Dear Statalist members,

I would like to run a SFA loop to get efficiency estimates for a worldwide sample for 28 years. I group firms based on Country and Fama and French classification and then run the following loop

egen group = group(CountryID FamaFrench)

generate eff=.

forvalue i= 1(1)20 {
sfpanel Sales CostofGoodsSold SellingGeneralAdminExpenses PropertyPlantandEquipment if group==`i'
predict temp`i', bc
replace eff= temp`i' if group==`i'
drop temp`i'
}

The loop runs for a number of groups but then stops and gives the following error:
"initial: Log likelihood = -<inf> (could not be evaluated)
could not find feasible values
r(491); "
I have dropped observations with missing values. I also think it is not because of the number of observations, because the loop has already run for a group with less observations.
I understand that it has to do with ML maximization but I don't know whether and how I can solve it. Any ideas regarding how I could solve it?

Yours sincerely
Periklis Boumparis