RESEARCH NO.05
2026.08.29
Behind the Coin
Are Treasuries
Stablecoins Do NotBuy America’s 30-Year Debt A New Channel Linking Global Dollar Demand to Treasuries Maturing in 93 Days or Less
- 01 The Law Is Not Yet in Force · The Market Is Already Moving
- 02 The Other Side of One Dollar · Reserves and Interest
- 03 Treasuries Were Bought · But Only at the Short End
- 04 The Difference Between Gross Purchases and Net-New Demand
- 05 The Dollar’s Two Faces · Reserve Currency and the Private Usage Network
- 06 In Korea, the Issue Is Not Treasuries but the FX Channel
The Law Is Not Yet in Force ·
The Market Is Already Moving
The United States had largely treated crypto assets as objects of regulation and enforcement. This was the day it created a separate federal issuance framework for payment stablecoins.
July 18, 2025 · a law was enacted to bring dollar-pegged coins— stablecoins—inside the regulatory perimeter · GENIUS Act.
“This technology will buttress the dollar’s global reserve currency status, expand access to the dollar economy for billions of people around the world, and lead to a surge in demand for U.S. Treasuries backing stablecoins.”
The claim: preserve for longer the world’s use of the dollar as its reference currency.
The claim: even people without a U.S. bank account can hold dollar exposure.
how much issuers bought in reserve assets creating a new source of demand.
The scoring starts with the claim that is easiest to quantify · claim numbers retain the order of the original statement.
Three pillars of the statute
Read together, the design becomes clear · the law does not require issuers to buy Treasuries · it narrows eligible reserve assets to cash,Federal Reserve balances·demand deposits· U.S. Treasuries maturing in 93 days or less · Treasury-backed overnight repo· government money market funds (MMFs)— and a limited set of similar instruments · the narrow menu can channel more assets toward Treasuries as the market grows, but Treasury purchases are not mandatory.1
The relationship with bank deposits also needs to be stated in the right order · the Act defines a stablecoins— payment and settlement digital asset distinct from a deposit and explicitly excludes deposits from the definition · separately, Section 4(a)(11) bars an issuer from paying interest or yield merely for holding, using, or retaining the stablecoin · it is not “not a deposit because it pays no interest”; it is defined as something other than a deposit, and holder interest is then prohibited on top of that.1
The implementation timeline moved one step on August 17
The Act itself is still not in force · Section 20 sets the effective date as the earlier of 18 months after enactment and 120 days after regulators issue final rules · the one-year deadline for rulemaking (July 18, 2026) has already passed · however, on August 17, 2026, the U.S. Treasury released a Notice of Proposed Rulemaking to implement Section 3 moving the process into formal public comment.3
| DATE | SIDE | WHAT HAPPENED |
|---|---|---|
| 2025.07.18 | LAW | GENIUS Act enacted · Section 20 effective-date provision established |
| 2026.04 | LAW | FinCEN ·OFAC joint proposed rule · would require issuers to maintain AML programs and technical capability to freeze and block transactions 4 |
| 2026.07.10 | MARKET | Circle · OCC final approval for a federal trust bank obtained |
| 2026.07.18 | LAW | one-year statutory rulemaking deadline passed · no final rules confirmed by that date |
| 2026.08.03 | MARKET | BlackRock · reserve assets tokenized products designed for stablecoin reserve use launched |
| 2026.08.13 | MARKET | Tether · 2025 annual financial statements released with KPMG audit · unqualified opinion released |
| 2026.08.17 | LAW | U.S. Treasury Section 3 Notice of Proposed Rulemaking · comment period closes Oct. 19 |
| 2027.01.18 | LAW | NPRM footnote states the expected effective date of the Act |
| 2028.07.18 | LAW | Section 3(b)(1) · from this date, U.S. handling of unlicensed stablecoins stablecoins restricted |
January 18, 2027 is the expected effective date of the Act itself(18 months after enactment); the document does not separately specify when the August proposed rule would become effective if finalized · Treasury announced it Aug. 17; it was published in the Federal Register Aug. 18.3
Those that did not wait for the law to take effect
- BlackRock · Aug. 3, 2026 · launched two tokenized products (BSTBL and BRSRV) holding cash, short-term U.S. Treasuries, and Treasury-backed overnight repo · the company said the investment strategy of both funds is intended to make them GENIUS Act under the “eligible reserve assets(eligible reserve assets)"”5
- Circle · July 10, 2026 · received OCC final approval for a federal trust bank · establishing a base for digital-asset custody10
- Tether · Aug. 13, 2026 · released KPMG U.S. audit opinion · unqualified on its 2025 annual financial statements · the company itself explicitly distinguishes this audit from its prior quarterly assurance engagements6
Regulators have not designated the products above as eligible reserve assets· BlackRock has only stated that the products were designed with that objective.
