Wednesday, April 8, 2009

Fed + yield curve = end of recession

Mark Perry had a nice post yesterday with an update of the Fed's model for predicting recessions and recoveries. The upward slope of the Treasury yield curve now says that the probability of recession this year is rapidly approaching zero: "the Fed's model shows a recession probability of only about 1% on average through the next 12 months, and below 1% by the end of the year."

This prompted me to update my own model, which also uses the slope of the yield curve, but which adds in the real Fed funds rate, since the latter is a good measure of just how tight or loose the Fed actually is. As this chart shows, the yield curve is always negatively sloped going into recessions and positively sloped coming out of recessions. That's because every recession in modern times has been preceded by a significant tightening of monetary policy, and a return to easy money has marked the end of every recession. So today it is clear that we have the essential monetary ingredients for a recovery. Indeed, given the rise in commodity prices and other signs of improvement that I've been noting for awhile, it seems pretty likely that the economy will be on the mend before mid-year, as I predicted at the end of last year.

Of course, when recessions end it is never immediately obvious, and it typically takes many months or even a year or more before the numbers confirm that the recession has ended. I recall how Bush Sr. lost his reelection bid in 1992 in part because of the widespread belief that the economy was hopelessly mired in recession; by the end of 1993, however, revised numbers came out which showed that the economy had actually enjoyed a decent recovery in 1992. Similarly, during the summer and fall of 2003 the mantra was that we were in a "jobless recovery," monetary policy was "pushing on a string," and deflation threatened the global economy. We later learned that the economy took off like a rocket starting in July of that year.


Mark A. Sadowski said...

The Fed's model will not work in our current predicament. The historical relationship between the slope of the yield curve and the economy’s performance is due to the fact that the long-term rate is essentially a prediction of future short-term rates. If investors expect the economy to decline, they also expect the Fed to cut rates, which tends to make the yield curve downwardly sloped. If they expect the economy to expand, they expect the Fed to raise rates, making the yield curve upwardly sloped.

But there's a problem with using the Fed's model right now (as even the Fed itself has acknowledged). It is only based on data collected since the 1950's. Thus the model does not include data from the last time we were in a liquidity trap (the Great Depression). The Fed can’t cut rates any more because they’re already at zero percent. All it can do is raise rates. So the long-term rate has to be above the short-term rate, because short-term rates might move up, but they sure can’t go down.

In fact, if one considers the most recent example of a country in a liquidity trap, Japan during the Lost Decade, you'll notice that the yield curve was upwardly sloped all the way through ten years of stagdeflation. In 1999-2000, when the zero interest rate policy (ZIRP) was in effect, long-term rates averaged about 2%, not too far below current rates in the United States.

The Fed's model almost certainly doesn't work in a liquidity trap. Those who are relying on it need to wake up to the fact that we are in ZIRP, and that is precisely where the Fed's model is completely useless.

Scott Grannis said...

I don't agree. The fact that the Fed is engaged in quantitative easing does not render the slope of the yield curve a meaningless variable. If the market thought that the US economy were really going to be facing a decade of deflation and depression, then it would not expect the Fed to raise rates by very much. In contrast, the current steepness of the curve implies a fair degree of tightening. It is also significant that the curve from 10 to 30 years is steeper than normal, as that part of the curve is especially sensitive to long-term expectations.

Comparisons with Japan are not necessarily valid. The BoJ only implemented quantitative easing during the 2002-2006 period. Prior to that the growth of the monetary base in Japan was extremely low. Quantitative easing successfully ended Japan's decade of deflation. Japan's other problem was that banks were not allowed or encouraged to write off their bad loans, so they remained "frozen" for a very long time. And then of course there is the huge difference between the demographics of Japan and the U.S;

Today's positively sloped yield curve is very beneficial for banks, since they can borrow at almost a zero rate and lend at a much higher rate, and the Fed is all but promising not to upset this money-making machine. I don't doubt that Wells Fargo's surprisingly strong results were powered at least in part by this regime.

Most importantly, we are not in a liquidity trap. The Fed has effectively abandoned its interest rate targeting policy and is injecting massive amounts of reserves and liquidity into the system. I think you ignore the potential of this to make a significant change on the margin in market conditions and investors' expectations at your peril.

I have never known a time when it paid to bet against the Fed, especially when it is trying very hard to achieve a particular objective.