FATF’s travel rule made each crypto transfer depend on a counterparty you could not see
FATF’s June 2019 standard said countries should require virtual asset providers to send originator and beneficiary details with transfers. That made each firm’s compliance depend on finding, reaching and trusting the institution at the other end, which no single firm could guarantee.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
FATF’s June 2019 standard said countries should require virtual asset providers to send originator and beneficiary details with transfers. That made each firm’s compliance depend on finding, reaching and trusting the institution at the other end, which no single firm could guarantee.
- The week On Friday, June 21, 2019, at its plenary in Orlando, Florida, the Financial Action Task Force (FATF) adopted an Interpretive Note to Recommendation 15 on New Technologies, known as INR.
- What the standard required Recommendation 16 was written for bank wire transfers.
- Part one: finding the institution on the other side A bank wire names the beneficiary bank.
- Part two: building the message channel Banks, as we see it, inherited their message channel: long-established interbank networks and standards already carried wire data.
- Part three: holding data that someone else sent Under the travel rule, a beneficiary institution receives personal data about someone who is not its customer: the originator’s name, account number and an address or identity number or date and place of birth.
- Part four: deciding what to do when the other side is silent Every institution subject to the rule needed a policy for the cases in which it could not comply in full.
- The strongest objections The objection holds that the direct burden fell on exchanges and custodians, not banks, and it has some force.
- How it differs by institution For US virtual asset providers registered with FinCEN as money services businesses, FinCEN’s May 2019 guidance said qualifying transfers may fall within the Funds Travel Rule.
- With hindsight More than seven years on, the June 2019 standard reads as the start of a long implementation rather than a settled rule.
The week
On Friday, June 21, 2019, at its plenary in Orlando, Florida, the Financial Action Task Force (FATF) adopted an Interpretive Note to Recommendation 15 on New Technologies, known as INR. 15. The United Nations Security Council had welcomed FATF’s work on virtual assets in its Resolution 2462 of March 28, 2019. FATF sets the international standards for anti-money laundering and countering the financing of terrorism (AML/CFT), and its member countries include the United States. The note built on changes FATF made in October 2018, when it amended Recommendation 15 and added two new definitions to its glossary to clarify that its standards applied to virtual assets and virtual asset service providers, or VASPs.
FATF’s public statement that day described the note as binding measures for both countries and providers. Countries had to assess and mitigate the risks of virtual asset activity, license or register providers, and supervise or monitor them through competent national authorities. FATF said countries would not be permitted to rely on a self-regulatory body for that supervision. Providers had to apply the full range of preventive measures: customer due diligence, record-keeping, suspicious transaction reporting and screening of all transactions against targeted financial sanctions. The statement also noted that some countries might decide to prohibit virtual asset activity altogether.
The same day FATF published updated Guidance for a Risk-Based Approach to Virtual Assets and Virtual Asset Service Providers, building on its 2015 paper. The guidance defined a VASP by five activities carried out as a business for another person: exchange between virtual assets and fiat currencies, exchange between forms of virtual assets, transfer, safekeeping or administration, and financial services related to an issuer’s offer or sale of a virtual asset. It confirmed a USD/EUR 1,000 threshold above which providers must conduct customer due diligence on occasional transactions.
The provision with the largest operational consequences, in our reading, applied Recommendation 16, the wire transfer rule, to virtual asset transfers. The guidance said countries should treat all such transfers as cross-border qualifying wire transfers. The ordering institution had to obtain and hold required and accurate originator information and required beneficiary information, and submit it to the beneficiary institution. The required items were the originator’s name; the originator’s account number where an account such as a wallet is used; an address, national identity number, customer identification number or date and place of birth; the beneficiary’s name; and the beneficiary’s account number. The information had to move “immediately and securely”, but it did not need to travel on the blockchain itself. Countries could adopt a de minimis threshold of USD/EUR 1,000 for these transfers.
FATF said it would monitor implementation and conduct a 12-month review in June 2020. Its statement added that G20 finance ministers and central bank governors, meeting in Fukuoka, had welcomed the action. In the United States the domestic ground had been laid the month before. On May 9, 2019, FinCEN issued guidance stating that transactions in convertible virtual currency qualify as transmittals of funds and may fall within the Funds Travel Rule, which reaches transmittals of $3,000 or more.
