Covariance between cross sections in panel data?

[wsjUNMC]

Hi,

I am currently using panel data to run a GARCH estimation by programming it myself with Eviews 9. However, I am currently having problems trying to specify my covariance equation.

I have seen examples such as
cov=@cov(y1-mu(1),y2-mu(2))
cov=@cor(y1,y2)*@sqrt(@var(y1)*@var(y2))

However, these only specify the calculation of covariance between variables y1 and y2.

I would like to specify a covariance equation BETWEEN CROSS SECTIONS, such as Equation (10.4) from page 261 of

Cermeño, R., Grier, K.B., 2006. Conditional heteroskedasticity and cross-sectional dependence in panel data: an empirical study of inflation uncertainty in the G7 countries. In: Baltagi, B.H. (Ed.), Panel Data Econometrics, vol. 10. Elsevier, New York, pp. 259–278

The specification in the paper is roughly as follows:

cov(i,j,t) = alpha(i,j) + beta*cov(i,j,t-1) + delta*resid(i,t-1)*resid(j,t-1) for every i does not equal j

Therefore, is it possible to specify a covariance equation between cross sections?

Any help is greatly appreciated. Thank you.