A banking licence is generally regarded as the most comprehensive authorisation in financial services. It allows an institution to accept deposits, make loans and conduct activities that less extensively regulated financial institutions cannot undertake.
Switzerland’s proposed stablecoin framework introduces an interesting qualification to that hierarchy.
Under the consultation draft, a Swiss bank wishing to issue the specific category of regulated stablecoins covered by the reform would have to conduct that activity through a separate legal entity holding a payment instrument institution licence. The bank’s existing licence would not cover issuance within the bank itself. The government’s explanatory report expressly describes this separate-entity requirement.
That gives the proposed licence a distinctive position: it would authorise an activity that a banking licence does not itself authorise. We explain the new licence and its stablecoin rules in Stablecoin Licence Switzerland.
Is Switzerland alone in taking this approach? No. But the international comparison requires some care.
Chapter 1: Where else must a bank use a separately authorised entity?
First, the Swiss mechanism deserves a closer look.
Three provisions work together:
- Proposed Article 6 lists the activities covered by a banking licence through the licensing hierarchy. Payment instrument institutions are omitted.
- Proposed Article 12a reserves issuance of the defined stablecoin category to licensed payment instrument institutions, subject to specified exceptions.
- Proposed Article 51a defines a payment instrument institution as an entity that does not hold a banking licence.
The third provision is particularly significant. The proposal does not merely ask a bank to obtain an additional permission. It establishes two different institutional categories that cannot simply be combined in the same legal person.
The banking group could own the issuer. The bank itself would not be the issuer.
Several other jurisdictions follow a comparable structural approach, although their authorisations have different names and scopes.
| Jurisdiction | Separate entity for bank-issued stablecoins? | Position and qualification |
|---|---|---|
| United Arab Emirates: CBUAE framework | Yes. | The regulation expressly prohibits a bank from acting as a Payment Token Issuer. A subsidiary, affiliate or other related entity may undertake issuance, subject to its own licensing requirements. |
| United States: GENIUS Act | Yes, for the insured-bank issuance route. | An insured depository institution participates through a subsidiary approved as a permitted payment stablecoin issuer. Certain uninsured banks have different routes, so this should not be described as a universal rule for every US bank charter. |
| United Kingdom | Yes, as the PRA’s supervisory expectation for retail issuance. | The issuer should be a separate, non-deposit-taking entity insulated from insolvency contagion, with differentiated branding. Wholesale-only proposals receive a separate, proportionate assessment. |
| Singapore: September 2026 proposal | Yes, proposed. | Banks and merchant banks wishing to issue MAS-regulated stablecoins would have to establish a separate non-bank entity with its own authorisation. This remains a consultation proposal. |
| European Union: MiCA | No general subsidiary requirement. | Article 48 allows a credit institution itself to issue e-money tokens, subject to the applicable stablecoin requirements, notifications and white paper. |
| Hong Kong | No general subsidiary requirement, but a separate stablecoin licence is required. | A bank can itself become a licensed stablecoin issuer. Additional authorisation does not necessarily mean incorporation of an additional company. |
The UAE provides the clearest direct parallel.
Article 2(8) of the CBUAE’s Payment Token Services Regulation states:
“A Bank may not act as a Payment Token Issuer”
It then permits the activity through a subsidiary, affiliate or other related entity, subject to licensing and other regulatory requirements. The regulation is in force. This is an express separation between the licensed bank and the licensed issuer, closely matching the structural point raised by the Swiss proposal.
The United States provides another substantial precedent, with an implementation qualification.
The GENIUS Act was enacted in July 2025, but enactment and commencement are different matters. A US Treasury press release of 17 August 2026 states that the Act’s effective date is expected to be 18 January 2027. It is therefore accurate to describe the subsidiary model as part of the enacted US framework, while recognising that implementation is still underway.
Singapore is especially relevant because its proposal is recent and explicit.
MAS’s consultation of 1 September 2026 proposes that banks and merchant banks wishing to issue MAS-regulated stablecoins establish a separate non-bank entity, itself licensed as an issuer. The consultation closes on 16 October 2026. This is a proposed institutional separation, not merely an additional compliance obligation imposed on the existing bank.
Hong Kong demonstrates the alternative particularly well.
On 10 April 2026, the HKMA granted stablecoin issuer licences to both Anchorpoint Financial Limited and The Hongkong and Shanghai Banking Corporation Limited.
HSBC’s licence was granted to the banking entity itself. Anchorpoint, by comparison, is a joint venture involving Standard Chartered Bank (Hong Kong), HKT and Animoca Brands. The same jurisdiction therefore accommodates direct bank issuance and issuance through a separate company.
That distinction is essential. There are three different regulatory choices:
- Allow the bank to issue under its existing authorisation, with additional product requirements.
- Require the bank itself to obtain an additional stablecoin authorisation.
- Require a separate entity to obtain that authorisation.
Switzerland’s proposal belongs to the third category. It has credible international precedents, but it is not the only available model.
Chapter 2: Does the Swiss government’s explanation justify this separation?
The government’s central argument concerns the legal treatment of the money.
Ordinary bank deposits are liabilities of the bank. The proposed stablecoin framework instead provides for protected client assets, restricted investment and segregation in insolvency. The explanatory report argues that combining these arrangements inside one company would create uncertainty and risks. It also identifies the possibility of customers rapidly switching bank deposits into the bank’s own protected stablecoin, destabilising its funding.
There is substance to both concerns. However, neither automatically establishes that a separate company is the only workable answer.
Banks already manage assets with different legal treatments. FINMA distinguishes between bank deposits and securities held in custody: the latter are segregated and returned to clients in a bank insolvency. The coexistence of different client rights within one banking entity is therefore familiar. Stablecoin issuance raises additional issues, but the existence of different legal concepts does not itself make coexistence impossible.