FATF’s June 2019 package required countries to license or register crypto exchanges and other providers and to supervise them, and required the providers to apply the same preventive measures as other entities subject to AML/CFT regulation. That mattered. For the technology and compliance leaders of banks, broker-dealers and fintechs, though, the provision with the longest tail was the application of Recommendation 16. It turned every transfer between two regulated providers into a data exchange between them. A bank could have perfect controls over its own customers and still fail the rule, because the rule depended on what happened at the other end.
Our argument is that the travel rule converted a payments compliance question into a counterparty and messaging problem that no single institution could solve on its own. Compliance turned on four things, in order. First, finding out whether there was a regulated institution on the other side of a transfer at all. Second, having a channel to send it the required data at the moment of the transfer. Third, holding and protecting data that another institution sent. Fourth, deciding what to do when the other side was absent, unregistered or silent. We take each in turn after an account of what the standard required. A sixth section tests the argument against the strongest objections, and the last looks at how the problem differed by type of institution.
One point of precision before we start. FATF’s standards reach firms through national law: the guidance is written as what countries should require. A US institution’s obligations came from US law and regulation: the Bank Secrecy Act, FinCEN’s rules, and its prudential regulator’s expectations. FATF’s standard shaped what US authorities and foreign counterparts would expect, and it shaped what other countries would require of the firms on the other side of US transfers. Institutions should check with counsel which rules apply to which of their activities.
What the standard required
A wire transfer rule applied to a different kind of transfer
Recommendation 16 was written for bank wire transfers. The guidance explained its purpose as preventing terrorists and other criminals from having unfettered access to electronically facilitated funds transfers, and detecting misuse when it occurs. In a bank wire, the originator and beneficiary information travels in the payment message. The ordering bank fills in fields, the intermediary passes them on, and the beneficiary bank receives them with the funds. The messaging network and the payment are, for practical purposes, one thing.
A blockchain transfer is different. The ledger records a movement of value from one address to another. It has no field for a customer’s name or address. Anyone can generate an address, and an address says nothing on its face about whether a regulated provider controls it. The guidance recognized this and did not ask for names on the chain. It said the required information need not be communicated as part of the transfer on the distributed ledger. Submitting it to the beneficiary provider could be an entirely distinct process. Any technology was acceptable, provided it let both sides meet the requirements.
That flexibility moved the problem rather than removing it. If the data does not ride with the value, someone must build a second path for the data. The second path must find the right counterparty, deliver the data at the right moment and match it to the right transfer. The guidance left the means to the industry and named no common channel for these steps, as bank wires have.
“Immediately” and “securely”
The guidance defined both words. “Immediately” meant that providers should submit the required information simultaneously or concurrently with the transfer itself. “Securely” meant that providers should protect the integrity and availability of the information, to support recordkeeping and use by the receiving institution, and protect it from unauthorized disclosure. The guidance added that the security language was not meant to impede the objectives of the rule.
The timing requirement is the demanding one. The guidance tied it to the cross-border nature, global reach and transaction speed of virtual assets. We would also plan on the basis that a confirmed ledger transfer cannot be recalled by the sender, so a firm that learns after settlement that the beneficiary is a sanctioned person may have no way to pull the funds back. So the data has to arrive, and be screened, before or as the value moves. That puts the data exchange in the critical path of the transfer, with all the availability and latency engineering that implies.
Obligations on both ends and in the middle
The guidance assigned duties to each role. The ordering institution must obtain and hold required and accurate originator information and required beneficiary information, and send it on. The beneficiary institution must obtain and hold required, though not necessarily accurate, originator information and required and accurate beneficiary information. Both must make the information available to authorities on request. Both must screen against targeted financial sanctions and take freezing actions. Intermediary institutions in a chain of transfers must pass the information along and keep records.
