Suit Up · Asia's institutional desk on tokenisation

The thesis

Six tokenisation-relevant regimes have crystallised over the last 36 months: the GENIUS Act, MiCA, the Hong Kong Stablecoins Ordinance, the Japan PSA Electronic Payment Instruments regime, the MAS Single-Currency Stablecoin framework, and the US OCC trust bank charter. They diverge meaningfully on reserve composition, issuer eligibility, retail-versus-wholesale scope, and redemption mechanics. The harmonisation that the 2018-to-2022 commentariat predicted is not coming. Practitioners should stop waiting for it and start designing for the divergence as the operating environment.

The setup

The standard prediction five years ago was that crypto-asset regulation would converge the way prudential regulation converged after Basel I, the way derivatives reporting converged after the post-2008 G20 process, the way FATCA pulled cross-border tax-information reporting into a workable global pattern (with caveats, and some persistent holdouts). The argument was structural: financial centres copy each other's homework when the alternative is ceding cross-border flow to the early mover, and the major central banks coordinate at BIS-level on prudential expectations that eventually filter into national rulemaking.

That prediction was wrong on stablecoins and on tokenised market infrastructure. Each major regime that has shipped reflects a different domestic political-economic answer to a different domestic question. GENIUS reflects the post-FTX US pressure to put a federal floor under retail payment stablecoins after years of state-by-state money-transmission patchwork. MiCA reflects the EU's commitment to a single rulebook approach across all financial instruments, the same instinct that produced MiFID II and PSD2. The Hong Kong Stablecoins Ordinance reflects HKMA's positioning as the regulated gateway between mainland China and global capital, with the offshore-renminbi corridor as the structurally distinctive case. The Japan PSA reflects FSA's preference for routing new instruments through existing legal categories (banks, FTSPs, trust law) rather than inventing fresh ones. The MAS SCS framework reflects MAS's narrow-perimeter discipline, restricting issuance to single G10 currencies through MPI licensees and refusing the basket-pegged designs that MiCA accommodates under the ART category. The OCC trust bank charter reflects an explicitly federal alternative to state-chartered crypto custodians, designed to give crypto-native firms a single supervisor rather than 51.

Different problems, different answers. The convergence pull at the prudential margin is real. The structural choices about who can issue, how reserves are constituted, and who can hold are domestic choices, and they are not converging.

The argument

Why harmonisation isn't coming

The political-economic drivers are local. Each regulator answers to a different legislature, a different political constituency, a different banking sector structure, a different geopolitical positioning. The OCC's relationship with the Bank Policy Institute and the GSIB lobby is a specific Washington fact pattern that produces specific regulatory outcomes (the BPI is reportedly considering a lawsuit over the trust-charter programme, on the argument that crypto firms enter the banking system under lighter oversight than full-service banks face). HKMA's positioning relative to Beijing's capital-account posture is a specific Hong Kong fact pattern that produces specific outcomes around offshore-CNH stablecoin perimeter. FSA Japan's three-route EPI structure is a specific Tokyo fact pattern reflecting the institutional weight of the megabank-trust complex (MUFG, Mizuho, SMBC routing through Sumitomo Mitsui Trust on the Progmat consortium). These structural facts do not negotiate themselves into a global compromise.

The international standard-setter pathway exists, and it matters at the prudential margin, but it cannot produce harmonisation at the structural level. Basel SCO60 is the worked example. SCO60 is a global prudential standard intersecting with each national regime; it sets minimum bank-side capital expectations for cryptoasset exposures. It does not replace, override, or harmonise GENIUS, MiCA, the Hong Kong Ordinance, or the MAS SCS framework. The Basel Committee, the IOSCO cross-border conduct working groups, and the BIS technology-side work via the Innovation Hub are coordinating mechanisms, not substitution mechanisms. Reading them as the leading edge of harmonisation misreads what they are designed to do.

The actual divergences are bigger than they look

Headline comparisons across the six regimes systematically understate the divergence. The standard observation is that GENIUS requires high-quality liquid asset reserves, the Hong Kong Ordinance requires HQLA reserves, MiCA EMT requires HQLA reserves, MAS SCS requires HQLA reserves, so the regimes are converging on reserve composition. They are converging at the asset-list level. They are diverging at the issuer-eligibility level, which is where the operational lift sits.

