Crypto
Home›Crypto›Market Structure›AI-run deposit switching could reduce banks’ interest-…
AI-run deposit switching could reduce banks’ interest-rate risk appetite
A Dallas Fed analysis estimates $7 trillion of interest-rate exposure in 10-year equivalents tied to deposit duration, and a sensitivity case suggests a $700 billion drop in duration-risk appetite if deposits become more rate sensitive.
AI-directed bank accounts could move deposits quickly across institutions, weakening a funding advantage that helps banks finance longer-term credit, according to a Federal Reserve Bank of Dallas analysis published Aug. 25 and summarized by CryptoSlate.
The Dallas Fed notes that while customers can withdraw demand deposits at any time, balances often stay put for years, and deposit rates typically rise by less than market rates. It models this behavior as partly like long-duration funding, using an effective duration framework based on weighted average life and deposit beta.
The analysis estimates about $7 trillion of asset-side interest-rate exposure in 10-year equivalents, with about $5.84 trillion tied to deposits other than large time deposits. In a sensitivity case, a modeled 10% increase in deposit price sensitivity, assuming a four-year weighted average life, reduced aggregate duration-risk appetite by about $700 billion in 10-year equivalents.
It also modeled a separate 10% reduction in weighted average life that cut maturity-transformation capacity by about $580 billion. The Dallas Fed said banks could respond by issuing more term debt, but wholesale funding could raise borrowing costs for consumers and businesses, while holding more reserves and Treasuries would leave less room for illiquid credit.
The report cites the June 2026 Clearing House initiative to develop 24/7 interoperable tokenized commercial-bank money and refers to a separate 2025 Central Bank of Brazil paper finding that heavier use of Brazil’s Pix instant-payment system increased liquid-asset holdings and reduced liquidity transformation.
sources_used: CryptoSlate, Federal Reserve Bank of Dallas analysis (Aug. 25), Clearing House (June 2026), Central Bank of Brazil paper (2025)