Why the ten-year yield no longer obeys the Fed

When a central bank holds its policy rate flat for many months, the long bond usually follows its lead. This year the opposite has happened. The Federal Reserve has kept its short-term rate at 3.50% to 3.75% since the start of 2026, yet the yield on the benchmark ten-year Treasury note has climbed to nearly 5%, with the thirty-year touching levels not seen since 2007. The move has been driven not by the Fed but by the market's own judgment about the American state. That is the striking part, and it is worth working out, because it tells a retail investor where the discount rate for nearly everything is being set.

For most of the past generation, the ten-year yield was best understood as a transmission of the policy rate. When the Fed cut, long yields fell; when it tightened, they rose. The central bank's guidance was effectively the price of long-dated money. That neat mechanical link has come apart. The policy rate sits still while the long end climbs on its own. The market is no longer pricing the Fed's intentions. It is pricing the term premium — the extra compensation demanded for holding a long government bond against an uncertain future — and that premium now has a clear, uncomfortable source.

Start with the arithmetic of the American state. The budget deficit is on track for about $2 trillion this fiscal year; July alone posted a shortfall of $432bn, the widest monthly gap since early 2021. Public debt has climbed above $40 trillion, roughly 100% of gross domestic product. Financing that pile is no longer cheap: interest costs reached $1.12 trillion through July and are projected to hit about $1.37 trillion for the year, more than the federal government spends on anything other than Social Security and Medicare. When the state borrows more at a higher price, and the term premium rises to reflect the risk, longer yields get pushed up regardless of what the Fed does with the overnight rate.

Inflation supplies the second leg. Core prices are running near 2.5%, stuck above the Fed's 2% target, while headline inflation is around 3.4% — and the conflict with Iran has pushed oil higher, raising the chance that energy-driven shocks keep recurring rather than fading. A third leg is supply from the private sector. America's five big AI hyperscalers have issued roughly $220bn of debt this year, more than double last year's total, and global corporate issuance has reached a record. A wall of duration from companies, not governments, is crowding out Treasuries in the market for long-dated money.

The question is where this ends, and on that point the forecasting community has already told the world it does not know. In the August Reuters poll of bond strategists, the median forecast had the ten-year falling to 4.50% in three months and drifting to 4.34% in a year — in short, that yields had peaked and would ease as the economy cooled. Yet when the same strategists were asked to bet against their own near-term forecast, 82% said the yield was more likely to come in higher than the three-month figure. The consensus expects rates to fall; it does not believe itself.

That confession is worth more to an investor than the forecast it accompanies. The ten-year yield is not an arcane policy variable. It is the base rate against which equities, housing and corporate borrowing are all priced: it has already pushed thirty-year mortgage rates to nearly 6.7% (a one-year high), and it sets the discount rate that decides how much a promised future dollar of earnings is worth today. When the term premium rises and long yields detach from the Fed, the cost of capital is being repriced by markets rather than managed by a central bank. A holder of stocks is, in effect, short these moves twice over — through the discounting of earnings and through the competing appeal of a risk-free bond that now pays.

The opposing case deserves its full weight. The selloff has so far been orderly rather than panicked, and Scott Bessent, the treasury secretary, argues that worries about debt overlook the strength of American growth. There is even a respectably liberal reading in which market-set long rates are a correction toward a healthy cost of capital after years in which the Fed held financing artificially cheap. All true, and none of it restores the old relationship.

The structural point survives the concession. The Fed's promise to keep financial conditions easy no longer reaches the long end of the curve, and a stable policy rate is no longer a floor under the discount rate for everything else. Until the deficit arithmetic or the path of inflation changes, the honest bet is that long yields are more likely to surprise on the upside than the downside. That is a reading supported both by the data and by the bond market's own forecasters, who write down a falling number while privately insisting, by a margin of four to one, that they expect it to go the other way.