Specify different initial values but output gives the same

[strypste]

Hi all

I would like to estimate an ar(1)arch(1) model with different initial values. In order to do that, I use the "s=" option. In the example below, eq1 uses 0 as initial values and eq2 uses the ols estimates. But the output, however, gives the same initial values.. How is this possible??

equation eq1.arch(1,0,showopts,s=0,z) y c ar(1)
equation eq2.arch(1,0,showopts,s=1,z) y c ar(1)

best
s

[EViews Gareth]

The initial value option only applies to independent, non-constant regressors in the mean equation. If you had an X in there, you'd notice the difference.

[EViews Gareth]

Might also be an issue with the way you're estimating the constant in the LogL object. Remember that your Y doesn't follow an AR process. The error term does. It won't impact estimation, other than your starting value won't be identical across estimations.