Tokenized deposits could reduce U.S. banks’ capacity to hold long-term interest-rate exposure by $700 billion under one modeled scenario, according to research published Aug. 25 by Dallas Fed economists Rosie Levy and Srini Ramaswamy.
Summary
- Dallas economists estimate 10% greater rate sensitivity could reduce banks’ duration capacity by $700 billion.
- A 10% shorter deposit life could reduce maturity transformation capacity by approximately $580 billion systemwide.
- The estimates measure ten-year equivalent interest-rate exposure, not deposits predicted to leave banking institutions directly.
- Tokenization may let depositors and AI agents move funds instantly toward banks offering higher yields.
- Banks could respond with higher deposit rates, larger liquidity buffers or additional wholesale debt issuance.
The figure does not represent $700 billion of deposits expected to leave banks or an equivalent guaranteed decline in lending. It measures a possible reduction in banks’ duration risk appetite, expressed as the equivalent exposure to ten-year Treasury securities.
The authors also stated that their views should not be attributed to the Federal Reserve Bank of Dallas or the Federal Reserve System.
Tokenized deposits could make bank funding less stable
Tokenized deposits are ordinary commercial bank deposits represented on a blockchain or another distributed ledger. They can support automated payments, programmable transactions and around-the-clock settlement while remaining liabilities of the issuing bank.
Their speed could weaken the practical barriers that make deposits relatively stable. Customers seeking higher yields could move money between institutions faster than they can through many existing banking systems.
“Instant settlement would allow deposit holders who prioritize yield to switch banks almost instantaneously,” the economists wrote.
Smart contracts could automatically transfer balances when another institution offers a better rate. Agentic artificial intelligence could theoretically monitor yields and initiate those transfers without requiring customers to act manually.
The authors did not predict how broadly depositors would use such automation. They described large-scale adoption as uncertain and evaluated what could happen under specific assumptions.
The $700 billion estimate measures duration capacity
Banks use relatively stable deposits to finance mortgages, business loans, securities and other longer-term assets. Although customers can withdraw demand deposits at any time, aggregate balances often remain with banks for years.
This behavioral stability gives deposits an effective duration. Banks also measure deposit beta, which shows how closely the interest rates they pay customers move with market rates.
Using Federal Reserve H.8 balance-sheet data, the economists estimated that U.S. banks held approximately $7 trillion of long-term interest-rate exposure on July 15. About $5.8 trillion, or 80%, was supported by the duration characteristics of deposits other than large time deposits.
Their analysis found that a 10% increase in deposit rate sensitivity could reduce banks’ duration risk capacity by $700 billion, assuming deposits have an average life of four years.
A separate scenario found that reducing average deposit life by 10% could lower maturity transformation capacity by approximately $580 billion.
These are back-of-the-envelope estimates based on assumed durations and aggregate balance-sheet matching. They are not forecasts of actual loan losses, deposit withdrawals or bank failures.
Banks could raise rates or hold more liquid assets
Banks could respond by offering higher deposit rates, reducing the incentive for customers to switch. That approach would increase funding costs and compress lending margins.
Institutions could also hold more reserves and government securities instead of long-term loans. Another option would involve issuing additional term debt to preserve existing lending levels.
Greater reliance on expensive wholesale debt would “likely adversely impact the cost of credit,” the authors estimated.
Research using Brazil’s Pix system provides an early comparison. A Central Bank of Brazil study found that increased instant-payment usage led banks to hold more liquid assets, particularly government bonds, while reducing the share of loans on their balance sheets.
The Brazilian findings do not establish that U.S. tokenized deposits will produce identical results. Pix is an instant-payment network rather than a tokenized deposit system, and the two markets operate under different banking structures.
U.S. banks continue building tokenized networks
Large American banks are moving forward with tokenized deposit infrastructure despite the possible funding risks. The Clearing House announced a shared network supporting automated workflows, interoperability and 24/7 settlement.
Bank of America, Citi, BNY, Wells Fargo and other institutions support the project. As crypto.news reported, JPMorgan and major competitors are building shared tokenized deposit infrastructure intended to connect blockchain activity with regulated commercial bank money.
Community and regional banks are also entering the sector. Thirty-nine state banking associations recently formed BankChain Alliance, which is targeting a nationwide blockchain launch during 2027.
The design of these networks will determine how easily deposits can move between institutions. Interoperability could improve payments while also increasing competition for funding, making deposit behavior, liquidity rules and bank-size differences central issues for regulators.