Hong Kong’s framework provides a more direct comparison. Its licensing framework expressly contemplates bank issuers while setting requirements for the management and safekeeping of reserve assets. That does not establish that the Swiss proposal should copy Hong Kong. It does show that a regulator can choose to address the problem without categorically excluding the banking entity from issuance.
The deposit-flight argument also needs qualification.
Suppose customers move substantial balances from a bank account into a stablecoin backed by assets outside that bank. The bank may lose funding and need to replace it. That problem can arise whether the stablecoin is issued internally, by the bank’s subsidiary or by an unrelated issuer.
A second company does not, by itself, prevent the transfer.
Moreover, the Swiss draft allows client funds to be held as sight deposits with a bank or another payment instrument institution, or as short-dated, high-quality liquid assets (draft Article 51i). It does not require all reserves to sit outside the commercial banking system.
The economic effect therefore depends on where the reserves are placed. Deposits may move between banks, return to the originating bank through the issuer’s reserve account where permitted, or leave commercial bank balance sheets. Central-bank analysis likewise recognises that the effects depend heavily on reserve management and the source of stablecoin purchases.
My assessment is that the stronger justification for a separate issuer is clearer responsibility, creditor protection and management of failure. The argument that incorporation of a subsidiary will itself prevent a bank run is considerably less convincing.
Chapter 3: Why this matters for the proposed licence
The commercial significance is easy to overlook.
A specialised financial licence is often assessed by comparing it with a banking licence: fewer permitted activities, fewer prudential obligations and a narrower business model.
Here, the comparison changes. The proposed Swiss payment instrument licence would occupy a distinct regulatory position. It would provide access to an issuance activity that the bank could not undertake within its own legal entity.
That does not make the payment instrument institution more powerful than a bank across financial services. It makes the authorisation relevant to a different institutional function.
For a banking group, this creates a structural decision: establish an issuer, participate in a jointly owned issuer, or work with an independently authorised issuer. Those arrangements would remain subject to the final rules and the permissions governing each participant.
For financial entrepreneurs, it raises the possibility of building an institution that serves banks as well as competing for payment business. The authorised issuer could become a specialised component of a wider financial infrastructure.
These are commercial possibilities arising from the proposed structure, not automatic consequences of receiving a licence.
There are defensible reasons to favour separation.
First, it can make failure easier to manage. A dedicated issuer can have its own reserve arrangements, redemption procedures, operational resources and wind-down plan. A regulator may find it easier to transfer or close that business without simultaneously restructuring the bank’s lending activities.
Second, it can clarify the product offered to customers. A stablecoin holder should understand which company owes the redemption obligation and which protections apply. The UK’s approach expressly connects separate issuance and product differentiation with the risk of confusion between insured deposits and other forms of digital money.
Third, it can support shared infrastructure. Several banks may prefer to own a common issuer rather than have one participating bank act as issuer for all the others. A separate company can provide a practical governance structure for such cooperation.
Fourth, and in my practice the most important: it ring-fences cyber risk and liability. The Swiss explanatory report does not mention it, but it is the reason I have always given to clients. In Europe, when e-money institutions wanted to add crypto activities, I advised them to do it through a separate company as soon as that became possible. Not because of deposits, but because of hackers.
A business that holds crypto-assets or runs stablecoin infrastructure is a target. Sooner or later, someone will try to break in and steal the coins. If that happens, it is far better that the loss sits in a separate company, which can be wound down if necessary, than inside the bank itself.
Separation also changes what attackers aim at. They go after the systems of the crypto or stablecoin company, not the core banking software. A separate entity with its own infrastructure keeps the bank’s systems out of the line of fire.
So if a client tells me “we are going to issue a stablecoin”, my advice is simple: do it in a separate company. Not because the law may require it, but for liability reasons.
There are also substantial objections.
Separate incorporation creates real costs. Governance, capital, reporting, contracts and operational arrangements must be maintained for another regulated entity. Existing banking supervision should therefore be recognised wherever it can avoid unnecessary duplication.
Legal separation does not eliminate group contagion. Common branding, shared systems, guarantees and expectations of parental support can connect the fortunes of the issuer and the bank. The effectiveness of separation depends on the arrangements behind the company structure. The same applies to a hack: a separate company limits the financial and technical damage, but not necessarily the damage to the group’s name. If the stablecoin subsidiary is hacked, the headlines will still carry the bank’s brand.
Protection against the issuer’s failure does not automatically protect against failure of its reserve bank. If reserves include ordinary deposits with a commercial bank, that exposure still needs to be assessed. A separate certificate of incorporation does not remove the underlying counterparty risk.
The alternative deserves a fair hearing. A bank may be able to satisfy appropriately designed reserve, redemption and insolvency requirements directly. The EU permits direct bank issuance under MiCA, and Hong Kong has actually licensed a bank as an issuer. Those approaches cannot reasonably be dismissed as legally inconceivable.
I would therefore assess the Swiss proposal by asking what the separate entity demonstrably achieves: whether it protects the reserve assets, clarifies the holder’s claim, limits contagion and makes orderly redemption or closure more credible.
For banks, the resulting question would be how to organise their participation in stablecoin issuance. For prospective payment instrument institutions, it would be whether they can build the governance, banking relationships and operational capability to perform a function that even a banking group would need a separately authorised entity to undertake.
A bank planning stablecoin issuance, or an entrepreneur building an issuer? Write to insight@fintechlex.com.
The Swiss discussion above concerns the consultation draft published on 22 October 2025. The international comparison distinguishes rules in force, enacted frameworks awaiting implementation, supervisory expectations and consultation proposals as at 14 October 2026.
