Stablecoin Divination: Banks, Risks, and the Blockchain Balancing Act
As 2024 ends, we unravel the hype surrounding stablecoins, tokenized deposits, and their true impact on traditional finance. How reliable are Coinbase's predictions or Bitwise's prophecies?
It's almost year-end, which means everyone is channeling their clairvoyant skills and making predictions about the future. Did they pass their exams at Hogwarts?
Stablecoin’s Fortune
I've read one too many times that stablecoins are the killer app and that every bank is eager to offer payment solutions based on such a token construct. I very much doubt it.
Stablecoins perform a vital role in crypto markets, but there is no need for a stablecoin to fill this gap in today’s financial markets because we already have Dollars and Pound Sterling instead.
I’ve written about this before.
But there are other reasons. Banks rely heavily on Net Interest Income (NII), which comes from the spread between deposit and lending rates. Stablecoins directly challenge this by removing customer deposits from bank balance sheets, assuming the client is using a stablecoin issued by a third party. And it’s not just about profit: if stablecoin adoption were to take off, it could impact capital and liquidity ratios, putting every regulator on high alert. This is no small change.
In addition, all the privileged rights the token issuer typically awards itself to block transactions or even invalidate the token if they need to make them very unattractive for banks because it brings a new risk that the immutable blockchain can make assets disappear.
If a token issuer invalidates a token via a smart contract, what happens depends on the implementation of the smart contract and the token’s design.
Many smart contracts emit events (logs) when specific actions happen, such as token invalidation.
Depending on the wallet or app, invalidated tokens might:
Disappear from a wallet (e.g., if the balance is set to zero by the smart contract).
Remain visible but unusable
Should a bank proactive monitor this or wait till it happens?
The problem is not so much that errors can happen but the fact that permissionless blockchain blur the line between internal operations and external environment yet we design operating models where each bank takes the view to make unilateral decision as if this distributed shared ledger made no difference. I think this will need some fine-tuning. On a blockchain, internal operations become externalized—transactions and decisions are visible and auditable by anyone. This makes actions like invalidating tokens or reversing transactions glaringly public creating new reputational risks. In a shared ledger, decision-making is no longer confined to a single entity. Actions by one participant affect others, requiring collective governance mechanisms rather than isolated fixes.
Having said this, why would a bank accept some stablecoin from Société Générale–FORGE (SG–Forge)—a subsidiary that carries higher default risk than the parent company, unless the parent guarantees its performance? Banks typically don’t do that, and SG–Forge is regulated as an investment firm, not a bank. Ugh! (Otherwise I have nothing negative to say about them of course).
Why would banks prefer that over holding a nice, safe Euro in Target2 with their local central bank?
Cost savings?
The transaction cost depends on the network. Right now, an Ethereum transaction costs about $0.62—not that cheap, actually. There are all sorts of alternatives that usually offer lower costs, as long as you're willing to accept more centralization (and we don’t account for the cost to implement this change).
Tokenized Deposits Are No Solution
What alternatives are there? Tokenized deposits do not offer a solution:
"The claim on the credit institution in the form of a tokenised deposit, by representing bank account balances that are nominative, typically cannot be transferred to another individual; only the underlying funds, once withdrawn, can be transferred. A payment using tokenised deposits involves destruction of the claim on the bank of the sender and creation of a new claim on the bank of the receiver, which is cleared by an interbank settlement frequently using central bank money." (EBA 2024)
Such a model can, of course, be designed to burn and mint tokens as they travel from one bank to the next. However, this solution is tokenization in name only. In fact, using tokens here creates new operational problems (e.g., how to reconcile that my burned token was actually minted on the other side) with no real benefit from blockchain technology. Since we’re still using an existing clearing system between two banks, it’s probably simpler to leave it at that.
The Travel Rule Problem
We also face an issue with travel rule data. A tokenized payment instrument bifurcates a single process step, creating new high-risk reconciliation issues.
Private tokenized systems could prove valuable for specific use cases, such as closed-loop networks for corporate treasury management or settlement layers within defined ecosystems. These applications minimize the need for external compliance checks, reducing some of the friction associated with public tokenized systems.
My prediction however is that we’ll see a renaissance recognising the benefit of a separate instruction message and value transfer process, rediscovering the lost beauty of ancient times (i.e., 1973, when SWIFT was born).
But until then, we must first endure the dark ages of ignorance and superstition. For example, the Coinbase Institutional Crypto Market Outlook 2025 Edition identifies stablecoins as the killer app and reflects the dogma of the present day:
"Ultimately, we think the market could potentially evolve towards having many interoperable stablecoins, much like how customers view dollars held at different commercial banks as interchangeable today. In a paper published in the Harvard Business Review (August 2024), Christian Catalini and Jane Wu argue that such an outcome could benefit consumers and businesses by facilitating lower-cost and faster payments, while simultaneously pushing leaders like Circle and Tether to differentiate themselves in a more crowded market."
