Bond primer: does an inverted yield curve indicate a recession?

This is a short primer, written in January 2019 as the yield curve flattened toward inversion for the first time since 2007, on a single question: does an inverted yield curve predict a recession? The answer from the record is yes, every time, and the reason is not mysterious. It is the Federal Reserve.

The record

The yield curve is the line drawn through the interest rates on government debt of different maturities. Normally it slopes upward: lenders want more to lock money away for ten years than for two. It inverts when short rates rise above long ones. The Federal Reserve Bank of St. Louis publishes a chart of the difference between the ten-year and two-year Treasury yields going back to the 1970s, with recessions shaded. Every time the line crossed below zero, a shaded area followed within a year or two. There is no inversion in the series that was not followed by a recession, and no recession that was not preceded by one.

Why it inverts

The curve inverts because the Fed makes it invert. At the end of every recovery the Fed raises short-term rates to head off inflation, and it keeps raising them until investors in the bond market conclude that a recession is coming. At that point the logic of the curve reverses. An investor who expects a recession within a year or two does not want a two-year note, because it will mature in the middle of the downturn and have to be reinvested at the lower rates the Fed will have cut to by then. He wants to lock in today’s rate for ten years. So demand shifts to long bonds, their price rises and their yield falls, and the long yield drops below the short one the Fed has just raised. The inversion is the bond market’s bet on what the Fed will do next, and the bond market has been right every time because the Fed has done it every time.

The inversion is also a symptom of something wider. If investors in the safest asset in the world are that worried about the next two years, every bank and lender making riskier loans is more worried still, and credit begins to tighten across the economy. That is the mechanism by which a curve becomes a recession: not the shape of the line but the seizing of lending that the shape reflects.

This time is different

Toward the end of 2018 the curve was visibly heading for inversion, and the Fed’s own chart showed it. Asked about it, the Fed’s former chair said there was “a strong correlation historically between yield curve inversions and recessions”, that “correlation is not causation”, and that there were “good reasons to think that the relationship between the slope of the yield curve and the business cycle may have changed”. The observation made here was blunt. If someone sits on your chest and you cannot breathe, it is possible you are having a heart attack; if they sit on your chest twelve times and you cannot breathe twelve times, the causation is not in serious doubt. A central bank that looks at forty-five years of its own tightening producing forty-five years of inversions and recessions, and expects a different result from the same behaviour, has offered a definition of something, and it is not prudence.

The inversion of 2018 to 2019

The two-year and ten-year yields came within a few basis points of each other in December 2018, when the Fed made its fourth rate rise of the year into a falling stock market. This page was written in anticipation of the crossing. The rest of the primer series, on bond yields, prices and what an inversion does to bank lending, is referenced from the original but not restored here.

Later note. The curve inverted in August 2019. A recession began in February 2020. Its cause was a pandemic, which the bond market did not foresee, and the record now contains one inversion whose recession arrived by a route the primer did not describe. The next inversion, in 2022, lasted over two years and, as of 2026, has not been followed by a dated recession, the first such case in the series.