I am testing time series data for cointegration with structural breaks using the Gregory-Hansen approach. The software am using to conduct this test is RATS, but the test only provides the ADF test statistics in the standard RATS software package. Can someone explain to me how to obtain the z-test statistics for the Gregory-Hansen cointegration test using any econometric software.
Thank you
First, you should use the Engle-Granger ADF critical values, as you are interested in cointegration testing and have to estimate the cointegrating relationship to construct the residuals. For these critical values, the number of regressors are decisive. Second, due to a potential structural break you can treat the dummies simply as additional regressors. As a reference, see Hassler (2004): Dickey Fuller cointegration tests in the presence of regime shifts at known time.
Hansen has posted the Gauss and Matlab code for this model (and several others) on his website at the Economics Department at the University of Wisconsin (Madison).
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Journalism