Not all “issuer data” carry the same level of verification · this report treats the four lines below as distinct categories.
| MATERIAL | VERIFICATION | AS OF |
|---|---|---|
| Tether 2025 annual financial statements | KPMG U.S. audit · unqualified opinion | released Aug. 13, 2026 |
| Tether Q2 2026 reserves report | BDO assurance engagement | 2026.06.30 |
| Circle USDC reserve assets reserve report | independent accounting-firm examination | 2026.06.30 |
| Tether cumulative frozen amount | Company self-reported figure | announced Apr. 23, 2026 |
An audit and an attestation/assurance engagement are different procedures · an audited annual financial statement should not be treated as equivalent to a quarterly reserve assurance report · a company’s own tally is a third tier, without external verification.679
the market is already preparing for the post-effective-date regime①②③ WHAT THIS REPORT DOES
stablecoins For how a stablecoin works and how one token is kept near one dollar, see Glossary 06· this report assumes that mechanism and starts with the next question.
An issuer receives one dollar and issues a token ·
where does that dollar go, and what does it do there?
The Other Side of One Dollar
The ban on paying holders interest is widely understood · the less-discussed question comes next.
Where does the Treasury interest that the user does not receive actually go?
The part we can verify in the numbers · Circle
| ITEM | AMOUNT |
|---|---|
| Reserve income | $668mn |
| Distribution, transaction and related costs | $412mn |
| USDC reserve assets USDC in circulation (June 30) | $73.3bn |
Reserve income rose 5% YoY · according to the company, a 25% increase in average USDC circulation was partly offset by a 66 bp bp decline in reserve yield· Treasury interest does not all remain with the issuer; amounts shared with distribution partners appear as expenses.8
And Tether
Major earnings sources include U.S. Treasuries and repo management income · excess reserves, after subtracting liabilities from $187.8bn of total assets, were $4.1bn.7 One caution when citing the amount outstanding · the assurance report’s primary table lists issued tokens at $183.6bn, while contractual redemption value is $184.6bn · the definitions differ, so this report does not combine them.
| reserve assets | TETHER · % OF TOTAL ASSETS | CIRCLE · % OF RESERVE ASSETS |
|---|---|---|
| U.S. T-bills Direct T-bills | 61.23% | 11.62% |
| Treasury reverse repo | 13.64% | 71.62% |
| Cash · deposits | 0.02% | 16.90% |
| Precious metals | 10.03% | N/A |
| Bitcoin | 3.09% | N/A |
| Secured loans | 7.16% | N/A |
Both describe themselves as 1:1 backed, but the jars contain different things · Tether holds Treasuries directly, while Circle is weighted toward repo Treasury-backed reverse repo · the denominators also differ · Tether uses total assets, Circle uses reserve assets reserve assets, so the two columns should not be added together.79
the yield on the reserves accrues to the issuer/distribution network.
Why does this structure tilt toward Treasuries?
When holders are not paid interest directly, reserve income tends to remain a core source of issuer economics · though neither revenues nor costs are limited to reserve management · within the permitted asset list, the instruments that most easily combine yield, liquidity, and scale are T-bill and Treasury-backed repo · cash, bank deposits, and government money market funds can also reserve assets serve as reserves · this is why funds can tilt toward the short end even without a statutory Treasury-purchase mandate.(INTERPRETATION)
※ The statute does not require the purchase of any particular asset · the paragraph above is Snowshagal’s interpretation of the eligible-reserve list and issuer economics.
With money flowing through that structure,
how many Treasuries did issuers actually buy?
Treasuries Were Bought ·
But Only at the Short End
According to a 2026 BIS study · dollar-stablecoins issuers held U.S. T-bills totaling $153bn at end-2025 · estimated net purchases were about $35bn in 2024 and $33bn in 2025.11
It sounds large, but the reading changes with the denominator.
