Page 1 of 1

FAVAR model estimation

Posted: Tue Jun 13, 2017 2:23 am
by sivek
Hi!!

Please help. I am trying to run a FAVAR model using the FAVAR add-in but its giving me a response which says "factor 1 is not the same length as sample." it my first time running a FAVAR. Secondly what the appropriate minimum number of variables if one is to successful run a FAVAR model.

Regards

SK

Re: FAVAR model estimation

Posted: Tue Jun 13, 2017 2:37 pm
by dakila
I think you should change the sample size because you may have the missing values after the transformation of the variables. Next time you should ask question in the Add=in support section.

Re: FAVAR model estimation

Posted: Mon Aug 06, 2018 11:48 am
by uurak
Hi;

I'm sorry for dumb question but I've needed some clearance. To determine the number of factors, there are two ways, if I'm not wrong. Principal component analysis and Nai&Bg (2002) test. However, is it possible to run Bai&Ng test in Eviews, or in any other programme?

I've found this code, but I'm not sure because he emphasizes panel data in the description below:
"This code computes the various panel criteria proposed by Bai and Ng (2002) to determine the number of factors in a multifactor model of large dimension. "

Thanks.