Three Tokenized Assets, Three Different Money Questions
The subject here is the contrast among tokenized deposits, tokenized money-market funds, and payment stablecoins. They may all be described as digital dollars, but they are not the same claim, do not give the holder the same path to cash, and do not arrive in the monetary statistics through the same reporting channel. Comparing them on those common dimensions makes the practical judgment clear: the settlement rail matters, but the liability, dominant use, conversion path, backing, and geographic reach matter more.
A bank deposit and a fund share do not become the same claim
Start with the two cases that are easiest to confuse because both can be represented on a blockchain. A tokenized deposit is a digital representation of a traditional depository-institution deposit. The bank still owes the holder the deposit. It remains on the bank’s balance sheet, is legally a bank deposit, falls under state or federal banking regulation, and is eligible for deposit insurance. The token changes how the deposit is recorded and transferred; it does not turn the bank’s liability into a fund share or a free-standing cryptoasset.
A demand-style tokenized deposit also preserves the functional traits associated with M1. The Federal Reserve note describes instant bank-mediated dollar payments, available 24 hours a day, 7 days a week, 365 days a year, and immediate conversion into cash without selling an asset or waiting for redemption. Institutional customers are using such deposits for continuous interbank settlement, corporate treasury operations, cross-border payments, and securities collateral management. Those are uses of a checking-like balance: the holder keeps the claim ready to pay, not primarily to earn a market return.
That is why a demand-style tokenized deposit can remain inside M1 rather than becoming a new aggregate component. The bank liability is the economic object being measured. A new rail may make the liability more useful for a particular kind of settlement, but it does not create an additional dollar of private money merely because the representation changed. In this case, the economically conservative conclusion is also the analytically useful one: keep the balance with deposits while monitoring whether its new circulation patterns change what the aggregate tells us.
The answer changes if the deposit is a time deposit. A tokenized time deposit can be moved on a blockchain, but its purpose remains short-term saving and its access conditions remain those of a time deposit. When it fits the existing small-time-deposit category, its natural home is the non-M1 portion of M2. A transfer of the claim may be quick; that does not make the underlying claim immediately spendable in the way a demand deposit is. The holder may still be choosing a return and accepting a constraint in exchange for that return.
Tokenized money-market funds are different again. They are digital representations of shares in money-market mutual funds, not bank deposits. The funds have the same low-risk backing as conventional money-market funds and target a one-to-one net asset value. They remain subject to the Securities and Exchange Commission’s Rule 2a-7, which governs the quality, liquidity, and diversification of the instruments such funds may hold. Their legal and economic relationship is therefore to a fund and its portfolio, not directly to a bank’s promise to repay a deposit on demand.
The key comparison is not whether the fund share can travel quickly. Tokenized money-market-fund shares can transfer instantly on blockchain networks, but converting a share into dollars still requires redemption, typically taking one to two business days. A recipient may receive a transferable investment claim immediately while still lacking ordinary dollars for a purchase. The token has improved the movement of the share; it has not removed the fund’s conversion step or changed its short-term-savings function.
That is why tokenized money-market funds are currently treated as part of the retail money-market-fund component of non-M1 M2. Their market-linked return and redemption lag point to store-of-value use. The classification could be reconsidered if settlement, collateral management, cross-border payments, and on-chain lending became the dominant uses, but that is a conditional possibility rather than a conclusion about every tokenized fund.
The reporting picture reinforces the distinction. Depository institutions already report tokenized deposits together with traditional deposits. The aggregate can therefore capture the deposit balance through an existing channel, although routine data do not reveal how much adoption is occurring specifically on blockchain rails. Money-market-fund data come through the Investment Company Institute, which distinguishes retail from institutional funds but does not identify whether a particular fund has been issued in tokenized form. The fund category can be counted while the tokenized subset remains invisible.
Imagine two institutional users choosing between these forms for the same dollar amount. A treasury team that needs to settle an obligation immediately is concerned with whether the balance can be paid directly and converted to cash without a sale or redemption. A team willing to wait for redemption in exchange for a market-linked return is using a different kind of claim, even if both balances can be displayed in digital wallets and transferred between approved parties. The comparison is not a recommendation about either product; it isolates the economic choice that the aggregate is trying to capture.
The backing also has different statistical consequences. A tokenized deposit is backed within the bank’s balance sheet by the bank’s assets, including liquid and less-liquid assets that are not themselves included in the monetary aggregates in the same way. The Federal Reserve note therefore identifies no double-counting risk from the tokenized deposit representation. The deposit is already part of the reported bank total. A money-market fund, by contrast, can hold bank demand deposits among its investments. Those deposits are already in M1, so the fund’s net assets can overlap with an existing component. The note regards the likely amount as small because deposits yield little and funds have an incentive to minimize them, but the overlap is still a property to evaluate rather than ignore.
