How to exclude one group from fixed effect when running forvalue loop

[Olivia Johns]

I ran the following regression with importer fixed efects

ppml trade exp_fe_* imp_fe_1-imp_fe_`=$NoC-1' LN_DIST CNTG import_cost_1, iter(30) noconst

The regression runs but it drops importer 9 FE, because of

note: imp_fe_9 omitted because of collinearity

So as a next step, when I try to run the following:

forvalues i=1(1)`=$NoC-1'{ // use fixed effects estimates (MR)
qui replace exp_fe_`i'=exp_fe_`i'*exp(_b[exp_fe_`i'])
qui replace imp_fe_`i'=imp_fe_`i'*exp(_b[imp_fe_`i'])
}

I get

[imp_fe_9] not found

How can I tell STATA to skip imp_fe_9 when running the forvalues loop?

Thanks!!

[Nick Cox]

capture replace

will eat any error.

[Clyde Schechter]

Or

forvalues i=1(1)`=$NoC-1'{ // use fixed effects estimates (MR)
    if `i' != 9 {
        qui replace exp_fe_`i'=exp_fe_`i'*exp(_b[exp_fe_`i'])
        qui replace imp_fe_`i'=imp_fe_`i'*exp(_b[imp_fe_`i'])
    }
}