Macroeconomic Policy Design Using Large Rational Expectations Models: Methodology and Applications

This paper proposes and applies to the London Business School (LBS) model a general methodology for the design of macroeconomic policy using large rational expectations models. Design proceeds through the following four stages: first, a small, linear representation of the original large, nonlinear model is obtained. Second, control techniques adapted to ration al expectations models are applied. These result in optimal, open-loop trajectories for both the instruments and the endogenous variables, plus a feedback rule which stabilizes the economy about its open-loop path. The optimal policy is only optimal ex ante. Ex post it ceases to be optimal and there exists an incentive to renege. This is the problem of time inconsistency. In the absence of a reputation for precommitment the optimal policy will lack credibility and the policy-maker must be constrained to a sub-optimal policy from which there is no incentive to renege. The paper examines the possibility that a concern to maintain reputation may be sufficient to sustain the optimal, but time-inconsistent, policy.