You don’t really need to know a lot about the economy or bond market to know that there’s one signal that investors live in fear of more than just about any other. It’s called the inverted yield curve — which just means a flippening of sorts in the relationship between long-term and short-term U.S. government bonds. Under normal conditions, the longer-term bonds – say, 10-year bonds – pay a higher rate of interest (aka yield) than shorter duration ones, like the two-year. When there’s an inversion, that relationship switches, so the 10-year temporarily pays a lower rate than the two-year. (The relationship of the yields on all Treasuries ranging from a one-month note to a 30-year bond constitutes the yield curve.) The phenomenon defies easy logic, but it has consistently predicted every recession for more than a half century.
Unfortunately, this is all very relevant to the health of the economy today: The yield curve inverted nearly a year ago, and it remains inverted. While some recession watchers have recently begun to declare the coast clear — predicting also that the indicator will fail this time — others note that the shape of the curve has been changing lately, and they are more worried than ever. To make sense of this high-stakes puzzle, we went to the definitive expert on yield-curve inversions: Campbell Harvey, a finance professor at Duke University, who originally discovered the pattern in 1986. I spoke with Harvey about what the current inversion means, whether there’s any reason to think this time could be different — and how he nearly made a guest appearance in the Sam Bankman-Fried trial.
Could I just ask you to explain the significance of an inverted yield curve, as if you were talking to your niece or nephew or your mother-in-law?
So the yield curve is just the difference between a long-term interest rate and a short-term rate. Think of a ten-year treasury bond as a long term rate and a three-month treasury bill as a short-term rate. So in a normal yield curve, which we see almost all the time, the long rate are higher than the short rates. It makes sense because there’s some risk of locking your money up for a longer period of time. But on certain rare occasions, this gets upended, and you get the short rates much higher than the long rates — that’s unusual, and it’s called an inverted yield curve and suggests a problem in the economy. And over the past eight recessions, from the late 1960s to today, every time we got that upending of the bond market, where the long-term rate was lower than a short-term rate, a recession followed without any false signals — so eight out of eight. And so this is perhaps the most reliable indicator of what will happen in the economy.
And does it usually happen because the Fed is raising rates?
So the yield curve can invert in many different ways. And given that there are only eight examples, there’s not a lot you can learn from each one. But I will say that the current inversion is what I consider the most dangerous type of inversion, where both the short rate and the long rate go up, but the short rate goes up more than the long rate. Let me tell you why. When the yield curve inverts, that’s bad for banks because the bank business model is that they pay depositors a short-term rate and then they receive interest from their loans that are longer term. So as the yield curve inverts, this damages their profitability, where they’re paying out more and receiving like the same or less. So an inversion, no matter how it’s structured, hurts the banks’ profitability.
But the particular type of inversion we’ve seen, where the long rates go up also, that damages not just profitability but the balance sheet. So the balance sheet of these banks has these longer-term obligations like these loans and mortgages or bonds, and they go down in value as these long rates go up. So this is exactly what happened with Silicon Valley Bank, where their government-bond portfolio had to be written down and that caused them to have negative equity and they went out of business. So this one is, like, not a good one.
The yield curve has been inverted for a while, and at first you thought that it might be a false signal. How has your thinking evolved on that?
Yeah, so the yield curve inverted in November 2022. On January 4, I wrote a LinkedIn post saying that this could be a false signal, and I had some credibility in saying that, given that it’s my model. But there was a very important caveat. I said, Yeah, I think we can dodge a recession, but it’s conditional on the Fed stopping their rate hikes. And that’s not what happened. I still believe that if the Fed had stopped their rate hikes in January, we could have avoided a recession. But given that the Fed has continued, all of this has really upended things.
And I think that pushing the rates up as high as they have has been counterproductive. They give the inflation reason, but it’s a false narrative. I believe that the true rate of inflation right now is more like 1.5 percent. So you can’t use the inflation excuse. The Fed has gone too far. They’re very late realizing that they need to pause or reverse course. It’s technically called overshooting.
There’s an interesting phenomenon with the yield-curve inversion now, which is that it’s starting to disinvert — which is to say, go back to a normal shape. In your model, what’s the significance of that?
