Bill Nelson on the Future of the Fed's Balance Sheet

Episode Summary

Executive Summary: David Beckworth and Bill Nelson discuss the Fed’s balance sheet, arguing that QE has created a persistent “ratchet” in reserve demand and that the current ample-reserves framework is costly. They explore reforms to shrink the Fed’s footprint by making the discount window and other ceiling facilities more usable, revising liquidity regulations, and better managing Treasury General Account swings.

Main Topics: The Fed’s reserve-demand ratchet (Priority: 5/5): Nelson explains that reserve demand has risen structurally after each QE episode, driven by bank behavior, supervision, and liquidity rules that normalize higher reserve holdings. Liquidity regulation and the discount window (Priority: 5/5): A major theme is that banks hold excess reserves partly because regulations and examiners discourage reliance on the discount window; reforming liquidity rules could reduce reserve demand. Shrinking the Fed’s balance sheet (Priority: 5/5): The conversation centers on how the Fed could sustainably reduce reserves and its balance sheet without impairing monetary policy implementation. Costs of a large balance sheet (Priority: 4/5): Nelson argues that a large balance sheet has political-economy costs, weakens market functioning, raises independence concerns, and may increase financial fragility. Treasury General Account and temporary operations (Priority: 4/5): Beckworth and Nelson discuss proposals to neutralize TGA swings and use temporary open market operations to avoid predictable reserve shortages and repo volatility. International central bank comparisons (Priority: 3/5): They note that other major central banks have formally reviewed their operating frameworks and often prefer smaller balance sheets and more active market functioning.

Key Arguments: Reserve demand has ratcheted upward after each QE round because banks, examiners, and liquidity rules adapt to abundant reserves as the new normal. The discount window should be incorporated into liquidity assessments so banks can count pledged collateral and reduce their need to hold idle reserves. A smaller Fed balance sheet would improve market functioning, revive interbank lending, and reduce stigma around borrowing from the central bank. Large reserves are not costless: they can weaken Fed independence by making the balance sheet a more attractive tool for political demands. The Fed’s large footprint suppresses price discovery in money markets and contributes to the decline of the federal funds market. QE may increase long-run fragility by encouraging banks to fund more with short-term liabilities while assuming reserves will always be available. Temporary open market operations could smooth predictable reserve shocks from quarter-ends, tax dates, and Treasury settlements without requiring a permanently large balance sheet. Other central banks have concluded that the costs of abundant excess reserves exceed the benefits, suggesting the Fed should re-evaluate its framework.

Data Points: Fed balance sheet peak: $8.9 trillion - Approximate high point before QT ended. Fed balance sheet after QT: $6.5 trillion - Size when QT stopped earlier in 2025. Structural reserve demand, April 2008: $35 billion - Fed estimate before the ample-reserve era. Structural reserve demand, November 2016: $300 billion - Fed estimate after QE-era changes. Structural reserve demand, November 2018: $1 trillion - Fed estimate showing the ratchet effect. Structural reserve demand, October 2019: $1.4 trillion - Estimate before COVID-era QE. Current reserve demand: around $3 trillion - Nelson’s estimate of reserve demand today. Bank balance-sheet share held for liquidity needs: about 25% - Nelson’s estimate of the share of bank balance sheets tied up in liquidity requirements. Discount-window collateral composition: about two-thirds loans - Current collateral pledged is mostly loans rather than securities. Discount-window collateral composition, historical: about 80% loans - Earlier share of collateral pledged to the discount window. Treasurer survey sample: about 40-50 member banks; roughly 25 responses - BPI survey of reserve demand drivers among large banks. JPMorgan reserve balances end-2023: $400 billion - Example of a large bank reducing reserve holdings over time. JPMorgan reserve balances end-2025: $100 billion - Illustrates banks can economize on reserves when incentives change. Riksbank corridor width: 20 basis points total - 10 bps below and 10 bps above the target rate. Fed operating losses: about $200 billion - Direct losses Nelson attributes to the Fed’s balance sheet operations. Total losses including foregone franchise income: about $500 billion - Nelson’s broader estimate of the cost of the Fed’s losses.

Pivotal Quotes: "It’s a bit deceptive this time because a lot of the runoff in the assets actually corresponded to a reduction in the ONRRP facility." — Bill Nelson: Explaining why QT’s apparent balance-sheet reduction did not fully translate into lower reserve demand. "It kills the interbank market because everyone is flush with reserves and everyone is getting paid the market rate on their excess reserves." — Bill Nelson: Arguing that abundant reserves suppress market-based liquidity allocation and price discovery. "If we’re stupid, then the RBA, the ECB, and the Bank of England, and the Riksbank, and Norges Bank, and the Bank of Canada are stupid too." — Bill Nelson: Responding to the claim that abundant reserves are obviously optimal and noting foreign central banks’ different conclusions.

Implications: The episode suggests the Fed may need a new operating framework: smaller balance sheet, stronger discount-window access, and more active liquidity management. If not, reserve demand, stigma, and market distortions may keep ratcheting upward.