On September 24, 2026, the US 10-year inflation-protected real yield hit 2.85% — its highest since November 2008. One number, doing enormous work.
It explains why gold can sit 20% below its January record while central banks buy at a record pace, why long-duration stocks carry a lasting valuation discount, and why credit markets, instead of celebrating the end of QT, stay watchful.
To see how that number got there, follow what the Fed's balance sheet did for four years — and what it stopped doing last December. Quantitative tightening drained more than $2 trillion of liquidity between June 2022 and late 2025, working in the background while rate hikes grabbed the headlines. Then, at its October 2025 meeting, the Fed voted to end balance-sheet runoff as of December 1, 2025, shifting to reserve-management purchases that keep bank reserves comfortably ample. The policy rate, meanwhile, is back as the active tool: the Fed raised it to 3.75%–4.00% on September 16, 2026. Investors now have to read a two-handed policy: an active rate on one side, a balance sheet that has stopped shrinking on the other.
The machine that stopped shrinking
Quantitative tightening is the mirror image of quantitative easing: instead of buying bonds to push money into the system, the central bank lets holdings mature without replacing them, passively draining reserves. For four years that drain was the quiet metronome behind every bond market move.
Ending it did not freeze the balance sheet in place. Since December 1, 2025, all principal from maturing Treasury securities has been fully reinvested, so the Treasury portfolio no longer contracts. The machine has not been switched off — it has been reprogrammed.
The October 2025 decision to stop was explicitly preemptive. As macro analyst Engin Demirel wrote in a widely circulated December 2025 analysis, the Fed ended QT slightly before its own models suggested reserves would approach a "danger zone" around January 2026, aiming to avoid a repeat of 2019 rather than waiting for funding stress. Read that choice carefully: financial-stability concerns overruled the theoretical target of shrinking the balance sheet to merely "somewhat above ample reserves." The Fed decided the last stretch was not worth travelling.
Two rehearsals, one scar
The Fed has run this experiment twice, and the two episodes could hardly have been more different. QT1 ran from October 2017 to August 2019, shrinking the balance sheet from $4.5 trillion to $3.8 trillion — roughly $700 billion over 22 months, starting at just $10 billion per month and ramping to $50 billion. Modest numbers. Modest pace. It still went too far: reserves were drained excessively, culminating in the September 2019 repo market crisis, when overnight rates spiked and the Fed was forced to restart balance-sheet expansion. As DollarLiquidity.com has documented, QT is not merely background noise — it directly reduces the liquidity available for risk-taking.
QT2 began in June 2022 at nearly double the pace: up to $95 billion per month, split between $60 billion in Treasuries and $35 billion in agency MBS. The balance sheet shrank from $8.96 trillion to roughly $6.8 trillion by early 2026. The faster pace coincided with the 2022 bear market in stocks and crypto, though separating QT's impact from simultaneous rate hikes is difficult — two hands squeezing at once, with no clean way to attribute the pressure to either.
Set the two episodes side by side and the October 2025 decision explains itself. The first run, at half the speed, still ended in a funding-market crisis. The second was faster, larger, and run against a banking system whose appetite for reserves the Fed could only estimate. And here is the asymmetry: stopping too late means a crisis — reserves cross an invisible line, funding markets seize, and the reversal happens in public, under pressure. Stopping too early means something far more forgiving — a somewhat larger balance sheet and a slower journey back to "normal." Faced with that choice, the vote was barely a close call.
The mortgage exit that won't hurry
The more interesting transformation is on the mortgage side of the ledger. The Fed continues to let agency debt and MBS run off, redirecting the inbound proceeds — from mortgage payments, refinancing, and home purchases — into Treasury bonds of like size. The central bank is converting its mortgage holdings into Treasuries, one maturity at a time, without changing its overall footprint. The composition changes. The total does not.
Progress has been much slower than policymakers hoped. As of September 10, 2026, the Fed held $1.914 trillion in mortgage paper, down from $2.707 trillion when QT began in June 2022, according to HSH.com. By the Fed's original design, $1.768 trillion should have been trimmed by the end of September 2026. The actual decline is only about $794 billion — far less than half the hoped-for pace. The reason is no mystery: refinancing and homebuying remain sluggish, and mortgage paper only leaves the balance sheet when households move, refinance, or pay down. The Fed can set the intention. Borrowers set the speed.
Had mortgages retired at the expected pace, September 2026 would have been the first month since February 2013 that the Fed held less than $1 trillion in mortgages. Instead, the milestone recedes, and the mortgage exit looks like a protracted, multi-year process.
Alongside the conversion, the Fed has resumed Treasury purchases through reserve-management operations — framed explicitly as plumbing, not stimulus. The stated purpose is to keep reserves comfortably in the "ample" zone, currently around $3.2 trillion, and to absorb seasonal swings such as tax-payment dates. The Fed insists on the distinction: these purchases prevent accidental tightening from reserve scarcity, not ease conditions the way QE did.
