Optimal Policy with Partial Information in a Forward-Looking Model: Certainty-Equivalence Redux

Download PDF file (204 KB)

Lars E.O. Svensson
Princeton University,
CEPR and NBER
and
Michael Woodford
Columbia University and NBER

First Draft: January 1999
This Version: June 2002

Abstract

This paper proves a certainty equivalence result for optimal policy under commitment with symmetric partial information about the state of the economy in a model with forward-looking variables. This result is used in our previous paper, “Indicator Variables for Optimal Policy,” which synthesizes what is known about the case of symmetric partial information, and derives useful general formulas for computation of the optimal policy response coefficients and efficient estimates of the state of the economy in the context of a fairly general forward-looking rational-expectations model. In particular, our proof takes into account that, under commitment, the policymaker can affect the future evolution of the observable variables, and thereby potentially affect the future information available.

JEL Classification: E37, E47, E52, E58

Keywords: Partial information, Kalman filter, monetary policy, discretion and commitment, certainty equivalence, separation principle.