State Space assuming zero covariance

[Chrissle]

Hello,

I was set up a Kalman model when running it everything seems fine, until I run @curr_statecov where I cant find a way to make eviews spit out decent covariances.

I know eviews, by default assumes zero covariance, but after appending @evar cov(e1, e2) = c(25) etc. I presumed getting results other than zeros.

Any ideas? I would be most greatful.

Chris

[trubador]

These are model-based problems and there is no single solution. Search the forum for discussions on state space estimation in general. You can find several useful tips...