Hi! Is it possible to determine Johansen test for co-integration with structural breaks from Bai-Perron method? Or is a Chow test with known break dates preferred? Both are built on the assumption of a break in one of the variables, and are single equation tests. But Johansen cointegration is built on a VAR. Are the tests still compatible?
I think it is creatively possible if the person wants more punishment: save residual from normalized Johansen vector and apply the methods mentioned in your question. However, one can use recursive estimation and plot the eigenvalues oR perform a test of the eigenvalues,, see my work in Journal of Economic studies
Thanks! Any idea on how to save the residuals from Johansen test in Eviews? What line of code should I use? And what do you mean by "punishment"?
See if you can get MICROFIT, it allows one to save the residuals from each normalized vector, hence allows you to do Bai or Chow without much difficulty. Using E-views requires creativity and absolute knowledge of multivariate cointegration. That is, the researcher must keep track of which residuals were saved (i.e,, the name)
Alternatively, one can generate the residuals by using parameters of the VECM. Right above the VECM parameters are the alphas for testing for weak exogeneity test with their associated t-values. Please you need to normalize the coefficients of the VECM on Eview by multiplying by -1.. .MICROFIT is more menu driven than E-Views. I still use MICROFIT 4 . These are what I refer to as punishment. For example, you have many things to keep track of, such as the Lag order 5 in MICROFIT is Lag order is 4 , that is, K-1 in Eviews (that is 4) . {Stuff like that} Play with first and feel comfortable with it before finalizing your exam. That is, is why I prefer the eigenvalue stability test recommended by Johansen.
Journal of Economic Studies 28(6) (2001) ,378-396 --See Emerald library I showed the Johansen eigenvalue test
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis