Fiat conversion and local payment rails drove the cost, not blockchain fees Stablecoin savings appeared in some corridors and disappeared in others Argentina’s cheapest route reflected a curr
Researchers at the Bank of Italy ran 200 real remittances through the stablecoin system and came away unconvinced by the industry’s central selling point. Their mystery shopping study, published in late July 2026 as Paper No. 86 in the bank’s Markets, Infrastructures and Payment Systems series, sent $200 worth of USDC across ten corridors linking Italy with Argentina, Brazil, South Africa, the United Arab Emirates and Japan, then measured what actually landed on the other end. Total cost ran from 0.30% on the cheapest route to almost 9% on the most expensive. That spread is wide enough to sink the claim that digital dollars are reliably cheaper than a bank wire or a money transfer operator. The blockchain leg, the part crypto marketing loves to quote, accounted for almost none of it.
Alberto Di Iorio, Enrica Di Stefano, Michele Mascioli and Giorgio Trebeschi did not model hypothetical fees. They moved real money on March 24 and 26, 2026, buying USDC on exchanges including Binance, Kraken, Ripio and Foxbit, sending it mostly over Ethereum, and cashing out into local currency at the far end. Every stage got timed and priced. They then held the results against the World Bank’s reported global average of 6.65% and against simulated transfers through Wise.
200 × $200 USDC transfers tested 10 corridors both directions 0.30%–8.96% total cost range ~0.4% blockchain’s share of cost 3 of 7 corridors beating Wise <20 min / 1–2 days with vs without instant railsThe authors note that their conclusions are their own and do not necessarily represent the Bank of Italy’s official position.
Sending USDC between two wallets is close to free. On modern Layer-1 and Layer-2 networks the on-chain fee is fractions of a cent, and the study confirmed it barely registers in the total. The expensive parts sit at the two ends. Funding a crypto account with a card or a wire commonly triggers a 1% to 3% processing charge. The heaviest hit comes at foreign exchange: converting dollar-pegged USDC into a thin or volatile currency such as the Argentine peso or the South African rand opens wide bid-ask spreads on local exchanges, and that spread quietly eats the value. Then the recipient still has to withdraw into a domestic bank account, which runs through local clearing systems that charge their own wiring fees. A migrant worker does not want a token. They want cash their family can spend, and every step back toward that cash adds a toll.
The Italy to Argentina corridor posted the study’s lowest cost at 0.30%, which reads like a stablecoin triumph until you look at why. Argentina runs an official exchange rate alongside a much weaker parallel market rate, and the gap between the two, not any blockchain advantage, produced the bargain. The same distortion punished anyone sending money the other way. Argentina to Italy came in at 8.96%, the worst result in the entire sample. One country, two directions, and a nearly thirty-fold difference in cost that had nothing to do with the technology.
Where a country runs a modern instant-payment system, the off-ramp is fast and cheap. Brazil’s Pix let researchers cash out in near real time at minimal cost, which is why Brazilian corridors performed well. Strip that infrastructure away and a stablecoin transfer becomes an expensive detour that still ends in a slow legacy bank wire. The efficiency belongs to the destination country’s own rails. Stablecoins ride on top of them and cannot substitute for them.
The Bank of Italy is not raising these questions in isolation. On July 30, 2026, the Bank for International Settlements reported that Project Agorá had completed real-value testing, with 28 private institutions including JPMorgan, Citi, UBS and Standard Chartered joining five central banks to settle roughly $1 million across 30 live transactions in about 80 seconds. The catch for the crypto sector is what the system runs on. Agorá tokenizes central bank reserves and commercial bank deposits, not public stablecoins, and it settled across six currencies without plugging directly into the banks’ legacy core systems.
Project Agorá · BIS 30 tokenized transfers, ~80 sec, six currencies Project Acacia · RBA Wholesale CBDC with tokenized private assets Bank of Japan CBDC pilot with private payment operators Bank Negara Malaysia Ringgit-backed B2B stablecoins, launched 2026The Bank of Italy made one point that the stablecoin camp will happily amplify: the savings would be “substantially higher” if users simply kept their money on-chain. If a worker sends USDC and the family spends it directly on groceries, utilities or a phone bill through a wallet app, the off-ramp fee falls to zero and the whole friction problem disappears. That is the closed-loop vision, a parallel dollar economy that meets the World Bank’s sub-3% remittance target by never touching local currency at all.
The industry’s own data cuts the other way. Borderless.xyz, which benchmarks live routes, found stablecoin cross-border payments priced below the interbank exchange rate across 260 corridors in the second quarter of 2026, and Western Union launched its own dollar-backed token in May, a sign that even legacy remittance firms expect the rail to hold.
Governments read the same scenario and see a threat. If households in Argentina or South Africa start settling daily transactions in dollar-backed tokens, they dollarize their own economy from the ground up, which strips the central bank of monetary policy control and raises the risk of capital flight. Some of the off-ramp friction the study measured is genuine inefficiency. Some of it is deliberate. Strict KYC and AML rules, local transaction taxes and quiet banking choke points all serve to keep money inside the sovereign currency, and they are unlikely to loosen while stablecoins look like a rival rather than a tool.
The likely outcome is two separate systems. Regulated corridor traffic drifts toward bank-interoperable, multi-currency ledgers of the Agorá type, where tokenized bank money moves under central bank oversight instead of on raw public rails. Fintech aggregators such as Thunes and Yellow Card keep building specialized local API corridors to smooth the last mile in individual countries. At the same time, peer-to-peer stablecoin use holds firm in hyperinflationary economies, where a family will accept high offshore trading costs rather than watch their savings dissolve in the local currency. Europe’s MiCA framework and the US GENIUS Act pull private stablecoins toward formal guardrails, while other blocs accelerate sovereign digital currencies designed to shut those consumer loops down.
Governor Fabio Panetta had already reached a similar verdict in May, arguing that stablecoins can help in selected corridors but offer no universal cure for expensive remittances, and that the real work lies in upgrading domestic instant-payment rails and wiring fast-payment systems together across borders. That prescription puts the burden back on public infrastructure rather than private tokens. With the stablecoin market now near $307 billion and up roughly 16% over the past year, the pressure to resolve that question is only building.
The post Stablecoin Remittances Aren’t Always Cheaper, Bank of Italy Finds appeared first on ETHNews.
