Tokenized deposits could reduce US banks’ ability to hold long-term interest-rate exposure by as much as $700bn under one modelled scenario, according to research from two Dallas Fed economists.
The estimate, published on 25 August by Rosie Levy and Srini Ramaswamy, does not mean banks are expected to lose $700bn in deposits. Nor does it represent a guaranteed fall in lending. Instead, it measures a potential reduction in banks’ appetite for duration risk, expressed as the equivalent exposure to 10-year US Treasury securities.
The authors stressed that their views should not be attributed to the Federal Reserve Bank of Dallas or the Federal Reserve System.
Tokenized deposits are conventional commercial bank deposits recorded on a blockchain or another distributed ledger. They remain liabilities of the issuing bank, but can support automated payments, programmable transactions and settlement around the clock.
The economists said the speed and flexibility of the technology could weaken some of the practical barriers that currently help keep deposits stable. Customers looking for better returns could potentially move money between banks more quickly than is possible through many existing banking systems.
Smart contracts could automatically transfer funds if another institution offered a higher interest rate. In theory, agentic artificial intelligence could monitor rates and begin those transfers without customers having to take action themselves.
The research did not attempt to predict how widely depositors would use these features. Instead, it examined the consequences under particular assumptions, while describing broad adoption as uncertain.
Banks rely on the relative stability of deposits to fund mortgages, business lending, securities and other assets with longer maturities. While demand deposits can be withdrawn at any time, the total amount held across a banking system often remains in place for years.
That behaviour gives deposits an effective duration. Banks also track deposit beta, a measure of how closely the interest rates paid to customers move in line with market rates.
Using Federal Reserve H.8 balance-sheet data, Levy and Ramaswamy estimated that US banks held about $7tn of long-term interest-rate exposure on 15 July. Deposits other than large time deposits supported approximately $5.8tn of that figure – around 80% – because of their duration characteristics.
Under one scenario, a 10% rise in the sensitivity of deposit rates would reduce banks’ capacity to take on duration risk by $700bn, assuming deposits had an average life of four years.
A separate calculation suggested that cutting the average life of deposits by 10% could reduce banks’ maturity-transformation capacity by about $580bn.
The figures are broad estimates based on assumed durations and the matching of assets and liabilities across aggregate bank balance sheets. They are not predictions of loan losses, deposit withdrawals or bank failures.
The report outlined several ways banks might respond. They could raise the rates paid on deposits to make customers less likely to switch providers, although that would increase funding costs and put pressure on lending margins.
Banks might instead hold more reserves and government securities rather than longer-term loans. They could also issue additional term debt in an effort to maintain current lending levels.
However, greater use of more expensive wholesale borrowing would “likely adversely impact the cost of credit,” the authors estimated.
Evidence from Brazil’s Pix payment system offers an early comparison, although it does not provide a direct forecast for the US. A study by the Central Bank of Brazil found that greater use of instant payments led banks to hold more liquid assets, particularly government bonds, while reducing the proportion of loans on their balance sheets.
Pix is an instant-payment network rather than a tokenized deposit system, and the two countries have different banking structures. The Brazilian results therefore do not show that tokenized deposits in the US would have the same effect.
Despite the possible funding risks, major US banks are continuing to develop tokenized deposit infrastructure. The Clearing House has announced a shared network designed to support automated workflows, interoperability and 24/7 settlement.
Bank of America, Citi, BNY, Wells Fargo and other institutions back the project. JPMorgan and other major banks are also building shared tokenized deposit infrastructure intended to connect blockchain activity with regulated commercial bank money, as reported by crypto.news.
Community and regional banks are beginning to take part as well. Thirty-nine state banking associations have formed BankChain Alliance, which is targeting a nationwide blockchain launch during 2027.
The way these networks are designed could determine how easily funds move between banks. Greater interoperability may make payments more efficient, but could also intensify competition for deposits.
That means deposit behaviour, liquidity requirements and differences between large, regional and community banks are likely to become important issues for regulators as the technology develops.
