Yield-curve confusion abounds.
The abrupt inversion of the Treasury yield curve over the last month -- in which the 30-year bond's yield has dropped far below the yields of shorter-maturity instruments -- has sparked the endless repetition on financial television of the notion that the economy is destined to slow. Because as
knows, an inverted yield curve is a time-tested predictor of recessions.
Hold it right there.
The first thing investors need to know is that there is a whole bunch of different ways of measuring the yield curve.
The yield curve is nothing more than a graph of Treasury yields, from the shortest maturities to the longest. This, for example, is yesterday's Treasury yield curve, from the 3-month bill to the 30-year bond.
Your Basic Yield Curve
As you can see, it is inverted in some places, but not in others. The portion between the two-year note and the 30-year bond is dramatically inverted. So is the portion between the five-year note and the 30-year bond and the 10-year note and the 30-year bond. The 30-year yield is a little lower, even, than that of the one-year Treasury bill.
Here's another way of looking at it: The difference between the 30-year yield and the two-year yield over time.
The Inverted Yield Curve
As you can see, the difference began collapsing aggressively in late January. The move was triggered mainly by the
disclosure that it will buy back from investors a large number of long-maturity issues this year, a debt reduction measure made necessary by the growing federal budget surplus. Still, people have wondered whether that's
all there is to the inversion.
The curve is also inverted between the two- and 10-year notes and between five- and 10-year notes.
Every other segment of the yield curve, however, remains positively sloped.
The Right Curve
The second thing everybody should know is this: Authorities on the relationship between the shape of the yield curve and future economic activity look at the segment of the curve between three months and 10 years. Alternatively, they look at the difference between the 10-year note yield and the
fed funds rate
pointed out: The
New York Fed
run the numbers on the difference between the 10-year and three-month yields. And the yield curve measure that goes into the
index of leading economic indicators
, which attempts to predict economic activity six to nine months in the future, is the difference between the 10-year yield and the fed funds rate.
The basis for the idea that the yield curve (whichever way it's measured) tells you something about future economic growth is this: Short-term interest rates are effectively set by the
. Long-term interest rates are set by the market, based on inflation expectations over the life of the securities. Long-term rates rise relative to short-term rates when inflation expectations rise, which they generally do when the Fed lowers short-term rates for the purpose of stimulating economic activity. And vice versa.
In either case, as this next graph shows, while those segments of the yield curve have flattened substantially over the last month, they remain positively sloped by a healthy margin.
The Not-So-Inverted Yield Curve
Why It's the Right One
The fact that, historically, the difference between the 10-year yield and either the three-month yield or the fed funds rate (let's lump them together and call them the money-market rate) has predicted recessions more accurately than any other segment of the yield curve is not open to debate. It's what the record shows.
Why is not really a mystery. Contrast the spread between the 10-year yield and the money-market rate with the spread between the 30-year yield and the two-year yield. At the long end of the comparison, the 10-year yield is preferable to the 30-year yield because it has more bearing on economic activity. Most corporate bonds, for example, mature in 10 years or less.
On the short end, the money market rate works better than the two-year note yield because the two-year note yield, like the 10-year note yield, contains inflation expectations. "The uncertainty involved in forecasting interest rates two years from now is not tremendously different than the uncertainty involved in forecasting them 10 years from now," says Arturo Estrella, senior vice president in the New York Fed's research and market analysis group and co-author of the research cited above. "Three months from now you know that the short end will be dominated by monetary policy moves or very short-run monetary policy expectations."
The third thing everyone should know is that many people care deeply about the difference in yield between the two-year Treasury note and the 30-year bond (as well as about every little segment of the Treasury curve between two and 30 years), but not because they care about the economy. They care because they trade those differences, both in the bond market itself and in the Treasury futures market.
For example, to bet that the difference in yield between the two-year note and the 30-year bond will shrink, one can buy the 30-year and sell short the two-year. As long as the yield difference shrinks, the trade is profitable, even if both issues go down in price.
Check out the
Chicago Board of Trade's
Web site each day for a list of the most popular so-called spread trades involving the Treasury bond futures contract.
Damned If You Do...
And the final thing investors should know is that many market analysts believe that, however you look at the yield curve, it's a lot less reliable than it used to be at predicting recessions. Some don't trust it as an indicator at all, even when the yield curve in question is the one that has successfully predicted past recessions.
The issue is complicated by the fact that over the last 30 years, interest rates have moved broadly lower and have become less volatile. As a result, the recessions of the 1970s and 1980s were preceded by much stronger yield-curve recession signals than the recession of 1990-91. While the yield curve signaled near-100% probabilities of the recessions of 1980 and 1981-82, it signaled about a 25% probability of the last recession.
And there have been false signals. Last year, for example, the curve signaled nearly as high odds of a recession as it did before the 1990-91 recession, and there hasn't been one. Estrella concedes that "as we gather more data, maybe it will show that the yield curve is not as strong a predictor of recessions as it was before."
The basic reason why many bond market analysts no longer put stock in the yield curve's ability to predict a recession is this: The last 20 years have seen such an explosion in the size of the bond market that perhaps intermediate- and long-term yields are more important to the economy than the fed funds rate. If they are high relative to the fed funds rate, producing a steep yield curve, then economic activity should be curtailed rather than stimulated.
In any case, it would clearly be folly to say either that the yield curve -- as measured by the difference between the money-market rate and the 10-year note yield -- is predicting a recession or that it is not predicting a recession. The curve isn't predicting a recession, and if it were, lots of people wouldn't believe it.