Business Cycles6 分钟阅读

Yield Curve Inversion and Recession

The side lending money for longer receives less interest—doesn't that seem backwards? Let's unpack, one piece at a time, why the news frets about recession when 'short- and long-term rates have inverted.'

What is rate inversion?

Usually, the longer you lend money, the more interest you receive. Lending to a friend for 10 years is more uncertain than for 1 year, so it's common sense to demand more interest for that risk. Treasuries are the same, so normally the 10-year yield is higher than the 2-year yield.

But sometimes this order flips. When a short-maturity Treasury yield, like the 2-year, rises above a long-maturity one, like the 10-year, that's called an 'inverted yield curve.'

The most-watched gauge is the U.S. 10-year minus the 2-year (2s10s for short). When this value turns negative, an inversion has occurred. Another gauge especially trusted by academics and the U.S. Federal Reserve is the spread of the 10-year minus the 3-month (10y3m).

A 'spread' means the difference between two rates. If the 10-year is 4.0% and the 2-year is 4.5%, the spread is -0.5 percentage points—an inverted state.

Why does this strange thing happen?

The reasons an inversion arises are explained along two broad lines.

First, the expectations theory. A long-term rate is really the market's average expectation of 'how short-term rates will move over the next several years.' If investors believe 'the economy will soon worsen and the central bank will cut rates sharply,' they price in those future low rates and accept today's long-term rate as low. As a result, the long-term rate drops below the short-term rate, producing an inversion.

Second, the term premium. This is the extra compensation demanded for holding long-term bonds for a long time; when it becomes very low or negative, the long-term rate can be pressed down and an inversion can appear even without a particular recession expectation. So keep in mind that an inversion isn't always a 'recession prophecy.'

Source: yield-curve explainer materials from the Bank for International Settlements (BIS), the U.S. Brookings Institution, and the Richmond Fed.

The basis for calling it a 'recession signal'

The yield curve inversion became famous because of its historical track record. Cross-referencing several sources, it's summarized that an inversion appeared first before almost every official U.S. recession since 1955.

That said, the timing is erratic. Various studies report that from the inversion to the start of a recession takes roughly 6 to 24 months, on average a little over a year (the median differs slightly by source, at about 15–16 months, or an average of about 13 months, etc.).

The New York Fed's Estrella and Mishkin researchers even built a model that calculates 'the probability of a recession within the next 12 months' from the 10-year-minus-3-month spread. It's widely cited that this gauge is better than other economic indicators at seeing 2–6 quarters ahead.

The historical lead time is compiled differently by source. It's safer to understand it as 'roughly this range' rather than a single fixed value.

Why you still shouldn't blindly trust it

A good track record doesn't make 'inversion = recession, always' true. Representatively, 1966–67 is often cited as a 'false signal' case where an inversion appeared but did not lead to an official recession.

More recently, the 2022–2024 inversion lasted the longest on record (about 26 months), yet no official recession immediately followed, bewildering many experts. Factors like the aforementioned shrinking term premium and large-scale quantitative easing complicated the interpretation of the signal.

Above all, an inversion doesn't tell you 'when' a recession will come or 'how' deep it will be. Because the lead-time variance is large, trying to predict market direction or time trades based on an inversion alone is dangerous. This isn't a prophecy but merely one of many signals.

This article does not recommend buying or selling any specific asset at any specific time, nor does it predict future prices. Yield curve inversion is treated only as an educational concept for understanding the business cycle.

So how should a long-term investor take it?

When news of a rate inversion comes out, it's easy to get scared first, but from a long-term investing perspective you can view it a bit differently. It's healthier to take an inversion not as a signal that 'you must sell right now,' but as a reminder that 'the business cycle may be in its later stage, so it's time to check how much decline my assets can endure.'

If you check with your own eyes how much stock prices fell in past recessions (the maximum drawdown) and how long they stayed below principal (the loss duration and recovery period), you'll be less shaken when the next downturn comes.

'The Return of Almost Everything' is a site I operate personally, and it shows the drawdown and recovery process during past crisis windows by asset. You can also compare how the results differed between putting money in all at once and contributing steadily in installments. Rather than reacting to inversion news, use it as a tool to check whether your investing habits can endure a crisis.

常见问题

Q. If rates invert, do stocks fall right away?

No. An inversion doesn't tell you exactly 'when' a recession will come. Historically, from the inversion to a recession took roughly 6–24 months, on average a little over a year, and during that time the stock market often kept rising. It's hard to use the inversion itself as a trade-timing signal.

Q. Has an inversion always been right?

Historically an inversion preceded almost every U.S. recession, but there are exceptions. The 1966–67 inversion did not lead to an official recession, and the record-long 2022–24 inversion was not immediately followed by a recession either. It's a 'signal that has almost always been right,' not a '100% prophecy.'

Q. Should I watch the 10-year-2-year or the 10-year-3-month?

Both are widely used. In the news, the 10-year-2-year (2s10s) is often cited, while academics and the U.S. Fed's recession-probability models prefer the 10-year-3-month (10y3m). Rather than looking at just one, it's safer to reference several gauges together.