Tuesday, November 29, 2022

A worthwhile Canadian stablecoin initiative

One interesting thing about stablecoins, the world's newest payments technology, is that they are almost all U.S.-dollar based. More than 99% of the $145 billion worth of stablecoins in circulation are denominated in dollars, the remaining 1% being mostly euro-denominated. 

Even though no significant Canadian dollar stablecoin has emerged to date, the Canadian government is beginning to think about these financial products. A financial sector legislative review of digital currencies -- including stablecoins -- was announced in the government's recent budget. I suspect that a big part of the review will involve trying to answer the question of how to regulate these new instruments.

A few quick thoughts on how we Canadians should regulate stablecoins.

1) There's nothing fundamentally new about stablecoins. All digital Canadian dollar balances are recorded on databases. In the case of a Bank of Montreal account or a PayPal C$ balance, those dollars are instantiated on an internal SQL or Excel database (or whatever database traditional institutions use). Stablecoin issuers opt for a different sort of database to record dollar balances: shared databases like Ethereum, Solana, and Tron. These blockchain-based databases are often described as decentralized, although it is disputable how decentralized they actually are.

But abstracting from the choice of database, stablecoins are just another instance of regular old finance.

Canadian financial regulations should, in principle, be database agnostic. And so in my opinion, all existing financial regulations that are currently applied to issuers of Canadian dollar balances should be passed on to Canadian dollar stablecoins, perhaps with a bit of pruning.

2) In the spirit of the database agnosticism that I set out in 1), OSFI-regulated banks and credit unions should be able to issue blockchain-based Canadian dollar balances (i.e. stablecoins) under all the same rules that they issue SQL-based Canadian dollar balances (i.e. deposits). Those stablecoins would be insured, too, up to $100,000.

Here's where the "pruning" comes in. Some thought will have to go into how to apply deposit insurance to failed stablecoin issuers. For instance, if $10,000 in failed Canadian dollar stablecoin units is locked up in a Uniswap contract, how will deposit insurance be applied? What happens if no one ever withdraws the coins to claim the insurance? How do the smart contracts of a failed stablecoin get turned off? What happens if the decentralized database itself fails?

Because smart contracts can be programmed, I think it's possible to solve most deposit insurance problems. Regulators like OSFI or CDIC might even go so far as to specify the exact code that issuers must include in their smart contracts in order to qualify for insurance.  

3) In addition to 2), non-banks should be allowed to issue uninsured stablecoins, perhaps under the emerging payment services provider license that the Bank of Canada will be administering.

There are some caveats. Non-bank stablecoin issuers should only be allowed to invest customer funds in safe short-term assets. They would also have to  keep customer funds ring-fenced in bankruptcy-remote structures, like trusts, so that if the issuer fails, customers will be guaranteed to get their money back rather than being treated like a regular unsecured creditor.

4) Lastly, regulators will have think about stablecoin anti-money laundering issues. 

Right now, popular stablecoin issuers like Tether and Circle only identify people who are redeeming stablecoins for "fiat" money or withdrawing stablecoins by depositing fiat. But the great majority of stablecoin transactions currently occur bilaterally between those who never go through a know-our-customer (KYC) process, much like physical cash. 

This "no-identity" model is a big part of what has made these stablecoins so popular. Users can rapidly deploy stablecoins across multiple decentralized financial protocols without having to go through the frictions of an onboarding process. Exchanges and other financial intermediaries can use stablecoins as a way to replicate U.S. dollar balances for their customers without having to establish formal banking relations.

But this cash-like treatment also makes stablecoins riskier. For instance, I recently wrote about a ponzi scheme called Meta Force which is using Dai stablecoins on the Polygon network for pay-ins and pay-outs. Thanks to the way that the stablecoin smart contracts have been deployed, and the lack of KYC, there is nothing to prevent the scammer who manages Meta Force from openly making use of these safe instruments to con his unwitting customers.

Canadian regulators will have to weigh the usefulness of a no-identity cash-like model against the risks of pseudonymity. 

There is one last risk to consider. Say that regulators choose to tolerate a cash-like model for Canadian dollar balances instantiated on blockchain-based databases like Ethereum and Tron while continuing to require full KYC on Canadian dollar balances instantiated on regular databases. The consequence could be mass regulatory arbitrage as financial institutions migrate over to the former in order to avoid the more onerous requirements of the latter.


  1. The stablecoins issued by a bank would I assume have some form of capital backing. Does your pruning envisage that this is not required so long as the assets backing the stablecoin are held in a trust? For what it is worth the Australian regulator seems to be proposing that stablecoin issuers would require capital equal to 4% of the liabilities.

    With respect to deposit insurance, I am not sure how Canada works but in Australia deposits in the banking sector have a priority (i.e. super senior) claim on the bank’s Australian assets. That has the effect of reducing the risk of providing deposit insurance in Australia. Apart from the question of capital buffer, a stablecoin arrangement will not have the benefit of super seniority so insurance would be riskier. In Australia the proposal is that stablecoins be regulated like a Purchase Payment or Stored Value facility and these do not benefit from deposit insurance. As yet this is all hypothetical but the policy was put in place (I think) in response to the Libra/Diem initiative

    1. "The stablecoins issued by a bank would I assume have some form of capital backing."

      Bank-issued stablecoins would be subject to all the same capital requirements as regular bank deposits.

      "Does your pruning envisage that this is not required so long as the assets backing the stablecoin are held in a trust?"

      Holding assets in trust is a rule that would be applied to the second type of stablecoin, those issued by non-banks. Stablecoins issued under existing banking regulations would not be subject to trust requirements, in the same way that a bank's deposit business is not held to that standard.

      "In Australia the proposal is that stablecoins be regulated like a Purchase Payment or Stored Value facility and these do not benefit from deposit insurance."

      Yep, the same for Canada. In 3), I outlined how they stablecoins could be issued by non-banks under our payments services regulatory framework, which is probably similar to the Australian framework. I also suggest that stablecoins could be issued by banks, which I outlined in 2).