Hello Statalist,
I am trying to investigate the impact of the depth of preferential trade agreements (PTAs) on trade in parts and components.
I am using a gravity equation and a panel of around 191 countries across 21 years (1995-2015). The dataset is unbalanced because I am not taking total imports or exports as the dependent variable but rather trade in parts and components (which is limited to a certain number of SITC Rev. 3 items).
I would like to include country-pair fixed effects, exporter-year fixed effects and importer-year fixed effects.
I used the following command to do this:
reghdfe ln_import depth_index contig comlang_off ln_distance ln_gdpcap_imp ln_gdpcap_exp ln_pop_imp ln_pop_exp, absorb(i.exp_n#i.year i.imp_n#i.year countrypair) vce(robust)
where ln_import is the log of trade in parts and components, depth_index is a number that ranges from 0 to 7, contig refers to contiguity, comlang_off to common language, and so on.
From what I understand, if you use country pair fixed effects, all variables that are specific to the country pair (here would be contig, comlang_off, ln_distance) are omitted and by using exporter-year fixed effects and importer-year fixed effects, variables that are specific to the exporter and importer are omitted.
However, my results were as follows:
I don't know if you can see but I got the results I wanted which is that a higher depth_index leads to more trade in parts and components. Also, I thought contig, comlang_off, ln_distance would be omitted because I used country-pair fixed effects and that ln_gdpcap_imp, ln_gdpcap_exp, ln_pop_imp, ln_pop_exp would be omitted because I used exporter-year and importer-year fixed effects.
However, as you can see, ln_gdpcap_exp and ln_pop_exp were omitted but the rest of the variables were not; their coefficients are merely statistically insignificant.
I was wondering whether there is a problem with my results, or whether this is ok.
I would be extremely grateful for any help!!
Thank you for taking the time to read my post.
I am trying to investigate the impact of the depth of preferential trade agreements (PTAs) on trade in parts and components.
I am using a gravity equation and a panel of around 191 countries across 21 years (1995-2015). The dataset is unbalanced because I am not taking total imports or exports as the dependent variable but rather trade in parts and components (which is limited to a certain number of SITC Rev. 3 items).
I would like to include country-pair fixed effects, exporter-year fixed effects and importer-year fixed effects.
I used the following command to do this:
reghdfe ln_import depth_index contig comlang_off ln_distance ln_gdpcap_imp ln_gdpcap_exp ln_pop_imp ln_pop_exp, absorb(i.exp_n#i.year i.imp_n#i.year countrypair) vce(robust)
where ln_import is the log of trade in parts and components, depth_index is a number that ranges from 0 to 7, contig refers to contiguity, comlang_off to common language, and so on.
From what I understand, if you use country pair fixed effects, all variables that are specific to the country pair (here would be contig, comlang_off, ln_distance) are omitted and by using exporter-year fixed effects and importer-year fixed effects, variables that are specific to the exporter and importer are omitted.
However, my results were as follows:
I don't know if you can see but I got the results I wanted which is that a higher depth_index leads to more trade in parts and components. Also, I thought contig, comlang_off, ln_distance would be omitted because I used country-pair fixed effects and that ln_gdpcap_imp, ln_gdpcap_exp, ln_pop_imp, ln_pop_exp would be omitted because I used exporter-year and importer-year fixed effects.
However, as you can see, ln_gdpcap_exp and ln_pop_exp were omitted but the rest of the variables were not; their coefficients are merely statistically insignificant.
I was wondering whether there is a problem with my results, or whether this is ok.
I would be extremely grateful for any help!!
Thank you for taking the time to read my post.
Comment