Issuer eligibility under GENIUS is the three-branch perimeter: insured-depository-institution subsidiary, federally chartered NBFI (the OCC trust-bank charter is the path most non-bank issuers are taking), or state-qualified issuer below USD 10 billion outstanding. Under the Hong Kong Ordinance, eligibility is a Hong Kong-incorporated body corporate meeting the HK$25 million paid-up capital floor, with banks eligible to apply but the licence shape itself non-bank. Under MAS SCS, eligibility is restricted to MPI-licensed entities under Part 3 of the Payment Services Act, with the MPI as the issuer of record (no thin-wrapper group structures). Under Japan PSA, eligibility is the three-route structure: banks, FTSPs, or trust companies and trust banks. Under MiCA EMT, eligibility is restricted to credit institutions or authorised electronic money institutions. Under the OCC trust bank charter pathway, eligibility is the chartering process itself, a multi-quarter exercise with conditional approvals preceding final charters.

These eligibility structures are not interchangeable, and the differences compound. An issuer authorised under GENIUS as an OCC-chartered trust bank does not automatically meet the MiCA EMT credit-institution-or-EMI test, does not automatically clear the Hong Kong Ordinance Hong Kong-incorporation requirement, does not automatically register as an FTSP under Japan PSA, and does not automatically hold an MPI licence under MAS SCS. Each is a separate authorisation track, with separate evidence burdens, separate timelines, and separate post-authorisation supervisory obligations. The reserve-composition headline convergence hides the issuer-side structural divergence, and the issuer-side divergence is what determines whether a particular operator can actually ship into a particular jurisdiction.

The other axis of underrated divergence is redemption mechanics. The Hong Kong Ordinance requires par redemption in the reference currency within one business day unless HKMA approves a longer period, the tightest statutory window of the six regimes. MiCA EMT requires par redemption at any time, free of fees. GENIUS requires par redemption with timing in the issuer's published procedures, no statutory floor. MAS SCS requires par redemption on a tight timeline set by operative MAS notices. Japan PSA inherits the parent-licence redemption framework, which differs across the three routes. These differences sound like edge cases until an issuer with a treasury sweep into longer-dated instruments has to fund a Hong Kong-licensed redemption inside one business day, against a US-licensed redemption that can run on the issuer's published procedures.

The four navigation patterns that work

Operators who have shipped successfully at scale follow one of four patterns. Treating these as alternatives to be selected, rather than waiting for a single global-licence outcome that resolves the choice, is the practical operating posture.

Pattern one is home-jurisdiction-first. Anchor the issuance under the regime that fits the product structure most cleanly, then export under that regime's third-country, Reg S, or comparability rules. BlackRock's BUIDL and Franklin Templeton's FOBXX are the worked examples on the tokenised-MMF side: launch under US 3(c)(7) for qualified purchasers with Securitize as transfer agent, then expand into Asia via Reg S distribution to APAC private banks that have already onboarded the operational pattern. The pattern is operationally efficient because the home-jurisdiction structure does the structural lift once, and each subsequent jurisdiction is a distribution exercise rather than a re-issuance exercise.

Pattern two is regulator-shopping. Pick the regime with the lightest licensing burden for your specific product structure, accept the resulting distribution constraint, and operate within the perimeter the chosen regime grants. The pattern is real but the upside is narrower than the 2018-era reading suggested: every major regime has tightened its distribution gating to keep arbitrage paths closed. MiCA's third-country-issuer rules force any non-EU stablecoin distributed in the EU to clear the EMT perimeter via an authorised EMI holder. The GENIUS Act's foreign-issuer comparability test means a Hong Kong or Singapore-licensed issuer cannot offer to US persons through US-licensed venues unless the home regime is certified comparable, which has not been granted to any APAC regime as of late April 2026. The Hong Kong Ordinance reaches extraterritorially for any HKD-referenced design regardless of where the issuer sits. Soft-regime issuance produces a footprint, not a global product.

