If you haven’t been worrying about the possibility of an inverted yield curve, now might be a good time to start.
If you’d purchased a 3-month U.S. Treasury bill at last week’s auction, you’d be earning an annual yield on your investment of 4.03%. A 6-month bill would be yielding 4.3%, about 25 basis points more. Does that mean that the 6-month security was a better investment for your money?
Not necessarily. Suppose you bought the 3-month bill and then, as many Fed-watchers are expecting, the Fed raises the fed funds rate another 25 basis points at each of its meetings in December and January. If that prediction turns out to be right, then when your 3-month bill matures in February, you could probably reinvest it at a new 3-month rate that’s 50 basis points higher than today’s 3-month rate and 25 basis points above today’s 6-month rate. If that indeed comes to pass, the 6-month return from rolling over two 3-month bills would be virtually the same as if you’d purchased a single 6-month bill right now.
|
A very simple model known as the expectations hypothesis of the term structure of interest rates posits that investors don’t particularly care which maturity they invest in, and as a result would always bid prices for different maturities so that the expected yield from rolling over securities of different maturities is identical. The blue line in graph at the right displays the current yield curve, which plots the annual yield of Treasury securities of different maturities on the vertical axis and the length of time to maturity on the horizontal axis. According to the expectations hypothesis, the current steep slope of the yield curve at the shortest maturities signals that investors are indeed anticipating a very rapid rise of 50 basis points or so in the short-term interest rate over the next three months, but expect subsequent hikes in interest rates to be more gradual and drawn out. The red line shows that the same curve was substantially steeper this time last year.
Although the expectations hypothesis has some descriptive power, a large number of academic studies have demonstrated that it is not a completely accurate quantitative description of how yields on different maturities are related. For one thing, the yield curve almost always slopes up, even in times when investors believe interest rates are likely to fall. A longer term security involves more risk of a capital loss and is less liquid than a shorter term security, so most investors would prefer the short rate even if they expect to earn a lower return. For this reason, it in fact is a fairly unusual development for the yield curve to become inverted, or actually slope down. A modified version of the expectations hypothesis holds that the long yield is an average of the expected returns from rolling over short rates plus a constant term premium. This modification, too, is not consistent with all the evidence, which shows that the term premia can change over time in significant ways.
|
The lower graph at the left records one quick summary of the slope of the yield curve at any given date, as provided by the spread between the yields on a 10-year Treasury bond and a 3-month Treasury bill that you would have seen in any given month. The spread became negative (signaling an inverted or downward-sloping yield curve) on only six occasions during the last half century. Five of these were associated with economic recessions (indicated as shaded regions on the graph), and the sixth (1966) with a significant economic slowdown.
Arturo Estrella, an economist at the Federal Reserve Bank of New York who has done a great deal of research on the term structure, has a a nice summary of the academic research on the relation between an inverted yield curve and economic recessions (hat tip: Economist’s View). One reason that the yield curve inverts prior to an economic slowdown is that economic recessions tend to mean lower interest rates, both because of lower demand for borrowed funds as well as the fact that the Fed is likely to lower rates aggressively in response to an economic downturn. According to the expectations hypothesis, to the extent that investors see this coming, the yield curve would start to slope down before the drop in economic activity.
In research with my former student Professor Dong Heon Kim, I found that the expectations hypothesis alone does not explain the ability of the yield curve to predict an economic slowdown, but that there is movement over the business cycle in the premium that investors require for long-term securities that also contributes to the ability of the yield spread to predict economic downturns. Research by another former student, Dr. Hibiki Ichiue of the Bank of Japan, suggested that this may be related to news about inflation, which seems to raise the market price of output risk and thereby lower the term premium prior to an economic downturn.
If the ten-year rate had stayed at the values below 4% that we’d seen at the start of the summer, we would be at another point of inversion right now. However, longer term rates have drifted upward over the last five months to keep the yield curve from acquiring a downward slope, though not fast enough to keep it from becoming more flat. With the 10-year over 3-month spread currently at 50 basis points, unless we see further hikes in long rates, we could face an inverted yield curve by the end of January.
The statistical evidence suggests that any time the yield spread gets below 100 basis points, it raises the probability of an economic slowdown even if not a full recession. And if two more FOMC hikes to 4.50 by January aren’t enough to do the job, four hikes that put us to 5.0 by May surely would. For all these reasons, I take comfort that some Fed-watchers are concluding that the Fed is likely to pause before then.
Let’s hope they’re right.
Most commentaries,including the op here, imply that an inverted yield curve causes a recession. That’s not so. What causes the recession is the tightening monetary policy by the Fed. In the natural course of things, short-term rates are higher than long-term rates. It requires interference by an external actor-the Fed-to disrupt the natural yield curve. So when the yield curve inverts it’s a sign that the Fed is at work.
The inverted curve is only a symptom of the Fed’s action, which is the thing causing a recession.
Fed Watch: Changing the Course?
Here’s Tim Duy’s latest Fed Watch: Nothing like the Fed minutes to liven up the debate! The minutes generated a buzz among bloggers as well as the financial community, with stocks and bonds gaining on news that the Fed may
another thing to consider is the absolute amount of interest rate increases – the fed went from 1 to 4%. 4% is 4x 1%. Granted, the Fed did it in 0.25% increments and not 0.5% increments.
The inverted curve is only a symptom of the Fed’s action, which is the thing causing a recession.
But isn’t the fed itself only reacting to other drivers (inflation for instance, or excessive money growth believed to drive inflation)… the fed isn’t really causing the recession then either but rather some other stimulus farther up stream… yes, no?
