A complete VARMA modelling methodology based on scalar components

This paper proposes an extension to scalar component methodology for the identification and estimation of VARMA models. The complete methodology determines the exact positions of all free parameters in any VARMA model with a predetermined embedded scalar component structure. This leads to an exactly identified system of equations that is estimated using full information maximum likelihood.