Asymmetry in Monetary Policy: An Asymmetric Objective Function and a New-Keynesian Model
This paper presents a theoretical examination of the possibilities of asymmetric monetary policy on inflation, in the light of the fact that inflation in Israel has been below the target for several years. The economic model used is the New-Keynesian one, which has featured very widely in the theoretical literature since the beginning of the decade. The model has two equations, one of aggregate demand and the other of aggregate supply, and both include "pure" rational expectations, without inertia from the past. The innovation in this study is the combination of the New-Keynesian model with the asymmetric objective Linex function, instead of the commonly used quadratic function. Unlike the quadratic function, which incorporates certainty equivalence, the Linex function, by including the variance, also relates to uncertainty that accompanies the variables in the objective function. The Linex function nevertheless does include the quadratic case as a special one, so that the results of the current study can be compared with those of other studies that used the same model but with the quadratic objective function. After developing the theoretical model to cover the case of policy under discretion, the paper present simulations, and several interesting results emerge: (a) in equilibrium (without shocks), adopting an asymmetric policy has no effect on the real variables, i.e., real expected monetary interest rate and the output gap, but only on the nominal variables, i.e., interest rate and inflation. In other words, in equilibrium, an asymmetric policy is neutral. (b) In the convergence process following a shock, an asymmetric policy has an effect only if the shock was a supply shock, not if it was a demand shock (apart from the effect on nominal interest rate). Hence identifying the source of the shock--supply or demand--assumes even greater importance. (c) The higher the degree of inflation aversion of the policy, the lower will be inflation; the real interest rate will also be higher, although nominal interest rate will be lower (d) Discretionary policy may have a deflationary bias, a finding which contradicts the generally accepted view that discretionary policy results in an inflationary bias. Other results are given in the summary.