The US economy hit an important milestone in December—the yield curve finally “uninverted,” meaning it returned to its typical upward sloping state, where yields on the 10-year Treasury were higher than yields on the 3-month T-bill. This was significant because the yield curve had been inverted since October 2022, and this two-plus year inversion was the longest stretch, going back at least 85 years.
During this long stretch of an inverted curve—where the yield on the 3-month had been higher than the yield on the 10-year—I often wrote on the historical connection between such an inversion and an impending recession. (Most recently I wrote about the Fed’s rate cuts in the fall of 2024 and how they would pave the way out of the inversion.) On the InFi podcast I recently covered the uninversion and what it portends for a US recession; the present post accompanies that discussion.
First let’s just document the current situation:
As the chart indicates, whenever the blue line—which is the spread between the yield on the 10-year Treasury and the 3-month T-bill—goes negative, meaning it dips below the horizontal black line, there soon follows a recession (indicated by the shaded gray bars). Furthermore, whenever there is a recession, there has always been a nearby, earlier inversion. Thus the yield curve as a “predictor” of recessions has no false negatives or positives, during the 45 years documented on this chart.
In fact, we can go back even further, but the data sets available at FRED force me to do the calculation manually. It yields this picture:
So for the purists, let me be clear that the apparent “near miss” in the early 1990s is spurious; as our first chart indicated, the 10yr-3mo spread really did go negative in that episode, but because the two data sets I’m mashing together in the second chart aren’t exactly apples-to-apples with their units, it doesn’t quite work.
However, if we go farther back on the second chart, it looks like a “false positive” occurs around 1967, where the blue line dips below the horizontal black line, and yet there isn’t a recession soon after. This particular incident isn’t a statistical illusion; it really does constitute a “false positive” as far as the 10yr-3mo spread works as a recession indicator. However, even in this case, the economy did slow down in this period, it just didn’t quite slow enough to be classified as a recession.
Now that the yield curve has broken back above the horizontal black line, it’s natural to ask: Is there a recession in the cards, and if so, how long before it hits?
If we use the prior four episodes as guides, we calculate the following lags:
Based on the above, if history repeats itself then we should expect a recession to begin during 2025. As I always caution on these matters, that doesn’t mean we will know in real time that a recession has begun, because the data always come with a lag. But my prediction is that, years down the road, looking back the National Bureau of Economic Research (NBER)—which has the task of officially dating the beginning and end of economic expansions—will put down in the record books that our recession began this year.
I assure the skeptical reader that my confidence in the yield curve indicator isn’t simply because of historical correlations. On the contrary, it dovetails nicely with the theory of the business cycle developed by the great Austrian economist, Ludwig von Mises. (To my knowledge, Mises himself never talked about the yield curve inverting.) Ryan Griggs and I have a paper in the Quarterly Journal of Austrian Economics spelling out the affinity of the Misesian theory and the yield curve’s empirical track record.
For our purposes here, let me simply comment on the following chart, which decomposes the 10yr-3mo spread into the two constituent series:
As the chart illustrates, what happens during a “boom” period—when the yield curve is “normal” and upward-sloping—is that the Fed/banking system push down the 3-month yield (red line), while the 10-year yield (blue line) doesn’t drop nearly as much. Then, presumably because of rising consumer price inflation and other warning indicators that the economy is “overheating,” the Fed and banks begin to tighten, which causes the red line to shoot up. It overshoots the blue line, which is what we see as an inverted yield curve. Then the economy soon after tips into recession, but the Fed begins cutting before recession actually starts, presumably because the Fed can see signs of the weakening economy.
To reiterate, this narrative lines up with the business cycle theory put forth by the Austrians, who argue that loose monetary policy leads to artificially low interest rates. This “easy money” policy causes an unsustainable boom, which inevitably leads to a crash once the banks get nervous and begin tightening. Thus the boom and recession don’t occur merely because of lines on a chart; the injection of new money during the early stages of a boom leads entrepreneurs to malinvest resources in projects that cannot be carried to completion.
As I write this post, Trump has just finished his first week in office. It has been a whirlwind, with many decisions—including the pardoning of Ross Ulbricht—that are clear signals that the US is poised to become the AI and crypto capital of the world, as Trump put it during a virtual address to the World Economic Forum. This of course is music to our ears here at infineo, where we are using AI and blockchain technology to revolutionize life insurance as an asset class. So although I’m optimistic about the economic benefits flowing from Trump’s policies, the US economy unfortunately is inheriting the legacy of the massive infusions of cheap credit in response to the Covid panic. Trump’s moves may cushion the blow, but I still think we will see an official recession starting this year.
Dr. Robert P. Murphy is the Chief Economist at infineo, bridging together Whole Life insurance policies and digital blockchain-based issuance.
Twitter: @infineogroup, @BobMurphyEcon
Linkedin: infineo group, Robert Murphy
Youtube: infineo group
To learn more about infineo, please visit the infineo website.