You do not have to guess, because the two parts are separately observable. A nominal yield is the inflation lenders expect over the term plus the return they want on top of it, called the real yield. The inflation-protected Treasury (TIPS) prices that second part directly, so expected inflation is simply the gap between the two.
The arithmetic is one subtraction. Nominal 10-year minus 10-year TIPS equals the market's expected inflation over ten years, the figure usually called the breakeven. Both series are free and daily, so you can run the split for any date and any horizon.
Why it changes the conclusion rather than decorating it. A rise driven by expected inflation is a story about prices, and a policy rate is built to address it. A rise driven by the real yield is lenders demanding more to part with money for ten years, and a policy rate does not reach that. Same headline number, two different problems, two different things to watch next.
What the current move is made of, measured. On 23 September 2026 the US 10-year was 5.11 per cent, the 10-year TIPS 2.76, leaving expected inflation at 2.35.
- Over one year the 10-year rose 0.99 points. The real yield rose 1.01 and expected inflation fell 0.02.
- Over two years the 10-year rose 1.36 points, and the real yield was 1.18 of it.
- Over three years the 10-year rose 0.67 points. The real yield rose 0.69 and expected inflation fell 0.02.
Almost the whole move is the real leg, and expected inflation is about where it was one and three years ago. That is a repricing of the cost of long money, not an inflation scare.
Two things to get right. Match the dates: a nominal quote from today against a TIPS quote from last week manufactures a move that did not happen. And a breakeven is a price, not a forecast - it is the level at which the two instruments are indifferent, so it carries risk and liquidity premiums and will not equal any economist's number.
One case where the whole method needs a caveat: a capped market is not a priced market. Japan's 10-year was held down by policy until 2024, near zero until late 2022 and below 1 per cent after, so most of the rise since is the market setting the price again. Decomposing a yield that was administered tells you about the end of the policy, not about lenders changing their minds.
Both series are on https://www.tigzig.com/tremor, so the subtraction is two clicks. Related: why long rates rose while the Fed was cutting, and whether a rate rise reaches supply-driven inflation. Full write-up: https://www.tigzig.com/post/bond-market-rout-sep2026.
Building something like this? How I work covers the rates, the availability and what I take on.