What Happens When Central Banks Shrink Their Balance Sheets

The Slow Unwinding of a Very Large Footprint

Between December 2005 and December 2025, the Federal Reserve's balance sheet grew from about $800 billion to roughly $6.5 trillion. That is an increase from around 6 percent of GDP to 21 percent. The expansion came mainly from two decisions: buying bonds to push down longer-term interest rates after the 2008 crisis and the pandemic, and, in 2019, shifting to an "ample reserves" system rather than the scarce-reserves setup used before 2008.

Then the Fed reversed course. It began shrinking its balance sheet in June 2022 and finished that process on December 1, 2025. Days later, on December 10, it announced it would start buying reserves again to keep the system ample. The steady-state size of a central bank's balance sheet remains an open question, with no consensus among economists or policymakers.

This matters to ordinary savers, borrowers and businesses because the size and shape of a central bank's balance sheet influences the amount of money in the banking system and, through that, broader financial conditions.

Reserves: The Oil in the Money-Market Engine

Banks hold reserves — deposits at the central bank — as a buffer for settlement obligations and internal liquidity needs. When that buffer shrinks, pressures build in money markets. The spreads that reflect the relative cost of liquidity rise, because banks demand more compensation to part with scarce cash.

In a Federal Reserve note published in January 2026, economists Burcu Duygan-Bump and R. Jay Kahn describe a "balance-sheet trilemma." A central bank can achieve only two of three goals at once: a small balance sheet, low volatility of short-term rates, and limited market intervention. The tension arises because the demand for reserves is nonlinear and uncertain — a phrase that would make a fine excuse for almost any market surprise.

When reserves are plentiful, modest changes in liquidity conditions barely move rates. When reserves are thin, the same modest changes produce outsized rate movements. The Fed's own research shows that as reserves fall relative to outstanding Treasuries, the spread of the Tri-Party General Collateral Rate — a broad measure of repo rates — becomes increasingly responsive to movements in the Treasury General Account, Treasury issuance, quarter-ends and foreign repo pool balances.

Repo, short for repurchase agreement, is a form of short-term borrowing where one party sells a security and agrees to buy it back later, typically overnight. It is the plumbing of the financial system, and when the plumbing rattles, the noise travels.

The Fed's Options and the Cost of Each

The trilemma forces a choice with distinct costs. A large balance sheet stabilizes rate control and reduces the need for intervention, but it expands the central bank's structural footprint, may crowd out private-sector lending, and creates duration risk. Duration risk means the risk that longer-term securities held by the central bank lose value when interest rates rise. This can lead to lower — and potentially negative — remittances to the Treasury, as the Fed pays interest on reserves at higher rates than the income from its older, lower-yielding bonds.

High rate volatility, on the other hand, can weaken a central bank's control over interest rates and complicate monetary transmission. Funding costs become less predictable, making investment planning harder. Persistent volatility can spill over to longer maturities as investors demand higher term premiums to compensate for short-rate uncertainty. A term premium is the extra yield investors require for holding a longer-term bond rather than rolling over short-term debt.

Frequent intervention also carries costs. Active open market operations require continual assessment of how many reserves are needed, and unanticipated shocks — such as large margin calls or natural disasters — are hard to gauge. Misjudging the direction or magnitude can amplify, rather than smooth, rate movements. Passive tools like the Standing Repo Facility and Overnight Reverse Repo Facility can place ceilings and floors on rates, but frequent use might weaken market discipline.

The United Kingdom: Selling Gilts, Watching Demand

The Bank of England has taken a different path, actively selling government bonds — known as gilts — from its Asset Purchase Facility. The Monetary Policy Committee voted in September 2025 to reduce the stock of gilts held for monetary policy purposes by £70 billion over the period from October 2025 to September 2026, to a total of £488 billion.

The Bank's approach shows how operational details matter. It aims to sell fewer long-maturity gilts than those at other maturities, to reflect demand conditions. In the third quarter of 2026, it planned to sell short-maturity bonds across three auctions of £725 million and medium-maturity bonds across two auctions of £600 million. In the second quarter, it included one long-maturity auction of £675 million alongside short and medium sales.

The Bank reserves the right to amend its schedule if market conditions warrant. In April 2025, it amended its schedule to carry out an extra auction of short-maturity bonds in place of a long-maturity auction. Flexibility, it turns out, is not just a personal virtue.

