Hi everyone!
I've been running this code: https://github.com/aptech/gauss-carrion-library/blob/master/examples/brkfactors_heterog1.e and I've got two questions.
- I've noticed in the literature that authors report Z-statistics along with AIC/BIC for each estimated model. I am not sure I understand how they get the AIC/BIC values but I'd like to have them because they seem to be pretty important.
- How can the above code be amended so that instead of getting heterogeneous structural break I receive one homogeneous break?
Thank you for your time and help.
1 Answer
0
Hello and thank you for your question,
The source code for the both the factcoint_iter and MQ_test procedures called in the brkfactors_heterog1.e file can be found in the factcoint.src file located in your GAUSS home directory in the "pkgs/carrionlib/src" directory.
In this file you will find:
- The BIC used to determine number of factors. This is done in the factcoint_iter procedure and is completed on line 585. While this value is not returned from the procedure, this could be printed to screen with a fairly simple modification. For example, add the following to line 586:
print "The BIC used to find the number of factors:";
minc(IC1);- The BIC used to determine the order of the VAR model. This is done in the MQ_test procedure on lines 883-894. Again, this value is not returned from the procedure but can be printed to screen by adding the following to line 895:
print "The BIC used to find the VAR order:"; minc(temp|m_bic);
In regards to your second question about the homogenous breaks, I have two papers I would suggest:
- A test to determine whether the break of each series in panel data models occurred at the same location.
Jiang, P., Kurozumi, E., 2023. A new test for common breaks in heterogeneous panel data models. Econometrics and Statistics https://doi.org/10.1016/j.ecosta.2023.01.005 - An estimation procedure of a common break date in panel data models.
Baltagi et al. (2016) "Estimation of heterogeneous panels with structural breaks" https://doi.org/10.1016/j.jeconom.2015.03.048
Your Answer
1 Answer
Hello and thank you for your question,
The source code for the both the factcoint_iter and MQ_test procedures called in the brkfactors_heterog1.e file can be found in the factcoint.src file located in your GAUSS home directory in the "pkgs/carrionlib/src" directory.
In this file you will find:
- The BIC used to determine number of factors. This is done in the factcoint_iter procedure and is completed on line 585. While this value is not returned from the procedure, this could be printed to screen with a fairly simple modification. For example, add the following to line 586:
print "The BIC used to find the number of factors:";
minc(IC1);- The BIC used to determine the order of the VAR model. This is done in the MQ_test procedure on lines 883-894. Again, this value is not returned from the procedure but can be printed to screen by adding the following to line 895:
print "The BIC used to find the VAR order:"; minc(temp|m_bic);
In regards to your second question about the homogenous breaks, I have two papers I would suggest:
- A test to determine whether the break of each series in panel data models occurred at the same location. Jiang, P., Kurozumi, E., 2023. A new test for common breaks in heterogeneous panel data models. Econometrics and Statistics https://doi.org/10.1016/j.ecosta.2023.01.005
- An estimation procedure of a common break date in panel data models. Baltagi et al. (2016) "Estimation of heterogeneous panels with structural breaks" https://doi.org/10.1016/j.jeconom.2015.03.048