In May of this year, the yield on three-month U.S. Treasury securities fell below that of the Treasury’s ten-year note; This yield has remained lower than the ten-year note, with the exception of a few days. In early September, it seemed as if the yield on two-year Treasury securities was also dipping below the long-term rate. Many analysts have taken alarm that the “yield curve” (a plot of yields versus the length of securities) is “inverting.” This is a fancy way of saying that long-term interest rates are lower than short-term rates.
Why would this be a concern?
It’s a simple, observable phenomenon. Whenever there has been a considerable period of several months during which the yield curve has inverted, almost always, a recession follows. The yield curve inverted in 2007 and then the recession started at the beginning of 2008. The yield curve inverted in 2000 and the recession started at the beginning of 2001. The only real exception in the last few generations was 1966: the economy did appreciably slow in 1967, but that “growth recession” (which hit housing as hard as the real thing) never got into the books as an official slump.
While there are many other measures that have been proposed and followed as “leading indicators” of recessions, the yield curve seems to be the most reliable. Much of the research on this subject has been done by my old New York Federal Reserve (Fed) colleague, Arturo Estrella. The New York Fed website contains a chart of the yield difference between the ten-year and three-month securities. Also expanded upon is an estimated probability, based on Estrella’s work, of the U.S. being in a recession in 12 months (it’s not a probability that a recession will begin in 12 months, it’s the probability that this month a year from now will be part of a recession. In other words, the recession could begin sooner, or we could be in one right now and it will last that until then). The charts are updated monthly. The recession prediction, based on data through September, was a 35% chance in September 2020 we will be in a recession. Perhaps that doesn’t seem so high, but that’s about as large of a chance as we’ve seen in the last 35 years. The recession signal wasn’t appreciably higher prior to the 1990, 2001, or 2008 recessions. There is no question; if one buys into these results, this suggests a substantial recession risk.
An “inverted yield curve” is a somewhat abstract concept. Why should the fact that the bond market setting a lower yield on long-term securities rather than short-term suggest that economic troubles are ahead? It’s pretty intuitive that a decline in orders received by manufacturers, or a dip in housing starts or permits, means some future trouble.
What’s so special about how two interest rates compare to each other?
To understand this, we have to think about why an investor will buy a long-term bond rather than a short-term bill (we are talking about U.S. Treasury securities, so we don’t have to worry about the possibility that the interest and principle won’t be paid). Of course, any long-term investor wants to be compensated for locking up their money, so there is a natural tendency for long-term rates to be higher than short-term rates (borrowers would normally be willing to pay more to get long-term funding, since it means they don’t have to scramble for funding so often).
Suppose investors think that short-term interest rates are likely to increase. To buy long-term securities, they will demand higher yields in compensation—after all, they could just buy and roll over short-term securities and profit from the expected future increases in short-term rates. So, the yield curve will “steepen” when investors expect short-term rates to rise from current levels.
The inversion occurs when investors think that short-term interest rates will be lower in the future. When that happens, they scramble to buy long-term securities to lock in their return, and long-term rates drop. An inversion is rare, simply because the expected drop in short-term rates has to be large enough to overcome the normal tendency for long-term rates to be higher, even when rates aren’t expected to change.
What would cause investors to expect short-term interest rates to fall in the future?
The usual answer is when they expect the Federal Reserve to act to reduce short-term rates. In short-term markets, the Fed is the proverbial 500-pound gorilla in the room. Why would the Fed act to reduce rates? The answer: a recession, when the economy is in trouble. The upshot is that an inversion of the yield curve is, reasonably, a reflection of bond market investors’ forecasts that the economy is heading toward the shoals.
There are some who question this conclusion, as well as those who will explain why it will be wrong. In the past, especially in the 1970s and 80s, recessions seemed to be the direct result of the Fed overshooting. By raising short-term interest rates high in order to bring down inflation, many financial markets became stressed, and a recession developed. Although the Fed did raise short-term rates from 2015 to 2018, the increase was a mere shadow of many earlier episodes, and not by itself, enough to bring on a recession. However, similar comments could have been made in respect to the 2000 and 2007 inversions. Perhaps, more importantly, the Fed itself looks at such recession signals more intensively than anybody else. This year, it has been rolling back some of the recent rate increases to try to prevent recession momentum from developing. In the past, the Fed, typically concerned about high inflation, was slow to reverse rate increases. Our evidence about the connection between rate inversions and future recessions is largely derived from those days. It may be that different Fed behavior these days has ruined the old relationship—the true probability of a recession coming from the inverted yield curve might be very much less than 35%. Unfortunately, it is quite hard to judge the merits of this argument. Until then, it’s reasonable to see the yield curve inversion as a serious, though perhaps not definitive, sign that a recession may be upon the economy in the next year.