So the average lead time from an inversion to a recession, over the past four, is 13 months. So we’re not even at the average yet — it’s way too early to say that it’s a false signal. But more importantly, you look at the past four recessions, and before the recession begins, every single time, you see the uninversion happening or whatever the word is. The steepening occurs before the beginning of the recession. So this is exactly what you’d expect.
Yes, “the bear steepening,” as people call it when the long end of the curve rises and the inversion ends. If that is happening now, when do you expect a recession?
I expect first quarter or second quarter of next year. It’s going to be very interesting because the third-quarter GDP is going to be very impressive, and people will think, Oh, well, this must be a false signal. But that growth has been driven exclusively by consumers drawing down their savings, the COVID-era savings, and those savings will run out in the fourth quarter. Plus the resumption of the student-loan payments affects 40 million Americans, and it’s all being taken away from disposable income. So many forces are suggesting that the consumer is not going to be able to bail out the economy in 2024.
So how worried are you? How bad do you think this time could be?
This is very important: I hope that my model is wrong. I certainly hope that it is a false signal, though I don’t believe it is, and kind of the good scenario for me is a mild recession, something like we had in 2001 or 1990. And ’91.
It’s important for people to remember that even though the 2020 recession was very short — because we had the pandemic and then we had all the aid — the yield-curve inversion predicted that too.
Be careful here. And some people say, Well, surely, the yield curve didn’t predict COVID. But in real time, when it inverted in 2019, there was a lot of other data that suggested that there was going to be a recession in 2020. So we’ll never know. It’s got kind of an asterisk on it because of COVID, but again, we likely would have had a recession anyway.
Does the severity of the inversion predict at all the severity of the recession?
No. But there is a very tight relationship between the duration of an inversion and duration of recession. And it depends when we uninvert, but it indicates that the recession could last about nine months to a year. So that’s obviously longer than the COVID one but shorter than the global financial crisis.
You haven’t had any false signals yet. But is there any reason to think this time is different?
Every time is different. And there’s many differences this time around. One difference is the excess demand for labor. So that’s kind of unique, and that I think is important because I think it will dull the blow. That excess demand has been decreasing, and it could go the other way very quickly. But nevertheless, I think that that is something that is different.
You mentioned that the yield-curve inversion hurts banks, and already we’ve seen some of the small-to-midsize banks fail. Do you think some of the big banks are at risk?
Given the further increases by the Fed, the banks are not in as good shape as they were, like, a year ago. This has really been punishing for the banks and frustrating because the Fed has taken the banks into a scenario that is far beyond the adverse scenario they get for their stress test. It’s a little bit unfair. So I think the financial system is at risk right now. Here’s something interesting that I’m frustrated the media has not picked up on: You know, you can get like 5.2 percent or something like that in a money-market fund? Most people don’t do that. They’ve got a savings account, and the average savings rate across all banks in the US is about 0.5%. So that is kind of dramatic — you can get in a money-market fund ten times what you’re getting from your bank. But that’s not the whole story. The too-big-to-fail banks are only paying like .01% to .05% on savings. That gap between what they’re paying on savings and what you could reasonably get from another vehicle, that to me is a red flag. And that suggests to me that the financial system is far more at risk than people think.
We’ve seen even with the recent jobs data that the economy keeps adding jobs. And it seems like some economists — at the Fed and elsewhere — sort of won’t be happy until we have more unemployment. What do you think?
Unemployment is a lagging indicator. So you need to be careful. It always is low before a recession, and we’re not there yet. We’re in strong growth.
You’re known for your work on the yield-curve inversion, but you’ve also written a book on DeFi and you follow crypto pretty closely. Have you been following what’s been happening in the regulatory landscape and the Sam Bankman-Fried trial?
Yeah, I was actually asked to be an expert for Sam Bankman-Fried’s defense. So I declined that. But I am following very closely. I think they wanted me to give an opinion on the whole exchange thing and be available to kind of teach the jury and the judge about some of the subtleties of the space. And again, that’s not decentralized finance. That’s centralized finance. He was running a centralized exchange and committing fraud by channeling customer money to a hedge fund. It had nothing to do with DeFi, other than trading some decentralized tokens.
What is your outlook for crypto in the US now?
I think the big growth area is with just regular companies using this technology. So I’m quite positive.
