As I write this, the yield curve is about to be inverted.
I was taken aback today when I saw that 3-month LIBOR is now sitting at 2.6%. As a reminder, LIBOR is the rate that other banks will lend funds to each other on a short-term basis. It is often used a benchmark interest rate for other loans, for example corporate loans will typically be set at LIBOR + some spread.
For a long time coming out of the crisis, LIBOR didn’t matter that much for these loans since there was a floor set at 1%. As such, and L+200 loan would be 1% (the floor) plus 2.0%, to equal 3% yield. Now however, that loan would yield 4.6% due to move up in LIBOR. It has moved up as the Fed has raised short-term rates which in turn impacts other rates of similar maturities.
If we look at the 10 yr rate today however, obviously a much longer-term bond/rate than 3-month LIBOR, we can see the yield is still just ~3.2%. What this tells us right off the back is that the spread between long-term bonds and short-term bonds is very narrow. Said another way, the term structure of interest rates, or the “yield curve” must be very flat.
In a healthy market, the yield curve will typically slope upwards and to the right as investors demand higher rates of interest for longer-term investments. You want to be paid for the unknown for a longer period of time.
However, when the yield curve flattens people often wonder whether the yield curve will become inverted — when short-term rates actually will exceed long-term rates.
This is often viewed as a harbinger of a recession. The reason it’s a recession predictor is that every recession since 1970 has been preceded by an inverted curve.
In addition, a rise in short-term rates pulls back on the capital available on the “fringes” of the economy (e.g. entrepreneurs). At the same time, expected deflationary pressures in the long-term may push long-term rates down. The combination of these factors results in an inverted yield curve and is easy to see why this would indicate a looming recession. (Note, though, this is a small sample size of predictors and there is no perfect predictor out there!)
Ray Dalio wrote in “Principles for Navigating Big Debt Crises” that,
“Typically, in the early stages of the top, the rise in short rates narrows or eliminates the spread with long rates, lessening the incentive to lend relative to the incentive to hold cash. As a result of the yield curve being flat or inverted… people are incentivized to move to cash just before the bubble pops, slowing credit growth…”
A look at prior 10-2 year spreads:
I decided to look at other instances when the spread between the 10 and 2 year were this tight and what the right investment decision at the time would be.
Below are four different instances when the spread between the 10 year and 2 year yield were as tight as it is now and what unfolded few years or months. Note, I start each the 10 and the 2 year yield at 100 and mark the change in yield over time (since price data isn’t available) which is charted on the left-hand side. The right-hand side and in gray is the spread between these two rates (the “10-2 spread”), essentially showing the data another way.
What may or may not be surprising is that in most cases, the yield on the 2 year tightens more than the 10 year by a dramatic amount in percentage terms. Why that may not be surprising is that the Fed uses short-term interest rates to dictate monetary policy and historically, that has had a less pronounced impact on long-term rates.
Since yields move in an inverse relationship to prices, this points to price appreciation for bonds in each of the cases.
This may not be a great predictor since you can imagine the spread widening as long-term rates gap wide and short-term rates move less, such as a long-term change in the inflation assumptions. However, given demographics and a technology changes resulting in long-term deflationary pressures, this currently seems unlikely to be the case to me.
Should we be adding long-term bonds or short-term bonds to our exposures:
Does it make sense to buy short term bonds today, which are yielding ~2.98% or the 10 year, which is yield 3.24%?
At first, I would clearly say buying a 2 year bond makes a lot of sense. You’re getting paid the same as the 10 year essentially and we know rates will go lower, driving bond prices up. However, we must remember that interest rate risk cuts both ways.
Let’s run through some scenarios by comparing the purchase of a 2 year bond and re-investing the proceeds after 2 years into another 2 year bond. We will compare this to holding a 10 year bond for a total of 4 years.
First, I’ll go through what I personally think is most likely to happen.
We undoubtedly are later in the business cycle and when we inevitably enter a recession, I think the fed will once again lower short-term rates to help the economy with a softer landing and to spur investment. I have written a bunch of posts on why I think lower for longer is here to stay, even though that is not a popular opinion (mentioned here and here and here).
As such, I show the 2 year rate moving to ~1.80% and the 10 year moving to 2.92%. I base these numbers off of the typical tightening we saw in prior cycle in the charts above. E.g. the 2 year yield comes in by about 40% and the 10 year much less (I assume about 10%, which is not as much as it has, but am being conservative). The result may surprise you.
The 2 year actually under-performs the initial coupon rate on an IRR basis given you must re-invest in a lower yielding security at maturity. On the flip side, the 10 year appreciates in value as rates move lower and you can sell it for a nice profit, for an IRR above the initial coupon rate.
I know what some of you are already thinking… OK that is great, but I think rates are going up, not down! Well for that scenario, we could look back to 1994 when the Fed abruptly raised rates.
I didn’t show this one before given the spread between the 10-2 was high and tightened over time (opposite than today). Recall in 1994, there was a great “Bond Massacre”, as Alan Greenspan who was the Fed Chair at the time, decided to let air out of the system to prevent an overheating of the economy (that really only had been out of a recession for a couple years).
Let’s look at the numbers if that were to occur. I assume the 2-yr rate moves higher to 4.47%, nearly 150bps higher than today and also 1.5x the current level and assume that the 10 year has a 80bps spread to the 2 year, which implies a 5.27% yield.
In this case, it is obviously advantageous to invest in the two year bond and roll over into a higher yielding bond for the next two years. However, it is also interesting to see that you don’t witness a negative IRR on the 10 year bond (and obviously if held to maturity, you still capture the 3.2% that you locked in a purchase).
To back this up, I used portfolio visualizer to show the results of either buying $10k of 10 year treasuries (Portfolio 1) or $10k of short-term securities in 1993 (Portfolio 2) and holding through 1995. The 10 year dips below on performance for a bit, and then returns to outperforming.
Rates moving like this is an unlikely scenario for me to bank on for a few reasons.
One, it is hard for me to see an abrupt change in the long-term inflation assumption, which will cap long term rates. For example, if the “real yield” i.e. the yield in excess of long-term inflation is 1.7% today (3.2% less 1.5% inflation assumption) then the inflation figure would have to move up to ~3.6% for the 10 year to be priced at 5.3%, keeping the real yield figure flat.
Second, it would then be tough to see how a sharp rise in short-term rates doesn’t cause a significant tightening on the economy.
Bottom line: I hope this is helpful analysis for those making long-term asset allocation decision for an inverted yield curve.
For me, interest rates could likely move up further, but it doesn’t mean that we should be significantly underweight long-term, high quality securities at this point in the cycle. Especially when it appears to be contrarian to do so…