First, compare it with the accumulated stock
$153bn is the end-2025 holding · the denominators below are also end-2025 stocks
$153bn out of $38.514tn
Of every $100 the U.S. owes, about 40 cents · not remotely large enough to alter the debt problem
$153bn out of $30.2713tn
still around one two-hundredth of the total
$153bn out of $6.5473tn
the footprint grows as the maturity window narrows
But the picture changes when measured against one year of net new demand
Here the comparison uses net increase, not outstanding stock · and the numerator is that year’s net purchases, not total holdings
issuer net purchases of $33bn versus a $360.4bn increase in T-bills outstanding
roughly T-bill outstanding net increase of its 9% in size
※ Not a share of primary issuance · the BIS $33bn estimate is the change in issuers’ T-bill positions and includes secondary-market purchases · it does not mean issuers subscribed to 9% of new Treasury issuance · it is a size comparison only.
Why this fourth ratio differs from the three above: the difference between what has accumulated and what is newly added · marginal Treasury pricing is sensitive to new net supply and marginal demand, while reallocations by existing holders, dealer inventories, and collateral demand also matter · the important shift is toward 9.16%.
| DATE | T-BILLS OUTSTANDING | SHARE OF MARKETABLE DEBT |
|---|---|---|
| End-2019 | $2.4169tn | 14.49% |
| End-2022 | $3.6974tn | 15.44% |
| End-2023 | $5.6758tn | 21.52% |
| July 2026 | $6.9889tn | 22.22% |
The U.S. Treasury has shifted the center of gravity of borrowing toward the short end · T-bills’ share of marketable debt 14.49% → 22.22% · the biggest step-up came in 2023 · stablecoins this is precisely the maturity bucket stablecoin issuers buy.12
Stablecoins Do Not Buy America’s 30-Year Debt
93 days or less
- the Act’s maximum maturity for Treasuries eligible as reserve assets
- BIS-estimated price effect concentrated in the 3-month bill
- same direction as Treasury’s expansion of short-term funding
20Y · 30Y
- long-term real yields and term premium
- fiscal credibility and long-run inflation
- demand from long-duration investors
Keep the conclusion in three separate sentences · FACT 1 stablecoins Stablecoins can increase demand for U.S. Treasuries · FACT 2 that demand is concentrated at short maturities · FACT 3 it is not a direct solution to the problems in 20- and 30-year Treasuries.
Did the flow leave a mark on yields? · BIS estimate
In the same study · a $3.5bn inflow into stablecoins lowers the 3-month T-bill U.S. T-bills yield by about 0.71 bp bp on impact, and by as much as about 4 bp bp within ten days · spillovers to longer maturities are limited.11
※ The May 2025 first draft estimated the same ten-day effect at 2–2.5 bp bp · figures in this report use the June 2026 revision · bp 1bp = 0.01 percentage point · because maturity-level details differ across versions, this report relies only on the revised abstract’s broad wording.
what matters is not merely that new Treasury demand exists, but where on the maturity curve it appears· when long-bond demand thins, long yields and the term premium rise, increasing Treasury’s incentive to refinance at shorter maturities · stablecoins stablecoins create demand inside 93 days, but they do not directly address weak long-bond demand,term premium or fiscal credibility.
※ Snowshagal applies Ray Dalio’s debt-cycle framework to the maturity question · Dalio did not directly analyze stablecoins stablecoins in the cited material.
not America’s long horizon, but its shortest one.
Promise ③ holds only at the short end.
One question remains · where did the money come from?
Gross Purchases vs.
Net-New Demand
The numbers on page 3 tell us how much issuers bought in T-bills—and nothing more.
Whether the U.S. financial system as a whole gained the same amount of new Treasury demand is a separate question · reaching that answer requires passing through six steps.
STEP 1 · Was a new coin actually minted?
Higher trading volume alone does not reveal how much reserves or Treasury demand actually changed · stablecoins the market has two layers · in the Federal Reserve researchers’ description, fiat-backed stablecoins issuers generally conduct minting and burning primarily with institutional customers while retail users typically access coins through the secondary-market secondary market.33
- EXISTING COIN secondary-market TRADE · investor A sells existing USDT to investor B · only token ownership changes · total supply is unchanged · issuer reserves are unchanged · the transaction itself creates no new T-bill Treasury demand
- NEW MINT · Mint · dollars or equivalent funds enter the issuer or an authorized primary-market channel and new coins are minted · supply rises · reserves can rise · reserve assets reserve assets must be acquired · funds can be allocated among cash, deposits, government MMFs,repo·T-bill and other eligible assets · only the amount that actually flows into T-bill Treasuries becomes direct Treasury demand
- REDEMPTION · BURN · Burn · tokens are returned to the issuer for dollars · supply falls · reserves decline · redemptions may be met with cash and maturing assets or by liquidating reserve assets T-bills · depending on conditions, this can move in the opposite direction from Treasury demand
But the two layers are not sealed off from one another · secondary-market price and flow pressure in the secondary market can induce primary-market primary-market minting or redemption · primary-market through arbitrage by eligible participants, helping keep the peg intact.3311 In the table below, “direct new T-bill Treasury demand” means the demand created immediately by the transaction itself.
