Page 1 of 1

MA models, how to estimate lagged values of the error term?

Posted: Sat May 23, 2015 1:04 pm
by E-Conman
How can I use lagged values of the error term in an equation? I tried the most common letters like e and u, and they are not working.

Re: MA models, how to estimate lagged values of the error te

Posted: Sat May 23, 2015 1:18 pm
by EViews Gareth
Could you provide context?

Re: MA models, how to estimate lagged values of the error te

Posted: Sat May 23, 2015 1:48 pm
by E-Conman
Could you provide context?
Apologies if I wasn't clear, I would simply like to estimate the following equation:

Y = C(1)*X + e(-1) + e

Where e(-1) is the lagged error term and e is the error term. Similar to MA models which use lagged error terms as regressors

Re: MA models, how to estimate lagged values of the error te

Posted: Sat May 23, 2015 3:13 pm
by startz
Estimate the equation including MA(1). Your "e" will be in the series resid.

Re: MA models, how to estimate lagged values of the error te

Posted: Sat May 23, 2015 3:39 pm
by E-Conman
Estimate the equation including MA(1). Your "e" will be in the series resid.
Is there a way to restrict the MA(1) to have a coefficient of one?

Re: MA models, how to estimate lagged values of the error te

Posted: Sat May 23, 2015 4:23 pm
by startz
I don't believe so.