Saturday, June 2, 2012

What Volcker Did

I've been thinking a little bit about the interest rate targeting approach to monetary policy for which New Keynesians like John Taylor and Greg Mankiw advocate -- Taylor had an op-ed in the Wall Street Journal on Thursday which somehow led me to do some serious work on this.

In particular, I've been thinking about the inadequacies of such an approach, which strike me as awfully baked into the cake. If you're targeting the instrument to achieve a result, then you've introduced an intermediate item which is distracting if not misleading -- a sort of "monetary parallax" which distorts the field of vision of the central bank.

In the late 70s, this parallax effect confused the Burns and Miller Feds about the appropriate level of the fed funds rate given high inflation and nominal GDP growth, because such nominal interest rates seemed unreasonable even when real rates were below zero. In the productivity boom of 90s, the Greenspan Fed consistently feared that such real growth and unemployment below the "natural rate" would inevitably induce inflation and often considered rate hikes. Today, the Bernanke Fed is stuck with this parallax problem, struggling to see beyond the zero federal funds rate to the fall-off from the old NGDP path.

And I've been mulling over the Volcker Fed, which is credited, rightly so I think, for bringing about disinflation in the United States. I think Volcker succeeded because he saw beyond short-term interest rates -- he saw monetary growth and longer-term Fed behavior as critical. He avoided the parallax problem.

Allan Meltzer writes in  A History of the Federal Reserve that:
Paul Volcker's major contribution stands out. Unlike 1966, 1969, 1973, and other times, he persisted in an anti-inflation policy long enough to bring the inflation rate down permanently.
Greg Mankiw, in fact, has a version of the Taylor rule, which estimates the federal funds rate based on core inflation and the unemployment rate. It breaks several times, depending on your fitting parameters, and these breaks actually tell us a lot. (Moreover, that is a flaw with Mankiw's version of the rule, in that it's not clear what years you are supposed to sample from.)

Toying with Mankiw's model, we can see why Volcker's efforts succeeded where those before him had failed.

It wasn't because he raised interest rates up high in '79 and '80; it was because he kept them there in '83, '84, '85, and for the remainder of his term as chairman. That runs directly contrary to the Mankiw model's recommendation, which calls for a massive cut of the federal funds rate in the early 80s, even into the negative nominal territory -- impossible, I know -- but such cuts would have likely meant that the U.S. did not reduce inflation expectations as it did under Volcker.

Below is a graph of the nominal federal funds rate since '68, versus what one might call the "structural federal funds rate." To calculate it, you derive the Mankiw rule from the last ten years of data on core inflation, unemployment, and the federal funds rate, and then plug in an environment of 2 percent inflation and 5 percent unemployment (that is, a Mankiw indicator of -3).What you see is the signature of the Volcker disinflation: a massive and prolonged increase in the structural federal funds rate, which began in '82 lasted all the way into '94, when slow disinflation stopped and the Fed established its then de facto target of 2 percent inflation.

The change was massive. Until 1982, my measure of the structural federal funds rate floats around 4 percent, which sounds like a reasonable estimate of where the Fed would want the fed funds rate to be. Then comes Volcker. The structural fed funds rate goes to approximately 8 percent; Volcker committed to disinflation by keeping rates far higher even as inflation fell and unemployment rose. After 1994, the structural federal funds rate returned to 4 percent, until recently -- and correctly in my view, the Bernanke Fed has kept the structural federal funds rate at around 3 percent to support demand. (It's the reverse Volcker situation today.)

Notice how little the structural federal funds rate is correlated with the actual federal funds rate. Volcker's genius lies in, and his results came from, the decision to keep the federal funds rate high. The structural federal funds rate didn't budge in '79 or '80, which is why inflation did not descend like it did in '82. The increase in the structural federal funds rate in '82 is what produced the disinflation.

This is why it's hard to read interest rates as a sign of tight or loose monetary policy -- you've got to do the reading in the macroeconomic context. Using the Mankiw indicator to work backwards and find the structural rate shows how Volcker succeeded, and perhaps the path forward for resuscitating aggregate demand today.

4 comments:

  1. Evan: this post looks interesting, but I think you need to spend more time explaining clearly what you mean by the "structural Fed Funds rate". I'm guessing that you mean something analagous to the "structural deficit" measure of the stance of fiscal policy?? But I'm not sure I see exactly what your red line means.

    ReplyDelete
  2. (I had trouble posting my comment. The "publish" button didn't appear. I had to hit "edit", then "publish". Maybe it was me screwing up, I don't know.)

    ReplyDelete
  3. @Nick Rowe:

    Let me try to clarify now.

    You take the last ten years of data on the federal funds rate (R), core inflation (I), and unemployment (U), and use that to calculate a Mankiw-style Taylor rule in the form of R = a*(I-U) + b through regression, where a and b are the determined constants. Then you assume that inflation is 2 percent, unemployment is 5 percent, so the "Mankiw indicator" I-U is -3. You plug that back into the rate formula, and what you have is a graph of where the Fed would have put the federal funds rate, assuming that macroeconomic conditions where exactly where they wanted them. What you see is that this rate is usually around 4 percent -- except for a dramatic change to 8 percent during the Volcker-Greenspan disinflation.

    Makes sense? If not, I can clarify further.

    ReplyDelete
  4. Re-reading my explanation in the post, I see it's pretty unclear and hasty. Hmm. Maybe I will go back and add this comment as an asterisked footnote if it makes things clearer.

    ReplyDelete