※ The statements that issuers primarily mint/burn with institutional customers and that retail users access coins through the secondary-market secondary market come from Federal Reserve researchers · the statement that secondary-market secondary-market trades do not change outstanding issuance or reserve assets reserve assets” is Snowshagal’s interpretation of that structure.(INTERPRETATION)
| TRADE | ISSUER RESERVES | DIRECT NEW T-BILL DEMAND |
|---|---|---|
| EXISTING COIN secondary-market Buy | No change | None |
| New USDT/USDC mint | Can increase | Can increase |
| Redemption · burn | Decrease | Can decrease · assets may be liquidated |
| Exchange Bitcoin for existing USDT on an exchange | No change | None |
Trading-volume metrics do not distinguish these four cases · therefore this report does not jump directly from “volume rose” to “Treasury demand rose.”
not when the coin trades, but
when reserves increase and that incremental money is allocated to Treasuries.
STEP 2 · Where did the new money come from?
| SOURCE | POSSIBLE EFFECT |
|---|---|
| U.S. bank deposits | More likely a change in the location and composition of deposits · different from genuinely new system-wide demand |
| Government MMFs | If the fund already held U.S. T-bills the same T-bills, the wrapper may change without much new underlying demand |
| Foreign local currency · cash | Closest to genuinely new offshore dollar and T-bill Treasury demand |
| Proceeds from selling crypto assets | May simply represent reallocation within crypto |
| Corporate payment · remittance funds | Creates reserve demand for as long as the funds remain in stablecoin form |
No public data were found that quantify the shares by source · this table maps paths rather than magnitudes.
STEP 3 · How much of the reserve increase becomes T-bills?
Not all incremental reserves go into T-bill T-bills · permitted destinations include cash,Federal Reserve balances·demand deposits· government money market funds·Treasury-backed repo· Treasuries maturing in 93 days or less T-bill · even when the same dollar enters reserves, the portion counted as Treasury demand depends on where it is allocated.
- It differs by issuer · as of June 30, 2026, Tether directly held 61.23% of total assets in T-bill· Circle held reserve assets of its 11.62% in direct Treasuries and T-bill 71.62% Treasury reverse repo (page 2)
- repo shows why reverse repo is closely tied to Treasuries but is not the same thing as a direct Treasury holding · reverse repo is a collateralized transaction, not an outright Treasury purchase · an increase in stablecoin supply does not imply an equal-dollar increase in Treasury holdings
- It changes over time · the same issuer can change allocation as rates and maturity structures shift
pure net-new demand is only partly observable①②③ GROSS IS NOT NET
Do Bank Deposits Disappear?
「stablecoins The claim that stablecoins destroy bank deposits one-for-one does not hold · money transferred from a U.S. bank deposit into a stablecoin can flow back into the banking system through the issuer’s bank deposits, Treasury dealers, repo or other financial institutions · the data fit better with a shift in the location and character of liquidity than with outright destruction of deposit volume.
※ Deposits use the Federal Reserve H.8 series; stablecoin market cap uses the BIS Annual Report · they are different datasets from different institutions and are not treated as one time series · the ratio is only a scale comparison.14
What was observed · New York Fed bank-level analysis
| ITEM | CHANGE |
|---|---|
| Interbank payment activity (9 months after relationship begins) | +67% |
| Additional reserve balances held | about $1.5bn |
| Loan-to-asset ratio (vs. control group) | -14%p |
The direction of interpretation is clear · stablecoins banks receiving stablecoin-related deposits tend to keep more liquidity to accommodate rapid minting and redemptions rather than aggressively expanding lending · the deposits do not disappear; the receiving bank operates them more conservatively.26
※ The New York Fed Staff Report is preliminary research and reflects the authors’ views · the paper explicitly states that it is not the official position of the New York Fed or the Federal Reserve System.
The counterevidence matters too
- Federal Reserve researchers’ historical comparison · banks have historically adapted to innovations such as money market funds online payment platforms through regulation and product innovation · their conclusion: when innovation responds to market demand, meeting that demand is more effective than resisting it27
- Separate aggregate volume from internal structure · another note in the same series says that even if aggregate deposit volume remains broadly unchanged, the underlying structure of deposits can shift substantially · with implications for funding stability, liquidity risk, and credit supply27
More reallocation of liquidity than disappearance of deposits · the evidence does not support a one-sided conclusion yet.
