J statistic

[maulubas]

I am estimating a system using GMM (the system is obtained from what is normally a panel data set, where each year forms a cross-sectional data set and these cross-sections are separate models in Spec - so it is pooled cross-sections over time).

I am trying to interpret the J statistic provided by Eviews for the purposes of testing the overidentifying restrictions. In the Help section, it is said that the J statistic in a panel differs from that of cross-sectional by a factor of number of observations. I believe this is not applicable to my case?? Do I have to multiply the provided J statistic by the number of observations or take it as given?

Thanks.