The guidance also addressed transfers with a regulated institution on only one end. When a provider sends value for its customer to an individual using an unhosted wallet, FATF did not expect the provider to send the data to that individual. When a provider receives value from an unhosted wallet, it should obtain the required originator information from its own customer. That carve-out became one of the hardest edges of the rule, as we discuss below.
Supervision without self-regulation
The statement’s line on self-regulatory bodies deserves attention from anyone who deals with counterparties. FATF said countries would not be permitted to rely on a self-regulatory body to supervise or monitor providers. Supervision had to come from competent national authorities with powers to inspect, compel information and impose sanctions, as the guidance put it. That choice matters for the travel rule because it determines what a sending institution can rely on. A license or registration issued by a public supervisor is a fact an institution can check and cite. Membership of an industry association is a claim the institution must test itself.
The guidance also stressed international co-operation between supervisors, given the cross-border nature of providers’ activities. For a US institution, that co-operation was out of its hands. What it could control was how it weighed a counterparty’s supervisory status. A provider supervised by an authority that had implemented the standard could be treated differently from one supervised nowhere. Writing that distinction into policy, with named jurisdictions and a review date, turned FATF’s country-level standard into a firm-level decision rule.
The US version
FinCEN’s May 2019 guidance reached a similar result through existing law. It said a transmittal order involving convertible virtual currency is an instruction to pay a determinable amount of money, so such transactions qualify as transmittals of funds. Under the Funds Travel Rule, a transmittal of $3,000 or more, or its equivalent in virtual currency, may trigger requirements on a money transmitter acting for the transmittor, for the recipient or as an intermediary. The money transmitter must obtain or provide the required information before or at the time of the transmittal of value.
FinCEN also anticipated the engineering problem. The parties need not use the same system or protocol for the value and for the information. If a protocol cannot carry the information, it may go in a separate message. And in a footnote, FinCEN said that a person who chooses a transaction system that makes compliance difficult may not invoke that difficulty to justify non-compliance. For institutions weighing whether to wait for industry tooling, that footnote was the answer.
Part one: finding the institution on the other side
An address is not a counterparty
A bank wire names the beneficiary bank. The sending bank knows, before it sends, which institution will receive the funds, and it can look that institution up. A virtual asset transfer names an address. The sending provider must work out whether the address belongs to a customer of another provider, and if so, which one. Only then can it know whether the travel rule applies and where to send the data.
We are not aware of any authoritative directory in June 2019 that mapped addresses to providers. Providers could ask their customers who controlled the destination. They could use analytics that cluster addresses and attribute them to known services. They could rely on bilateral arrangements with providers they dealt with regularly. Each method had error rates and gaps, and none of them was a regulatory register. The rule assumed a fact, the identity of the beneficiary institution, that the system did not readily supply.
Due diligence on the counterparty institution
Knowing that a provider sits at the other end is the start. The sending institution also needs to know whether that provider is licensed or registered, where, and whether it can be trusted with customer data. FATF’s standard required countries to license or register VASPs. Where a country had not yet done so, a sending institution had no register to check. It had to perform something closer to correspondent due diligence on each counterparty provider: who owns it, where it is supervised, what its AML program looks like, and how it protects data. The guidance pointed the same way: where relationships between providers had characteristics similar to cross-border correspondent banking, countries should apply Recommendation 13’s correspondent measures to the providers that develop them.
For a bank, that is familiar work. Correspondent banking due diligence exists for the same reason. The difference was scale and speed. A bank has a bounded list of correspondent relationships, reviewed on a cycle. A provider serving retail customers might see transfers to hundreds of other services, some of them new, some of them offshore and some of them gone within a year. The due diligence had to be continuous and partly automated, with a policy for transfers to providers not yet reviewed.
Unhosted wallets
The unhosted wallet carve-out meant that a large class of transfers had a regulated institution on one end only. For those, the obligation fell on the one institution present, and it reduced to collecting information from its own customer. That information was self-reported. A customer withdrawing to a wallet could say the wallet was theirs, or a friend’s, or a merchant’s. The institution had limited means to verify the claim.
