What does the U.S. yield curve tell us about recessions?

Today, the vast majority of the U.S. yield curve is negatively sloped. Every time this has happened in the past, a recession soon followed. This is no doubt troubling for those adversely affected. But, as an investor, I am not unduly concerned – any market turmoil could present opportunities to invest in quality businesses at bargain prices.

The U.S. Federal Reserve’s Chairman, Jerome Powell, addressed the U.S. Senate Banking Committee on Capitol Hill in Washington DC last week, warning investors that rates might have to rise in larger increments and the previously anticipated terminal rate of 5.25 per cent was likely too low.

The reaction from U.S. bond markets was predictable. Shorter-term rates went up ­reflecting the expectation that rates in the near term need to be higher, but longer-term bond rates went down in acknowledgement that the higher short-term rates would presage a recession, ultimately forcing rate cuts.

When securities with their interest rates on the y-axis are plotted against time on the X-axis, the result is the yield curve. Typically, yield curves are sloped positively, with longer-term yields higher than short-term rates. This is because longer-maturity investors demand a return premium to compensate for the additional risk associated with making longer-term investments.

Today, however, the vast majority of the U.S. yield curve is negatively sloped and there are some important implications. Perhaps the most important implication of the majority of the yield curve being negatively or inversely sloped, as it is now, is that it foreshadows a recession.

In the late 1970s, early 1980s, late 1980s, early 2000, in 2007 and in 2019 more than 55 per cent of the U.S. yield curve was inversely sloped.  And on every occasion, without exception, the event was followed by a recession.

Perma-bear and GMO founder Jeremy Grantham had this to say recently about the presence of an inversion between the 10-year and three-year part of the curve: 

“The most ancient and effective predictor of future recession, the 10-year minus 3-month yield spread, is now clearly signalling recession within the next year. This spread has gone negative only 8 times in the past 50 years, and all 8 times have been followed by recessions. To rub it in, there have been no other recessions. That is, every one of them was preceded by a negative reading. Not bad.”

Ok, so let’s say the current prevalence of inversions maintains the historical record of predicting a recession. So what? Does it mean the stock market falls from here?

Well, it depends on whom you speak to.

According to RBC Global Asset Management, “There has never been a consistent relationship between the stock market and the economy. While the two tend to loosely move in the same direction, they often act in widely different ways – particularly over shorter time periods.”

Jeremy Grantham’s analysis, however, is slightly different. He looked at recessions and market troughs, noting, “You can look at all bear markets, routine to savage, and conclude that the market low has lagged the start of recession by an average of 7 or 8 months. I think though that doing so is to make a classic error. The great psychological bubbles are very different from ordinary bull markets and produce very different results for everything. In this case, I am only interested in 1929, 1972, 2000, and the 2006 housing bubble. In these cases, the market lows took their time.”

In other words when averaged across all bear markets, the market bottomed eight months after the recession commenced. But looking at only the bear markets that followed a serious bubble, which Grantham says 2021 was, the lows take longer to be set.

Grantham again: “In 1929 it was years after the recession started – let’s dismiss this as an extreme case. But even the other three took 11, 15, and 19 months. If we conservatively assume 1 year, and we believe a recession will likely not start for 6 months to a year, you can see how easily we can conclude that the final low for this market might be well into 2024.”

If you believe all of that, the only thing left is to define a recession and its starting date. And here’s a bit of a cat amongst the pigeons: The U.S. recorded two negative quarters (a conventional definition of a recession) last year. The U.S. economy shrank in the three months to June 30, 2022 by 0.9 per cent. In the quarter preceding it, to the end of March 2022, GDP, or gross domestic product, decreased at an annual rate of 1.6 per cent.

So if, as Grantham suggests, the market bottoms a year after a recession commences, the low should be in now. Of course, conveniently, the U.S. non-profit, non-partisan organisation called the National Bureau of Economic Research did not determine the U.S. had a recession.

I will leave it for you to decide.


View all funds