What is an inverted yield curve?

Factual Questions
21

Interest rate patterns can’t be taken at face value right now.

For about 10 years, we’ve seen excessive suppression, then a bit of enthusiasm in raising at first hint that we should do so.

Sort of in the middle of a reset and creating a new baseline, because, quite frankly, the whole baseline was screwed up for too long and when they got a chance to reset it, they did it wrong. SMH.

Recessions matter to people who lose jobs. To those who maintain jobs/income, it’s a way ahead… EXCEPT… if we label what we have – irresponsibly – a recession, when it really isn’t, and people change behavior, including corporations slowing down their growth and adjusting their risk, then the typical counter punches can’t be used, because we still sit on interest rates that are lowish.

If we have a cyclical recession, it had better be a straight up legit one, and not a media-fueled one. I fear it will be the latter though.

Sad.

22

Talking about this:

This is completely backward. When short-term yields are higher, that means the “collective wisdom” isn’t demanding short-term bonds so much anymore. Assuming that “collective wisdom” is seeking the best returns, this means “collective wisdom” isn’t confident about what’s happening in the short term. In the long term we know trends average out, we know the stock market returns 10% over time if you’re willing to bet long enough. That’s very predictable. It’s hard to beat that in the short term except in the right conditions with very good information.

Predictability depends on whether your information is pointing toward any clear conclusion. If your information about the short-term seems more conflicted than known trends about the long term, then the short term is more predictable than the long term.

Which is what we’re seeing today. Investors are turning away from short-term instruments, and buying longer-term instruments. Assuming they are seeking best returns, it means the short-term bet doesn’t look so good.

23

I’m still not sure I understand what you mean by “backward”, and you seem to persist in assuming that the only factor in yield is risk, which is incorrect.

Again, the yields discussed reflect assessments about the relative returns of the market as a whole and its various investment options. Higher yields reflect an expectation of higher returns. Lower yields reflect an expectation of lower returns. When the economy tanks, then yields go down because (leaving aside central bank manipulation) there are not that many opportunities to make higher returns. When the economy is doing well, the opposite is true. So yield rates for various terms are a proxy for market expectations about market returns in those periods.

To call this “backward” you would be saying that people accept lower yields when they think the market is going to do well, and demand higher (default-risk free) yields when they think the economy is going to tank. This is counterintuitive, and demonstrably incorrect.

I assume you mean “less” predictable, which is the point you’ve been making and appear to be making here.

This is logically impossible, as described above (again, unless there is some negative correlation between the different periods).

24

Yes.

The yield is higher because the price is lower. The price is lower because demand is lower.

But demand can be lower for more than one reason. There’s more than one possible substitution here. You can say demand can be down because people are substituting between different maturities on the yield curve (selling 1-month, buying 2-year) as you seem to be arguing here, but demand can also be down because they’re substituting between private market assets and government bonds. I’m not sure you’re fully considering this aspect of it, because it works directly counter to what you’re saying here. And it is arguably the more important substitution.

Treasuries tend to be the realm of refuge from the private markets in times of insecurity.

If yields are high (the price is low) for the 1-month and 2-month, then that means people aren’t demanding them, exactly as you say. Your logic is solid on that part of it. But the entire reason people aren’t trying to flood into 1-month Treasuries, at the current price, is quite possibly that they aren’t worried about the economy on such a short horizon. Confidence is fine along such a short interval, so they’d prefer short-term private market instruments over buying up government bonds.

Their worry is for private market investments along the 1-year or 2-year range, and so traders who are looking at broad market trends along that longer horizon, and who are afraid, will decrease their demand for private assets with payoffs along that horizon, and start bidding up the price of 1-year and 2-year Treasuries instead. Which will push up their price, and push down the yield.

This is again assuming a substitution along government bond maturities, selling the short-bonds and buying the long-bonds, and so pushing down longer yields compared to the short.

But there’s an issue here. The problem with assuming that the relevant substitution is among different maturities is that that long-bonds are inherently the expected (risk-weighted!) value of repeatedly buying the short. The price of a two-period bond will be equal to the expected (risk-weighted!) value of buying a one-period bond, and then buying a second one-period bond after the first matures. (After all, a two-period bond purchased next period should have a price exactly identical with a one-period bond issued next period: they are effectively the same instrument.)