From a risk perspective, the carve-out created an incentive. Value that a criminal wanted to move without a data trail could route through unhosted wallets between regulated hops. Institutions that treated every unhosted transfer the same way would either block legitimate activity or wave through risky activity. The practical answer was risk-based treatment: which wallets had been seen before, what analytics said about their history, how large the transfer was, and whether the customer’s explanation fit their profile. Each of those is a judgment the institution must be able to defend to an examiner.
Sanctions screening on both ends
The guidance required both ordering and beneficiary institutions to screen against targeted financial sanctions and to take freezing actions. In a bank wire, the beneficiary bank screens the originator data that arrives with the payment. For a virtual asset transfer, the beneficiary institution can screen only what it receives on the separate channel, and only if it arrives in time. If the data comes after the value, the institution has credited funds before it knows who sent them.
That sequencing problem has an operational answer, though an imperfect one. An institution can hold inbound value in a suspense state until data arrives and clears screening, then release it to the customer. That protects the institution and slows the customer. It requires the ledger and customer-facing systems to show a pending state, with clear time limits and a path for manual review. It also requires a decision on what happens if the data never comes. Each of those is a system change, and each should be specified before the first transfer, not discovered on it.
Part two: building the message channel
Why the channel did not exist
Banks, as we see it, inherited their message channel: long-established interbank networks and standards already carried wire data. Virtual asset providers in 2019 had no equivalent that we know of. More than a year later, commenters on the 2020 US proposal still pointed to the lack of an international messaging standard for virtual currency transactions, according to the agencies’ 2026 withdrawal notice. FATF said any technology was acceptable, which meant the industry had to choose, build and agree on one. Until it did, two providers that wanted to comply had no common way to talk.
That is the heart of our argument. The obligation was individual, but the means of meeting it was collective. A single provider could build an excellent outbound messaging service and still have no one to send to. The same was true in reverse. Compliance depended on a network effect, and network effects take time.
What a workable channel has to do
A channel that meets FATF’s description has to do several distinct things. It must discover the counterparty: given an address, find the provider and the endpoint to send to. It must authenticate both parties, so that a provider does not send customer data to an impostor. It must carry a defined set of data fields in a format both sides understand. It must deliver the message before or as the transfer settles, and confirm receipt. It must let the beneficiary institution reject a transfer, or flag a mismatch, before value moves when possible. And it must leave records both sides can produce to authorities.
Each of those functions has an owner inside an institution. Discovery and authentication belong to security and architecture. The data format belongs to the compliance data team. Timing belongs to whoever runs the transfer engine, because the message has to sit inside the transaction flow. Recordkeeping belongs to records management and legal. We think obligations that cut across that many owners need one named accountable executive, or each team may optimize its own piece while the seams fail.
Data standards
A channel also needs a shared data format. If one provider sends a name as a single string and another expects separate given and family names, screening and matching suffer. If one sends an address and another expects a date and place of birth, the beneficiary institution must accept whichever of the permitted identifiers arrives. The guidance listed the information required but did not prescribe a format. Institutions therefore had to map each counterparty’s format to their own, test the mapping, and decide how to treat records that did not map cleanly. That mapping work is invisible to customers and easy to underfund, and it determines whether screening on inbound data works at all.
Interoperability between channels
If several competing channels emerge, a provider must connect to each, or accept that transfers to counterparties on other channels cannot meet the rule. Interoperability between channels then becomes its own project. For a bank entering the space, the choice of channel is a procurement decision with lock-in risk, much like the choice of a payments processor. The contract should address data handling, availability commitments, audit rights and exit, because the channel will carry regulated personal data and sit in the critical path of customer transfers.
Testing the channel
A channel in the critical path needs testing like any payment system. Before go-live, an institution would want end-to-end tests with each counterparty: a transfer sent with complete data, one with missing fields, one to a wrong address, one that times out. It would want to know what the customer sees in each case and what the compliance queue receives. After go-live, it needs measures it can watch: the share of qualifying transfers with data sent, the share with data received, the median delay between data and value, and the number of exceptions open past their deadline.
Those measures do two jobs. They tell operations where the channel is failing. And they give the compliance officer an honest picture to report to the board and to examiners. A program that reports only that a channel exists, with no coverage figures, tells a reader nothing about whether the rule is being met.
