Stats › Valuation › Yield curve
The yield curve as a recession signal
Normally you get paid more for lending money for longer. Ten years of risk should cost more than two, so long rates sit above short ones. Every so often that flips over, short rates end up higher than long rates, and the financial press calls it an inverted yield curve and starts talking about recessions.
Okey so, the version everyone quotes is the 10 year Treasury yield minus the 2 year. It's the most cited recession indicator there is. What makes this page worth reading is that the Federal Reserve's own researchers have published a paper arguing that particular spread is the wrong one to look at, that a different measure carries the actual information, and that the whole mechanism works backwards from how it's usually described. The curve isn't causing anything. It's reporting what the market already thinks.
So this page is mostly about that paper and what it says. What the near term forward spread is, why the term premium muddies the popular measure, and why the 2019 inversion followed by the 2020 recession is much weaker evidence than it looks. It's less a statistics page and more a page about reading an indicator honestly. The Fed note is linked in full at the bottom and it's readable without a finance degree. If you click one link off this site, make it that one.
The 10 year minus 2 year Treasury spread is the most quoted recession indicator in finance. The Federal Reserve's own researchers have argued it should not be. Their 2022 paper found that a near term forward spread, reflecting expectations for Fed policy over the coming eighteen months, carries genuine predictive power for recessions, GDP growth and equity returns, while the 2 to 10 spread adds no incremental information once that is accounted for. The mechanism is also not what most people assume: term spreads predict recessions through reverse causality, reflecting pessimism that already exists rather than causing anything.
Members only
The rest of this page is for members
Below this point there are 7 sections and 2 named sources, roughly 1300 words of it. Every figure carries the source it came from and the date the data is from.
You can keep browsing every statistic in the library for free. The intro and the summary are always open.