Monday , April 23 2018
Home / John B. Taylor / Show Us Your Model and Your Method

Show Us Your Model and Your Method

Summary:
In a Wall Street Journal op-ed today I addressed claims made by Neel Kashkari in an earlier op-ed about rules-based monetary reforms, showing that his claims that the reforms were mechanical or computer-run were simply false and misleading. Kashkari mentioned the Taylor rule a lot in his op-ed. For example, he reported that “my staff at the Minneapolis Fed,” found that unemployment after the 2008 financial crisis would have been higher with a Taylor rule. However, he gave no reference to the study, its methodology, or even its authors, unlike many other Fed officials who write or speak on policy. In my response I cited research that got opposite results to those reported by Kashkari. That research is published and publicly available. So many people have asked me:  What model did Kashkari use, and how did he use it?  I hope that someday the Minneapolis Fed study will see the light of day, but in the meantime we do not know the answers. From the op-ed it sounds like the staff did an old fashioned “alternative path” simulation with some model rather than a more realistic “alternative rule” simulation. (For an historical review of the old and new models and methods, see my recent paper for the Bank of Canada.) Thus the calculation did not take account of expectations or other systems effects normally considered essential when evaluating policy.

Topics:
John Taylor considers the following as important:

This could be interesting, too:

John H. Cochrane writes Basecoin

John H. Cochrane writes Fiscal theory of monetary policy

John Taylor writes Still Crazy After All These Years—And What About the Next 50

Bradford DeLong writes Now That John Williams Is President of the New York Fed, He Really Should Convene a Blue Ribbon Commission on What the Inflation Target Should Be

In a Wall Street Journal op-ed today I addressed claims made by Neel Kashkari in an earlier op-ed about rules-based monetary reforms, showing that his claims that the reforms were mechanical or computer-run were simply false and misleading.

Kashkari mentioned the Taylor rule a lot in his op-ed. For example, he reported that “my staff at the Minneapolis Fed,” found that unemployment after the 2008 financial crisis would have been higher with a Taylor rule. However, he gave no reference to the study, its methodology, or even its authors, unlike many other Fed officials who write or speak on policy. In my response I cited research that got opposite results to those reported by Kashkari. That research is published and publicly available.

So many people have asked me:  What model did Kashkari use, and how did he use it?  I hope that someday the Minneapolis Fed study will see the light of day, but in the meantime we do not know the answers. From the op-ed it sounds like the staff did an old fashioned “alternative path” simulation with some model rather than a more realistic “alternative rule” simulation. (For an historical review of the old and new models and methods, see my recent paper for the Bank of Canada.) Thus the calculation did not take account of expectations or other systems effects normally considered essential when evaluating policy. Rather than evaluating a monetary rule over time they looked as a one-time change in the policy instrument.

If so, this is pretty bad economics. Nearly all of the hundreds of studies of monetary policy rules over the past 30 years have taken this systematic approach. Kashkari is welcome to weigh in on this debate, but to do so he needs to report the model and the methodology that form the basis of the results.

John Taylor
John B. Taylor is the Mary and Robert Raymond Professor of Economics at Stanford University. He formerly served as the Director of the Stanford Institute for Economic Policy Research where he is currently a Senior Fellow. He is also the George P. Shultz Senior Fellow in Economics at the Hoover Institution.

Leave a Reply

Your email address will not be published. Required fields are marked *