Cross panel Chow tests

[mscarlatos]

I have an overall panel including 7 countries across 120 quarters which I break into two subgroup panels. I understand Eviews does not perform stability tests on panel data refressions (fixed effects). However, is it wrong or right to regress the overall panel (fe) and record the residual sum of squares, number of observations and the number of estimated parameters and then do the same with each of the subpanels and then to calculate a Chow F? Rather than indicating whether there is a break across time, could this be viewed as a structural break between the 2 subgroups?