EViews.com

how do I remove the serial correlation in eviews? Most the methods I have found I cannot use because I have a lagged dependant variable as an independent variable..

thanks

If you're doing a regression, add AR(1) to the variable list.

If you're doing a regression, add AR(1) to the variable list.

thank you for your reply. How would this remove the serial correlation? Correct me if I'm wrong, but is this the 'cochrane-orchutt iterative procedure'? I read about this in a book today, however it didn't make it clear whether it was possible to use this to correct serial correlation when there was a lagged dependant variable as an independant variable - I'm guessing I can use it then?

thanks

This is almost (not quite) identical to the Cochrane-Orcutt iterative procedure, and it is okay to use with a lagged dependent variable.

This is almost (not quite) identical to the Cochrane-Orcutt iterative procedure, and it is okay to use with a lagged dependent variable.

thanks for your help :D what is this procedure called?

Also, do i need to pay attention to how many iterations it takes? (i.e, on a regression that i tried of the form [ log(y) log(x1) log(x2) log(y-1) C @trend ar(1)] it said 'convergence achieved after 6 iterations') is this important and worth elaborating on? Or, should i just accept the fact that the serial correlation is removed and concentrate on my new coefficients?

thanks again

The number of iterations doesn't matter.

You might read the section in the Help system under "How EViews Estimates AR Models."

The number of iterations doesn't matter.

You might read the section in the Help system under "How EViews Estimates AR Models."

thanks that was very helpful, when referring to this is it ok if I refer to it as "the iterative procedure"? - or would a different name be more appropriate?

Probably safer to say "an iterative procedure"

I have a related question:

I have specified a partial adjustment model, but there is some serial correlation (looks like AR(1) disturbances).

If I include an AR(1) in the estimation line...does this change my interpretation of my partial adjustment model (and the coefficients that I'm backing out)? How do I interpret the coefficient on the AR(1) coefficient that shows up in the estimation output?

Also, can AR(1) disturbances be used in a system e.g. a SUR system?

Cheers