User’s Guide : Advanced Single Equation Analysis : Cointegrating Regression
Cointegrating Regression
This section describes EViews’ tools for estimating and testing single equation cointegrating relationships. Three fully efficient estimation methods, Fully Modified OLS (Phillips and Hansen 1992), Canonical Cointegrating Regression (Park 1992), and Dynamic OLS (Saikkonen 1992, Stock and Watson 1993) are described, along with various cointegration testing procedures: Engle and Granger (1987) and Phillips and Ouliaris (1990) residual-based tests, Hansen’s (1992b) instability test, and Park’s (1992) added variables test.
Notably absent from the discussion is Johansen’s (1991, 1995) system maximum likelihood approach to cointegration analysis and testing, which is supported using Var and Group objects, and fully documented in “Vector Autoregression (VAR) Models” and “Cointegration Testing”. Also excluded are single equation error correction methods which may be estimated using the Equation object and conventional OLS routines (see Phillips and Loretan (1991) for a survey).
The study of cointegrating relationships has been a particularly active areEViews 13a of research. We offer here an abbreviated discussion of the methods used to estimate and test for single equation cointegration in EViews. Those desiring additional detail will find a wealth of sources. Among the many useful overviews of literature are the textbook chapters in Hamilton (1994) and Hayashi (2000), the book length treatment in Maddala and Kim (1999), and the Phillips and Loretan (1991) and Ogaki (1993) survey articles.