Darrell Duffie, professor of finance at Stanford University, presented a paper at the Brookings Papers on Economic Activity Spring 2026 Conference on March 26 arguing that Fedwire's settlement design creates a structural floor on the Federal Reserve's balance sheet. The paper, titled 'The Payment System Puts a Floor on the Fed's Balance Sheet,' proposes adding a liquidity savings mechanism that would allow banks to fund outgoing payments with incoming ones, reducing the reserve balances the banking system must hold at the Federal Reserve.
Fedwire currently operates as a pure real-time gross settlement system. Each of the roughly 870,000 daily transactions settles individually, requiring the sending bank to hold sufficient reserves before submitting a payment order. CHIPS, the other major USD large-value payment system, operates a continuous bilateral and multilateral netting algorithm that achieved a 29:1 liquidity efficiency ratio in 2024. Its 42 direct participants collectively settle approximately $2.2 trillion in daily gross payments using a fraction of that amount in prefunded positions. Duffie argues that applying a comparable mechanism to Fedwire could substantially reduce aggregate reserve demand across the banking system.
Fed Chair nominee Kevin Warsh has advocated reducing the Fed's $6.6 trillion balance sheet, giving the proposal immediate policy relevance. Reserve balances stand at approximately $3 trillion, the largest Fed liability after currency in circulation. Banks' operational need to pre-fund Fedwire payments contributes to demand for these reserves. The Federal Reserve has not publicly responded to the proposal. Implementation would require changes to Fedwire's core settlement architecture, Operating Circular No. 6, and the existing intraday credit framework.