Taylor, the Taylor Rule, and Marsupials

July 16th, 2013 at 8:45 am

There’s an old Gary Larson cartoon where this drunk is standing at a bar next to a kangaroo, pointing a finger in its face and saying, “Now let me tell you a little something about marsupials.”

This is by way of saying, “don’t argue about the Taylor rule with a guy named Taylor.”  That’s John Taylor, the Stanford economist who derived the rule.  He’s got a post up on his blog saying that the rule I used in a post the other day isn’t the one that he endorses.

However, the one I used has become a common version, one that’s frequently applied by Janet Yellen (as Taylor mentions), the vice-chair of the Fed, and since my post was about the Fed’s stance right now, it seemed legit to use her version (though I did tweak it a bit in a way I’ll show in a moment—and it’s a way that gets results closer to Taylor’s).  Taylor’s point that I should have shown alternative results, including his own version, is a perfectly good one…I’ll post his alternative later.

But here’s the Yellen paper where she uses the version I used (see footnote 15); note also her discussion as to why she prefers this version, which made sense to me.  I did, however, use 2 as opposed to Yellen’s 2.3 as the Okun coefficient on the output gap and I used the CBO’s NAIRU instead of the 5.6% constant used by Yellen.  For 2013q1, the rule I used returns a fed funds rate of -1.6%; Yellen’s version gets you -2.0%, which actually goes further in the non-Taylor-endorsed direction.

John says that his version would put the fed funds rate in positive territory, so the version of the rule that you choose makes an important different to this analysis (I did calculate his version for the most recent quarter–2013q1–using CBO potential GDP for the output gap and got -0.1% for the rate, but prior quarters were positive).  I do object to Taylor’s comment that the Okun coefficient should be less than two.  As noted, Yellen uses 2.3, a value that’s consistent with recent work by Ball on calculating the Okun coefficient (see his table 1 results—the coefficient Yellen and I both use is the inverse of his beta).

A final point, the original rule as derived by the man himself may need tweaking in unusual periods like the current one, where the all-important zero-lower-bound on nominal interest rates is in play.  Over to Yellen for the last word (I’ll let her tell the kangaroo about marsupials):

While the Taylor (1999) rule can serve as a useful policy benchmark, its prescriptions fail to take into account some considerations that I consider important in the current context. In particular, this rule does not fully take into account the implications of the zero lower bound on nominal interest rates and hence tends to understate the rationale for maintaining a highly accommodative stance of monetary policy under present circumstances.

Print Friendly

3 comments in reply to "Taylor, the Taylor Rule, and Marsupials"

  1. Phil Perspective says:

    Hasn’t Krugman proven, lately, that Taylor doesn’t even endorse his own prior work?


  2. WNY-WJ says:

    DeLong has a nice take-down of Taylor w.r.t. inflation in this post: http://delong.typepad.com/sdj/2013/07/john-taylor-asks-you-disapprove-of-my-methods-but-i-see-no-method-at-all-here.html .

    One would hope that politics is not trumping the “science” of economics, but that seems to be the case today. Current controversies will not be sorted out for years (if ever), and even if a consensus is reached, there will be a rump minority with opposing views just waiting to foment a counterrevolution… just like Friedman, and then Lucas, did against Keynesian consensus.

    Economic “schools” resemble religious sects today, with each one selectively choosing data to support its assumptions – no matter how far removed from reality. A sad state, really. The discipline is losing credibility fast among nonspecialists.


  3. Edward Lambert says:

    Mr. Bernstein,
    Briefly, I still have a problem with 2 variables in the Taylor rule. One is the CBO potential real GDP which is above the effective demand limit of $14.1 trillion.
    The second variable is the NAIRU. My work and that of others points to a value of 7.0% plus or minus 0.2%.
    http://effectivedemand.typepad.com/ed/2013/04/predicting-the-end-of-an-expansion-using-the-unemployment-rate.html

    The Fed funds rate will calculate higher when the constraints on these variables are taken into consideration.


Leave a Reply

Your email address will not be published.

Current month ye@r day *