Term premium is back, and it is reshaping the long end of the Treasury market

Bob Sheehan put it plainly: “Term premium is going to matter to the long end again.” That observation, which might read as incremental to anyone who tracks fixed income only casually, carries more weight than its brevity suggests. It marks a structural break from the conditions that defined the Treasury market for most of the past decade, a period during which term premium was essentially compressed to zero or negative, and the long end of the curve was driven by expectations alone.

The external record now confirms the shift Sheehan described. The New York Federal Reserve’s ACM model, which is the most widely cited academic framework for decomposing Treasury yields into expectations and term premium components, shows the 10-year and 30-year term premium turning positive and reaching multi-year highs through 2025 and into 2026. Reuters, RSM, Vontobel, and PIMCO have each noted the same development, with Reuters posing the question in May 2025 of how much further the term premium can rise. The fact that multiple independent analytical frameworks are converging on the same directional finding makes this harder to dismiss as noise.

Term premium, for readers who follow equities more closely than bonds, is the extra yield investors demand for holding long-duration debt rather than rolling over short-term bills. When it is near zero, the yield curve flattens and long rates track the market’s expectations of future short rates almost mechanically. When it rises, long rates can move independently of those expectations, and the long end becomes sensitive to supply, inflation uncertainty, fiscal trajectory, and the broader appetite for duration risk. In practical terms, a positive and rising term premium means the 30-year Treasury can sell off even when the Fed is expected to cut.

I think we're bringing back kind of this um this new thing where term premium is going to matter to the long end again. Bob Sheehan

That is not a minor technical distinction. It changes the analytical framework that portfolio managers, corporate treasurers, and mortgage market participants need to apply. Models built on the low-term-premium era, where the 10-year yield was a reasonably clean read on the Fed’s expected rate path, are less reliable in the environment Sheehan identified. Duration management becomes a separate decision from rate-path positioning in a way it simply was not during the prior decade.

The sources Reuters assembled in May 2025 did not treat term premium’s return as a surprise. They treated it as a feature of a fiscal and inflationary environment that differs meaningfully from the post-financial-crisis period. Sustained deficit spending increases Treasury supply. Supply at the margin requires incentive to absorb. That incentive takes the form of a higher term premium. Add residual uncertainty about the inflation path and the Federal Reserve’s long-run reaction function, and the compression that characterized the 2010s becomes harder to reproduce.

What makes Sheehan’s framing useful is not that it predicted the specific level of the term premium at any given date. It is that it correctly identified the directional regime change before the multi-year high readings became widely reported. The ACM model data and the analyst commentary that followed confirm the direction. The question the market is now working through, which Reuters captured in its headline, is not whether term premium has returned but how far the re-rating runs from here. That is a harder question, and the evidence does not settle it. What the evidence does settle is that the decade-long assumption of a structurally suppressed term premium deserves to be retired.