WALCL: Composition by Instrument, Treasuries, MBS, Fed Assets Breakdown

As of June 10, 2026, the $6.73 trillion WALCL splits into $4.48 trillion of Treasuries (67%), $1.96 trillion of agency MBS (29%) and $0.28 trillion of other assets (4%), per the Federal Reserve’s H.4.1 release. QT ended on December 1, 2025: since then, the balance sheet’s composition has been moving more than its size.

Three asset families structure the contemporary Fed balance sheet. Each obeys distinct absorption dynamics that the aggregate-level reading entirely masks.

Reading the aggregate WALCL level without looking at instrument-level composition is like watching a thermometer without reading the scale. Composition determines what the Fed can do mechanically at each phase of the cycle, how it can do it, and at what pace. Before 2008, the Fed balance sheet was almost entirely composed of short and medium-maturity Treasuries, held against currency in circulation. The introduction of agency MBS via QE1 in January 2009 inaugurated a doctrinal rupture that no other G7 central bank has replicated at this scale, and which still structures today the constraints on WALCL as a monetary-stance indicator.

This analysis runs parallel to but distinctly from the technical definition of the FRED ticker. The definition lays out what WALCL aggregates; composition describes what is inside it, and why this internal structure changes the operational analysis of the balance sheet. For direct time-series access to the underlying data, the FRED WALCL dataset for time-series access provides the canonical entry point.

Treasuries in WALCL: maturity distribution and runoff mechanics

As of December 31, 2025, the Fed’s Treasury portfolio stood at $4.23 trillion per the Federal Reserve’s H.4.1 release. The bulk is nominal Notes and Bonds ($3.57 trillion), ahead of TIPS ($312 billion at face value, plus $114 billion of inflation compensation) and Bills ($234 billion).

This maturity distribution is not neutral for the runoff mechanism. The Fed does not actively sell Treasuries to remove them from the balance sheet: it lets these securities mature and chooses, within the authorized monthly cap, not to reinvest the repaid principal. The actual Treasury runoff pace is therefore a direct function of the portfolio’s maturity calendar. That link ended with QT: since December 1, 2025, the Fed rolls over at auction all principal from maturing Treasuries.

The Treasury monthly cap, set at -$30 billion in June 2022, -$60 billion from September 2022, -$25 billion from June 2024 then -$5 billion from April 2025 until QT ended on December 1, 2025 per the FOMC Statements, was not always saturated. When the volume of securities maturing in a given month exceeds the cap, the Fed rolls the excess; when it falls below the cap, actual runoff is reduced. Measured on the H.4.1 release, Treasury holdings fell by $14 to $34 billion a month in the first quarter of 2025, then by $1 to $6 billion a month from May to November.

This mechanics has a rarely explicit consequence: a reduced Treasury cap does not necessarily signal an intentional QT slowdown, but may reflect the effective maturity of the observed window. The distinction matters for how composition mirrors doctrinal regimes in reading successive phases. Bills (under one year) served as the balancing item: they filled the cap whenever maturing coupons fell short, which brought holdings down from $326 billion in April 2022 to $195 billion in September 2025. Since the FOMC implementation note of December 10, 2025, they carry the purchases meant to maintain ample reserves: $560 billion on September 30, 2026.

The Treasury portfolio’s sensitivity to long-rate moves is substantial. According to the Federal Reserve Banks’ 2023 combined financial statements, unrealized losses on the SOMA portfolio (Treasuries and MBS) stood at $1,080 billion at end-2022 and $948 billion at end-2023, mainly reflecting the 2022-2023 rise in yields. These losses are not recognized in WALCL, and a security held to maturity is redeemed at par: they are realized only if holdings are sold before maturity. Remittances to the Treasury are suspended when the Fed’s interest expenses exceed its earnings, the shortfall being booked as a “deferred asset”.

Agency MBS: the specificity of passive runoff via prepayments

The Fed’s MBS portfolio totaled $2.04 trillion on December 31, 2025 ($1.96 trillion on June 10, 2026), mainly composed of agency MBS guaranteed by Fannie Mae, Freddie Mac and Ginnie Mae, issued on the 30-year and secondarily 15-year segments. Nearly all of these securities were acquired during the four successive QE programs between 2009 and 2022, mostly at the low mortgage-rate levels that prevailed during that period.

