29 MARCH 2019


By Raul Elizalde

This post also appeared in forbes.com

The shape of the yield curve has a lot of information about where the market thinks interest rates and the economy are going.

One often-discussed measure is the difference between 10- and 2-year US Treasury rates. In normal times it is positive (10-year rates are higher than 2-year rates) but it seems to turn negative before a recession.

Critics object to it, saying that it has flashed false positives in the past. Spring, for example, has followed winter before every recession too.

Critics may have a point, because the 2-year rate is influenced by policy as the Fed raises or lowers short-term rates, but the 10-year rate is almost purely market-driven. If the difference can be distorted by central bank intervention, it may not be a good indication of what the market truly thinks, and advocates may be putting too much credence on a measure that compares apples to oranges.

Either way, while the 2-10 difference has narrowed considerably in recent months, it is still positive by about 0.15%-0.20%, so it has not flashed the “official” warning sign that many observers are waiting for to formally announce that a recession is coming.

There is another way of looking at rates that might reduce significantly the apples vs. oranges problem. Under this measure, the market is already in recession warning.

Interest rates at any maturity can be thought of as a sequence of shorter rates. For example, a 5-year rate is a sequence of a 3-year rate and a 2-year rate starting three years from now. The latter is a “forward” rate, and when calculated it reflects where the market thinks 2-year rates will be three years from now. Comparing forward rates with today’s rates is a good way to measure whether the market believes that rates will be higher or lower in the future.

We looked at the U.S. Treasury rate database and calculated forward 2-year rates, three years ahead. Then we compared them with current 2- and 3-year rates at each available date to see whether the market expected rates to fall or rise.

Typically, one would expect forward rates to be higher than today’s rates given the normal upward slope of the yield curve. But we found that ahead of each of the last seven recessions, 2-year forward rates fell below both 2-year and 3-year current rates.

Unlike the 2-10 difference, which is still positive, our alternative measures have just reached negative territory. It may well be that critics are right that policy rates distort 2-year rates and not 10-year rates, but our indicator compares roughly the same section of the curve today with its future expectation, so such distortion should largely cancel out.

It is important to notice that entering negative territory does not signal the immediate start of a recession. Most often, it took around 8 months for it to arrive, by when the difference between forward and current rates turned sharply positive. That exit from negative territory is what seems to take place almost at the same time as the start of a recession.

If these conditions were to repeat, it looks like the most likely time for the expansion to end would be somewhere between the end of 2019 and the first part of 2020.

One should take these conclusions with caution, because a seven-count sample is exceedingly small to make a solid statistical inference. Still, when put into context with other indications, it is hard not to see the mounting evidence that the longest economic expansion on record is coming to an end.