That completes the Treasury side of the ledger ·
what happened to the dollar itself?
The Dollar’s Two Faces
“Dollar dominance is eroding” and “the dollar remains overwhelmingly dominant” can both be true · because the two statements refer to different rooms.
One is the vault of central banks; the other is the network through which trade and remittances flow.
Central-Bank Vaults
The dollar’s share of global foreign-exchange reserves held for official purposes
declined gradually over a quarter-century · among foreign holders of U.S. Treasuries, the official-sector share—central banks and similar institutions—also fell from 69.6% in 2015 to 45.1% in 2025.17
Trade and Transaction Network
Share of global FX trades with the dollar on one side
actually rose · the highest level in the history of the BIS Triennial Survey · the currency used to invoice trade remained broadly stable through 2023.18
The two lines move in opposite directions · so which room do stablecoins belong to? · the BIS addressed this directly in May 2026.
“Stablecoins are unlikely to directly affect the functions of unit of account, intervention currency, or reserve-currency in the near term. Their impact is most likely to emerge in the private-sector functions of store of value and medium of exchange.”
The reason is straightforward · stablecoins stablecoins are held by the private sector, not central banks · roughly 98% of stablecoin value is dollar-denominated, and the same paper describes the effect as reinforcing the existing currency hierarchy · holding dollars without a bank account is labeled “stealth dollarization19
“Distributed ledger technologies and tokenized assets, such as stablecoins, are creating new channels for global dollar intermediation that operate alongside—and at times together with—traditional banking and payment systems.”
In the same remarks, Waller separately identified the foundations of the dollar’s role · the size and depth of the U.S. economy and financial markets, and trust in U.S. institutions and the rule of law · those conditions remain critical even as the surrounding environment changes rapidly.20
they are new plumbing laid on top of the existing foundation.
The Three Layers Inside $35 Trillion
This is where Promise ② is scored · the most frequently cited market-size figure is annual transaction volume · but that number contains three layers.
※ The 1% figure compares one specific classification of payment-related flows with gross volume and classification standards vary by institution · adjusted volume is not the same as payment volume · adjusted volume can still include exchange flows, decentralized exchange activity, lending, mint/burn activity, and conversion gateways.1314
The Financial Stability Board puts it more simply · “crypto-assets and stablecoins are not widely used in financial services supporting the real economy.”15
But the other side is also true
- Rankings reverse when flows are measured against GDP · North America had the largest absolute inflow at $633bn · relative to GDP, Latin America and the Caribbean 7.7%, and Africa and the Middle East 6.7% rank higher (2024 data)
- Nigeria alone received $59bn · crypto-asset inflows over the year beginning July 2023 · about 60% of sub-Saharan Africa’s stablecoins inflows
- Capital controls do not appear effective · in a BIS study of more than 130 countries, account restrictions reduced deposit dollarization dollarization by about 1.3 percentage points, but had stablecoins no statistically significant effect on stablecoin inflows21
- But the scale remains small · in the same study, Latin American stablecoins stablecoin inflows were 0–4% of GDP · foreign-currency deposits in those countries were several multiples larger
Why the totals differ—and why this report uses only one ratio
Transaction volume changes dramatically depending on what is counted · the key issue is how much internal movement and automated trading are removed · the BIS Annual Report estimates 2025 gross volume at about $28tn· BCG narrows more than $62tn of raw flows to $4.2tn of adjusted volume and estimates real-economy payments at $35bn–$55bn· Visa’s dashboard adjusts 30-day volume from $3.9tn to $817.5bn.14
The headline ratio in this report uses one matched pair from the BIS General Manager’s speech ($35tn and $390bn) only · mixing a numerator from one source with a denominator from another manufactures a ratio that neither source actually reports · the ratio changes with the methodology · but all methodologies point in the same direction: flows closest to payments for goods and services are much smaller than total stablecoin on-chain activity.
Are remittances really cheaper? · Bank of Italy field test
On March 24 and 26, 2026, the Bank of Italy actually sent 200 USDC across ten corridors linking Italy with Argentina, Brazil, South Africa, the UAE, and Japan · total cost ranged from 0.30% to 8.96% · the blockchain leg itself averaged just 0.4%; most of the cost came from the fiat-to-coin and coin-to-fiat on/off ramps· the central bank’s conclusion: there is “no systematic cost advantage.”16
Compared with World Bank country-average remittance costs, the Brazil corridor was far cheaper (2.21% vs. 9.96%), while the UAE corridor was more expensive (8.95% vs. 2.65%) · but the first numbers are measured costs on specific test corridors and the second are country averages · the paper itself notes that this is not a like-for-like comparison.