The analytical specificity of MBS in WALCL lies in the runoff mechanism. Unlike Treasuries that leave the balance sheet on a known date per their maturity calendar, MBS leave the balance sheet via prepayments, i.e. early repayments of the underlying mortgage loans by US households. These prepayments are themselves a function of mortgage rates prevailing in the secondary market.

The mechanics works as follows: when 30-year rates fall significantly below the average rate on existing loans, households refinance and prepay, which accelerates MBS runoff. When 30-year rates remain elevated, as has been the case since late 2022, households retain their existing low-rate loans and prepayments collapse. The MBS cap, set by the FOMC at $17.5 billion a month from June to August 2022 and $35 billion from September 2022, was almost never hit in practice. Measured on the H.4.1 release, actual MBS runoff has hovered between -$15 and -$20 billion per month over 2024-2025, roughly half the authorized cap.

This asymmetry between theoretical cap and actual runoff has a structural consequence. If Treasuries leave the balance sheet faster than MBS, the relative MBS share in WALCL mechanically rises over time. The MBS share in WALCL thus moved from 30.6% at the April 2022 peak to 32.7% in February 2025, despite a nominal reduction in the overall balance sheet. The move reversed once QT ended: since December 1, 2025, MBS principal has been reinvested into bills, and the MBS share fell back to 29.2% on June 10, 2026, then 28.1% on September 30, 2026. The balance sheet’s exposure to the mortgage market now shrinks at the pace of household repayments.

Active MBS sales (“outright sales”) have never been implemented: the May 2022 balance-sheet reduction plan worked through non-reinvestment caps, and since December 2025 all MBS principal has been reinvested into bills. Current doctrine privileges exclusive passive runoff, but this choice is not unconditional: it reflects the Fed’s preference to minimize disruption of the secondary MBS market, where the Fed has become the largest historical holder. Active sales would alter the price equilibrium on agency MBS, with direct consequences for US mortgage financing conditions.

The residual line: emergency facilities and support liquidity

The Fed balance sheet’s residual line, $374 billion on December 31, 2025 (including $75 billion of year-end repos) and $281 billion on September 30, 2026, groups several heterogeneous categories: unamortized premiums on securities bought above par ($209 billion on September 30, 2026), repurchase agreements overnight and term resulting from routine open-market operations, primary credit extended via the discount window to banks demanding liquidity, historical gold reserves and special drawing rights, and balances of special facilities still active. This envelope is small relative to Treasuries and MBS, but its volatility during stress periods is incomparable.

The most recent example is the Bank Term Funding Program (BTFP), opened on March 12, 2023 in response to the Silicon Valley Bank and Signature Bank failures, and closed to new draws on March 11, 2024. The program carried roughly $167 billion at its peak in January 2024 per the weekly BTFP reports published by the Fed, and its balance returned to zero on March 12, 2025, as the last loans matured. This temporary addition to the Fed balance sheet did not reflect a QE intent: it answered emergency funding demand from bank counterparties. The same logic is developed in the monetary plumbing and its market effects.

Emergency support facilities share a common feature: their WALCL impact is non-discretionary in the monetary-policy sense. The Fed announces a program and sets its parameters; counterparties then decide whether to use the facility. WALCL reflects the resulting aggregate but does not distinguish in its level what stems from monetary intent versus what stems from market response. Rigorous reading requires unpacking the residual line when it becomes material.

The historical March 2020 facilities under the Covid framework (PMCCF, MLF, MSLP and others) were mostly closed between late 2020 and early 2021. Some accounting traces persist in affiliated accounts, but their impact on WALCL is now negligible. These facilities are however reactivable in case of a future crisis, and the March 2020 doctrinal precedent remains a structuring element of the Fed’s monetary palette for the monetary plumbing and Fed transmission in future stress phases.

Tying these three asset families back together: the contemporary Fed balance sheet is not a homogeneous block but a heterogeneous portfolio whose internal dynamics differ markedly. Treasuries dominate by size and provide the calendar-driven runoff backbone. MBS introduce a structural dependence on the mortgage market that no rate decision alone can override. The residual line absorbs unpredictable demand shocks from emergency facilities. Reading WALCL as a single number aggregates all three dynamics into one, masking the fact that a policy move on any one of them does not propagate identically across the others.

This heterogeneity is also what explains why successive Federal Reserve chairs since Bernanke have devoted substantial communication effort to the composition question, not just to the size question. The composition is the practical lever; the size is the headline result.

Last updated — 8 October 2026