Or at least not a desirable thing...
There’s a growing belief that regulatory clarity in the US will lead to a lot more tokenization, including for bank deposits. I agree with the first half of that thesis. Tokens are a superior form factor for a digital economy. But just because you can tokenize something doesn’t mean that you should. For too long, people have treated tokenization like it’s a supply side problem. If you tokenize it, they will come. What actually matters is demand, and that’s where tokenized bank deposits will fall short.
The success of stablecoins has proven the demand for tokenized dollars. Most of the legacy stablecoins (Tether, USDC, etc) are based on a narrow bank design. The newer yield bearing stablecoins are closer to tokenized money market funds (MMFs). But aside from a few legal and regulator considerations, they are effectively the same thing — a liability against a corporate entity that holds a basket of Treasuries. My guess is that competition will eventually force every coin to pass on most of the yield. I also anticipate a new class of issuers willing to lose money on the stablecoin to use it as a loss leader for some other activity.
The theory behind tokenized deposits goes something like this: like stablecoins, bank accounts are just dollar liabilities people use for savings and payments. Since tokens are superior to accounts, we should have banks tokenize their deposits so people could use this on-chain, too.
The problem is that a tokenized bank deposit, from a traditional levered bank, is inferior along 3 crucial axes.
To wit:
Yield: Traditional deposits are a relatively cheap source of funding for banks, but tokenized deposits can’t be.
Most banks pay relatively little interest to most depositors. They can get away with this because people and companies have to use banks for savings and payments regardless. MMFs usually pay more interest, but you can’t use them for payments and there’s usually a delay in getting in and out. As a general rule of thumb, the more utility a bank account has (e.g., unlimited withdrawals from a checking account) the less interest it pays.
This means that deposits are a cheap source of funding for banks. Despite rates being near multi-decade highs, the average interest paid on a dollar bank account is still near 1%. MMFs (and their tokenized counterparts) pay close to 4%. Pretty good deal if you can get it!
Tokenization changes this dynamic because tokenized MMFs have the same utility of a tokenized deposit. They both offer 24/7 real-time payments, instant liquidity via DeFi, and so on. But the MMF pays more interest, making it a clearly superior product.
Of course a bank could always pay up to compete, but that sort of defeats the point of deposits for the bank. Depending on the rate environment they might get cheaper financing elsewhere. The other option is for the bank to start taking more risks on the asset side of its balance sheet, but now users are holding a riskier token that offers the same yield and utility.
Speaking of which:
Risk: Fractional reserve banks are fragile, making a tokenized claim against them less desirable.
Levered banks (and their tokens) are more dangerous than FinTechs, narrow banks, or money market funds, because they have run risk. There are protections like deposit insurance but the $250k limit is too low to matter for large institutions. Approximately half of all US bank deposits are uninsured.
It’s also unclear how the limit would be enforced if deposit tokens are permissionless. Users can split their holdings into multiple addresses. A bank could restrict access to whitelisted or KYC’d users but that, once again, diminishes the utility. Why would a large institution for whom the insurance is negligible to begin with opt for a token that both pays less yield and is more restricted?
The caveat to this argument is a tokenized deposit from a Too Big to Fail bank. As I’ve argued before, deposits at banks like JPM are implicit CBDCs from a safety point of view. There’s no plausible scenario under which the government would allow depositors at major banks to lose money — they wouldn’t even entertain the idea of a tiny haircut for large deposits at SVB. So the safety issue for tokens issued by some banks might be moot.
But this logic also applies to any sufficiently large MMF, bringing us back to the basic point that narrow banks are always safer, ceteris paribus.
Reg Arb: All tokenized dollars require greater regulatory accommodation to go mainstream, but tokenized bank deposits need a ton more
Banks are among the most regulated entities on the planet, and tokenization challenges how they are regulated. From AML supervision to macroprudential management, the framework falls apart on chain. This isn’t an argument against updating the rules: I think all regulatory frameworks should be reinvented periodically. But doing so is going to be a lot harder for levered banks because any discrepancy with how things are done today opens the door to regulatory arbitrage.
There are also open questions about what the greater transparency and 24/7 real-time operation of a token means for the always fragile public perception of a bank. Every stablecoin depegs slightly from time to time, if for no other reason than short term liquidity. But any kind of depeg, even if measured in basis points, is problematic for a bank. The negative signal could cause a run.
Narrow banks and tokenized MMFs don’t have this issue because they can be unwound to zero. Tether famously withstood billions of redemptions post FTX without a hitch. A bank couldn’t tolerate that.
In Conclusion
Tokenized bank deposits make sense if we take the current architecture of finance and project it on chain. But disruption seldom works that neatly. Today, we rely heavily on levered banking because we have no other choice. When it comes to important activity like making payroll or buying stuff, they are often the only option. Public blockchains change the paradigm, exposing banks to new competition.
Compared to other types of dollar liabilities, a tokenized bank account is inferior in almost every way. It will pay less interest, be more restricted, and carry more risk. This doesn’t mean that no bank will have success with them — I’m sure there will be smaller ones that figure out some novel way to use them as a loss leader. But the big money center type banks are really going to struggle to make this work. As I’ve joked before, banks like JPM and Citi are “the poor man’s Ethereum.” Their entire claim to fame (as stated in their own marketing docs) is their ability to let clients move money all over the world. But the user experience and cost is abysmal compared to what’s possible with tokenized MMFs on a public blockchain. If you are a big bank that relies on deposits, you are on the wrong side of history.