Pattern three is multi-regime issuance. Launch under multiple regimes in parallel via separate issuance vehicles per jurisdiction. HSBC Orion's bond programmes are the worked example: UK gilts via DIGIT, multi-currency green bond series via the HKSAR Hong Kong issuer, Luxembourg-domiciled institutional bonds for European distribution. Each issuance vehicle is a complete legal and operational stack, which is operationally heavy. The trade-off is a cleaner regulatory perimeter per issuance and the elimination of the cross-border-passport problem entirely, because there is no cross-border-passport claim being made. The pattern is available to issuers with the operational depth to absorb running multiple parallel stacks; a single-issuer programme cannot reasonably absorb that overhead.

Pattern four is partnered distribution. The home regime authorises the issuance; a locally-authorised distributor in each target market handles the customer-facing layer. Japan's PSA EPI intermediary registration is the explicit channel for this on the stablecoin side: an offshore exchange or wallet distributing JPYC or a Progmat EPI to Japan residents must register separately as an EPI service provider, with its own conduct obligations and the deposit-taking prohibition. The pattern works for issuers that prefer to keep issuance in one home jurisdiction and rent local distribution rather than build it.

The wrong pattern is "wait for harmonisation." That is a free option being purchased at premium. Every quarter spent waiting is a quarter of unauthorised distribution, deferred customer relationships, and competitive ground ceded to operators who chose one of the four patterns and shipped.

Implications for practitioners

For founders structuring a multi-jurisdictional launch, pick the home jurisdiction first based on which regime fits the product structure most cleanly, then treat each distribution market as a separate licensing exercise. The structural fit question (single-currency vs basket, payment vs deposit-substitute, retail vs wholesale, bank-issuer vs non-bank-issuer) determines the home regime. The distribution markets are downstream choices.

For in-house legal teams running the regulatory tracker, stop running quarterly convergence trackers. Convergence is not happening at the structural level, and the convergence-tracker as a deliverable creates the impression of imminent simplification that the underlying data does not support. Start running quarterly divergence trackers instead, focused on the specific axes where regimes are moving apart: issuer-eligibility changes, reserve-composition adjustments, redemption-cadence tightening, distribution-gating updates. That is where the operational risk lives.

For institutions comparing structures, the choice of regime is downstream of the choice of asset class and counterparty profile. Lead with the asset-class question (a tokenised deposit conversation differs structurally from a tokenised MMF conversation differs structurally from a payment-stablecoin conversation), then the counterparty question (is the customer a bank, a corporate treasurer, an asset manager, a retail wallet), then the regime question. Leading with the regime question produces the regulator-shopping anti-pattern in disguise.

For analysts and commentators, the harmonisation prediction was wrong, and the divergence-as-feature thesis is the right next frame. The interesting analytical question is no longer "when does this get simpler" but "which combinations of regime fit which combinations of product structure and customer base, and where are the operating frictions when an operator runs across multiple combinations". That is the question that determines whether a particular tokenisation programme actually ships.

Where this could be wrong

Two scenarios where the divergence thesis breaks. First, a major political event forces fast harmonisation. US-EU systemic-risk negotiations have produced rapid harmonisation in adjacent areas before (the Volcker-rule extraterritoriality climbdown, the post-Lehman OTC-derivatives reporting alignment, the FATCA reciprocity arrangements). It is not impossible for a stablecoin-specific event to do the same. The trigger would likely be a cross-border issuer failure with retail consumer impact in two or more major jurisdictions.

Second, the BIS / IOSCO pathway accelerates beyond minimum prudential standards into more prescriptive harmonisation on issuer eligibility or reserve composition. The current standard-setter posture is firmly minimum-floor, but a sufficiently sharp policy shock could shift it. Both scenarios are tail-cases on a three-year horizon. Our read is that divergence persists, and the practitioner posture should be designed for divergence as the operating environment rather than for a harmonised endpoint that may or may not arrive.

[Suit Up]

Regulatory navigation in a six-regime world: the harmonisation that isn't coming