ABSTRACT
Cointegration analysis is perhaps the most significant development in econometrics since the mid 1980s. In simple words, cointegration analysis refers to groups of variables that drift together, although each is individually non-stationary in the sense that they tend upwards or downwards over time. This common drifting of variables makes linear relationships between these variables exist over long periods of time, thereby giving us insight into equilibrium relationships of economic variables. Cointegration analysis is a technique used in the estimation of the long-run or equilibrium parameters in a relationship with non-stationary variables. It is a new method popularised in response to the problems inherent in the specific to general approach to time series analysis. It is used for specifying, estimating and testing dynamic models, and it can be used for testing the validity of underlying economic theories. Furthermore, the usefulness of cointegration analysis is also seen in the estimation of the short-run or disequilibrium parameters in a relationship, because the latter estimation can utilise the estimated long-run parameters through cointegration methods. In this chapter we provide an introduction to the methodology of cointegration, focusing on two-variable regression models. Key topics
Spurious regression and modern time series econometrics
The concept of cointegration
The Engle-Granger (EG) methodology
The estimation of the error correction short-run models
