Links
Abstract
I propose a formal method for decomposing the frequency domain information about latent variables in dynamic models. A model describes the joint probability distribution of a set of observed and latent variables. The amount of information transferred from the former to the latter is measured by the reduction of uncertainty in the posterior compared to the prior distribution of any given latent variable. Casting the analysis in the frequency domain allows decomposing the total amount of information in terms of frequency-specific contributions as well as in terms of information contributed by individual observed variables. Using the proposed spectral decomposition can help researchers identify where information about shocks and other latent variables in structural macroeconomic models come from, thereby making the estimation of such models more transparent. I illustrate the usefulness of the methodology with applications to three recent articles.