VARMA models and macroeconomic modelling: some new methodology and algorithms

Abstract

This thesis studies the usefulness of vector autoregressive moving average (VARMA) models in macroeconomic modelling and forecasting. This thesis aims to provide more evidence on the empirical performance of VARMA models, such as forecast evaluations and impulse response analysis. By doing so, it will contribute to the growing body of literature which uses VARMA models for macroeconomic modelling, and suggest that VARMA models can be both beneficial and relatively straightforward to estimate.