All of this is to say that when the 1-year is higher than the 1-month, it is a firm market prediction that the yield of the 1-month bond is going to drop soon. Which is another way of saying that it’s an explicit prediction that the price of the 1-month will soon increase. An inverted yield curve of this sort is an explicit prediction that there is a class of assets out there, right now at this very instant, which will soon see an increase in price. (I.e. short interest rates will drop soon.) Yet people aren’t yet buying those assets which should soon increase in price. Why? Quite often a firm prediction that an asset price will increase tomorrow is enough to ensure that the price increases today. But that is not enough inducement in this case. Why?

Part of that is the timing of Fed policy changes, which complicates every issue in this thread. But another part of it is that people are very likely sufficiently confident in the private markets for the next month or two that they’re not willing, yet, to substitute away from the private markets and into the relative security of short-term bonds.

I don’t mean to imply in any of this that an individual trader will never substitute along different maturities of the yield curve. Traders do this all the time.

But the aggregate behavior is going to be different. The market as a whole is unlikely to do such substitutions along the yield curve, because a 10-year bond can easily be synthesized from ten consecutive 1-year bonds. The (risk-weighted) expected value of each purchase should be identical. They are effectively the same thing as each other. This is exactly why a sharp drop in the 1-year or 2-year yield is highly predictive of a coming drop in shorter-term interest rates. The price of the 1-month will increase whenever those interest rates finally drop. Everyone is expecting this.

In the aggregate, for the market as a whole, the relevant substitution isn’t along the yield curve, but the substitution between private market assets and government bonds.

25

The wording there though is a little vague and I believe the other poster took it to mean something different than what I think you mean, and what’s correct.

A lower yield on the 10 yr than the 2 yr doesn’t mean an ‘expectation of lower return’. It means you will get a lower return on US govt bonds for 10 yrs if you buy a 10 yr than you’ll get for 2 yrs by buying a 2 yr (give or take reinvestment of coupons, or assume they are a zero coupon 10 and a zero coupon 2 yr).

It also says nothing directly about expected return in the stock market, which the other poster seemed to assume you meant.

But let’s speak specifically of expected future yields not ‘returns’. If the 10 yr yield is lower than the 2 yr yield, it means on what’s call a ‘risk neutral’ basis that the market expects future 2 yr yields, like the one starting in 2 yrs and going to 4 yrs, the one starting in 4 yrs and going to 6 yrs etc, to be lower than today’s 2 yr rate.

What ‘risk neutral’ means is, excluding any risk premium above or below what market expects the future 2 yr rate to be. In general it’s assumed that, let’s take 4 yr yield and 2 yr yield to simplify, that if the market expects the 2 yr yield 2 yrs from now to be the same as today’s 2 yr yield, the 4 yr yield today will still be slightly higher than today’s 2 yr yield. That’s called a ‘positive term premium’, meaning that lenders (people buying bonds) get a little more yield by locking in the term for a longer period than by lending for a shorter period then rolling it over, IOW borrowers (the govt in this case) have to pay a little to lock in a rate for a longer term rather than borrow for a shorter term then roll it over, even if the market expects the future short rate to be the same as now’s. Lenders are relatively more averse to locking in long term rates and borrowers relatively more willing to pay to do that, historically.

So, a downward sloping yield curve between 2 and 10 yrs will generally imply the market expects the 2-4, 4-6, 6-8 and 8-10 yr rates, when they set in the future, to be even lower than the current 2 yr than the yield curve implies. But, and this is where ‘this time could be different’, many analytical models of the yield curve now say the term premium, which can’t be seen directly, is negative. The reason for that might be what another poster mentioned, ‘suppression’ of rates by central banks buying longer term bonds (doesn’t necessarily matter if they buy other countries’ bonds, that buying pressure still transmits into the US bond market).

If the term premium is negative, a slightly down sloping curve might still mean that the market expects future 2 yr rates to be as high or slightly higher than now’s, and it wouldn’t be as good a recession predictor.

The reason it ‘predicts’ a recession being, if it’s the market’s expectation that the Fed will keep short term rates low in the future (the Fed only sets very short term rates, but near term Fed policy directly affects the 2 yr much more than it does the 10 yr). And the Fed tends to keep rates low when the economy is weak. Hence lower future short term rates=weaker future economy, or that’s the tendency.

This signal could be ‘jammed’ somewhat though if the term premium has shifted from historically being general positive to being negative. Then a down sloping YC isn’t as clear a predictor of lower future short term rates.

26

If that’s what the other poster assumed then they understood correctly, because that’s pretty much what I meant. The only quibble is that it’s not the stock market specifically, but the spectrum of alternative investments, including real estate and whatever else.

