The Ultimate Playbook for Reducing The Fed’s Balance Sheet | Professor Darrell Duffie on 4 Tools For Federal Reserve To Shrink Reserve Demand In Banking System

Episode Summary

Executive Summary: The episode examines how the Fed could shrink its balance sheet without triggering funding-market stress, focusing on Darrell Duffie’s four proposals: temporary open market operations, easing liquidity-regulation stigma, adding a liquidity savings mechanism, and tiering reserve remuneration. Duffie argues the real constraint is reserve demand from payments and bank behavior, not just asset sales.

Main Topics: Why the Fed’s balance sheet is constrained by liabilities, not assets (Priority: 5/5): Duffie explains that shrinking assets requires shrinking liabilities first. Currency is essentially fixed, Treasury’s account can only be trimmed modestly, so reserve balances are the main adjustable liability. Reserve scarcity and funding-market stress (Priority: 5/5): The discussion centers on how reducing reserves too far causes repo and other funding rates to spike, as seen in September 2019 and again in late 2025, because banks are unwilling to dip below perceived liquidity thresholds. Temporary open market operations as a buffer (Priority: 4/5): Duffie’s first proposal is for the Fed to actively offset short-term reserve swings from taxes, quarter-ends, and government cash flows, smoothing reserve demand and modestly lowering the average needed balance. Stigma from liquidity regulations (Priority: 5/5): Reg YY and RLAP make large banks reluctant to use the discount window, standing repo facility, or overdrafts, raising reserve demand. Duffie argues the Fed could change communications or the rules to normalize facility use. Liquidity savings mechanism for payment netting (Priority: 5/5): A software-based netting system could cancel offsetting payments inside Fedwire, reducing how many reserves banks need intraday. Duffie notes other central banks already use similar systems. Tiered remuneration of reserves (Priority: 5/5): The Fed could pay full interest only on the reserve amount banks truly need, while paying a lower rate on excess balances. This would reduce reserve demand while preserving monetary policy control. Political and institutional tradeoffs (Priority: 4/5): The interview closes on the tension between a smaller, more politically palatable balance sheet and the costs of implementation, including operational complexity, bank pushback, and possible higher rate volatility.

Key Arguments: The Fed should analyze reserve demand on the liabilities side of its balance sheet, because asset reduction alone is not feasible without threatening payments and funding stability. A small reserve shortfall can rapidly destabilize repo markets because large banks avoid using their own reserves or Fed backstops when they fear regulatory scrutiny. Temporary open market operations can smooth daily reserve shocks, but they likely only reduce reserve demand by a few hundred billion dollars. The bigger reductions require structural changes: reducing stigma around Fed liquidity tools, adopting payment-netting software, or tiering reserve interest rates. Liquidity savings mechanisms and tiered reserve remuneration have worked at other central banks and could materially cut U.S. reserve demand. If the Fed wants a dramatically smaller balance sheet, it likely needs to lower the demand for reserves first rather than simply letting assets run off. A smaller balance sheet would be politically easier to defend and would leave more room for future emergency expansion if a crisis or Treasury-market dysfunction returns.

Data Points: Fed assets / liabilities scale: About $6.5 trillion - Approximate Fed balance sheet size discussed as the starting point for shrinkage Currency in circulation: About $2.5 trillion - Hard-to-reduce Fed liability because cash cannot be forcibly withdrawn from the public Reserve balances: About $3 trillion - Main adjustable liability and the key target for balance-sheet reduction Treasury General Account: About $1 trillion - Government deposits at the Fed; can only be reduced modestly Fedwire daily payments: $4.5 trillion per day - Illustrates the enormous payments demand that requires reserve balances September 2019 repo spike: ~300 basis points above the Fed’s rate; intraday ~1,000 basis points above - Example of funding-market stress when reserves became too scarce Reserve growth pace: About $40 billion per month - Fed purchasing Treasury bills to rebuild reserves after determining they had fallen too low Temporary OMOs impact: Roughly $100 billion to $200 billion - Estimated reduction in average reserve demand from smoothing short-term reserve fluctuations Quarter-end reserve swings: $200 billion to $500 billion - Foreign banks reduce reserve balances at quarter-end for capital optics Payment-loop example: $20 billion - Illustrative amount netted out by a liquidity savings mechanism in the example chart and explanation Bank of England estimate: 20% to 30% savings - Approximate reduction in reserve needs from liquidity savings mechanisms in the UK system Reserve Bank of New Zealand rate spread: 100 basis points lower on excess reserves - Example of tiered remuneration to discourage unnecessary reserve holdings Fedwire stress timing: Payments arrived over 150 minutes later than normal - Observed delay on September 17, 2019 when banks held back payments during stress Historical use of tiering: A few weeks after Lehman’s failure - Fed briefly operated a two-tier reserve remuneration system after 2008

Pivotal Quotes: "Massively stupid and inefficient" — Governor Chris Waller, paraphrased in the interview: Used to criticize shrinking reserves without first reducing reserve demand "We shouldn't simply sell assets. Assets today and avoid problems because all of the liabilities on the Fed's balance sheet are serving very important roles." — Darrell Duffie: Core framing that the liability side determines how far the Fed can shrink "The payment system puts a floor on the Fed’s balance sheet." — Darrell Duffie / paper title: Central thesis: payments infrastructure and bank behavior limit how small reserves can go

Implications: If the Fed wants a meaningfully smaller balance sheet, it will likely need operational reforms, not just asset runoff. The biggest changes would come from payment-netting software and reserve-tiering, but both require years, testing, and political coordination.