The Curve Never Inverted in 2026. It Flattened Anyway.

The gap between two-year and ten-year Treasury yields is 0.38%, against 0.72% on the first trading day of the year. That is 34 basis points of flattening, and it happened without a single inverted close. The last time the spread finished a session below zero was September 2024.

Flattening has two forms and they are not interchangeable. In a bull flattening, long yields fall faster than short ones, usually because growth expectations are being marked down. In a bear flattening, short yields rise faster than long ones, because the policy path is being marked up while the distant outlook is not. This year is the second. The two-year yield is 4.61%, up 114 basis points this year; the ten-year is up 79. The front end did the work.

That follows from what each end prices. The two-year is mostly a bet on the policy rate over the next eight quarters, and that path has been revised upward: the effective funds rate stands at 3.75% after the quarter-point increase in September. The ten-year prices expected short rates over a decade plus a term premium, and it moved less because the market did not carry the tightening out to the far end of the horizon.

The direction is not the same thing as the level, and this year makes the distinction plain. The spread reached its high of 0.74% in January, bottomed at 0.20% in September — a level it had also touched in 2025 without going below — and has since widened 8 basis points in October alone. Plotted across the year it is a convergence with a bounce in it, not a straight line.

The inversion record matters here because the signal was never the shape on its own. An inverted curve earned its reputation by pricing a central bank that would be forced to cut; the flattening that stops short of zero prices a central bank that is close to done without saying it has overshot. That is a weaker claim, and it is the one the data supports. If the two-year's year-to-date increase ends the year smaller than the ten-year's, the bear label is wrong and this piece should be rewritten rather than patched.