Long-term Association in Time-Series through Cointegration Analysis: A Case Study

The association among variables, especially when endogeneity is not defined, can be modelled using Optimum Least Square. In the case of endogeneity, Two Stage Least Square Method (TSLS) can be deployed. However, presence of Cointegration can jeopardise the existing association among the variables in both the cases. Even if, either Optimum Least Square (OLS) or TSLS can be estimated but the estimated coefficients in the presence of Cointegration will not be appropriate and may be misleading. Testing of cointegration is done by any of the three methods, which is further elaborated by Error Correction Models (ECM) to explore the nature of Cointegration relation.