I found that a bit trivial for Harvard. How should an incumbent respond to more competition? Differentiate! What kind of differentiation strategy would Harvard recommend to Tether? Revise its legal framework, perhaps, to create legally enforceable accountability? We shall never know because it seems Coinbase has summarized and extrapolated from the article to fit its narrative. The original article doesn’t frame the argument the same way.
While the article discusses lower-cost and faster payments, it doesn’t robustly argue or provide evidence that a stablecoin market inherently delivers these benefits. It does note that stricter regulation could limit how issuers create and capture value, but it stops short of suggesting that market competition would necessarily drive innovation. In fact, it leans more toward the idea that regulation might homogenize stablecoins.
And Harvard doesn’t even have to wait for fully interoperable stablecoins (likeCoinbase does)—they argue that all stablecoins are, by definition, interoperable. But what exactly is an interoperable stablecoin, and how does it differ from non-interoperable ones? Any ERC-20 token on the Ethereum network is technically interoperable with any other ERC-20 token. This interoperability is protocol-level and not unique to stablecoins.
The article may conflate technical interoperability (tokens can be used within the same ecosystem) with fungibility(tokens are interchangeable and viewed as equivalent by the market). Even if stablecoins are technically interoperable, compliance requirements (e.g., wallet whitelisting, AML checks) can restrict their use. For example:
A wallet holding USDT on Ethereum may not meet the compliance standards of a bank that only accepts USDC issued by Circle.
Binance’s BUSD is integrated deeply into its ecosystem, encouraging users to stay within Binance rather than moving assets to competing platforms. The more dependent users are on Binance's platform for trading, staking, and DeFi, the less likely they are to move their assets elsewhere. Here is successful differentiation strategy for you Harvard. Perhaps that’s what banks would do leading to a market that becomes fragmented. Customers would face interoperability issues across different bank ecosystems, defeating the purpose of tokenization.
Paxos, which issues BUSD, has positioned itself as a neutral infrastructure provider, but its stablecoins are not universally accepted across all blockchains.
Misunderstandings in the Article
The article, however, doesn’t address these nuances. Statements like “all stablecoins are interoperable” or “fungible” are misleading because they gloss over critical distinctions and make it sound as if stablecoins are inherently superior to other crypto assets, when in reality, they share the same interoperability framework as any ERC-20 token.
Harvard, making a mistake of this caliber? No quality controls, Harvard?
It’s not impossible. The article does show lapses here and there, such as falsely labelling SWIFT a payment network or speculating if Coinbase may one day enter the stablecoin market, when Coinbase has already helped set up USDC and owns a minority stake in Circle.
This article presents a strangely ambitious and somewhat over-optimistic view of stablecoins’ potential to disrupt the financial system without much insight how the financial system works. The authors misunderstand the difference between crypto-native applications (e.g., DeFi) and traditional financial systems. Stablecoins might work well in crypto trading ecosystems, but their integration into regulated financial systems involves entirely different challenges (e.g., AML/KYC, liability for reserves, operational standards).
Finally, they postulate things without evidence or rely on implausible claims—such as banks having greater distribution power than Circle or Tether once they enter the market—describing the outcome as a "doomsday scenario." I’m not so sure this holds true when it comes to crypto assets. At least they stick to the theme of the dark ages: Domesday Books and doomsday scenarios (both share the same linguistic root from Old English dōm, meaning "judgment").
The Golden Age of Crypto
But my favorite fortune teller is actually Bitwise Asset Management, foreseeing that the year 2025 will usher in the Golden Age of Crypto. And of course, a bullish outlook on stablecoins finds its way onto their list. Their quarterly market update (p. 65) includes an interesting impressive graph about crypto adoption by financial institutions (they compare across 5 activities so 1/5 means Bank of America only covers 1 activity):
Bank of America (1/5)
BlackRock (4/5)
BNY Mellon (3/5)
CBOE (1/5)
Charles Schwab (2/5)
Citi (1/5)
CME (1/5)
Deutsche Bank (1/5)
Deutsche Börse (1/5)
Fidelity (4/5)
Franklin Templeton (4/5)
Goldman Sachs (3/5)
HSBC (2/5)
Interactive Brokers (2/5)
JPMorgan Chase (5/5)
London Stock Exchange (1/5)
Mastercard (3/5)
Morgan Stanley (2/5)
PayPal (3/5)
Société Générale (2/5)
UBS (3/5)
Visa 2/5)
JP Morgan winnning the crypto crown? Ha!
