Cross-border retail CBDC complications

The thousands of small decisions in retail CBDC design highlight the impracticality of the concept.

If any of you have been wondering why the dozens of retail CBDC trials currently ongoing are taking so long, it’s because the technology layer is the relatively easy part. Much more complicated are design decisions and legal considerations, which overlap but also have their own buckets of debates and details.

Even current pilot and live retail CBDC programs such as those in China, Nigeria and India don’t have this all figured out, and are adding on new features and rules as they go along. Yet what is not embedded from the beginning adds complexity which in turn hinders adoption, especially when the tokens don’t offer a clear advantage. So far, initial results suggest all existing platforms are struggling to gain traction.

One potential advantage is that of more efficient cross-border payments, which in theory could be almost instantaneous with blockchain-based tokens rather than the many steps required to send digital fiat. Yet private payments platforms have been working on smoothing international transfers, with some success. Why would a central bank need to get involved?

A key reason could be capital flight, or rather, wanting to prevent it. Most jurisdictions have restrictions on the amount of digital fiat that can be moved across borders. If these transfers are done with a CBDC, the central bank has both more insight and more control over how much is flowing out or in, when, by/to whom, etc. It’s notable that the most advanced retail CBDC project, China’s e-CNY, is actively seeking international partners for its digital currency (Hong Kong, Singapore), but only for tourism purchases and limited cross-border shopping, not peer-to-peer transfers.

Last month, the IMF published a report on some of the key considerations, which I’ll briefly outline here, in the interests of transmitting just how complex any new retail currency introduction would be. The same issues are also relevant, in a slightly simpler form, for wholesale CBDCs.

Access

Who can use the retail CBDC? If it’s only residents, there’s unlikely to be much demand for cross-border transfers. Allowing non-residents to hold a country’s CBDC implies a relaxation of currency control, and may conflict with the rules of other jurisdictions. And how would KYC of non-residents be handled? On which platforms would non-resident holders swap a country’s CBDC for their local currency?

This conversion could be done by the non-resident’s domestic bank – but then decisions would need to be taken as to which foreign banks can access a country’s CBDC. These banks would technically be under “foreign supervision” (regulated by their own central banks), which some authorities may feel adds political risk. And anyway, would there be enough demand for these foreign banks to go through the onboarding hassle?

For security and/or capital control reasons, limits could be imposed on how much of a foreign CBDC any entity or individual can hold. But this impacts potential demand and introduces another wave of design questions: who supervises? What happens when limits are breached? Automatic conversion is possible, but this adds another layer onto the question of where conversion would happen, and when – 24/7, or in traditional banking hours?

Another possibility is allowing unlimited balances during the day, but insisting on conversion and settlement resulting in a much lower limit overnight. But, overnight for whom? And would that impact exchange rate volatility at close, whenever that is?

Messaging

The messaging around a payment is usually one of the most overlooked aspects, because we all tend to think that payments just “happen”. And if we were to give it any thought, we’d probably assume it was as simple as telling one computer to deduct quantity X here and add over there.

It’s more complex than that: the instructions need to be in a certain format, with no exceptions. Banking systems around the world are moving toward the ISO 20022 standards in which all payments will follow the same messaging format, allowing for faster settlement and more detailed reporting. But it is designed for traditional payment rails; a method to link distributed ledger-based payments with traditional rails (necessary for reconciliation) will be needed.

What’s more, decisions will be needed about the messaging process. Should they be sent by the originating bank? Or should they be pulled from the network itself? The latter option would be better for synchronous efficiency, but would require all CBDC networks to either harmonize or find a suitable messaging bridge. This is a lot harder than it sounds, given likely disagreement about which technology to use.

Currency conversion

This is already a complex web involving many steps; a new type of transfer rail introduces new complexities. A deceptively simple question is: who initiates and who executes the conversion from one CBDC into another? Should the sender convert to the recipient’s currency before sending? Should the recipient convert after receiving? Or should there be middlemen to handle the transaction? If so, who, and where would they be domiciled?

