Dear Researchers
I have not question but need some suggestions. I have estimated the export supply functions for more than 25 export categories. However, if i report the dynamic stability test results, which i have obtained from cumus test, it needs a huge space i think i will be put more than 100 graphs in my paper for both short and long-run models
Here, Please suggest me the easy way to handle this problem.
Can i allow to report only one test results instead of both cusum test and cusum sum of squre test. if Yes, which one is good.
Secondly, i have already reported the Ramsey Reset test results for model specification. So, in this case can i allow to not report the dynamic stability results as i think the Reset Test also cover the stability
Please suggest me some other test with detail which is easy and for a results a little space is required. I am doing my analysis in eview please suggest also how i obtain it in eview
thanks and best regards
Irshad
Are you sure that you are concerned with your models' DYNAMIC stability? The cusum / cusumsq tests are concerned with the stability of the model SPECIFICATION, which is not the same thing. As to reporting cusum results, you could simply report whether or not the plot of recursive residuals is contained within the acceptance region for accepting the null of (specification) stability.
Professor Vince Daly
You said that CUSUM only tells about the model Specification then i already Reports the RESET test results for model Specification. Then i Think It is enough and no need to reports the cusum test
Second, how i check the model please suggest me
Regards
Irshad
Dear Sayed Irshad Hussain: Professor Vince Daly is correct, but he is implicitly suggesting a nice way to summarize the Cusum family of tests: just report YES is the plot implies specification stability (meaning parameter stability) or NO if that is not the case. Such a table (whose rows can be the whole set of SR & LR equations you have) may then include as many columns as RESET tests you have (together with the Cusum family of parameter stability tests), and even some additional tests and estimated coefficients for key variables. This is just a nice format for a results summary. If event such a table is too big, do it anyway as for an appendix and then go ahead and include a summary of that summary in the main text of your paper. Best, Carlos.
Have you reviewed either of these:
http://people.irisa.fr/Michele.Basseville/
http://journals.sagepub.com/doi/10.1177/096228020301200205
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series