Table 3. Single-equation and multivariate cointegration tests with and without structural break(s).
Single-equation cointegration tests | ||||
Engle and Granger (1987) | Gregory and Hansen (1996) | Hatemi (2008) | ||
No breaks | 1 (endogenous) break | up to 2 (endogenous) breaks | ||
Tau statistic | Z-t statistic | Break date | Z-t statistic | Break dates |
−3.83 | −4.82 | 2009M11 | −5.29 | 2006M1 2009M1 |
Single-equation and multivariate cointegration tests with and without structural break(s). The null hypothesis for all tests is the absence of cointegration. All the tests considered the case of no deterministic trend in the data and an intercept in the cointegration equation (CE), centered seasonal dummies outside the CE, while the number of lags is chosen using the Schwartz criterion. The tests allowing for break(s) considered the case of a level shift. * Significant at the 5% level.