The U.S. economy is seemingly on solid ground, but the Treasury bond market is sounding its most severe recession alarm in decades.
Many economists expected the U.S. to suffer a recession last year. Economists surveyed by The Wall Street Journal in late 2022 put the odds of a recession at 63%. The prevailing logic was that the Federal Reserve would raise interest rates too much, causing a substantial decline in spending that would snowball into higher unemployment and an economic downturn.
Instead, the economy remained rock-solid in 2023 despite aggressive rate hikes and elevated inflation. Economic growth actually accelerated last year, supported by strong (albeit slower) increases in consumer spending and business investments. The economy is currently projected to expand at an annualized 2.9% in the first quarter of 2024, above the 10-year average of 2.5%.
In short, recession fears have melted away. Many economists now believe the Federal Reserve will thread the needle and achieve a soft landing, meaning policymakers will tame inflation without triggering a recession. To quote Morgan Stanley analysts, “The U.S. economy is humming along, with nearly all data validating the soft landing.”
However, the Treasury bond market — a recession forecasting tool with a near-perfect track record — continues to sound its most severe alarm in decades. Recessions have typically coincided with a substantial decline in the S&P 500. Here’s what investors should know.
Treasury bonds have predicted past recessions with near-perfect accuracy
U.S. Treasury bonds are debt securities issued by the government. They pay a fixed interest rate until they mature, at which point the bondholder recoups the principal. The interest rate (or yield) is normally higher on long-date bonds as compared to short-dated bonds. So, the yield curve normally slopes up and to the right.
However, the yield curve becomes inverted (starting high and sloping down as it mores right) when long-dated bonds pay less than short-dated bonds. That can happen during periods of economic uncertainty. Some investors hedge against recession risk by purchasing long-dated bonds to get guaranteed returns over an extended time period. Demand for those long-dated bonds drives prices higher and yields lower.
For instance, the 10-year Treasury currently pays less than the 3-month Treasury, meaning that portion of the yield curve is inverted. What makes that noteworthy is the near-perfect accuracy with which those bonds have forecasted past recessions. Specifically, an inversion between the 10-year and 3-month Treasury yields has preceded every recession since 1969, with only one false positive in the mid-1960s.
Treasury bonds are doing something investors haven’t seen in decades
The chart below lists the start date for each yield curve inversion involving the 10-year and 3-month Treasuries since the late 1960s, and the start date of the subsequent recession. The chart also shows how much time elapsed between the two events.
YIELD CURVE INVERSION | RECESSION START DATE | TIME ELAPSED |
---|---|---|
December 1968 | December 1969 | 12 months |
June 1973 | November 1973 | 5 months |
November 1978 | January 1980 | 14 months |
October 1980 | July 1981 | 10 months |
June 1989 | July 1990 | 13 months |
July 2000 | March 2001 | 8 months |
August 2006 | December 2007 | 16 months |
June 2019 | February 2020 | 8 months |
The 10-year and 3-month Treasury yields have inverted before every recession since 1969, with no more than 16 months between the inversion and subsequent recession. For context, the current inversion began 15 months ago in November 2022, implying that the U.S. could slip into recession by the end of next month.
There is another point investors should consider. The current yield curve inversion was most severe in May 2023, when the average yield spread (10-year Treasury yield minus 3-month Treasury yield) dropped to -1.71%. The yield curve has not been so steeply inverted since June 1981, when the average yield spread was -2.04%.
The curve has flattened since May 2023. The average yield spread was -1.3% in January 2024, but that still represents the lowest reading since August 1981, when the average yield spread was -1.43%. In that context, Treasury bonds are doing something investors have not seen in decades. In fact, a blog post from the Federal Reserve Bank of St. Louis says the implied “recession probability would be unprecedentedly high for a false positive.”
Stocks typically decline sharply during recessions, but tend to rebound sharply before recessions end
The S&P 500 is commonly seen as a benchmark for the broader U.S. stock market. Since its inception in 1957, the U.S. economy has been hit by 10 recessions, during which time the S&P 500 declined by an average of 31%. In other words, if the economy does slip into a recession this year, history says the stock market would suffer a substantial drawdown.
That sounds quite alarming, but investors should avoid selling their stocks. In fact, the most prudent course of action is to stay invested and continue buying good stocks at reasonable valuations. I say that for three reasons. First, the bond market has been wrong in the past. The 10-year and 3-month Treasury yields inverted in January 1966, but that event was not immediately followed by a recession. The current yield curve inversion, despite its severity, could be another false positive.
Second, even if the economy suffers a recession, investors who sell wouldn’t know when to buy again. The S&P 500 usually rebounds about four or five months before a recession ends, and the index has historically returned a median of 30% between its bottom and the end of a recession, according to JPMorgan Chase. Investors who attempt to time the market will probably miss some of those gains.
Third, despite the occurrence of several recessions since 1994, the S&P 500 returned 10.3% annually over the last three decades. Investors can assume similar returns over the next three decades even if the economy suffers a recession this year. But any attempt to time the market could backfire, setting investors up for long-term underperformance.