Measuring the strength of cointegration and Granger-causality

This paper proposes a methodology that combines the use of Schwarz's BIC in subset autoregression and subset transfer function identification along with the posterior odds ratio test developed by Poskitt & Tremayne (1987) in the context of testing for Granger-causality and cointegration tests. This approach provides a measure for the strength (decisiveness) of causality and cointegration between the variables of interest. As an illustration of our methodology, we reexamine the case of bivariate relationship between money and income in Canada.