Stablecoins are testing capital controls as digital dollarization reshapes currency crises, challenges central banks and raises new risks.
For decades, a central bank defending a weak currency watched its banks for the first signs of a run: queues at branches, shrinking deposits and a widening gap between the official exchange rate and the street rate. Some of that pressure can now travel through a smartphone, via dollar-linked tokens that move beyond the traditional banking and foreign-exchange networks central banks were built to supervise. A new International Monetary Fund (IMF) working paper by Marco Reuter estimates that when a country enters a currency crisis, its residents' stablecoin holdings rise by roughly 75 percent over the following year relative to comparable countries, and that the new holdings do not fall back. The finding falls well short of showing that central banks are losing control of their economies, yet it suggests that dollarization, the old habit of moving savings into dollars when the local currency fails, has found a faster channel that official statistics barely see.
The figure is an aggregate across 21 crisis episodes, most of them in 2021 and 2022, and the holdings behind it are estimates: the paper attributes wallets to countries by modeling their on-chain behavior, such as time-of-day activity and the exchanges they use, rather than by observing who owns them. In dollar terms, the crisis countries added about US$7.9 billion to pre-crisis holdings of about US$10.4 billion, roughly 0.07 percent of their pre-crisis output, a magnitude the author calls "modest on any macroeconomic yardstick." The paper belongs to a series of research in progress, so it states its author's findings rather than an institutional IMF view. Its importance lies in speed. In Reuter's words, "a channel that barely existed in 2020 delivers first-year dollar accumulation at a pace the deposit margin has not been documented to match."
That speed is the problem emerging-market central banks now face. Dollar stablecoins make it cheaper and quicker to hold something that behaves like a dollar, and the pull is strongest where inflation is high, the currency is sliding, access to foreign exchange is rationed or trust in local institutions is thin. Capital-flow-management measures (CFMs), the legal limits on moving money into or out of a country, were designed around intermediaries that can be identified, licensed and inspected. The danger lies in the combination of digital access, macroeconomic weakness and regulation that reaches the bank but not the wallet. Can a central bank keep effective control when residents can shift into dollar-linked assets faster than the authorities can measure the shift? The answer depends more on data and supervision than on prohibition.
A stablecoin is a crypto asset designed to hold a fixed value against a reference asset, almost always the US dollar. In remarks in Cape Town in August, IMF First Deputy Managing Director Dan Katz set out the market's scale and the policy choices facing emerging markets: market capitalization nearly tripled between 2021 and 2025 and has stayed around US$300 billion for a year, nearly 99 percent of stablecoins are denominated in dollars, and reserves sit mostly in short-term Treasury bills and reverse repurchase agreements, although "some issuers also use less liquid and riskier backing assets." Tether's USDT and Circle's USDC account for roughly 90 percent of the market, according to an IMF departmental paper on foreign-currency stablecoins and macro-financial stability published on October 9. The peg rests on the issuer, which creates tokens when a wholesale counterparty pays in dollars and destroys them on redemption. Ordinary users rarely deal with the issuer; they buy on an exchange and exit by selling to someone else, so a retail holder's exit price depends on market liquidity, not only on the reserves. The United States' GENIUS Act, enacted in July 2025, requires permitted issuers to hold reserves "on an at least 1 to 1 basis" in specified liquid assets and bars them from paying holders interest, which raises the floor for regulated issuers without covering every token in circulation.
Stablecoins perform three functions. The departmental paper finds that settling crypto-asset trades remains the dominant use, while payment use accounts for less than 1 percent of stablecoins outstanding in the estimate it cites. The third, a dollar store of value for people who cannot easily obtain dollars, matters most here. This is digital dollarization: residents using a foreign currency instead of their own to save, pay or price goods, carried out through tokens rather than banknotes or bank deposits. The Bank for International Settlements (BIS), in the stablecoin chapter of its 2025 Annual Economic Report, notes this newer use "as a cross-border payment instrument for residents in emerging market economies lacking access to the dollar" and lists loss of monetary sovereignty and capital flight as major concerns for these economies.
Reuter's public panel gives that concern its first broad measurement. It covers 188 countries and US$21.4 trillion of gross USDT and USDC transfers between 2018 and 2025. Gross means a token entering and later leaving a country counts twice, so the total measures how intensively the rails are used, not how much capital moved on net. Emerging and developing economies receive about 56 percent of gross flows but hold about three-quarters of the roughly US$200 billion in holdings the panel covers at the end of 2025, itself about 78 percent of the US$256 billion then outstanding. Estimated holdings reach US$13.3 billion in Türkiye, US$5.1 billion in Argentina and US$4.1 billion in Nigeria, and the largest relative to output are in Venezuela (5.6 percent of gross domestic product) and Ukraine (4.1 percent), against 0.04 percent in the United States.
