Conditional Welfare Comparisons of Monetary Policy Rules

Literature on monetary policy can be broadly classified into two categories. The first category involves the construction of dynamic stochastic general-equilibrium models for monetary policy. The solid micro foundations built into these models are important because they facilitate interpretation of outcomes and cross-validation with the results of other studies: the hope is that better micro foundations will yield better positive macroeconomics. However, solid micro foundations are also important because they provide an appealing basis for the second category of monetary policy analysis which involves welfare evaluations of various monetary policy rules. Welfare analysis of monetary policy rules requires researchers to make two inevitable decisions. They are what kind of criterion to use in ranking different policy rules and what kind of policy rules to consider. In comparison to the practices of normative monetary policy analysis, we propose that welfare evaluations of monetary policy follow the following practice. First, to be consistent with microfoundation based on the consumer problem, we used the conditional welfare---rather than the unconditional welfare---as a criterion. Furthermore, unlike the timeless perspective, we claim that the conditional welfare should be evaluated at the time of adopting the rules. Second, due to the problems associated with time inconsistency of Ramsey solutions, we propose that we find the best policy among a class of `reasonable' time-invariant policy rules. We think that `reasonable' policies are implementable and easy to communicate. We present example economies and investigate the optimal policy under the good practice we propose. After finding the optimal policy, we evaluate the potential costs of other policies such as discretion and timeless perspective. The first example is a simple New Keynesian model, such as the benchmark case by Clarida, Gali and Gertler (1999). In the case when there is no distortion due to monopolistic competition, we show that the optimal policy according to our proposed approach is different from other time invariant rules such as the timeless perspective as far as the current state of the economy is not at its deterministic steady state. The other example is the model by Christiano, Eichenbaum and Evans (2004) which incorporates the monopolistic distortion and so induces another different between our proposed approach and other approaches.