An Introduction to the Cointegrated VAR Model

We give an introduction to the cointegrated vector autoregressive (VAR) model. We briefly consider the vector autoregressive (VAR) model and the link to the vector error-correction model (VECM). We discuss cointegration and the common stochastic trends as a reduced rank restriction on the coefficient matrix Pi. We illustrate how the Pi matrix can be decomposed into an alpha and a beta matrix, where the latter defines the cointegration coefficients and the former the error-correction coefficients. Finally, we consider a simple examle.