The sunrise problem
FATF’s standard applied to countries, and countries adopted it at different speeds. A provider in a jurisdiction that had implemented the rule might send to a provider in one that had not. The receiving provider had no legal obligation to accept or protect the data, and perhaps no means to receive it. The sending provider still had its own obligation. This mismatch, which we call here the sunrise problem, was built into a standard that relied on national implementation. It was foreseeable on the day the note was adopted.
Part three: holding data that someone else sent
Inbound data is a liability as well as an obligation
Under the travel rule, a beneficiary institution receives personal data about someone who is not its customer: the originator’s name, account number and an address or identity number or date and place of birth. It must hold that data, screen it and produce it on request. It did not collect the data, cannot easily verify it and did not agree with the originator on how it would be used. FATF’s statement pointed to the tension when it said the preventive measures include coordination with relevant authorities to ensure AML/CFT requirements are compatible with data protection and privacy rules.
For a US institution, the practical questions are concrete. The procedure should say where inbound originator data lives, and whether it sits inside the same security boundary as customer data, with the same access controls, encryption and monitoring. It should say how long the data is kept and under which records schedule. It should name who may query it and confirm that access is logged. And it should say how data is corrected when a provider sends something wrong, or about the wrong person. Each answer will be tested if the data is ever breached or subpoenaed.
Retention and production to authorities
The guidance required both institutions to make the required information available to appropriate authorities on request. That means the data must be retrievable by transfer, by customer and by counterparty, long after the transfer settled. It must also be linked to the on-chain record, so that an investigator can move from a ledger entry to the institution’s data and back. In a bank wire, the payment reference ties the two together. For virtual assets, the institution must create and keep that link itself.
The link is easy to lose. Data may arrive on one system, the transfer may be recorded on another, and the customer record may sit in a third. If the identifiers do not match, a production request becomes a manual reconstruction. Designing the data model so that every transfer carries a stable internal identifier, cross-referenced to the transaction hash and to the travel rule message, is cheap at the start and expensive to retrofit.
Accuracy runs one way
The guidance placed the accuracy duty asymmetrically. The ordering institution must hold accurate originator data; the beneficiary institution must hold required originator data, but not necessarily accurate. The beneficiary institution must hold accurate data about its own customer. That allocation makes sense: each side verifies whom it knows. It also means that a beneficiary institution’s sanctions screening on originator data is only as good as the sending institution’s onboarding. A provider receiving transfers from weakly supervised counterparties is screening weakly verified names.
Security in transit and at rest
The “securely” requirement made the channel a security system. Messages carrying customer identity data between institutions are an attractive target. An attacker who could read them would learn which customers moved value, to whom and when. An attacker who could forge them could make a transfer appear to come from a clean originator. Institutions needed mutual authentication, encryption in transit, integrity checks and an audit trail. Those controls are routine in bank messaging, but they had to be built and assessed for the new channels.
The security review should also cover the counterparty’s handling of outbound data. When a bank sends its customer’s name and identifier to another provider, that provider’s security becomes part of the bank’s exposure. A customer whose data leaks from a counterparty will hold the sending institution responsible. Counterparty due diligence therefore needs a data protection component alongside the AML one.
Part four: deciding what to do when the other side is silent
The decision rules
Every institution subject to the rule needed a policy for the cases in which it could not comply in full. The counterparty might be unidentifiable, unlicensed, in a jurisdiction without the rule, unreachable on any channel, or might send incomplete data. Each case calls for a decision: send anyway and record the gap, delay, reject, return, or escalate for review. For inbound transfers, the beneficiary institution must decide whether to credit the customer when originator data is missing.
Those decisions have direct customer consequences. A rejected withdrawal is a complaint. A held deposit is a call to customer service. A pattern of incomplete data from one counterparty may be a reason to stop dealing with it, which may cut customers off from a service they rely on. The policy must therefore be owned jointly by compliance, operations and the business line, and approved at a level that can accept the commercial cost.