The road widened · the number of people who actually traveled it cannot yet be counted.
No reversal in official reserve-currency share yet · but the direction of the private dollar network is clear.
The Freeze Button on
Permissioned Dollars
One concern follows naturally: U.S. sanctions have traditionally worked through the banking system · if dollars can circulate globally outside banks, does Washington lose enforcement power?
But permissioned dollars with a centralized issuer have produced evidence in the opposite direction.
U.S. Treasury Office of Foreign Assets Control (OFAC) added Iranian central-bank wallets to the sanctions list, and the issuer immediately froze the stablecoins in those wallets · the wallets had received $165mn in total.
GENIUS Act Section 4 permits issuance only by issuers with the technical capability to comply with lawful orders · the definition includes seizure, freezing, burning, and blocking transfers · an April 2026 joint proposal from FinCEN and OFAC would codify this capability alongside AML obligations.
Company self-reported figure · Tether says cumulative frozen funds exceed $4.4bn, of which more than $2.1bn is linked to U.S. authorities, and that it has cooperated with 340 law-enforcement agencies in 65 countries · these figures are not government-verified and differ from on-chain analytics estimates · their scope also differs from joint task-force figures.23
In its release on the disruption of the Russian exchange Garantex, the U.S. Department of Justice formally thanked the issuer for proactive cooperation · the exchange processed at least $96bn in crypto transactions since April 2019.22
| WHAT WORKED | WHAT DID NOT | WORKAROUNDS |
|---|---|---|
| Issuer freeze · burn capability | Complete suppression of offshore trading | Non-dollar stablecoins |
| Law-enforcement cooperation | Control of self-custody wallets | Offshore issuance networks |
| U.S. licensing gate | Universal enforcement beyond U.S. borders | Ruble-linked coins and other alternatives |
The freeze button depends on permissioned dollars with a centralized issuer a centralized, permissioned dollar issuer · it does not imply control over every stablecoins stablecoin or every offshore transaction.
The record of a workaround—and where that workaround ran into a wall
The stablecoin heavily used for Russian sanctions evasion was not dollar-linked but stablecoins A7A5, pegged to the ruble · cumulative processing above $100bn was reported in January 2026 · but what followed moved in the opposite direction.
- New issuance stopped · after July 2025
- Trading volume collapsed · down about 96% by June 2026 from the July 2025 peak, according to analytics estimates
- Distribution channel shut down · its main exchange effectively ceased operating after an April 2026 hack · reported losses range from $13.7mn to $15mn depending on the outlet
- Regulatory strategy changed · in a package effective May 24, 2026, the EU shifted from designating individual platforms to a blanket ban on transactions with Russian and Belarusian crypto-asset service providers · ruble-linked coins and the digital ruble were added to the list as well24
※ A7A5 volume estimates differ by about 63%: the issuer claims roughly $205mn per day, while an analytics firm estimates about $75mn · the analytics firm classifies about 34% of observed activity as circular trading · neither figure has been independently verified · the $100bn cumulative figure from January and the 96% decline through June refer to different dates.
The U.S. Treasury’s official logic sits in the same place · in a February 2026 speech, Bessent said reserve-currency status “strengthens sanctions actions,” while arguing that a well-regulated dollar stablecoins stablecoin market can reinforce the dollar’s global role.25 The design is not a trade-off between reach and control, but an attempt to pull wider circulation inside a regulatory perimeter and expand enforcement alongside it.
non-dollar workarounds are growing.
The gaps that remain · evidence on the other side
In April 2024, then-Deputy Treasury Secretary Wally Adeyemo told the Senate that Russia was using Tether to evade sanctions and asked for secondary-sanctions authority targeting offshore actors · a record of the U.S. government itself acknowledging gaps in its existing powers.25
In a March 2026 report, the Financial Action Task Force warned that a significant share of illicit crypto activity in 2025 involved stablecoins stablecoins · saying sanctions can work is not the same as saying there are no gaps.24
※ No U.S. primary-source document was found that specifically identifies North Korean money laundering as conducted via stablecoins stablecoins · Treasury releases use only the broader term “crypto-assets” · there is also no officially confirmed amount for crypto settlement of Russian oil payments.
