Gita Gopinath, the Gregory and Ania Coffey Professor of Economics at Harvard University and former First Deputy Managing Director of the International Monetary Fund (IMF), delivered the annual Per Jacobsson Foundation Lecture on June 28, 2026, in Basel, Switzerland, on the occasion of the BIS Annual General Meeting (AGM).
The lecture, titled “Stablecoins and Anonymous Money,” laid out what may be the most data-driven challenge yet to the assumption that new stablecoin laws will meaningfully curb financial crime.
The central thesis is simple but striking: public policy has spent decades pushing traditional money toward transparency, but the stablecoin ecosystem is drifting rapidly in the opposite direction, toward the most anonymous form of holding and transacting.
The anonymity gap between traditional money and stablecoins
Gopinath’s lecture drew on original research co-authored with Kyle Calder, Wenxin Du, and Jeremy Stein, using data from blockchain analytics platform Artemis Analytics. The research quantifies, for the first time, the relative importance of self-custody wallets versus centralized exchanges (CEXs) in stablecoin holdings and transfers.
The numbers paint a stark picture. In the traditional dollar system, roughly $2.5 trillion exists as physical cash, the most anonymous form of money. Bank deposits in U.S. institutions, the least anonymous form, total over $19 trillion. Dollar deposits held in banks outside U.S. jurisdiction add another $14.5 trillion. Cash, in other words, accounts for roughly 7% of total traditional dollar money. The system is overwhelmingly dominated by its most transparent layer.
Stablecoins flip this ratio entirely. Of the approximately $315 billion stablecoin market, between 70% and 75% of USDC and USDT holdings sit in self-custody wallets, where the holder controls the cryptographic keys and the owner’s identity remains pseudonymous. For Tether’s USDT alone, around $130 billion of its $190 billion total supply is held in self-custody wallets.
Roughly $60 billion sits on exchanges, and of that, approximately $50 billion is held on exchanges outside U.S. regulatory jurisdiction. For Circle’s USDC, about $61 billion is in self-custody and around $20 billion at exchanges, with the vast majority on non-U.S. platforms.
The Tron blockchain tells the most extreme version of this story. Of the $89 billion in USDT on Tron, approximately 95% is held in self-custody wallets.
Transfers are following holdings into anonymity
The data on how stablecoins move reinforces the holdings picture. Analyzing USDC and USDT transfers on Ethereum from January 2023 to March 2026, Gopinath’s research found that wallet-to-wallet transfers, where neither end passes through an identified intermediary, have grown from a modest share a few years ago to roughly half of all volume in the most recent data.
For USDC, self-custody-to-self-custody transfers rose from 15.4% of Ethereum volume in 2023 to 46.8% in 2026 year-to-date (YTD), while exchange-to-exchange transfers fell from 45.1% to 14.5% over the same period. For USDT, self-custody-to-self-custody transfers reached 48.7% by 2026 YTD, while exchange-to-exchange transfers dropped to just 16.5%.
Total Ethereum gross transfer volume across the study period was $5.1 trillion for USDC and $4.6 trillion for USDT.
The jurisdiction breakdown sharpens the point further. On Ethereum in 2025, U.S.-regulated exchanges and issuers appeared on one side of only about 8% of USDT transfer volume and about 27% of USDC volume. Self-custody wallets appeared on at least one side of more than 60% of USDT transfers and more than half of USDC transfers. The chains carrying the highest volume, Gopinath noted, are precisely the ones where address identification is weakest.
Only 11% of stablecoin holdings can be geolocated
Perhaps the most consequential data point in the entire lecture concerns the geographic visibility problem. In the Artemis dataset, only about 11% of stablecoin holdings could be assigned to a specific country, and even that limited attribution relied on network address information that is easily masked.
The indirect proxies other researchers use, inferring countries from exchange web traffic or time-zone signals, recover only broad regions and fail precisely when a user takes deliberate steps to hide.
This is not a theoretical concern. Sanctions enforcement, capital controls, and tax residency determinations all depend on knowing where a holder is located. If regulators cannot see where stablecoins are held, those instruments become structurally difficult to apply.
Why GENIUS and MiCA fall short, according to Gopinath
Gopinath directed specific criticism at both major regulatory frameworks currently governing stablecoins.
The U.S. Guiding and Establishing National Innovation for U.S. Stablecoins Act (GENIUS Act), signed into law on July 18, 2025, regulates stablecoin issuers and CEXs but explicitly leaves self-custody wallets, peer-to-peer (P2P) transfers, and offshore issuers and exchanges outside its perimeter.
While stablecoin issuers retain the technical ability to freeze or destroy tokens even in self-custodied wallets, these actions happen after a crime has been detected, not before. The law’s Know Your Customer (KYC) requirements, in other words, apply only to a shrinking slice of total stablecoin activity.
The European Union’s (EU) Markets in Crypto-Assets Regulation (MiCA), which reached its hard enforcement deadline on July 1, 2026, extends one hop further than the U.S. framework and is more restrictive in its approach to foreign issuers. But it, too, leaves transfers between two self-hosted wallets that never touch a regulated provider outside its regulatory perimeter. While more transactions require anonymity checks under MiCA, Gopinath’s research shows this still excludes a large share of activity.
