Postby hraouf » Mon Jul 04, 2011 1:30 am
Hello
I am running eviews 7.1 and am trying to run a rolling regression using the simple roll addin. The columns have 6080 rows and each 40 rows is for a certain group, so I need to do a regression for each 40 rows. When i run the first set of rows the regression works fine however I then go to add ins, use a window size of 40 and a step size of 40 and a get perfect multi colliniarity warnings (even though the equation is designed to prevent this) and meaningless coefficient output files with 6080 rows. What am I doing wrong? not least what is the sample size I need to enter into the equation box (e.g. to do the first set of rows I entered 1 40)
I have enclosed the output from my first run in case it sheds some light
Many thanks
Sam

Attachments

 IMG_0002.pdf
 (105.06 KiB) Downloaded 809 times