So
What Did the U.S. Actually Reinforce?
OFFICIAL reserve-currency share has not rebounded · the BIS concludes that a direct near-term effect is unlikely · meanwhile, the dollar remains at record-high levels as a private transaction and settlement currency, and stablecoins push in that direction.
Counterpoint · expanding the private dollar network is not the same thing as a rebound in the official reserve-currency reserve-currency share.
Access to digital dollars for users outside the U.S. has expanded materially · however, while trading volume has grown sharply, flows closest to payments for goods and services remain a small share · public statistics that count actual users by country and real-economy usage remain limited.
Counterpoint · a wider access channel and the scale of actual use are different things · in some emerging markets, inflows are already meaningful relative to GDP.
Issuer T-bill holdings and net purchases increased · but they amount to only 0.40% of total federal debt, are concentrated inside 93 days, and not all gross purchases can be shown to represent new external money · the long-duration demand problem remains.
Counterpoint · compared with the 2025 T-bill outstanding net increase in T-bills, issuer net purchases were equivalent to 9.16% in size · not negligible at the short end.
Issuers did buy Treasuries · but we still cannot establish that all of that money represented genuinely new demand.
The buyer base for short-dated paper has grown · but these are not buyers of America’s long-duration debt.
OFFICIAL reserve-currency Stablecoins have not reversed the decline in the official reserve-currency share · instead, they have widened the private channels through which dollars are used.
There are Treasuries behind the coin · but they are less the long bonds that would rescue U.S. fiscal financing than
the shortest securities widening the dollar network.
The paragraph above is Snowshagal’s interpretation of public data and should be distinguished from the official positions of the U.S. government and international organizations · figures and quotations on the preceding pages are drawn from primary sources.
What would falsify this view
In Korea, the issue is not Treasuries but the
foreign-exchange channel
What assets does the money buy?
- T-bill demand and reserve assets
- expansion of the dollar distribution network
- regulatory and sanctions-enforcement capacity
How easily can the money cross the border?
- dollar access and cross-border capital movement
- the channel between the won and the dollar
- bank deposits · FX markets · monetary sovereignty
First, three datasets that must not be mixed
| SERIES | WHAT IT MEASURES | LATEST VERIFIED FIGURE |
|---|---|---|
| Holdings | Market value held by investors on five Korean exchanges | KRW 607.1bn · end-Feb. 2026 |
| Net cross-border outflow | outbound transfers minus inbound transfers | KRW 560.3bn · June 2026 |
| Gross inflow · outflow | gross transfers in each direction | KRW 35.3tn each · Nov. 2024–Feb. 2025 |
The three series differ in institution, reference date, and definition · stocks are not read as flows, and flows are not read as stocks · holdings come from the Bank of Korea; net outflow and gross in/out flows come from FSS/FSC datasets.282930
The money did not simply leave; it moved both ways
In an April 2025 explanatory release, the Financial Services Commission said domestic inflows and overseas outflows of dollar stablecoins were both about KRW 35.3tn and similar in size making it difficult to characterize stablecoins as a one-way capital-flight channel.28 It is closer to a two-way corridor between Korea and overseas markets than to a “dollar exodus.”
Still, the net direction has leaned outward · according to FSS data submitted to the National Assembly, there were 18 consecutive months of net outflow from January 2025 through June 2026, totaling KRW 14.9246tn · monthly average KRW 829.1bn · June 2026 net outflow alone was KRW 560.3bn.29
※ Net outflow is the difference between cross-border outbound and inbound transfers not a decline in Korean users’ holdings · this figure is not linked causally to the KRW 607.1bn holdings figure above · the two numbers differ by institution, reference date, and definition.
A different path built in Korea
Earlier, on page 5 we noted one BIS finding: across more than 130 countries, account-based restrictions do not stop stablecoins stablecoin inflows · Korea, however, shows a different path · the Bank of Korea’s 2025 Financial Stability Report says that a real-name verified account KRW trading structure has become entrenched, so stablecoins stablecoins are barely used as an intermediary in domestic crypto trading31
The difference is the type of regulation · the BIS found little effect from restricting accounts · Korea did something different: rather than locking the account, it laid the entry rail into exchanges in won · once users are accustomed to buying and selling directly in KRW, there is less reason to route through a dollar stablecoin · this is less a blockade than the prior construction of another road.(INTERPRETATION) One country does not overturn an average across more than 130 countries; establishing causality would require separate evidence.
