ABSTRACT
This chapter provides an introductory coverage of cointegration analysis when more than two variables are investigated. The multivariate case is a bit more complex and it does require the use of matrix algebra to derive results. However, the methodology is now routinely carried out via time series econometric packages (e.g. Microfit 5 and EViews) and is straightforward to use in practice. In what follows we have tried to keep the level of mathematics to a minimum and have used a number of examples to illustrate key ideas and procedures. Key topics
The Engle-Granger (EG) methodology
Vector autoregression and cointegration
The Johansen approach
Granger causality test
