A negative figure means the payment landed better than the mid-market rate.

Where a company's own negotiated rate is compared with what companies of its size pay, the tiers above USD 10 million a year follow regulatory transaction data for FX (Hau, Hoffmann, Langfield and Timmer, Management Science, 2021).

3. What the reference is built from

The reference is built only from inputs that anyone could, in principle, assess for themselves:

• Official exchange rates published daily by several independent public institutions, in more than one jurisdiction.
• For currencies whose own monetary authority publishes a daily rate, that home rate, where it can be read or is supplied by the publisher's own download.
• For a live assess between daily publications, prices from public order books on several trading venues, used only after each venue passes the quality assesses in section 4.

Where several institutions publish the same currency on the same day, the reference is their median, so that no single publisher can move it on its own. A currency published by only one institution is used, and its figures carry a wider margin of error to say so.

Prices advertised by banks and providers, estimated business prices and negotiated ranges are never inputs to the reference. They are compared against it, and every such figure is labelled published, reported or estimated.

4. The margin of error

Every figure measured against the reference carries a margin of error, because a daily rate is one observation and a conversion may happen hours away from it. The margin reflects:

• how closely the publishing institutions agree on the day;
• whether only one institution published the currency;
• whether the latest rate is from an earlier day, and whether the publisher is weekly rather than daily.

A measured cost smaller than its margin of error is reported as not material, never as a cost or a saving. Headline totals are stated with the margin already taken off, so that they hold when assessed.

5. What is excluded, and when

• A trading venue's prices are used only while the venue passes quality assesses on every pair: a valid response, both sides of the book, a sane spread, and a mid close to the published reference. A venue that has not passed recently does not count.
• A rate older than 36 business hours is marked stale, the margin of error is widened, and every figure that uses it says so.
• A stablecoin's price on one venue is never taken as the currency it tracks: one exchange can print far below the peg while every other market holds it. A route through stablecoins is measured from the currency paid in to the currency paid out. Where a stablecoin swap is measured against its peg, only coins backed by fiat reserves and redeemable one-for-one are valued at par; others carry a wider margin, and synthetic, hedge-backed dollars are not measured as dollars at all.
• A conversion for which no reference exists is returned as unpriced, with the reason. It is never estimated from a nearby currency or date.
• Home rates outside a plausible range for their currency are rejected on arrival rather than corrected.

6. Labels, publishing and route scoring

• Every figure carries its source: measured (from real payments, or from stablecoin conversions read from public chains), published (by a provider), reported (in official data) or estimated (with the basis stated).
• No measured figure is published for a provider on a corridor until it rests on at least five measured payments. Customers are never named; only the corridor and the cost are ever published.
• Routes are scored per corridor and per amount band on four factors: landed cost, speed, reachability (can the recipient receive and cash out locally at that size), and operational overhead (wallets, compliance and audit questions). Price is one of the four, not the only one.
• Providers can correct the data listed about them. Operational specifics, such as venue counts, thresholds and excluded pairs, are kept private so the benchmark cannot be gamed.

7. Independence and conflicts of interest

• SwapEazi does not move, hold or exchange money, and does not trade. It has no position that the reference could favour.
• SwapEazi takes no fees, commissions or referral payments from any bank, payment provider, exchange, pool or token issuer it measures, and no one can pay to change how they are measured or ranked.
• Provider verification, which providers pay for, tests their own claims against the same reference and the same rules as everyone else, and is labelled as such.
• Any interest a director or employee holds in a provider or venue that SwapEazi measures is recorded, and they take no part in decisions about how it is measured.

8. Corrections

When an error is found in the reference or in a figure measured against it, SwapEazi corrects it, records what changed, when and why, and tells customers whose results were affected. Signed records are never altered after signing: a correction is a new record that refers to the one it corrects, so both remain verifiable.

9. Changes to this method

A material change to how the reference is built, how the margin of error is set or what is excluded is published here at least 30 days before it takes effect, with the reason, except where a change is needed at once to correct an error or to stop a compromised input from being used. Each version of this page is dated.

10. Assessing it yourself

Every verification SwapEazi signs can be verified by anyone, free and without an account, against SwapEazi's published public keys. Customers who need to test a figure can ask for the reference rates it was measured against, and any figure on the free verification can be reproduced from the amounts and dates given.

11. The standard we hold ourselves to

This method is organised along the lines of the principles for financial benchmarks published by IOSCO, the international body of securities regulators: clear governance, a stated method, input quality, conflicts managed, corrections recorded and changes announced. SwapEazi is not a regulated benchmark administrator and does not claim to be. It publishes this so that anyone relying on the reference can judge it, and it intends to have the method reviewed independently as its use grows.

Contact

Questions or requests under this policy: hello@swapeazi.io