They don’t say why this panel is a representative group of the wider financial sector. If the selection was based on public press releases or institutions frequently mentioned in tokenization news (which I suspect it was), it creates a skewed and potentially misleading narrative. Without transparent criteria for inclusion and proper evidence of substantive activity in each category, the graph risks overrepresenting or underrepresenting certain institutions.
Other than Citi, Mastercard and Morgan Stanley, everyone is as of September 2024 in the business of "Crypto Trading and Custody." A footnote clarifies that this includes the trading of crypto spot, futures, and derivatives products. Custody is fundamentally about holding and safeguarding property—assets with clear ownership rights—while contractual rights like those involved in cash-settled derivatives don't require custody in the same sense.
Grouping these activities under the same label (e.g., "Crypto Trading and Custody") conflates fundamentally different aspects of finance.
It could mislead readers into believing that entities offering trading of cash-settled derivatives also provide custodial services for the underlying assets—or that these services are related, which they are not. If that sort of thing interests you, I recommend this world class publication explaining the ins and outs of digital custody.
Let’s have a look. BONY? Based on various articles like this one, it’s unlikely they currently provide general crypto custody services beyond the narrow scope of ETP-related products. This makes Bitwise’s classification questionable unless their dataset includes future commitments or plans rather than current capabilities.
Visa? They operate as a payment processor and network, facilitating the movement of money between banks, merchants, and consumers. It doesn’t engage in trading, custody, or asset management—crypto or otherwise. Visa’s crypto-enabled payments initiatives could justify its inclusion in a category related to payment innovations, but it does not belong under "Crypto Trading and Custody." Hm?
Citi? They don’t seem to do anything other than tokenization. A focused business strategy—there’s nothing wrong with that.
And this time, Bitwise seems to have got it right. Citi went live with a tokenization solution, and if we can believe this press article, they had two projects under Citi Token Services. The cash service is live, while the other is still in ‘beta.’
How does this look? A private internal blockchain network (Hyperledger Besu) is used to facilitate liquidity movement between their New York and Singapore branches—nowhere else for the moment. Clients don’t need to hold or manage the tokens, and their first client said nothing changes in terms of processes or interfaces.
They may or may not use blockchain, but is this an example of ‘tokenization’? The functionality described wouldn’t necessarily require a blockchain—I think that’s a fair statement.
The journalist who wrote this article also seems to have a knack for making bad jokes:
“The bank has also helped to initiate multiple interbank tokenized deposit networks, including the UK’s Regulated Liability Network and the Regulated Settlement Network stateside.”
As if one RLN/RSN wasn’t enough, now there are two of them. Fortunately, I think that was just a misunderstanding.
I do like this press article, though. Citi is quoted with the required dramatic flair:
“By using distributed ledger technology (DLT) and smart contracts, Citi has created a patented programmable payment and liquidity platform, which will reduce costs and streamline processes.”
Duly noted. The article continues:
“Blockchain technology addresses the inefficiencies of traditional cross-border payment systems by using DLT to allow direct transactions with no intermediaries, reducing costs and speeding settlement times, with transactions potentially completed in seconds rather than days.”
Is anybody listening or reading anything anymore?
By the way, Hyperledger Besu is an open-source blockchain platform designed for enterprise applications, with a permissive Apache 2.0 license. While the license allows commercial use, including the creation of proprietary software, patenting something derived from open source often feels a bit at odds. Open-source projects thrive on shared development and collaboration, and moves like this might alienate developers or partners who view blockchain as a commons rather than a space for proprietary control.
Citibank’s patent filing
There is patent filed (May 2024) dealing with distributed system and trade finance "Distributed Database Methods and Systems" which seems to detail a methodology for managing tokenized assets through distributed systems, with features such as partitioning assets temporally or conditionally.
While the patent provides a formal structure for splitting, replicating, and managing assets on a distributed ledger, much of its functionality overlaps with existing blockchain technologies like Ethereum smart contracts or advanced token standards (e.g., ERC-721 or ERC-1155 for NFTs). The primary novelty appears to be in:
Structuring assets on non-value axes (time and condition).
Combining this with a distributed network for secure and transparent ownership transfer.
As with many blockchain-related patents, the question is whether the specific implementation details differ significantly from open-source solutions already in use. Citibank isn’t patenting the concept of escrow, time-based payments, or conditions-based payouts—they’re patenting a particular system or method that implements these ideas using specific technologies, such as distributed ledgers (blockchain) or stateless applications.
In essence, they’re patenting how the old concept of escrow is applied in a blockchain environment—but they didn’t invent escrow, blockchain, or conditional payouts.
On that note, I wish you Happy Holidays and a great start into 2025!