And one other quickie… a lot of things are as always before, no doubt… but never have we had as much intervention in the markets from central banks around the world… are there ‘realistic’ (and not just overly hopeful or optimistic) theories or explanations that maybe this time it will be different… that an inversion shouldn’t be predictive of a recession or slow down this time because… (fill in the blank).
The yield curve is some way connected to the overall investment activity. Strong investment activity (household and corporate, housing and manufacturing) increases the demand for longer term money and makes the “normal” curve. Inversed curve might tell about decreasing demand for longer term money and hence about weaker investments. This is a sure sign of recession.
Dryfly, yes the Fed is reacting to events, but it also has some choice in the degree to which it reacts to those events, and it could certainly choose (and I believe should choose) to wait a bit before moving the funds rate to 4.5 or above.
The most common story I’ve heard for why the yield curve inversion won’t mean a recession this time is that it’s caused by large foreign purchases of long-term Treasuries. I discussed this hypothesis in my post on the puzzle of long-term bond yields.
Another reason people say not to worry this time is that previous yield curve inversions were associated with high real interest rates. If you calculate the real interest rate as the current Tbill rate minus the CPI inflation over the last 12 months, they would have a case. If you calculate it from inflation-indexed Treasuries or the fed funds minus median CPI, there’s reason to be worried.
how often have we heard,”It’s different this time”.
Thanks JDH… and I’m kinda in Big Al’s camp on this one, in that I’ll believe it is different when it turns out to be different. But at least your description gives me a play card to sort of track the events. Thanks again.
The sky is falling!
Well, no, but the yield curve may be inverting, which is almost as bad. What does an inverted yield curve mean? It means that the yield on short term debt is higher than the yield on longer term debt. Historically, this has been a harbinger of recessio…
cum hoc ergo propter hoc
JS, you really should see someone about that cough.
It is interesting that Greenspan’s oblique reference to globalization (and wasn’t he just recently travelling to and from that same reference?) is not mentioned here. When AG claims that we need not worry this time about yield curve inversions, that globalization together with our flexibilityandresilience (the razamatazz of Fedspeak) and superlative management will maintain our fine trajectory…(ok maybe ‘path’ or ‘course’ for those of you who need more of a sense of control and less abandonment to the forces of gravity)…nobody listens (not here anyway).
It’s those tiny little 25bp steps, that conundrum talk (see? –I have a memory) and maybe that Bernanke is already in there as Chairman and we are waiting for his assessment. Maybe Greenspan’s commentary is not all that enlightening nor intended to be anything more than pacification for the worryworts who took his conundrum talk too seriously.
He was just kidding.
Is he still just kidding, is what we all want to know. And we will soon find out, possibly earlier than the end of January if recent mortgage rate movements constitute a trend.
At the end of 2001 treasuries with maturities of 10 years & longer accounted for 40% of all marketable treasuries ($1,168 T of $2.915 T). At the end of 2004 they ammounted for only 31%. At the same time BUYER NEED for longer Treasuries INCREASED substantially (hedge & pension funds, fx washing $, etc.) This pending inversion, like the housing run, has been influenced by Treasury & our (independent) Fed.
JDH,
During normal times, a banker (or anyone for that matter) has to include in the price of the loan the growing risk of getting hit by inflation (and others) while exposing his loan for a longer period.
Could it be that bankers, worrying about a potential downturn, conveniently and cautiously restrict longer term lending in favor of short term loans, crowding the short term demand, and therefore spiking the price of these loans?
Of course, the long term loan restrictions are totally independent of interest rates, making borrowers avoid being turned down for the long term loans -precipitating their price-, and flocking to short term loans where they’re applications are approved, but, at higher prices b/c of higher demand.
If this is so, the PBoC pushing for long term low rates and the Fed pushing for short term higher rates is a totally different state of affairs.
I guess, the PBoC’s attitude stems from its charter to promote employment to the other 800 million Chinese: but also, b/c they may have no other -safer- or better place to put their huge surplus.
The idea that the yield curve can safely invert because of the foreign purchases of the 10 year treasury doesn’t ring true for me. I would expect the market to adjust for the extra demand at the 10 year by some investors shifting to shorter maturities – offsetting the curve affecting impact of the extra buying from abroad. Investors accept a lower rate for longer duration when they expect the higher short duration rates to be temporary.
But the question of recession is more a tale of reduced investment and spending, which manifests in reduced demand for longer duration borrowing. The removal of such borrowers (demand) from the market reduces longer rates. I believe the inverted yield curve largely reflects the reduced investment/demand which is an early phase of a slowing economy. Shortly thereafter we have an observed recession.
One of the best rule of thumb in economics is that we get a negative yield slope every time time and only when the inflation rate is higher then the unemployment rate.
Jim Miller, dryfly and JDH are using different sorts of arguments: normative versus empirical. As a bond investor, I care more about the accuracy of the conclusion than which intellectual process (both can be valid!) was used. You may recall seeing this article
http://gsbwww.uchicago.edu/fac/monika.piazzesi/research/economist.pdf in The Economist, and I’m eagerly awaiting publication of the authors’ work.
They found that lagged short rates have been better predictors of recessions over time than the yield curve, though the curve was better in the 1990s. As The Economist concludes, however, both short rates and the yield gap are pointing in the same direction.
More complication and obfuscation…
spencer, are you following the Aussie experience?
http://www.ing.com.au/public/pdfs/ma_houseview.pdf
claims that unemployment is 5.2% with inflation about 3-3.5%. Not too different from the US picture but that invertion has been with them for a year or so.
Historically how long has an inverted yield curve lasted (time) before a recession has occurred?