unbalanced panel in spatial panel data models

Much thanks again.

Now, returning to my raw data, I realize that in fact the unbalanced
character comes mainly from a number of regions for which I do not have
any data for the main dependent variables, for none of the years. I guess
in that case imputation would be hard to calculate.
So I will dig this further but maybe in the end I will return to deletion
of missing observations.
Anyway I hope that this may be also helpful for others like Caleb
Gallemore which apparently had similar issue.

Thanks again, I will keep the list informed if in the end I use the
multiple imputation,

Patrick