These distinctions show why a new “tokenized money” line would be a poor substitute for analysis. It could mix an insured bank liability, a regulated fund share, and a private dollar-linked token merely because their records use similar technology. It would also hide the reason an asset is counted: direct payment access, a savings function, or a conditional use that has not yet become dominant. A more informative statistical system can preserve the existing aggregate placement while adding indicators of tokenized form or adoption where those indicators answer a separate analytical question.
The different relationships can be held in view at once:
The diagram’s practical lesson is that “tokenized” is not a common economic category. It describes a representation shared by claims with different issuers, protections, conversion paths, and purposes. The same word can cover a bank balance that is already M1, a savings product that remains in M2, and a stablecoin whose placement has not been settled.
Payment stablecoins sit at the unresolved boundary
Payment stablecoins raise the harder question because they combine direct on-chain transfer with a private reserve and redemption promise. The Federal Reserve note limits its analysis to payment stablecoins issued by permitted issuers under the U.S. GENIUS Act. Within that scope, the coins are designed to maintain a one-to-one value with the dollar and to be backed by relatively safe assets such as Treasury bills and bank deposits. Unlike a deposit, the coin is not simply the same bank liability in a new format. Unlike a money-market-fund share, it is designed to circulate as a token on a public blockchain and can support peer-to-peer dollar payments continuously.
Yet direct transfer does not settle the classification. The note says payment stablecoins cannot pay interest directly to holders, although indirect compensation may be possible. Their use still determines the relevant function. If holders mainly use them to move in and out of crypto markets or to hold a short-term dollar-like position, that behavior resembles a store of value and points toward non-M1 M2. If households and businesses use them broadly for ordinary payments, the token’s immediate payment function could support M1. The asset can have payment capability before payment use becomes its dominant economic role.
Redemption adds another uncertainty. Payment stablecoins are redeemable for dollars, but the GENIUS Act does not establish how quickly conversion must occur. The Federal Reserve note discusses a proposed OCC rule under which “timely redemption” would mean two business days, with an automatic extension to seven calendar days if redemption demands exceeded 10 percent of outstanding issuance in a single 24-hour period. Those provisions describe a proposal and a possible stress condition, not a settled operating fact for all permitted stablecoins. A token that transfers instantly on-chain may still have a slower or more conditional route back to bank money.
The reserve question is equally important. A stablecoin balance may look like an additional dollar claim, while its reserves include assets already represented elsewhere in the monetary aggregates. Bank deposits and money-market-fund assets are examples of backing that may already be captured in M1 or M2. The right question is not whether the stablecoin exists—it does—but whether adding its full outstanding balance would count both the private token and the monetary assets supporting it. Treasury securities create a different overlap question from bank deposits or fund shares, so the answer depends on reserve composition rather than on a generic “fully backed” label.
The reporting challenge follows from that composition. The GENIUS Act’s monthly reserve reports could provide a starting point for estimating the amount in circulation, but they would need standardized fields compatible with the Federal Reserve’s compilation systems. Someone would also need to decide whether reports should go directly to the Federal Reserve or pass through another federal or state regulator. A reserve disclosure that is adequate for legal compliance may not yet be adequate for a monetary statistic designed to be timely, comparable, and consistent across issuers.
Geography prevents a final shortcut. A stablecoin issued by a U.S.-regulated entity can circulate through public blockchains beyond the United States. Reserve data can show something about issuance, but not automatically identify the country in which a token is being used or held. The Federal Reserve note therefore treats geographic scope as an unresolved practical consideration. Counting every outstanding token as U.S. money would measure a global private dollar claim, not necessarily money available through the U.S. financial system.
The comparison gives each criterion a distinct job. Legal liability tells us what kind of claim the holder owns and which protections or obligations surround it. Dominant function tells us whether the claim behaves mainly as payment money or savings. Cash conversion tells us whether transferability is genuinely immediate spending access. Backing tells us how stability is supported and whether existing assets may be counted twice. Reporting and geography tell us whether the balance can be placed into a meaningful statistic at all.
No single criterion dominates every case. A tokenized demand deposit can pass the legal, functional, conversion, and reporting tests while still needing a separate adoption indicator. A tokenized money-market fund can be easy to categorize but difficult to identify separately from conventional funds. A payment stablecoin can have a strong payment interface and one-to-one reserves while remaining conditional because its dominant use, redemption practice, reserve overlap, and geographic reach are still developing. Comparing those differences is more informative than ranking the three by transaction speed. The next useful step is to test the model in a compact set of cases, especially the tempting but mistaken inference that an instantly moving token must be M1.