The capital-controls evidence covers 217 tightenings of restrictions in 182 countries between 2021 and 2024. Six months after a tightening, the restricting country's cross-border stablecoin activity is about 20 percent higher than in countries that did not act. After tightenings aimed at outflows, the rise reaches about 30 percent and holds near that level a year later, and later loosenings did not reverse it. The author reads this as "consistent with circumvention," and the wording is careful. The outcome is total activity, not net outflows, and the data record where tokens move rather than why. Statistical tests cannot distinguish the outflow response from responses to other kinds of restriction. Run on a commercial dataset from the analytics firm Chainalysis, the same design finds no response, and the two earlier studies that looked directly at stablecoin flows found none either. Against the obvious objection that controls arrive in crises that would raise dollar demand anyway, the paper shows no upward drift before activation and a modest tilt toward corridors where residents already hold bank deposits abroad.
The crisis result carries similar qualifications. Crises are dated with a mechanical rule, a 30 percent twelve-month depreciation that also accelerates, which cannot tell a disorderly collapse from a deliberate float. Fourteen of the 21 onsets fall between mid-2021 and the end of 2022, during one global dollar tightening, and twelve are in Africa and the Middle East. Excluding Russia and Ukraine, whose crises came with war and sanctions, lowers the average effect by about a fifth. The paper cannot separate new dollar demand from demand that would otherwise have gone into cash or offshore accounts. What it does show is timing: holdings did not move before the onset, built through the first year and showed no reversal across the two years observed.
The incentives are ordinary. In Reuter's data, a one-standard-deviation rise in exchange-rate volatility is associated with about 9.7 percent larger inflows, an association detectable only in emerging and developing economies. The departmental paper links larger inflows to high inflation, weak institutions and limited access to foreign-currency accounts. Much of this use is legal: a household protecting savings from depreciation is acting rationally. The same rails also carry money laundering and sanctions evasion, which the departmental paper treats as a separate risk, and treating every purchase as illegal flight would be a category error. Network effects are possible but conditional. Katz noted that dollar stablecoins have gained limited traction in South Africa and rand-linked tokens even less. In Bolivia, according to an IMF working paper by Brandon Joel Tan on stablecoins and fragility in fixed exchange-rate regimes, the central bank reported a twelvefold rise in virtual-asset transactions between July 2024 and May 2025 after lifting restrictions, and the price of USDT in bolivianos became a public reference for the parallel dollar that the central bank itself began publishing.
Follow the money. A resident usually pays an exchange in local currency to buy tokens through the on-ramp and later sells tokens for cash through the off-ramp. Both steps touch a bank or a licensed exchange. Between them, tokens can sit in self-custodial, or unhosted, wallets that no intermediary administers, or move between accounts on one exchange without reaching the blockchain at all. Purchases can also run through over-the-counter desks, peer-to-peer platforms or software-run decentralized exchanges. Brazil shows how incomplete the official picture is. Its central bank, which Reuter believes is the only one that systematically records and publishes such activity, logged US$52.3 billion of resident purchases over 2021–2025 but only US$1.9 billion of sales, against US$34.9 billion of redemptions the paper attributes to Brazil, because sales through global platforms often generate no foreign-exchange contract. Katz warned that local-currency stablecoins could move the conversion into dollars onto the blockchain itself: "Harder to monitor, harder to control." Controls still bite where tokens meet bank accounts, and the departmental paper calls applying the rules for foreign-exchange dealers to crypto intermediaries "central to CFM enforceability." A joint IMF and Financial Stability Board (FSB) synthesis paper cautions that blanket bans "tend to increase the incentives for circumvention."
| Channel | How pressure travels | What the evidence shows | When risk rises |
|---|---|---|---|
| Exchange rate and parallel premium | Large token purchases push the local price of a dollar token above the spot rate | Causal spillovers to spot rates and dollar funding costs in a 27-currency sample | Shallow foreign-exchange markets, stressed intermediaries |
| Bank deposits and credit | Savings leave domestic deposits for tokens whose reserves are invested abroad | Plausible mechanism; Katz reports no clear evidence of disintermediation yet | Tokens funded from deposits rather than cash |
| Fixed and managed exchange rates | A visible token price becomes a common signal of how overvalued the peg is | Theoretical model; Bolivia as a motivating case | Large misalignment, thin reserves |
| Issuer redemption and depegs | Doubts about reserves or claims trigger sales below par | USDC briefly at 90 cents after Silicon Valley Bank failed in 2023; Terra/Luna collapse in May 2022 | Weak reserve rules, no domestic backstop |
| Synchronized exits through ramps | Exchanges and ramps become bottlenecks when holders move at once | Model-based amplification when intermediary capital is depleted | Concentrated, lightly supervised ramps |
The strongest causal evidence concerns the first channel. A working paper by Iñaki Aldasoro, Paula Beltran and Federico Grinberg on stablecoin inflows and foreign-exchange spillovers, using daily data on four dollar tokens traded against 27 currencies on 64 exchanges from January 2021 to November 2025, isolates inflows driven by shocks in other currencies. A 1 percent exogenous increase in net inflows widens the gap between the token price and the spot rate by about 40 basis points, weakens the local currency by about 5 basis points and raises the premium on synthetic dollar funding by 5 to 10 basis points. Roughly one-sixth of the price pressure reaches swap-market funding costs, affecting banks and firms that never touch crypto, and the effects grow when intermediaries have suffered losses. The authors also find that easing market access through regulatory clarity or better on-ramps would narrow the premium without destabilizing traditional markets. For deposits, the departmental paper warns that a shift from domestic deposits into foreign tokens could raise bank funding costs and reduce credit, but Katz pointed to El Salvador, where bank credit and deposits continue to expand. For pegs, Tan's model shows a token price that improves access in calm periods but can coordinate a run when misalignment is large, and the departmental paper notes that pegs may require larger monetary and fiscal adjustments or a new parity if reserves come under too much pressure. For depegs, it adds a constraint no domestic reform can fix: a central bank's ability to provide liquidity backstops for a foreign-issued token in a foreign currency would be "severely constrained," if it exists at all.
