Estimation of Autocovariances

[keef1992]

Hi there,

I am an absolute beginner in EViews (no programming experiences for this software so far).
However, I have to estimate the following:

Given a time series x_t, I want to estimate autocovariances (--> vector of autocovariances dependent of lag)
Formula: 1/T sum( (x_t - x*) ( x_t+h - x*) )
where T = sample size, h= lag, x*= arithmetic mean (= 1/T sum(x_t) )

Ex.
lag h, x_t+h, x*,
0, x_t, ? ,
1, x_t+1, ?, --> ? auto covariance(1)
2, x_t+2, ?, --> ? auto covariance(2)
3, x_t+3, ?, --> ? auto covariance(3)
4, x_t+4, ?, --> ? auto covariance(4)
5, x_t+5, ?, --> ? auto covariance(5)

**EDIT: the series is mean stationary, so forget about x*. shall be constant ****

How can I implement a code (or maybe there is a build in function) to obtain values for the question marks ?

Thanks in advance!

Best regards.

[EViews Glenn]

How do you want these presented?

If you want these for further use, you can always just compute the covariance between Y and Y(-1), Y(-2), etc. and save the results.

If you just want the autocorrelations, EViews has a built-in view of a series (View/Correlogram...) which displays consistent approximations of the autocorrelation values (holding the estimated mean fixed).

So if you let us know how you want to use these, we'll better be able to help.