Telling customers what will happen
Customers do not see the travel rule; they see a withdrawal that did not go through or a deposit that is pending. An institution applying the rule needs customer-facing language that explains, in brief and direct sentences, why it may ask who controls a destination wallet, why some transfers are held, and what the customer can do. Agents need scripts and escalation paths. Complaints about held or rejected transfers should be tracked as a category, because they are an early signal of a counterparty that has stopped sending data or a rule that is set too tight.
The customer’s own statements also become records. When a customer says a wallet is theirs, or names the person who controls it, that statement is part of the institution’s evidence. It should be captured in a structured form, timestamped and kept with the transfer, so that a later review can see what the institution was told and when.
Missing data as a risk signal
We would treat missing data as a signal in its own right. A counterparty that never sends data, or a customer who repeatedly receives from unidentifiable sources, may warrant enhanced review or a suspicious activity report. The institution needs monitoring rules that use the presence, absence and quality of travel rule data as inputs.
Evidence for examiners
Because the rule depends on counterparties, an examiner will want to see how the institution handled the gaps as well as the clean cases. The evidence file should show the counterparty inventory, the due diligence on each, the channels used, the share of transfers for which data was sent and received, the decision taken on each exception, and the reviews that followed. We would expect that file to be the thing that distinguishes a defensible program from an undefended one, because no program in 2019 could show full coverage.
The strongest objections
This was a crypto industry problem, not a bank problem
The objection holds that the direct burden fell on exchanges and custodians, not banks, and it has some force. But banks were exposed in three ways. They banked VASPs, and needed to understand their customers’ compliance. They faced counterparties that might later become customers or partners. And they could expect to offer custody or transfer services later, under a regime shaped by these choices. A bank that ignored the June 2019 standard would face it later with less time.
The technology would solve it
A second objection holds that industry messaging solutions would emerge and make the problem routine. FATF’s statement had said it would set up a Contact Group to engage industry and monitor industry-led compliance efforts, so a technical answer was expected. But technology could not solve the parts of the problem that were legal and jurisdictional. A message channel cannot make an unlicensed provider licensed, or make a country implement the rule. As long as national implementation was uneven, an institution would face counterparties it could not lawfully or safely exchange data with. That residue is a policy problem, and it needs a policy answer from each institution.
The thresholds made it marginal
A third objection holds that thresholds, $3,000 in US rules and a permitted USD/EUR 1,000 de minimis under FATF’s guidance, would leave small transfers outside the rule. That is true for the data exchange obligation. It does not remove sanctions screening, monitoring or suspicious activity reporting below the threshold. And thresholds create their own risk: customers may split transfers to stay under them, and monitoring has to catch that. A program that treats the threshold as a boundary of concern, rather than a boundary of one obligation, misreads it.
The rule would push activity offshore
A fourth objection holds that strict data exchange rules would drive customers and providers to jurisdictions without them, leaving compliant institutions with less business and no less risk. That risk is real, and FATF’s 2025 reporting on offshore providers, discussed below, suggests it had not gone away. But it argues for a counterparty policy, not against one. An institution cannot stop activity moving offshore. It can decide whether to transact with providers there, on what terms, and with what monitoring. Choosing not to decide is itself a choice, and an undocumented one.
Where the argument is weakest
Our argument assumes that institutions on each side want to comply. Where one side does not, a better channel does not help. The travel rule then works mainly as a filter: compliant providers can identify non-compliant ones and restrict dealing with them. That is a weaker outcome than full data exchange, and it depends on enough providers applying the filter consistently. The standard’s value then rests on supervision and enforcement, which no single institution controls.
How it differs by institution
Exchanges and custodians registered as money services businesses
For US virtual asset providers registered with FinCEN as money services businesses, FinCEN’s May 2019 guidance said qualifying transfers may fall within the Funds Travel Rule. FinCEN noted that a non-exempt money transmitter must register with it within 180 days of starting to engage in money transmission. These firms carried the whole problem: counterparty identification, the channel, inbound data and exception policy. State money transmitter licensing could add its own examination layer.
