Hi all, quite new to Stata.

I am doing research on Excess CEO Compensation and I want to regress the natural log of CEO's total compensation on the control variables of CEO compensation. However, I want to perform a separate regression for each company for each year, and use each result as variable for that particular observation.

I've tried the following commands, however it doesn't provide me with the desired results:

sort CIK YEAR

gen expected_comp=.
egen id=group(CIK)

forvalues i=1(1)5696 {
I id CIK if id=='i'
reg logTDC1 logSALES logFS if id=='i '
predict p if id=='i'
replace expected_comp=p if id=='i'
drop p
}

Can anyone help me with this?