Yield Curve’s Predicted Recession Is Arriving

Elon Musk is famously having a feud with President Trump this week, which is getting a lot of media attention.  As one of the salvos in this feud, Musk posted on X that, "The Trump tariffs will cause a recession in the second half of this year."

I agree with the forecast, although not necessarily the causation he sites.  If a recession was going to happen anyway, and if tariffs make it worse, then did the tariffs "cause" the recession?  That is a rhetorical question, for which there never can or will be a satisfactory answer.

The inversion of the yield curve which occurred from November 2022 to November 2024 has been telling us to expect a dip in the economy.  This usually works with about a 15-month lag time, although the actual lag can vary slightly from that 15-month period.  The current economic slowdown which is appearing in the latest numbers for real GDP is a little bit late in getting started.

If we count 15 months forward from the last month of inversion in the 10-year to 3-month Treasury yield spread, that takes us to February 2026.  So a recession appearing anytime between now and February 2026 would still be right on time.  And remember that the actual lag time can be a few months off, and that is normal.

Back in 1974, Julian Shiskin published an article in The New York Times in which he proposed that a formal definition of a recession should be 2 consecutve quarters of negative GDP growth. Shiskin was a prominent economist and former Commissioner of the Bureau of Labor Statistics, and he proposed this rule of thumb as a practical way to identify recessions based on GDP data.  The concept was not entirely new, though, as economists had long used declines in economic output to describe recessions, but Shiskin’s formulation popularized the specific two-quarter metric.  It has never become the "official" definition of a recession, though, because economists do not want to abandon their prerogative of getting to declare when a recession occurs according to their own preferred metrics.

Q1 of 2025 saw a tiny drop in "real" GDP, meaning adjusted for inflation.  The "nominal" GDP numbers still showed growth, but not as much growth as the inflation rate used in those calculations.  So if Q2 were to see another shrinkage, this would meet Shiskin's definition (which was last met in 2022, although the NBER elected not to call that an official recession), and would renew the discussion about whether Shiskin's definition is the proper one.  It would assuredly also get political types all riled up about whether Trump is screwing up the economy, or instead that he inherited a screwed up economy.  That is not an argument which the GDP numbers will be able to settle to everyone's satisfaction.

The yield curve thus far in history is undefeated in terms of forecasting economic recessions when it sees an inversion.  So it would be quite unusual if we somehow did not get a recession in 2025.  That is an important point to keep in mind as we all evaluate the various tariff proposals, and the commentary about those proposals.  The politicians in Washington DC have a small ability to change that, for better or for worse, but the Fed has more of an effect when it allows the yield curve to invert.  It has thankfully disinverted now, but we are still waiting for the 15-month lag to go by.