Interest rates on treasuries don’t exist in a vacuum, but compete against other forms of investment. Generally, treasury rates will have a lower rate of return than other forms of investment because they are perceived as being zero (default) risk, but they’re nonetheless influenced by the marketplace of available investments. If treasuries are at (say) 2% against perceived alternative investment rates of 5%, and those other rates suddenly increase to 6%, then all else being equal treasury rates will increase as well. Because some percentage of the marketplace which favored the security and other features of treasuries at a 3% spread will find the alternatives more compelling at a 4% spread. That new 4% spread would decrease demand for treasuries, which would then decrease in price, which would raise the yield until they move to the point where the treasury vs alternative finds a new equilibrium.

Central bank manipulation is also a factor, as I’ve mentioned and as you also state, but that’s not the only factor. Even absent any central bank intervention (or expectation thereof), the economic cycle alone would produce changes in interest rates, as the perceived returns on the spectrum of investments changes. So if the perception of likely returns on other forms of investment changes, that itself also changes the yield on treasuries.

Therefore, if there is an expectation that the economy is about to tank, this translates to an expectation that in the future investment returns will deteriorate on all sorts of investments, including treasuries. Therefore, an expectation of a deteriorating economy will have the impact of flattening or even inverting the yield curve.

[Of course, it should go without saying that when referring to collective market expectations it doesn’t mean “a whole bunch of guys deciding they think there’s going to be a recession”. It means the collective market activity of people contemplating individual longer term investments finding the various individual opportunities not compelling, to the point that they would take a lower return from treasuries. The yield curve is a rough way of boiling down the collective expectations regarding millions of individual investment decisions and assessments.]

27

Yes everything is tied together, but if there’s any dominant direction of causality about the govt bond yield curve it’s what it shows about expectation of future monetary policy then what that says about the real economy, then real economy influence on stock and other risk assets. It’s not that govt bond rates exist in a vacuum, it’s just a general idea of what causes what means we want to understand what the market may see in the bond market before going downstream to talking about the stock market or other assets.

And on considering my previous too wordy post, there’s not substitute for introducing math in terms of the concept of forward rate to understand what an inverted yield curve means. If the 2 yr rate is 2% and 4 yr rate is 3% (assume them annual rates), the forward rate between 2 yrs and 4 yrs (the 2x4 rate) is approximately 4%.((1.03)^4/(1.02)^2)^(1/2)-1 IOW the yield for years 2>4 that combined with the known yield for years 0>2 would gives the known yield from 0>4. If the 2 and 4 yr rates are both 2%, the 2x4 rate is also 2%. If the 2 yr rate is 2% and 4 yr rate 1.5%, the 2x4 rate is 1%.

So, assuming borrowers and lenders are both indifferent to term risk, we could say the market ‘expects’ the 2 yr yield in 2 yrs time to be 4% in the first case, 2% in the second case, 1% in the third case. Since 1% would tend to imply looser monetary policy (again the Fed doesn’t set the 2 yr rate but it’s relatively directly sensitive to the Fed Funds rate the Fed does set), and loose monetary policy implies a weak economy, and a weak economy implies lower profits, and lower profits lower stock valuations…so yes it’s all connected. But the effect further upstream from the stock market is the apparent expectation of lower short term govt bond rates in the future than now, v higher or the same as now.

The twist though is that those 2x4 rates, 3%, 2% or 1% are not actually the market’s true expectation of the 2 yr rate 2 yrs from now. They are the expectation plus or minus a risk premium which reflects the relative willingness of lenders v borrowers to lock in longer term borrowing/lending, which is called the term premium. Usually the term premium is assumed to be positive, meaning that if the 2x4 rate is 2% the market really expects the 2 yr rate 2 yrs from now to be even less than 2%, ie even a flat yield curve implies lower expected rates in the future than now. But it could be now that the term premium is negative because of extensive (worldwide) central bank attempts to stimulate growth by buying longer term govt bonds. In that case a 2x4 forward rate of even 1% might not imply much actual reduction in expected rates. IOW an ‘inverted’ curve now may not imply as a big an expectation of actual future rate decreases as it normally would.

28

I don’t understand what point you’re making. The only thing I can figure is that you are focused on future rates as being the result of monetary policy - which reflects economic conditions - as opposed to being the result of general economic conditions themselves. If so, I don’t know why. Is it really your contention that absent any central bank intervention that treasury rates would be the same in a recession as in an expansion?

I don’t disagree with any of your math (I had typed up something along those lines myself but deleted it as being potentially confusing). But I’m unsure what point you’re making. ISTM the only issue is why the expectations are that the 2x4 rate would be lower than current 2 year rate, as above (with me emphasizing general economic conditions and you Fed policy in reaction to economic conditions). If you’re making some other point, I’ve missed it.

