Model uncertainty and the design of robust monetary policy rules in a small open economy: A Bayesian approach.

摘要

This dissertation examines the performance of simple monetary policy rules for a central banker that faces uncertainty over the model's structural parameters. I lay out a small open economy model featuring habit formation and derive the output gap. The model is estimated using Bayesian techniques with data from Switzerland and the Euro Area. The estimated uncertainty is then used to evaluate the performance and robustness of simple interest rate feedback rules.; One finding is that the persistent deviations from the law of one price featured in the model affect output and the output gap in different ways, implying a different policy response if one or the other is included in the interest rate feedback rule. Also, the Bayesian estimation procedure using three sets of priors with varying degrees of information shows that the data is not informative about several key parameters. Thus, choosing informative priors for these parameters will predetermine the estimation results, affecting the dynamics of the model and the choice of optimal policies. The estimated monetary policy rule of the Swiss National Bank is compared to the robust policy rule, and the results suggest that the Swiss National Bank could reduce the volatility of both inflation and the output gap by smoothing the interest rate less, and being more aggressive against inflation and the output gap. Finally, the optimization results indicate that increased uncertainty translates into robust policies that are less aggressive.