Is there a simple way to extract the X matrix from a reg?

[abossie]

Basically, I need to calculate E(xx') so that I can get confidence intervals for a structural break. I know I can group the variables then convert that to a matrix but this involves several steps and suspect it will give me trouble if I try to change the number of variables. Anyway I was wondering (hoping!?) there would be something like an @ command for this.

I'm using version 7.2

[EViews Gareth]

Make regressor group then convert to matrix is the only way.

Of course if you're just using least squares you could work with the covariance matrix to get X'X.