The Vector Autoregressive Model

Abstract Deals with the classical statistical analysis of the unrestricted vector autoregressive model. We give a necessary and sufficient condition for stationarity and a representation for the stationary solution. We derive the ordinary least squares estimators as maximum likelihood estimator and find the asymptotic properties of the estimators for stationary processes to compare them with the results for non‐stationary processes. Finally, we give a brief description of some misspecification tests for the unrestricted model and analyse two real‐life examples.