How to estimate Modified Jones by each country from a cross-country dataset

[Jack Kitiwong]

I tried to use the following command, but it does not work.

gen preta_c=.
xtset c_id_new year
tempvar ta
levelsof two_sic_code,local(levels)
foreach x of local levels{
foreach z of numlist 2009/2020{
capture bys country_id: reg ta mdj_x1 mdj_x2 mdj_x3 if two_sic_code==`x' & year==`z'
if !_rc {
predict `ta'
replace preta_c=`ta' if e(sample)
drop `ta'
}
}
}

[Carlo Lazzaro]

Jack:
welcome to this forum.
Surely what follows does not explain why your code did not get you what you wanted, but why -xtset- your dataset and then go -regress-?

[Jack Kitiwong]

Carlo:
Thank you. I want to generate the coefficients by country, industry, and year. However, I did know how to write the command.
Kind regards,
Jack

[Carlo Lazzaro]

Jack:
1) see https://www.statalist.org/forums/for...r-syntax-error ;
2) consider the community-contributed module -reghdfe- if your want to estimate multiple fixed effects.

[Jack Kitiwong]

I appreciate your help. Thank you.

[Jack Kitiwong]

I tried to use this command, but it did not work.

vallist two_sic_code
local a =r(list)
vallist year
local b =r(list)
vallist country_id
local c =r(list)
gen uhat=.

xtset
foreach i in `a' {
foreach x in `b' {
foreach j in `c' {
capture noisily reg ta mdj_x1 mdj_x2 mdj_x3 if two_sic_code==`i' & year==`x' & country_id= `j', nocons
capture noisily predict uhat_2, resid
capture noisily replace uhat= uhat_2
capture noisily drop uhat_2
di `i'
di `x'
di `j'
}
}
}

invalid syntax
last estimates not found

[Rich Goldstein]

first, please read the FAQ on why "it did not work" is not sufficient information and what you should say instead

second, see

h trace

for a way to find out where in your code the problem is (note that you should also read about "tracedepth" (same help file))