Description

This paper analyzes the quantitative releva nce of additive, multiplicative and data uncertainty in the implementation of Chile’s monetary policy. For the analysis of data uncertainty, we focus on the uncertainty associated with the estimation of the output gap using real-time data and various well-known methods to estimate the output trend. We find that the revisions to the output gap are important and persistent and that the unobserved components method performs better with r eal-time data than with others more commonly used, like the HP filter. In the case of additive and multiplicative uncertainties, we estimate the equations that govern the behavior of the economy with time-varying parameters and with state-dependent variances in the shocks of the model. This allows us to analyze the contribution of these two types of uncertainty on the total uncertainty. We find that additive uncertainty is the most relevant to explain total uncertainty and that shocks to the model are state-dependent.