All options are related to the decision on who can hold another jurisdiction’s CBDC, which I touched on above. And on whether cross-border CBDC payments have to be routed through central banks, local banks, or can go wallet-to-wallet.

The decisions also carry the risk of introducing more layers, with each additional step adding cost and counterparty risk. This could potentially reduce any efficiency gains – and, given the colossal cost of any shift to CBDCs, put the point of the initiative in doubt.

Settlement

For the transfer of value within a jurisdiction, settlement generally involves a ledger adjustment. This is relatively simple in the grand scheme of financial flows. It gets more complicated when a retail payment leaves its originating country, as now there are at least two payments involved: one to the currency conversion platform, another to the recipient. Of course, this assumes conversion is desired, which is part of the access decisions touched on above. And there could be many more than two payments, depending on whether or not CBDC transfers have to go through local banks.

This introduces potential complications regarding “settlement finality”, which is the moment a transaction is deemed complete. Even in traditional systems, which involves centralized ledger adjustments, there are jurisdictional differences as to the legal definition of this term. In a distributed ledger, the concept gets more complex, depending on the underlying technology used. For some blockchains, a transaction is deemed “final” when a given number of blocks has passed and rewriting previous updates has become prohibitively expensive. But this is technically not “final”, since there is no strict definition of “prohibitively expensive”. In other blockchains, settlement can be “instant” and yet still wound back; for retail participants, this reduces certainty.

In a CBDC system, smart contracts could play a role by initiating all legs of a transfer but “holding” final release until certain conditions are met. This introduces a new layer of trust, as well as of decisions including the jurisdiction of the code.?

Compliance

The world is moving towards standardized KYC procedures at the insistence of the Financial Action Task Force. Yet some regions are deemed “risky” and require additional information transfer. Unfortunately, these tend to be regions that largely rely on remittances and/or tourism – in other words, the potential impact of cross-border retail CBDCs is a bigger deal for these jurisdictions than for larger economies. Would they find themselves shut out of digital global finance, as they are with traditional rails? Or would tokenization solve some of their information risk issues?

If the latter, would the additional information be embedded in the tokens? If so, would it apply for everyone, making the system more cumbersome, or only for a few, making CBDCs less fungible? Or would the additional information be gathered separately, as with the current system, adding inefficiency and eliminating a significant chunk of any CBDC advantage?

Wholesale CBDC

Obviously, most of the above considerations also apply to the idea of wholesale CBDCs, with blockchain-based central bank money moving between financial institutions. Should foreign banks be allowed to hold a country’s digital currency? When is settlement final? How are transfers communicated? And so on.

But bank-to-bank transfers are relatively simple; the relationships already exist, banks are used to multi-step transactions and documentation requirements, and there is no expectation of privacy.

This is one of the main reasons I don’t think we’ll see widespread adoption of retail CBDCs – some use, perhaps, especially if they are accompanied by stimulus measures (if you spend by Friday, you get a 10% airdrop; a tax deduction if you buy furniture in this region by the end of the year). But privacy concerns are real, as is spreading distrust of institutions. And those who want access to convenient electronic payments can already, usually, get it; those that don’t for whatever reason aren’t going to be convinced by a central bank offering.

What’s more, the above highlights just how complex the idea is. Are the expected benefits worth the significant cost? For wholesale CBDCs, the equation is somewhat simpler – the advantages are more obvious (streamlined cross-border trade, updated financial systems, greater flow transparency) and the complications, while considerable, are somewhat more manageable.

Retail CBDCs don’t have what tech entrepreneurs would call “product-market fit”. This may change, especially as wholesale CBDCs start interacting with the real world. But, for now, their main value is as a fascinating intellectual exercise, and the opportunity to question what people, rather than governments, want from their money.

(This is an excerpt from the Crypto is Macro Now newsletter, where I look at the growing overlap between the crypto and macro landscapes — I hope you’ll consider subscribing!)