- Five KRW markets · Upbit, Bithumb, Coinone, Korbit, and GOPAX each operate with one partner bank and real-name verified accounts · partner-bank renewals have proceeded sequentially during 202632
- The motive to hold dollar exposure remains · the Bank of Korea also notes that some investors hold dollar stablecoins stablecoins to obtain dollar-denominated assets31
- reserve assets Reserve-asset direction is the same · among major stablecoin stablecoins reserve assets reserve assets, $172.8bn (81.5%) was managed primarily in U.S. Treasuries as of end-Q1 2025, including repo 31
Domestic USDT holdings · domestic USDC holdings · total dollar stablecoins stablecoin holdings of Korean residents · no public statistics break holdings out by token, so this report cannot state how much dollar stablecoin Koreans actually hold · balances in self-custody wallets outside exchanges are absent from all of these statistics.
For a country that uses the dollar: a new foreign-exchange channel.
The same coin, two different ledgers.
but they are less the long bonds that rescue U.S. fiscal financing than
the shortest securities that widen the dollar network①②③ THE CLOSING LINE
Glossary & Sources
Tap the underlined terms in the body for definitions · the sections below provide the full glossary and supporting sources · primary reliance is placed on international organizations, central banks, government disclosures and statutory text; where a point could not be cross-checked against a primary source, that limitation is stated in the body · each source is labeled by type.
24 Terms · Meanings Used in This Report
33 Sources · Including Source Type and Verification Limits
govinfo · PLAW-119publ27
home.treasury.gov · sb0197
home.treasury.gov · sb0605
federalregister.gov · 2026-16796
federalregister.gov · PPSI AML/CFT
federalregister.gov · OCC 2026-04089
businesswire · BlackRock
tether.io · audit
tether.io · Q2 2026
circle.com · Q2 2026
circle.com · transparency
circle.com · OCC approval
Version Note · the May 2025 first version reported a 2–2.5bp effect within ten days; the body uses the revised version · maturity-specific detail differs across versions, so the body uses only the broader language in the revised abstract · authors’ views; not the official position of the BIS
bis.org/publ/work1270.pdf
fiscaldata.treasury.gov · MSPD
bis.org/speeches/sp260420.pdf
Caution · 2025 total transaction volume varies across sources at $28tn, $35tn and $62tn, with different adjustment methodologies · the body uses numerator and denominator from the same source and presents the others only in the expanded notes
bis.org · AER 2026 ch3
visaonchainanalytics.com · methodology
fsb.org · P240326
Nature of Comparison · in Brazil 2.21% vs. 9.96%, and UAE 8.95% vs. 2.65%, the first number is the measured cost for a specific corridor in this experiment and the second is a World Bank country average · the paper itself says this is not a strict like-for-like comparison
bancaditalia.it · N.86-MISP
data.imf.org · COFER
bis.org/statistics/rpfx25_fx
bis.org/publ/bppdf/bispap170.pdf
federalreserve.gov · waller20260622a
bis.org/publ/work1370.pdf
justice.gov · Garantex
tether.io · freeze
Measurement Dispute · average daily volume differs sharply: about $205mn claimed by the issuer versus about $75mn estimated by the analytics firm; the analytics firm classifies about 34% of observed transactions as circular activity · neither figure has been independently verified by a third party · FATF’s March 2026 report warns about the role of stablecoins in illicit crypto-asset activity
fatf-gafi.org · Offshore VASPs
home.treasury.gov · sb0403
home.treasury.gov · jy2243
Disclaimer · "This paper presents preliminary findings … The views expressed in this paper are those of the author(s) and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System." · preliminary research; not the institution’s official position
newyorkfed.org · sr1185
federalreserve.gov · FEDS Notes 2025.12
federalreserve.gov · FEDS Notes 2026.05
fsc.go.kr/no010102/84385
Verification Limit · the original National Assembly submission was not obtained; figures were verified through August 2026 reporting, a secondary-source route
Verification Limit · the original National Assembly submission was not obtained; figures were verified through May 2026 reporting, a secondary-source route
bok.or.kr · Financial Stability Report
Verification Limit · exchange-by-exchange contract status was verified through August 2026 reporting, a secondary-source route
Scope of Attribution · the sentence “secondary-market trading does not change tokens outstanding or reserve assets” is not stated verbatim in the source · it is Snowshagal’s synthesis of the mechanics described above, and page 4 explicitly labels that distinction · authors’ views; not the official position of the Federal Reserve Board or the Federal Reserve System
federalreserve.gov · FEDS Notes 2024.02.23