Brazil Brought Stablecoins Into the FX Market. Then It Said Where They Cannot Go.

It reads as a crackdown. It is the second step of one policy, and the first step made it possible.

Share
Brazil Brought Stablecoins Into the FX Market. Then It Said Where They Cannot Go.

Resolution 561 reads as a reversal. It is the second half of one policy, and the first half is what made it possible.

On 30 April the Central Bank of Brazil issued Resolution 561, and the reaction in the sector was close to unanimous: Brazil is cracking down on stablecoins.

Five months earlier the same central bank brought virtual assets into the foreign exchange market, and the reaction then was equally unanimous in the opposite direction: Brazil is embracing crypto.

Both readings are wrong, and reading the two resolutions separately is what produces the error. They are one policy executed in two steps, and the second step was only possible because of the first.

What 561 actually says

Resolution 561 amends Resolution 277, which governs the eFX service — the arrangement for small-value, high-volume international payments. Foreign stores selling to Brazilian consumers. Marketplaces remitting funds abroad. Digital services contracted overseas. This is the retail cross-border channel, not one-off currency exchange.

The new article 50 governs how an eFX provider settles with a foreign counterparty. Settlement is permitted through a foreign exchange operation, or through movement in a non-resident's real-denominated account in Brazil. And then the operative clause: sendo vedado o uso de ativos virtuais — the use of virtual assets is prohibited.

So virtual assets may not be the settlement medium on the foreign leg of an eFX payment.

Three other provisions travel with it. Purchases of goods and services through a provider's digital payment solution are capped at US$10,000, or the equivalent, where there is no integration with an e-commerce platform. Providers not already listed must request central bank authorisation by 31 May 2027, operating in reduced scope in the meantime under the same cap and the same prohibition, with thirty days to stop if they miss the deadline or are refused. Institutions already authorised must register the modality in Unicad by 30 October 2026.

The resolution takes effect on 1 October 2026.

Note what it does not say. It does not ban stablecoins in Brazil. Virtual asset activity outside this particular arrangement remains governed by its own rules — the ones issued in November. What 561 does is specify which settlement instrument is permitted in one defined channel.

Why this is not a reversal

Here is the sequence, and the logic becomes obvious once it is laid out.

November 2025. Bring virtual asset operations that cross the border inside the foreign exchange market. Define who may do them, require authorisation, impose reporting, assign purpose codes. Make the flow visible and supervised.

April 2026. Within that now-supervised market, specify that one particular channel — retail-scale eFX — may not settle its foreign leg in virtual assets.

The first step is what makes the second step possible, and it is also what makes the second step narrow. A regulator that had never brought these flows inside a perimeter would have had only two instruments available: permit everything, or prohibit the category. Both are blunt and both are evadable.

A regulator that has the flow inside a supervised market can do something far more precise: leave the category permitted, and rule on which settlement medium may be used in which arrangement.

That is regulating by channel rather than by category, and it is a meaningfully better technique. A category ban invites relabelling — the same economic activity reappears under a different description. A channel rule attaches to the arrangement, which is defined by what it does rather than by what the instrument is called.

I think this is the most underappreciated thing a central bank has done in this region in two years, and it has been reported as a crackdown.

The trade Brazil made, named honestly

I want to be clear that this has a real cost, because the people celebrating it as prudence are skipping that part.

Small-value, high-volume cross-border payment is precisely the use case where stablecoin settlement economics are strongest. Pre-funding is expensive, correspondent chains are slow, weekend settlement does not exist, and the fixed costs of a conventional FX operation sit badly on a forty-dollar purchase. A settlement leg that clears in seconds, around the clock, with no pre-funded nostro, is genuinely better technology for exactly this flow.

Brazil has removed that option from exactly the channel it suits best. That is not an accident and it should not be glossed over.

So why would a central bank do it?

The honest answer, I think, is consumer protection at the far end of a chain nobody can see. A Brazilian buying from a foreign store is a retail consumer. If the settlement leg runs on a token held by an intermediary somewhere offshore, that consumer's recourse depends on an asset they never chose, a custodian they cannot assess, and an insolvency regime in a jurisdiction they have never heard of. Requiring settlement through an FX operation or a non-resident real account puts the leg back onto instruments with established legal treatment and a known creditor hierarchy.

There is also a supervisory argument. Brazil's purpose codes and reporting obligations for virtual asset flows do not take effect until 4 May 2026. The eFX prohibition takes effect on 1 October 2026. In other words, the central bank is closing this particular channel at almost exactly the point where it would otherwise have had to supervise it at scale — and before it has any of the data the November resolutions were designed to produce.

That is the part I would push on if I had the floor. The sequence was: make it visible, then restrict it before you can see it. I would rather have seen a year of purpose-code data and then a rule. A regulator may reasonably answer that the risk was too immediate to wait, and that is an argument, but it is an argument rather than an answer.

Two caps, one instinct

There is a pattern across both resolutions worth isolating, because it is the most portable idea in either of them.

In November: international payments with virtual assets capped at US$100,000 where the counterparty is not an FX-authorised institution.

In April: purchases through a provider's digital payment solution capped at US$10,000 where there is no integration with an e-commerce platform.