Banks with crypto-related customers
For a bank whose customers included VASPs, the June 2019 standard changed what customer due diligence on those customers should cover. A bank would want to understand how its VASP customer met the travel rule, how it handled unhosted wallets and what its exception rates looked like. A VASP customer that could not explain its approach was a riskier customer. That review fits within existing bank customer due diligence and enhanced due diligence expectations, and we think it is where a bank was likeliest to meet the travel rule first.
Broker-dealers and wealth managers
Broker-dealers and wealth managers considering digital asset services faced the same counterparty problem, with their own regulators’ expectations on custody and recordkeeping layered on top. For them the June 2019 standard was a design input: any service they built would need a travel rule capability from the start, or a decision to restrict transfers to counterparties they could identify. Building it later into a live service would cost more.
Credit unions and smaller banks
We would expect smaller institutions to be the least likely to transfer virtual assets themselves. Their exposure ran through members and customers who used exchanges, and through any fintech partners offering crypto features through them. For them, a useful first step was knowing which partners touched virtual assets and asking those partners for their travel rule approach in writing. Where a partner’s program depended on channels or counterparties the institution could not assess, that was a risk to record and escalate.
Fintechs operating across borders
Fintechs with users in several countries met the sunrise problem directly. They might be bound by stricter rules in one jurisdiction and looser rules in another, for the same transfer. Their compliance design needed to follow the strictest applicable rule for each corridor, or segment services by jurisdiction. Either choice has product and cost consequences, and either should be documented as a decision with an owner.
With hindsight
More than seven years on, the June 2019 standard reads as the start of a long implementation rather than a settled rule. FATF kept monitoring. Its sixth targeted update on virtual assets, published on June 26, 2025, assessed jurisdictions against Recommendation 15 and its interpretive note. It reported progress since 2024 on regulation, supervision and enforcement. It also said jurisdictions still faced difficulties identifying the people and companies conducting VASP activity, and reported challenges with mitigating the risk of offshore providers. Those are the counterparty problems described above, still open in 2025.
On the travel rule itself, FATF reported that 99 jurisdictions had passed or were in the process of passing implementing legislation. It published Best Practices on Travel Rule Supervision the same day. The update noted that jurisdictions with materially important VASP activity accounted for approximately 98 percent of the global virtual asset market. It also flagged new risks: increased use of stablecoins by illicit actors, and a theft of $1.46 billion from the provider ByBit that FATF attributed to North Korean actors, with only 3.8 percent of the stolen funds recovered by the time of the update.
In the United States, the attempt to tighten the domestic rule did not survive. On October 27, 2020, the Federal Reserve Board and FinCEN proposed lowering the recordkeeping and travel rule thresholds from $3,000 to $250 for transfers that begin or end outside the United States. They also proposed clarifying that the rules apply to convertible virtual currency and to digital assets with legal tender status. On November 9, 2026, the agencies withdrew that proposal. The withdrawal notice said they had received nearly 3,000 comments. It said more than five years had passed, that further public feedback would be useful, and that developments in the virtual currency industry, including the stablecoin statute enacted on July 18, 2025, should inform any future amendments. The agencies said no final rule would issue without another opportunity for comment.
FinCEN’s 2019 guidance had already made the operating point that later events confirmed. A firm that set up a transaction system in a way that made compliance difficult could not use that difficulty as a justification for not complying. The withdrawal of the 2020 proposal did not touch that guidance, and the notice still describes the rules as applying to transfers of $3,000 or more. We think the practical effect of more than five years of pending rulemaking was uncertainty about the threshold, while the obligation to pass data on larger transfers stood throughout.
So, as checked on November 14, 2026, the US thresholds remain at $3,000, the proposal to write virtual currency into the rule definitions has been withdrawn, and FinCEN’s 2019 guidance is still posted on its website. What held up from June 2019 was the structure of the problem: obligations on individuals, means of compliance held in common, and implementation by countries at different speeds. What did not happen was a fast, uniform sunrise. Institutions that built their programs to work with incomplete counterparties, rather than waiting for complete ones, chose the design that matched how things turned out. On FATF’s June 2025 count, uniform implementation had not arrived, and by November 2026 the United States had withdrawn its own proposed tightening.
Spotted an error? Tell us and we will check it and correct the article where needed.