Regression Technique of Cointegrated Processes

Abstract

Many financial series or microeconomic data are serially correlated, nonstationary and are found to be integrated. Cointegration has become a major tool in the analysis of financial and microeconomic time series since its introduction and has changed the way econometric modelling is carried out. This paper deals with the concept of regression cointegration. A detailed review of cointegration processes in the context of simple linear regression models is done. The tests on the existence of cointegration using Durbin Watson statistic and Augmented Dickey – Fuller unit root test are examined. Various data sets are used to demonstrate the determination of order of integration and test for the existence of cointegration.

Journal of the Ghana Science Association Vol. 9 (2) 2007: pp. 105-116