Cash flows between the APF and HM Treasury also matter. Between 2009 and 2022, the APF generated positive net cash transfers to the Treasury, peaking at a cumulative £123.9 billion at end-September 2022. Since then, payments have flowed the other way. By the end of the second quarter of 2026, the APF had transferred £16.2 billion to the Treasury in net terms, with further cash flows from the Treasury to the APF likely necessary until the unwind is complete. The Bank's projections show a wide range of outcomes — between minus £70 billion and plus £65 billion in present-value terms — reflecting uncertainty around both cash flows and debt-servicing savings.

The Bank's own analysis suggests that different unwind paces do not necessarily alter lifetime cash flows on a net present value basis, and that the path of interest rates has a more important influence. Holding onto bonds for longer does increase interest rate risk, however, meaning a faster pace of QT reduces the risk borne by the public sector.

The Euro Area: A Demand-Driven Approach

The European Central Bank has taken yet another route. In a November 2024 speech, Isabel Schnabel, a member of the ECB's Executive Board, described a framework where the marginal unit of reserves is provided elastically on demand through standard refinancing operations. Excess liquidity had fallen by more than a third from its 2022 peak and dropped below €3 trillion.

Schnabel noted that the ECB's balance sheet reduction was progressing smoothly. The phasing-out of reinvestments had not led to bottlenecks in bond absorption despite increased government issuance. Foreign investors absorbed the largest share of net issuance, with households playing an important role in some economies. Long-term interest rates did not rise excessively; the term premium initially increased from negative territory but then fell back and stabilized at low levels.

The ECB also adjusted its operational framework, reducing the spread between the main refinancing operations rate and the deposit facility rate from 50 basis points to 15 basis points. A basis point is one-hundredth of a percentage point. This narrow spread limits upward pressure on money market rates and their volatility. A "soft" floor means the ECB tolerates deviations from the deposit facility rate in both directions, provided they do not blur the signal about the intended policy stance.

Repo volumes between euro area counterparties grew by nearly 25% as excess liquidity declined, with the strongest growth in cross-border transactions. Banks issued record amounts of covered and senior unsecured bonds to replace maturing central bank funding. The share of reserves in banks' holdings of high-quality liquid assets fell from a peak of 78% in late 2022 to 56%. These are signs of adaptation, not disruption.

Why This Matters Beyond Borders

Tighter monetary policy in one major economy spills over to others. A Federal Reserve note from December 2022, by Dario Caldara, Francesco Ferrante and Albert Queralto, describes three channels. The exchange rate channel: a surprise rise in U.S. interest rates tends to strengthen the dollar, lowering the prices of foreign goods relative to U.S. goods. The demand channel: slower U.S. demand reduces imports, dampening foreign activity. The financial channel: higher U.S. longer-term yields lead international investors to rebalance from foreign to U.S. assets, tightening foreign financial conditions.

The dollar's dominant role in trade and finance amplifies these effects. In the Fed's "dollar dominance" model, foreign GDP falls by roughly twice as much as in a benchmark model, partly because foreign exporters set prices in dollars and because dollar-denominated borrowing tightens foreign financial conditions when the dollar appreciates.

This matters for currency markets because balance sheet runoff is one of the forces that can influence term premiums and exchange rates. But it is not the only force, and its effects are uncertain. The Fed note concludes that central banks risk overtightening if they underestimate spillovers, but they must also guard against de-anchoring inflation expectations. It is a genuinely difficult balancing act.

The Long Road Ahead

Federal Reserve Governor Stephen Miran argued in March 2026 that shrinking the balance sheet is desirable and solvable. He estimated that the balance sheet could potentially be reduced by $1 trillion to $2 trillion without returning to scarce reserves, by taking steps to lower reserve demand — such as easing liquidity coverage ratio requirements, reducing stigma around standing repo and discount window usage, and conducting more active open market operations around quarter-ends and fiscally significant dates.

But he also cautioned that the process would be slow. He noted that administrative steps would likely take well over a year, possibly several years, and that once reductions begin, the pace should be gradual to ensure the private sector can absorb the securities. He also suggested that resuming balance sheet reduction might warrant additional reductions in the federal funds rate, though he declined to put magnitudes on those effects.

The common thread across all these approaches is uncertainty. No one knows the optimal size of a central bank's balance sheet. The demand for reserves is inherently nonlinear and subject to uncertain shifts. Different jurisdictions have different market structures, regulatory frameworks and operating procedures. What works in one may not work in another. For anyone watching funding markets, currencies or bond yields, the lesson is not a prediction but a lens: how much liquidity is in the system, who needs it, and what happens when the cushion gets thin.