Hi all,
I have been struggling a great deal with this matter and as such I hope you could provide me with answer. I have added a dummy variable to my ARDL model to improve its stability (Eviews). However, if I estimate the CUSUM and CUSUMsq tests afterwards, they start from the point in time where the dummy variable turns 1. However, I am interested in reviewing the stability of the entire time period in scope of my analysis. Does anybody know how to enlarge the scope of the CUSUM and CUSUMsq tests? Curious to hear your thoughts. Many thanks!
Perhaps you included your dummy in the dynamic specification box. Try to put your dummy and any other fixed effect variables in the box " list of fixed regressors" and it should work fine.
Hi Sir, thanks for your suggestion, but I did put the dummy variable In the “list of fixed regressors box.“
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwimofjqy7HtAhUO0uAKHaSLBcUQFjAAegQIAxAC&url=https%3A%2F%2Fwww.researchgate.net%2Fpost%2FDummy_variables_in_ARDL_cointegration&usg=AOvVaw1rafpsQ3OcfGbCxg7cEBuP
Dear @Adeyemi, thanks for your sharing your input. However, I was not able to find the answer in the post you shared. I have already put the dummy variables in the section for exogenous variables and the cusum test still covers only the period after the dummy variable turns 1. Would you mind guiding me a little bit more?
to Owen de Boer
I think you are asking something that cannot be done, if you have an intercept
dummy EViews will not ignore it and always does the QUSUM and QSUM SQUARES AFTER THE DUMMY.
If you have a large sample you could then do The stability tests separately in each sub sample
best
George K Zestos
- Badges
- Science topic
- Similar topics
- Cognitive Science
- Thinking