Account-based vs. token-based CBDC? Stay focused on what matters.

In several analyses of CBDC infrastructure and models, the issue of whether to use an account-based or token-based (aka value-based) CBDC is discussed. However, for any practical purposes I fail to see the difference. Both models can provide the same functionality and both require transaction validation by a third party, or the update of a centralized ledger to preempt the double spend problem and ensure robust security. 

My notion is supported in Bank of England's CBDC discussion paper where they note: "We do not see any inherent reason that token‑based systems would automatically provide anonymity. Both account‑based systems and token‑based systems can be configured with various identity solutions, ranging from fully anonymous to pseudonymous and to a fully transparent, identifiable solution... In an account‑based system, the accounts of the payer and payee need to be debited and credited by the operator(s) of the ledger. And in a token‑based system, in order to prevent double-spending, ownership of tokens needs to be recorded in a ledger, which will need to be updated to reflect any changes in ownership."

In another thoughtful analysis by Alistair Milne, he concludes that "cryptocurrencies and crypto assets are account-based (recorded in supporting ledgers) not tokens or objects... Transfer of cryptocurrencies and crypto assets depends, just as with any other electronic transfer of an account-based money, on compliance with the transfer protocols of a payment scheme."

So what's the big deal? Why is this even an issue?

If you’re a central banker, I suggest you don’t waste time thinking about whether token-based or account-based. Instead, focus on what capabilities are required for issuance and use of the CBDC and its ecosystem. Whatever your requirements are, I’m almost certain you’ll find that whether a token-based or account-based system won’t matter at the end of the day.

After all, when your IT department chose an email system, I’m pretty sure they compared functionality and total cost of ownership rather than what database technology was used as a message store.