1. How to read a yield curve and what each shape means

Macro 3: The yield curve and the business cycle

The previous level left an idea half-finished: there is no such thing as the interest rate, there is one for every maturity. Plot all of them at once and you get the yield curve, probably the chart with the most information per square centimeter in the whole market. It does not predict the future, but it tells you precisely which future people are paying for with their own money right now: the aggregate opinion of millions of investors about the coming decade, condensed into a single line.

What a yield curve actually is

A yield curve puts maturities along the horizontal axis, from one month to thirty years, and yields up the vertical one. Every point refers to debt from the same issuer, so the only thing changing from one to the next is time and not credit quality. The global benchmark is the US Treasury curve; the euro area, which has no single issuer, uses the German curve as its equivalent and looks separately at the gap against Italy or Spain.

What the vertical axis shows is the yield to maturity: the annualized return you would earn by buying that bond today at the market price and holding it to the end, collecting every cash flow along the way. It is not the coupon. A bond issued with a 1% coupon and bought at a discount can yield 4%, and what the market quotes is always the yield.

In practice the curve gets summarized into a single spread: a long maturity minus a short one. The two most quoted are the 10-year minus 2-year and the 10-year minus 3-month. When the press says the curve has inverted, it almost always means one of those two numbers has turned negative.

The four shapes

• Upward sloping, the normal one. The long end pays more than the short end. Investors demand something extra for tying up their money for longer and usually expect some growth and inflation ahead.
• Flat. It barely matters whether you lend for two years or for ten. It tends to be a transition: the market reckons rates are close to their peak.
• Inverted. The short end pays more than the long end. It is counterintuitive, and that is exactly why it is informative: someone would rather lock in a mediocre yield for ten years than keep rolling over every few months, because they think there will soon be nowhere better to put the money.
• Humped. It rises to an intermediate maturity and falls after that. This is the typical look of a regime change caught halfway: the central bank still hiking and the market already pricing cuts.

What is inside the slope

A ten-year yield is, in the standard decomposition, two things added together: the average expected path of short rates over those ten years and the term premium. The slope of the curve is therefore also two things added together, and confusing them is the most expensive reading error made with this chart.

If the curve flattens because the market expects cuts, the signal is economic: someone believes monetary policy is too tight. If it flattens because the term premium has compressed, say because a pension fund is buying long bonds for reasons that have nothing to do with its growth forecast, the same shape means something else entirely. And here is the problem: the term premium cannot be observed, it is estimated with models, and different models give different numbers for the same day. Every reading of the curve carries that ambiguity with it.

The vocabulary of the trading desks

Four expressions combine two axes. Bull means yields falling, that is, bond prices rising; bear means the opposite. Steepening means the spread widens; flattening, that it narrows.

• Bear flattening. All yields rise, but the short end rises more. The classic signature of a central bank raising rates.
• Bull flattening. All yields fall, but the long end falls more. The market is starting to buy insurance against recession.
• Bull steepening. All yields fall, and the short end falls more. This is what happens when the cuts actually arrive.
• Bear steepening. All yields rise, and the long end rises more. It usually reflects higher expected inflation, heavier debt issuance or a bigger term premium being demanded.

Two curves with the same slope may have got there by opposite routes. The slope on its own is not enough: you have to look at which way the level has moved.

A real case: the US curve between 2022 and 2024

In January 2022 the 10 minus 2 year spread sat comfortably in positive territory, of the order of three quarters of a point. The Federal Reserve raised its policy rate by some 425 basis points in nine months and the two-year point, which is the one that responds to monetary policy, shot up; the ten-year rose as well, but far less. Textbook bear flattening. In July 2022 the spread crossed into negative territory.

What came next is the interesting part. This was not a matter of weeks: the inversion lasted more than two years, the longest on record, and by the middle of 2023 it had gone beyond a full percentage point deep, something not seen since the early eighties. And the exit did not come because the ten-year bond paid more, but because the two-year collapsed once the Federal Reserve began cutting in September 2024: bull steepening. The curve undid the inversion by doing the opposite of what it had done on the way in.

Keep the method, not the episode. Knowing that the curve is inverted is not reading it. Reading it means knowing which part has moved and in which direction, because that is where the difference lies between the central bank is tightening and the market is frightened.