Hi,
I need to develop the following formula:
the first stage will predict city level FDI schock using bartik shock: (delta) Foreign_output_sharect= (sum) employment share ct-1 * (delta)foreign_output_share t
the second stage regression will use the specification below, instrumenting real foreign output share with predicted value from bartik shock:
(delta)Yct=(delta)foreign_output_share ct = Foreign_output_share ct-1 + X + FEyear
where c is province level and t the time and Y is the share of employment in agriculture.
I did:
1 stage
gen inst_weight=sh_jobs_ipums[_n-1]*d.sh_Foutput
egen inst_F_output=sum(inst_weight) if inst_weight!=., by(year geolev1)
2 stage
ivregress 2sls d.sh_agri_jobs_ipums (sh_Foutput=inst_F_output) $xlist, robust cluster(geolev1) first
is it correct?
thanks
I need to develop the following formula:
the first stage will predict city level FDI schock using bartik shock: (delta) Foreign_output_sharect= (sum) employment share ct-1 * (delta)foreign_output_share t
the second stage regression will use the specification below, instrumenting real foreign output share with predicted value from bartik shock:
(delta)Yct=(delta)foreign_output_share ct = Foreign_output_share ct-1 + X + FEyear
where c is province level and t the time and Y is the share of employment in agriculture.
I did:
1 stage
gen inst_weight=sh_jobs_ipums[_n-1]*d.sh_Foutput
egen inst_F_output=sum(inst_weight) if inst_weight!=., by(year geolev1)
2 stage
ivregress 2sls d.sh_agri_jobs_ipums (sh_Foutput=inst_F_output) $xlist, robust cluster(geolev1) first
is it correct?
thanks
