I have this command:
foreach v in y x1 x2 x3{
xi: reg `v' i.id*year
predict r_`v', r
winsor2 r_`v', cuts(1 99) replace
replace `v'=r_`v'
drop r_`v'
}
I wonder it's reasonable in research other than financial or accounting.
More basically, what's the command doing?
Any literature use similar methods?
How to explain I cut the influence of id fix effect and time trend in each variables but not control it in fix effect model directly?
foreach v in y x1 x2 x3{
xi: reg `v' i.id*year
predict r_`v', r
winsor2 r_`v', cuts(1 99) replace
replace `v'=r_`v'
drop r_`v'
}
I wonder it's reasonable in research other than financial or accounting.
More basically, what's the command doing?
Any literature use similar methods?
How to explain I cut the influence of id fix effect and time trend in each variables but not control it in fix effect model directly?
