The Bank of Canada published a staff working paper exploring the anonymity of central bank digital currency (CBDC). Instead of looking at it from a consumer perspective, they explore it from a corporate borrowing angle. A small retail company might borrow money from a bank, but the bank can see the borrower’s debit card receipts used to repay the loan. This isn’t dissimilar to Square, which provides point-of-sale equipment, but another subsidiary grants loans.
A CBDC could have features that put it on a spectrum of anonymity. The paper concludes that the optimal anonymity is at either end of the spectrum, not in the middle. Additionally, competition between deposits and CBDC could hinder low anonymity for a CBDC.
The Canadian game theory paper models two types of entrepreneurial borrowers. There’s a high quality one that is happy to refinance a loan and use its revenues to repay it. And there’s a low quality business that prefers to be opaque about its retail receipts and hence favors cash because it probably won’t make a full repayment. Banks are aware of the different behaviors and allocate more loans to the higher quality companies.
Mid-range CBDC anonymity is less likely
If you introduce a CBDC with modest anonymity, then both types of companies are likely to adopt it. That makes it harder for banks to know which type of business the companies are, so they might allocate more loans than they should to the weaker companies.
This inefficient allocation of resources could be rectified by making the CBDC less anonymous, encouraging the low quality firms to switch to cash.
The paper then explores what happens if CBDC is attractive to merchants because of far lower fees. The model finds if there are weak sales benefits (less attractive fees), there should be weak anonymity similar to deposits, so low quality merchants choose cash and higher quality companies choose CBDC.
On the other hand if the fees are low, making CBDC is attractive for sales, low quality merchants might use it despite the lack of anonymity. This means that the low quality merchant attracts more lending than it should. Hence, the model suggests a CBDC with high sales benefits should have higher anonymity. That discourages banks from financing any loans if CBDC is used. So eventually the higher type chooses debit and the lower type cash.
Finally, the paper explores how banks might anticipate a CBDC and alter their behavior, such as making debit cards more similar to a CBDC. The authors show that the first mover – the bank or the central bank – achieves its favorable outcome. Hence, “if the bank can move first, the competition may hinder the most socially desirable outcome.” It also concludes that competition between CBDC and deposits could hinder low anonymity for a CBDC.
Critical assumptions
Some critical assumptions underpin the model. For example, it effectively assumes that banks cannot bundle services. Hence, they can’t require the borrower to use only debit for acceptance, which would give the bank full visibility.
There’s another similar real world scenario we wondered about. Cash is anonymous because there is no acquirer. Someone can always put money in their pocket rather than in the till. However, a CBDC always has to have an acquirer. Hence, rather than bundling services, a bank could request a borrower to opt into providing access to the acquirer’s processing records, even if the CBDC is anonymous by design. From a bank’s perspective, this is the equivalent of a CBDC with low anonymity.
