The Term Premium Quietly Became the Most Important Number in Markets
For most of the past decade, the yield on a ten-year government bond could be explained with one question: where do investors think the central bank's policy rate will be, on average, over the next ten years? Add a little, subtract a little, and you had the number. That is no longer a complete answer. A second component, the term premium, has moved from a footnote to the centre of the page.
The term premium is the extra return investors demand for locking money into a long bond rather than rolling short ones. It compensates for uncertainty: about inflation, about how much debt governments will issue, about who will be there to buy it. In the years of quantitative easing it was often negative; central banks were such reliable buyers of long bonds that investors accepted less than the expected path of short rates. Models maintained by the Federal Reserve Bank of New York put it below zero for long stretches of the 2010s.
That era has ended in most large bond markets. Central banks have stopped adding to their holdings and, in several cases, are letting them run off. Governments are issuing more long-dated debt to fund deficits that are structurally larger than they were before 2020. The marginal buyer is now a pension fund, an insurer, a foreign reserve manager or a hedge fund, and each wants to be paid for the risk.
Why it matters is arithmetic. If the ten-year yield is 4.5 percent and the market expects the average policy rate over the decade to be 3.5, the remaining point is term premium. A central bank cutting its policy rate by a quarter point lowers the first number by a fraction of that. It does nothing directly to the second. Mortgage rates, corporate borrowing costs and the discount rates used to value long-lived assets all follow the ten-year, not the overnight rate. A world where term premium is rising can see easier policy and tighter financial conditions at the same time.
The drivers are visible in the data. Debt issuance is one: the volume of long bonds a treasury must sell each year, net of what central banks absorb, has roughly doubled in the United States since 2019. Inflation uncertainty is another; investors do not need to expect high inflation to demand compensation for not knowing. Correlation is a third, and the least intuitive. For twenty years bonds rallied when equities fell, so a long bond was insurance and investors paid for it with a lower yield. When inflation shocks make bonds and equities fall together, the insurance value disappears and the premium returns.
None of this makes the number easy to read. The term premium is not observed; it is estimated, and different models disagree by half a percentage point or more on any given day. The New York Fed's ACM model and the Kim-Wright model at the Federal Reserve Board are the two most cited, and both publish daily series. Their direction over a year is more informative than their level on a Tuesday.
For investors the practical shift is in what to watch. Central bank meetings still matter, but so do auction results, the maturity mix a treasury chooses, and whether foreign official buyers are net purchasers or net sellers in a given month. A weak long-bond auction now moves the ten-year more than a hawkish speech does.
For governments the message is sharper. The cost of running a large deficit is no longer set only by the central bank. It is set by the willingness of private balance sheets to hold long paper, and that willingness is priced every day.
