Quantitative tightening (QT) is the deliberate shrinking of a central bank's balance sheet, the reverse of quantitative easing (QE). In its modern U.S. form it is "passive runoff": the Federal Reserve simply stops fully reinvesting the principal from maturing Treasuries and agency mortgage-backed securities (MBS) it holds in the System Open Market Account (SOMA), and lets those holdings shrink. The core tension is that QT drains an asset (securities) by retiring a liability — and exactly which liability shrinks (bank reserves, the reverse repo facility, or the Treasury's cash balance) determines whether QT is barely felt or whether it cracks the money markets. The Fed controls the speed but not the distribution, which is why QT is "predictable in size, unpredictable in incidence."

How it works (mechanics)

QT operates through redemption caps, not asset sales. The Fed sets a monthly dollar cap; principal payments below the cap are allowed to run off (not reinvested), while principal above the cap is reinvested to keep the pace orderly. In the 2022–2025 cycle the Fed announced (May 4, 2022 press release) an initial Treasury cap of $30 billion/month rising to $60 billion after three months, and an MBS cap of $17.5 billion rising to $35 billion. If maturing Treasury coupons fell short of the cap in a given month, the difference was made up by redeeming Treasury bills. The Fed later slowed the Treasury cap to $25 billion (effective June 2024) and then $5 billion (effective April 2025) to glide toward its target, and announced on October 29, 2025 that net runoff would cease on December 1, 2025 (per the Cleveland Fed and CRS).

The accounting is double-entry. When a bond the Fed holds matures, the Treasury pays it off; the Treasury's cash account at the Fed (the TGA) falls, and the Fed's securities asset falls by the same amount — but no new reserves are created to replace the maturing claim, so the consolidated banking system ends up with fewer reserves once the TGA is refilled by issuing new debt to the public. MBS run off more slowly and unpredictably because they amortize via prepayments, which collapse when mortgage rates rise — a key reason the Fed has rarely hit its MBS cap.

Crucially, QT shrinks the asset side, but the liability side decides the impact. Total Fed liabilities = bank reserves + reverse repo (ON RRP) + the TGA + currency + a few smaller items. As assets fall, some liability must fall too. From 2022–2024 the runoff was absorbed mostly by the ON RRP facility draining from a ~$2.5 trillion peak toward near-zero — so bank reserves stayed roughly "ample" and markets barely noticed. Once the RRP buffer was exhausted, further runoff began biting into reserves directly, which is when repo-rate pressure appears (the New York Fed and Joseph Wang/fedguy describe this transition).

How it's used in practice

QT is a monetary-policy tool for removing accommodation and normalizing the balance sheet — secondary to the policy rate, which remains the Fed's primary lever. The Fed's own framing (and outside estimates) treats QT as a modest tightening: a 2022 Atlanta Fed working paper (Bin Wei, using the Vayanos–Vila preferred-habitat model) estimates ~$2.2 trillion of passive runoff over three years as roughly equivalent to ~29 basis points of rate hikes in normal times, rising to ~74 bp during a crisis when risk aversion doubles — a model-dependent estimate defined by equal impact on the 10-year yield, not an official FOMC figure.

For markets analysts, QT matters mainly through the net-liquidity lens. Many practitioners track net liquidity ≈ Fed balance sheet − TGA − ON RRP as a proxy for cash available to private markets, and correlate it with risk-asset prices (especially equities and crypto). Under this view, what mattered in 2022–2023 was less the headline runoff than whether the RRP drain was offsetting it. This is a correlational, contested framework — useful for regime awareness, weak as a precise timing signal.

Adoption, debate & evidence

QT is now a standard, openly telegraphed central-bank tool (the Fed, ECB, and Bank of England have all run versions). Its mechanics are not controversial; its effects are. The honest evidence picture:

• Asymmetry vs QE. Research and Fed commentary (e.g. Cleveland Fed) note QE and QT are not mirror images: QE works largely through signaling and portfolio-balance effects during stress, while QT in calm markets is designed to be "like watching paint dry" (Yellen's phrase). The Macroeconomic Policy Nexus and others argue QT cannot cleanly reverse the liquidity claims QE created.
• The 2019 cautionary tale. The first QT cycle ended in 2019, yet reserves kept falling as currency and the TGA grew. In mid-September 2019, secured overnight rates spiked — SOFR reached ~5.25% and intraday repo traded as high as ~10% (New York Fed, OFR). The lesson: "ample" reserves can become scarce and unevenly distributed faster than expected, and the Fed cannot pinpoint the minimum comfortable reserve level in advance.
• Magnitude is uncertain. Estimates of QT's rate-equivalent effect vary widely and are model-dependent; treat any single number as illustrative, not settled.

The genuine dispute is whether QT meaningfully tightens financial conditions at all in an ample-reserves world, or only matters at the margin once reserves approach scarcity — at which point it matters a great deal and abruptly.

Strengths & limitations

Strengths: QT is transparent, rules-based, and slow — it lets the Fed normalize without selling assets into the market or shocking rates. It runs in the background while the policy rate does the active work.

Limitations and the #1 misuse: conflating the headline runoff pace with the actual liquidity impact. The same $60 billion/month was nearly invisible while the RRP absorbed it and became dangerous once reserves were the marginal liability. QT also has a hard floor — the Fed must stop before reserves become scarce, and it cannot reliably forecast that floor, so QT tends to end reactively (slowed in 2024, ended Dec 2025; the 2019 episode forced an emergency restart of operations). MBS runoff is slow and rate-dependent, leaving the Fed's portfolio more mortgage-heavy than intended.

Sources

• Federal Reserve, "Plans for Reducing the Size of the Federal Reserve's Balance Sheet," press release, May 4, 2022 (cap figures and bill-redemption rule).
• Federal Reserve, "Policy Normalization" page.
• Cleveland Fed, "QT, Ample Reserves, and the Changing Fed Balance Sheet" (Economic Commentary 2025-05) — runoff totals (~$1.6T Treasuries, ~$600B MBS), QT end Dec 2025.
• Congressional Research Service, "The Federal Reserve's Balance Sheet" (IF12147) — balance-sheet size (~$4T → ~$9T → ~$6.5T).
• New York Fed, "The Market Events of Mid-September 2019" (EPR 2021) and OFR working paper 23-04 — repo spike magnitudes.
• Richmond Fed, "The Fed Is Shrinking Its Balance Sheet" (Q3 2022) — mechanics overview.
• Joseph Wang (fedguy.com), "Balance Sheet Dominance"; Macroeconomic Policy Nexus, "Déjà Vu at the Federal Reserve" — liability-absorption and QE/QT asymmetry (analyst commentary).
• Bin Wei, "Quantifying 'Quantitative Tightening' (QT): How Many Rate Hikes Is QT Equivalent To?" Federal Reserve Bank of Atlanta Working Paper 2022-8 (July 2022) — the ~29 bp (normal) / ~74 bp (crisis) rate-equivalent estimate, based on the Vayanos–Vila (2021) preferred-habitat model (model-dependent, not an official FOMC figure).

Disputes flagged: the rate-equivalent magnitude of QT and the net-liquidity-to-asset-price link are both contested and model-dependent; treat as estimates.