Neither is a prohibition. Both say the same thing in different registers: you may do this at retail scale through a lighter channel, and above that you must come through the supervised door.

That is a genuinely good regulatory instrument and it is underused in this region. It does not require the regulator to decide whether a technology is good or bad. It requires only a judgment about the scale at which supervision becomes necessary — which is a question supervisors are actually equipped to answer.

Most regional regulation of new payment instruments is binary: authorised or not, permitted or prohibited. A calibrated threshold with a channel attached does more work and generates less evasion than either.

From a Panamá desk

The thing to take from this is not the prohibition. It is the category.

Brazil has a defined, named, supervised arrangement called eFX for small-value, high-volume international payments. Foreign merchants selling to residents. Marketplaces remitting abroad. Digital subscriptions bought overseas. It has its own rules, its own authorised providers, its own caps and now its own settlement requirements.

Most of this region has no such category. The same economic activity happens here — a Panamanian buying from a foreign store, a local marketplace paying a supplier abroad, somebody subscribing to a foreign service — and it runs through whatever instrument is nearest, mostly cards, with no particular arrangement defined around it.

That absence has a cost that only shows up when you want to do something. You cannot write a settlement rule for a channel you have not defined. You cannot cap it, supervise it, or require anything of it. You can only reach it through the general rules that apply to whoever happens to be carrying it.

So the useful question for a market like ours is not whether to permit stablecoin settlement. It is whether small-value cross-border digital payment is a defined arrangement here at all. If it is not, then every future question about it — settlement medium, consumer recourse, caps, reporting — has no object to attach to.

Defining the category is unglamorous, costs almost nothing, and is the precondition for every decision that comes after. Brazil did it first and is now three moves ahead, not because it decided more wisely but because it had something to decide about.

The second regional point is the one I raised about the November resolutions, and April has made it sharper. Brazil's text will be copied. A US$10,000 cap and a prohibition on virtual asset settlement will appear in draft rules elsewhere in this region within two years, lifted from a framework that makes sense because of the eFX category and the purpose codes sitting underneath it. Copied without those, the same words are a prohibition attached to nothing, in a market with no way to see whether it is being complied with.

My judgment

Read together, November and April are a central bank building an instrument panel and then adjusting one dial. Read separately, they look like enthusiasm followed by a crackdown. The sector has mostly read them separately.

The technique is the lesson, and it generalises well beyond stablecoins: bring the flow inside a supervised perimeter first, define the arrangement, then make granular rules about instruments within it. That sequence gives a regulator options that neither prohibition nor permission ever provides.

Where I would push back is on the timing. The data the November resolutions were designed to generate starts arriving on 4 May. The restriction takes effect on 1 October. Five months is not enough to learn anything from purpose codes, which means this rule was written from judgment rather than from the measurement the same central bank had just commissioned. That may well have been the right call. It is still worth noticing that the instrument panel was built and then not consulted.

And my prediction, so it can be checked: there is a third step coming. Once the reporting data exists, expect differential treatment by corridor or by purpose — not a single rule for all virtual asset settlement, but different treatment for a remittance corridor than for a marketplace payout. That is what the purpose codes are for, and a regulator that went to the trouble of assigning them did not do it to leave the rule binary.

Three questions for the people who own this

1. In your market, is small-value cross-border digital payment a defined arrangement with its own rules, or does it run under whatever licence the carrier happens to hold? If the second, what would it take to define it, and who would have to want to?

2. Brazil capped one channel at US$10,000 and another at US$100,000, each tied to whether a supervised counterparty is involved. Where would those numbers sit in an economy the size of yours — and has anyone in your market ever argued for a calibrated threshold rather than a yes or a no?

3. If your regulator restricted stablecoin settlement for retail cross-border payments next year, which of your institution's flows would have to be redesigned, and how long would that take? If the answer is that none of them use it, are you certain, or is it that nobody has looked?

Reply to this email with a line or two. The sharpest answers will shape the next issue.

Sources

FCM Law, analysis of Resolução BCB nº 561 of 30 April 2026 — the amendment of Resolução BCB nº 277/2022 governing the eFX service; the new article 50 permitting settlement with foreign counterparties only through a foreign exchange operation or movement in a non-resident's real-denominated account in Brazil, with the express prohibition sendo vedado o uso de ativos virtuais; the US$10,000 cap under article 49, I, "b" on purchases of goods and services through a provider's digital payment solution absent integration with an e-commerce platform; the authorisation requirement by 31 May 2027 under article 56-B for providers not listed in article 49, § 2º, with reduced scope in the interim and thirty days to cease if they miss the deadline or are refused; the Unicad registration deadline of 30 October 2026 for institutions already authorised; and the effective date of 1 October 2026 under article 4º. Link

Lefosse Advogados, alert on Resolutions BCB 519, 520 and 521 of 10 November 2025 — the November framework referenced throughout, including the US$100,000 cap where the counterparty is not an FX-authorised institution, and the reporting obligations and purpose codes effective 4 May 2026. Link

Both are law firm analyses, not the official texts. I have not read Resolution 561 or Resolution 277 in the original. The reading of November and April as two steps of one policy, the argument about regulating by channel rather than by category, and the criticism of the timing relative to the purpose-code data are all mine.