Approximation of partial correlation ?

[DBel2012]

Hello,

I'm not sure I'm am in the right forum, but anyways here is my problem.

I wonder if someone could help me understand something about the partial autocorrelation approximation made by Eviews.

In the formula found on page 326 of the Users Guide 1, Eviews 6, it is said that

Φ k,j = Φ k-1,j - Φ k Φ k-1,k-j

I know what Φk refers to (single subscript), but I cannot figure out what, for instance, Φ 1,1 would refer to?

Thanks!

[EViews Glenn]

It's just a recursive function for the partial autocorrelation. The double subscript phi terms are just defined in the line below.