Having run the ARDL model using STATA 11.2, I need to test for the stability of the model using the Cusum-Cusum squares test. I don't seem to know how I can go about it. Any help? Suggestions?
Thank you!
Hello!
In Stata 14, you need to use command "cusum6". For example:
cusum6 y x x x, cs(cusum) cs2(cusum2)
cs(cusum) and cs2(cusum) create 2 variables containing the values used in the cusum and cusumsq graphs, respectively.
To visualizate cusumsq, you will need to press enter after the cusum graph is showed.
Regards.
Hi Nip Nip,
I thank you so much for your response.
You wont believe I have tried this! But I keep getting error messages saying "no variable defined for cusum", and the same goes for "cusum2" and "cu".
I went further to do these, even though I was unsure of the outcome:
reg y x1 x2 year
cusum6 y x1 x2 year, cs(cusum) lw(lower) uw(upper)
gra cusum lower upper year, s([year]) c(.mm) xlab ylab
Still didn't work! What can I do, please?
I use STATA 11.2.
Many thanks!
Nip,
Yeah, I did that already.....maybe I'll just stick with EViews......
Thanks a lot!
cusum and cumsq are very simple to do in the gretl GUI. Gretl cam import data in a variety of formats, is free and very easy to use. It can be downloaded from gretl.sourceforge.net/.
You have to install cusum6 on the stata before the command.
Type
ssc install cusum
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression