I have unbalanced data for N=1227 and T=142. My data is severely unbalanced and when I run the fisher test in Stata it does not compute it for the majority of the panels. Can I rely on the results? If no, how can I overcome this problem without having to drop panels?