Gopinath summarized the regulatory gap bluntly in her lecture: policies that seek to improve KYC and Anti-Money Laundering (AML) regulations at the CEX and issuer levels can, at best, hope to have only modest effects in curbing illicit activity, given the predominance of wallet-to-wallet transactions.
The illicit finance picture
The lecture cited the Chainalysis 2026 Crypto Crime Report, which estimated that illicit addresses received over $150 billion in 2025 and that stablecoins accounted for 84% of illicit crypto transaction volume.
Gopinath emphasized that the identification of on-chain illicit activity mainly occurs after the fact, once entities have already been flagged by regulators or law enforcement off-chain, rather than through the kind of proactive monitoring that traditional bank compliance enables.
The asymmetry with traditional banking is also financial. Gopinath cited a LexisNexis Risk Solutions estimate putting the annual cost of financial-crime compliance in the United States and Canada at roughly $61 billion.
Banks bear that cost because identity verification and transaction monitoring are priced into every link of a payment chain under the Bank Secrecy Act (BSA). Stablecoin compliance obligations, by contrast, attach only to a subset of transactions.
The tax evasion dimension
Gopinath extended her argument into fiscal territory. Citing the work of economist Kenneth Rogoff, she noted that at least 40% to 45% of the U.S. tax gap is attributed to cash transactions. The latest Internal Revenue Service (IRS) estimates for 2022 point to a tax non-compliance rate of around 14%.
Applying that same rate to 2025 projections, the federal tax gap is $732 billion, or 2.44% of U.S. Gross Domestic Product (GDP). Including state and local taxes at a similar non-compliance rate brings the total to approximately $1 trillion, or 3.3% of GDP.
Gopinath then offered a hypothetical: if stablecoins grow to trillions of dollars in circulation and the anonymity they provide pushes the non-compliance rate just 10% higher, from 14% to 15.4%, the additional tax revenue lost by 2030 would amount to roughly $130 billion, or 0.35% of projected GDP.
While she was careful to frame this as a hypothetical, she argued that the potential fiscal impact of self-custody stablecoin holdings warrants serious policy attention.
Broader implications for monetary sovereignty and capital controls
The research also pointed to the growing use of crypto instruments to circumvent capital controls. Gopinath referenced work by Graf von Luckner, Reinhart, and Rogoff showing that crypto volumes in economies like Argentina rise in lockstep with parallel-market exchange-rate premiums. She noted that while dollar dominance may increase through stablecoin adoption, the ability to sanction such dollar holdings could simultaneously become more difficult.
These considerations also bear directly on a live policy question: whether crypto-affiliated institutions should be granted access to central bank infrastructure. Gopinath warned that such access would connect the central bank’s payment rails, at one remove, to an identity-free network in a way that no traditional bank does.
Three policy prescriptions
Gopinath closed with three recommendations.
First, rules that reach only exchanges and issuers will likely be insufficient. The identified parts of the system are the natural focus of regulation, but activity is migrating wallet to wallet, beyond them.
Second, because the phenomenon is inherently cross-border, common international standards are essential to prevent regulatory arbitrage. Where rules differ, activity migrates to the least demanding jurisdiction. The GENIUS Act and MiCA are progress, she said, but they differ in reach.
Third, sound domestic fundamentals remain an important part of the answer. Empirical evidence shows the demand for crypto instruments is strongest where domestic money is least trusted. Low and stable inflation, deeper financial systems, and better domestic payment infrastructure can all help preserve monetary sovereignty.
Timing and context
The lecture arrives at a pivotal moment for global stablecoin regulation. The GENIUS Act’s final implementing rules from six federal agencies are due by July 18, 2026, exactly one year after the law was signed. Meanwhile, MiCA’s hard enforcement deadline arrived on July 1, with the European Securities and Markets Authority (ESMA) warning that unauthorized firms must immediately stop onboarding EU clients.
In India, the Reserve Bank of India’s (RBI) Financial Stability Report (FSR), released on June 30, 2026, formally acknowledged the global regulatory momentum around stablecoins, referencing the GENIUS Act, MiCA, and the United Kingdom’s (UK) evolving framework. The stablecoin market now sits at approximately $309 billion in total capitalization, with transaction volumes exceeding $33 trillion in 2025, surpassing Visa and Mastercard combined.
The Financial Action Task Force (FATF) warned in a March 2026 report that P2P transfers via self-custody wallets represent a “key vulnerability” in AML oversight, and that stablecoins have become the most popular virtual asset in illicit transactions, including by actors in Iran and North Korea.
Gopinath’s research adds the quantitative foundation that has so far been missing from this debate. The question is no longer whether stablecoins present an anonymity problem relative to traditional money. The question, as she framed it, is whether policymakers can design frameworks that capture the benefits of the technology while preserving enough observability for public finances and the monetary system to function.
“Stablecoins are a genuine innovation,” Gopinath said in closing, “but they are held and used in their most anonymous form. The task for policymakers is to manage the resulting trade-off.”
Also Read: The New Stablecoin in Town: How Could OUSD Challenge or Replace USDC?