29

Assuming I’m interpreting this correctly, I cannot personally agree with the notion of a “dominant direction of causality”. Not at all.

Expectation of future monetary policy is not some independent thing either. It is itself based on the expectation of the real economy, and there is feedback between those two things. A lower rate can mean a weaker economy, but a lower rate can also potentially prevent a weaker economy. When we are forming these expectations about the real economy, trying to figure out what direction things are going based on whether monetary policy is influencing or being influenced (or both…), our expectation is built on many things, especially including forward-looking market signals such as the stock market and other risk assets. So we might look at the stock market, for example, to try to feel out whether the Fed is reacting to an inevitable slump, or getting out ahead of a possible slump.

Human language descriptions have to be linear, one sentence after the other, and that muddles our explanations of things, but our deeper understanding of market structures does not have to be so rigid and one-dimensional. Looking at this thing first, and then this thing second, and this thing third, implies a dominant direction of causality that simply does not exist. The winds are always shifting. What was once the dominant cause becomes the effect.

Today everything happens. Then because of that, tomorrow everything happens.

The yield curve doesn’t happen in isolation. The stock market moves, unemployment numbers are released, forex markets shift, the Fed releases minutes of its meeting, etc. Starting an interpretation in one particular place, and then looking to the next place, and then the next, can lock in a potentially mistaken interpretation of what’s happening that can be hard to shake. Getting the direction of causality wrong was, I’d strongly argue, a big reason for the aftermath of 2008. Better to look at multiple things simultaneously. Each individual piece we can look at is contextualized by the bundle of other pieces that are out there.

30

So I’m trying to think through why the term premium might flip from the normal state of affairs since we’re been studying economics. As it stands now, we assume a lender who ties up his money for a longer period of time is taking a greater risk than investing in short-term debt and rolling it over every time period. What risk is that exactly? Part of it is the risk that interest rates will rise. But if interest rates might also fall just as easily, then it might be considered a blessing to tie one’s money down for longer at a lower rate now. So that part of the difference is mainly that interest rates are going to in general be rising as the economy keeps trucking along normally. Recessions normally happen much more quickly than recoveries and growth periods, and recessions are when we expect the interest rates to go down, so when we consider the term premium to be positive, are we not just saying that averaged over time, interest rates tend to be more likely to rise than fall?

There’s also default risk to consider. But if the lender is planning on always lending to the borrower and will always extend the term regardless of economic conditions, just possibly charging a different rate, then how is default risk an issue for a longer time period?

I don’t know the answer to this, I’m just mulling it over.

31

A positive term premium generally indicates some sort of risk, especially one that compounds over time. The most obvious risk for government bonds is inflation risk, and that’s pretty one-sided given a fiat currency. There’s a lot more room going up than down. The US won’t “default” in a traditional sense, but there’s always that outside shot that any given government could inflate the burden away. In a different market, liquidity risk would also be an issue, but that’s not the case for standard Treasuries. They’re dead easy to unload quickly.

The most likely candidate I’ve seen for a negative term premium is simple insurance for the future stream of income. Given a strong desire for consumption smoothing (strong fear of volatility), a class of investors might actually pay more to receive a nice stable predictable coupon for a very long time – especially if they have lost all fear of massive unexpected inflation. When down is more scary than up is tempting, the short-term volatility could be unappealing.

Srsly, “insurance” is pretty much always a solid guess for why an asset price gets pushed up past where you might otherwise expect. But I’d also be interested in other plausible explanations.

32

This sounds very similar to a negative interest rate. If you make sure I have my money at some later time far in the future, you can keep some of it. Now instead of having absolute negative interest rates, its merely the term premiums that are negative - if you make sure I get my money 30 years from now, I won’t require you to pay me back as much, because I’m worried about there being any other safe place to put my money for that long. It’s not the lack of default risk per se that drives down longer term yields, it’s just that the default risk with any other investment is high enough. But then that’s exactly what you’d expect when you’re talking about the economy going into recession. The lower long-term yields are signs that people are very unsure about making any other kind of shorter-term investment.

33

Paul Krugman says something a little different. “An old economists’ joke says that the stock market predicted nine of the last five recessions. Well, an ‘inverted yield curve’ — when interest rates on short-term bonds are higher than on long-term bonds — predicted six of the last six recessions.” If I understand the graph he linked to, it does show that an inverted yield curve occurred before previous recessions.