2.85% prices everything — and spreads hide the risk
Return, then, to that 2.85%. The real yield — the inflation-adjusted return on government bonds — is the true opportunity cost of holding any asset that pays no income. When the Fed raises its policy rate, as it did on September 16, 2026, lifting the target range to 3.75%–4.00%, it pushes up nominal yields across the curve. If inflation expectations stay anchored, those increases flow straight through into higher real yields. Cash and short-term paper pay more in real terms, the present value of long-duration cash flows falls, and the hurdle for non-yielding assets rises — which is how gold can sit a fifth below its January record even as central banks accumulate gold at a record pace, official buying running into private arithmetic that says: why hold the metal when the risk-free alternative pays 2.85% after inflation?
There is a counterweight. The end of QT removes one source of upward pressure on term premiums — the extra yield investors demand for holding longer-dated bonds. With the runoff over, that pressure is gone. But the September rate hike and elevated real yields have more than offset the relief at the long end. Watch the threshold: any sustained move above 3% would mark genuinely restrictive territory not seen since before the financial crisis. At 2.85%, the real yield is fifteen basis points from a line that changes the conversation entirely.
Who feels it first? Anyone whose asset pays later rather than now. A long-duration equity is a claim on earnings years in the future; discount those earnings at a higher real rate and the claim is worth less today. A corporate borrower feels it as a higher refinancing bar: rolled-over debt must compete with a risk-free real return of 2.85%, and lenders price from that floor upward. Gold feels it as a standing invitation to sell, renewed daily. The real yield is the one price every other price must justify itself against — which is why fifteen basis points here carry more weight than a headline rate decision.
Corporate credit spreads — the extra yield investors demand to hold corporate bonds over Treasuries — have remained remarkably contained through the QT wind-down and renewed hiking. The containment is not a mirage: corporate cash piles remain large, default rates subdued, and demand for yield-bearing paper intense with money-market rates near 4%. But contained spreads deserve scrutiny, not celebration. Spreads are tightest precisely when investors feel safest, and 2.85% real yields, renewed Fed tightening, and a remodeled balance sheet leave a narrower margin for error than the headline numbers suggest. If funding-market stress were to reappear — the very scenario the Fed ended QT to avoid — credit spreads would show it first, before equities notice.
There is a feedback loop buried in that calm. Money-market rates near 4% keep yield-hungry buyers arriving for corporate paper, compressing spreads further, making borrowing look cheap, encouraging more of it. The interruption the Fed fears is not a default wave. It is a funding-market accident, the plumbing kind. And tight spreads leave no compensation for being wrong tomorrow.
Track three plumbing indicators the way the Fed does: bank reserve levels against the roughly $3.2 trillion comfort zone; usage of the Fed's Standing Repo Facility, which Wolf Street noted hit zero around year-end; and the overnight reverse repo facility, whose near-zero balances signal excess liquidity has largely been drained. These gauges flash first if the post-QT framework ever comes under strain — quiet dials that move before the loud ones.
Monetary policy in 2026 operates through two channels. The policy rate is the active tightening tool, flowing into real yields and valuations. The balance sheet has shifted from a tightening lever to a stability instrument, managed to keep reserves ample and funding markets calm. Three things belong on every investor's dashboard — real yields, the MBS-to-Treasury conversion pace, and the funding-market plumbing. Two hands, two different jobs. In 2026, markets ignore either one at their peril.
Fed policy does not stop at the US border: for countries negotiating with the IMF, Washington's real yields set the price of every alternative, which is why Pakistan's review talks, with $1.2 billion at stake, are hostage to the same data prints. Risk appetite further out tells the same story in miniature: Bitcoin cleared $86,000, then reversed when the jobs data landed.
Disclaimer: This article is general information for educational purposes only and is not investment advice, financial advice, or a recommendation to buy or sell any asset. Markets are volatile and past patterns do not guarantee future results. Consult a licensed financial professional before making investment decisions.
Sources
- HSH.com — "The latest move by the Federal Reserve – September 16, 2026" (QT ended Dec 1, 2025; MBS holdings $1.914T as of Sept 10) — hsh.com
- DollarLiquidity.com — "What Is Quantitative Tightening (QT)? How the Fed Balance Sheet Shrinks and Why It Matters" (QT1 vs QT2 history, $8.96T to ~$6.8T) — dollarliquidity.com
- 8figures.com — "Quantitative Tightening Explained: How the Fed Shrinks Its Balance Sheet" (QT ended Dec 1, 2025; ample reserves framework) — 8figures.com
- Engin Demirel / Medium — "A Silent Turning Point: How the End of QT Is Shaping the 2026 Financial Cycle" (Oct 2025 decision, preemptive rationale) — medium.com
- Respect Investment — gold price analysis September 2026 (10-year real yield 2.85%, highest since Nov 2008) — respectinvestment.com
