The Bank of Italy published research this week that should not have needed to exist. To find out what it actually costs to send money abroad using a stablecoin, its researchers ran a mystery-shopping experiment. Not a survey of published fee schedules. Not a spreadsheet model. They posed as ordinary customers, pushed real money along the full route, and measured what came out the other end.
The finding: stablecoin remittances are often no cheaper than the traditional rails they were supposed to undercut, once you count exchange fees, foreign exchange spreads and the banking legs at either end.
The finding is interesting. The methodology is the story.
TL;DR
- Bank of Italy researchers used mystery shopping, posing as customers and sending real money, to price stablecoin remittances end to end, and found them often no cheaper than traditional rails once exchange fees, FX spreads and banking legs are counted.
- The near-zero on-chain transfer fee was never the real price. It was simply the only leg of the journey anybody could see.
- When the only way to learn a system’s true cost is an undercover experiment, that system is unverifiable by design, however prominently it advertises a number.
- Crypto gaming publishes the identical shape of claim: RTP percentages, house edge, “zero fees”, headline prize pools. All self-reported, none decomposable, and nobody is mystery-shopping them.
- On-chain gaming turns the all-in cost into a subtraction rather than an investigation: stake, payout, VRF request and gas are events in the same transaction, queryable by anyone, with the platform out of the path.
The advertised number was never the delivered number
The advertised number for a stablecoin transfer is the on-chain fee. On a decent Layer 2 that is fractions of a cent, and it is completely real. It is also, for somebody actually moving money from one country to another, a rounding error inside a much longer journey.
That journey starts in a bank account, passes through an exchange to buy the stablecoin, crosses the chain, hits another venue to sell it, and lands in a different bank account. Four of those five legs are priced by intermediaries, and the two most expensive are usually not labelled as fees at all. They are spreads: the gap between the rate you were quoted and the rate the venue actually got.
A spread is not hidden in the sense of being concealed. It is worse than that. It is a cost with no line item, so there is nothing to disclose, nothing to compare, and no number to put on a marketing page. The only way to measure it is to complete the transaction and subtract.
Which is precisely why a central bank had to go shopping.
What it means when measurement requires an experiment
Strip away the remittance specifics and a general rule falls out. If the only way to establish what a system charges is to run a controlled experiment through it, the system is not transparent. It might be honest. It might be a genuinely good deal for some corridors. But its honesty is unverifiable by anyone who lacks a research budget and the patience to send real money through a route they suspect is overpriced.
Nobody in that corridor was lying. The transfer fee really is near zero. The claim was true about the one leg the claimant chose to talk about, and silent about the four that determine the answer. That is not fraud. It is the ordinary behaviour of any system where the interested party picks which number gets published.
And note what the Bank of Italy’s experiment could not do. It priced the corridor on the days it ran. It cannot tell you what the spread was last Tuesday, or what it will be next month, or whether your particular transfer landed on the good side of the quote. Even the undercover study produces a snapshot, because a snapshot is all an outsider can ever get from a system that does not publish its rows.
Crypto gaming publishes the same shape of number
Every crypto game and crypto casino you can name publishes headline figures of exactly this class. Return to player percentage. House edge. “Zero fees.” Prize pool totals. Player counts. Every one of them is an aggregate, self-reported by the party with the strongest interest in its value, covering a window that party selected, computed from rows that are never published.
Gaming is structurally worse than the remittance corridor, and it is worth being precise about why. A remittance has a beginning and an end the customer can see. Money left this account, money arrived in that one. The customer knows the delivered number even if they cannot decompose it, and enough customers comparing notes eventually produces pressure. A game outcome has no such anchor. You do not know what the fair result would have been, so you cannot compute the gap between advertised and delivered from one play, or a thousand. There is no arrival figure to compare against the quote.
Consider what a published RTP can quietly contain: which games it aggregates, the window it covers, bonus wagering requirements, maximum win caps, withdrawal minimums, token conversion at operator-set rates, and network fee markup on the payout leg. A platform can advertise 97% and be arithmetically correct about a number that describes almost nothing a player will experience. “Zero fees” is gaming’s near-zero transfer fee exactly: true about one leg, silent about the rest.
The difference is that a central bank went and measured the corridor. No equivalent institution is mystery-shopping crypto games, and the affiliate review sites that claim to are paid by the operators they rank.
On-chain makes it arithmetic instead of an investigation
Here is the thing worth taking from the Italian study. The problem was never dishonesty. It was that the cost was distributed across legs owned by different parties, none of whom published the leg they controlled. Fix the publication problem and the measurement problem dissolves.
When a game settles on-chain, the stake, the Chainlink VRF request, the gas and the payout are all events in the same transaction, written to a ledger the operator does not own. The all-in cost of a Satoshie coinflip is not a claim to evaluate. It is a subtraction: what left your wallet, what came back, everything in between itemised on BaseScan. Compute it over a window you choose, not one the platform chose. Compute it over every player, not a sample. Run it a year from now against the same immutable record. The platform cannot block the query, cannot see that it ran, and cannot publish a different aggregate to a different audience, because there is only one set of rows and everybody reads the same ones.
The randomness gets the same treatment, which is the part most platforms skip. VRF outcomes carry a cryptographic proof verified on-chain before the result is delivered, so fairness is not a percentage the operator reports at quarter end. It is a check that either passed or did not, on every single draw.
The honest caveat
The Bank of Italy’s finding applies to on-chain gaming’s front door too, and pretending otherwise would be exactly the selective quoting this post is complaining about. Buying the stablecoin you play with costs what it costs. The exchange fee is real, the FX spread is real, and no smart contract fixes either. Anyone who tells you on-chain gaming is free is quoting one leg at you.
The claim is narrower and it is the one that matters. From the moment funds are on-chain, every cost and every outcome is itemised in the same public record, so the all-in number is arithmetic any player can do rather than an experiment somebody has to fund. You still pay to get in. You no longer have to take anyone’s word for what happens after.
Three questions for any platform
- Could you calculate the all-in cost of your last hundred plays without asking the platform for anything? If the answer requires their dashboard, their export or their support desk, you are being handed a summary, not evidence.
- Who chose the window your RTP describes? A percentage over an unstated period, computed from unpublished rows, is a marketing figure wearing a statistic’s clothes.
- What is not in the headline number? Every advertised rate in crypto has legs it excludes. The trustworthy platforms will tell you which ones. The rest will change the subject.
The Bank of Italy had to go undercover to price a wire transfer. That is a reasonable thing for a central bank to do about the global remittance market. It is an absurd thing to require of somebody who just wants to know whether a coin flip is fair, and on-chain gaming is the only corner of this industry where nobody has to.
📷 Photo by Am (@mahathirr) on Unsplash
