Published for Greg Mankiw’s Blog, December 13, 2012
My friend and sometime coauthor Larry Ball sends me his quick analysis of the Federal Reserve’s recent announcement:
I think the FOMC announcement is big news: for the first time, the Fed clearly says it will be more dovish in the future than the pre-crisis Taylor Rule (TR) dicates.
In my estimation, the pre-crisis TR is something like the following for the real interest rate r:
r = 2.0 – (1.5)(u-u*) + (0.5)(pi-2.0).
Let’s say u* is still 5.0. Then if u=6.5 and pi=2.5, the TR says r = 0, which implies the nominal interest rate is i = 2.5. Yet the Fed says that i will still be zero!
Some argue that u* has risen above 5.0. That would raise the i implied by the TR, strengthening the conclusion that the Fed’s new rule is more dovish than the TR.
Some argue that r* [the constant term in the TR] has fallen from 2.0 to 1.0. I doubt it, but even with that change, the TR still implies i = 1.5. My conclusion about dovishness is robust.
This deviation from the TR has not happened since the TR was discovered. In particular, the Fed was NOT more dovish than the TR in 2003. I believe the numbers for 2003 are roughly u=6.0, u*=5.0, and pi=1.0. For the TR shown above, the 2003 numbers imply r =0 and i=1.0, which is about the same as the actual i.
It is not clear whether the Fed’s announcement of future dovishness will have significant effects today. The efficacy of announcements about future monetary policy is unproven.