The benefits are real. The World Bank's Remittance Prices Worldwide database puts the global average cost of sending money at 6.36 percent of the amount sent, more than twice the Group of Twenty (G20) target of 3 percent. Research summarized in the departmental paper estimates that a US$200 stablecoin transfer through centralized exchanges costs about 2 percent when users choose the cheapest ramps, yet the cheapest traditional provider still wins in roughly one-third of matched corridors, and the token premium pushes costs below zero in some corridors and above 8 percent in others. Investors expect competition: an IMF working paper by Alexander Copestake and coauthors on stablecoins and the future of payments estimates that passage of the GENIUS Act cut the market value of listed incumbent payment firms by about 18 percent, with the largest losses at firms focused on cross-border payments.
Monetary sovereignty, in the BIS definition, is "the ability of a jurisdiction to make decisions and exercise influence over the monetary system within its borders." Stablecoins can erode it in three ways, none yet severe. Policy rates lose traction when savings and credit move into a foreign unit. Seigniorage, the income from issuing money, shifts from the state to private issuers. Liquidity support cannot reach a foreign token. A country can tolerate dollar tokens replacing physical dollars, a substitution Katz judged to carry limited macroeconomic risk, and still need to act when tokens replace local deposits or start to price local goods.
Measurement is the first gap to close. Under the G20 Data Gaps Initiative, the IMF ran a pilot collection on stablecoins between July and December 2025. Of 11 jurisdictions hosting issuers, five reported total issuance and two reported resident holdings and flows; none reported holdings by sector or cross-border breakdown. The South African Reserve Bank compiled custody data from major exchanges and found that households hold most of the coins, while also learning how much sits outside its statistics. Reporting can respect privacy: the departmental paper would share tax-related transaction data only under "strong confidentiality safeguards and clear limits on onward disclosure," and it sets a sensible goal of "not perfect visibility, but a reporting perimeter broad enough to identify currency substitution, capital flow pressures, and financial stability risks in time to respond."
| Horizon | Priority measures |
|---|---|
| Immediate, within existing powers | Reporting by exchanges, custodians and ramp providers on resident holdings and flows; stablecoin scenarios in central-bank stress tests; consumer disclosure on redemption rights and depeg risk |
| Requires legislation or international agreement | Licensing ramps and exchanges serving residents; bringing stablecoins into CFM and foreign-exchange law where circumvention is material; issuer standards on reserves and redemption, kept separate from flow management; information sharing with authorities hosting foreign exchanges and issuers |
| Dependent on fundamentals | Credible monetary policy, sustainable public finances, cheaper domestic payments and adequate reserves, which reduce demand for dollar tokens at the source |
Meanwhile, international coordination is lagging. The FSB's October 2025 peer review of its crypto and stablecoin framework found that, as of August 2025, only five jurisdictions had finalized frameworks for global stablecoin arrangements. Table 1 of the full review lists Argentina, India, South Africa and Türkiye among the 11 with none in place. Katz's warning applies: without cooperation, "activity will simply migrate to jurisdictions with weaker oversight or into unhosted wallets outside the regulatory perimeter."
Three paths are visible, each with observable markers. In regulated integration, ramps are licensed, resident holdings appear in balance-of-payments data and the gap between token and official rates stays narrow. In accelerating digital dollarization, estimated holdings rise as a share of output outside crises, merchants quote prices in tokens, as some Bolivian shops began to do, and local-currency deposits stagnate while token balances grow. In crisis amplification, a depreciation or a tightening of controls is followed within weeks by a widening token premium, rising swap-market dollar premiums and congestion at exchanges, the sequence the Aldasoro and Tan papers describe as possible.
Bans alone will not stop the shift towards dollar stablecoins while inflation erodes savings, access to foreign exchange remains restricted and confidence in local institutions weakens. For central banks, the challenge is no longer simply to regulate a new form of money, but to detect when its growing use begins to undermine domestic deposits, complicate exchange-rate management or amplify a currency crisis. The next test will come when financial pressure builds faster than official data can capture it. By then, the question may no longer be how much digital dollarization has grown, but whether authorities can still contain its consequences.
Apple News, Google News, Feedly, Flipboard, and WhatsApp Channel
