PPMLHDFE with multilateral resistance terms and sector fixed effects

[Anna Schleithoff]

Joao Santos Silva
Dear Joao Santos Silva,

I am examining the impact of trade facilitation on sector exports. As a proxy for trade facilitation I use the Logistics Performance Index (LPI) published by the World Bank. In Stata, I run the following command for each sector and for each LPI index :

cap egen imp = group(country)
cap egen exp = group(exporter)
ppmlhdfe export100 gdp border comlang distcap customs, a(imp#year exp#year imp#exp) cluster(imp#exp)
ppmlhdfe export100 gdp border comlang distcap infrastructure, a(imp#year exp#year imp#exp) cluster(imp#exp), nest save(Regression)
ppmlhdfe export100 gdp border comlang distcap shipments, a(imp#year exp#year imp#exp) cluster(imp#exp), nest save(Regression)
ppmlhdfe export100 gdp border comlang distcap logistics, a(imp#year exp#year imp#exp) cluster(imp#exp), nest save(Regression)
ppmlhdfe export100 gdp border comlang distcap tandt, a(imp#year exp#year imp#exp) cluster(imp#exp), nest save(Regression)
ppmlhdfe export100 gdp border comlang distcap timeliness, a(imp#year exp#year imp#exp) cluster(imp#exp), nest save(Regression)

ppmlhdfe export101 gdp border comlang distcap customs, a(imp#year exp#year imp#exp) cluster(imp#exp)
..etc...

I would like to add sector fixed effects via:

cap egen sector100 = group(export100)

As soon as I run this command, I get an error message saying "insufficient observations".

Additionally, I was wondering how I could apply ppmlhdfe in a panel setting and how to add multilateral resistance terms.


I appreciate the help! Thank you already so much in advance.

Kind regards,
Anna
Dear

[Joao Santos Silva]

Dear Anna Schleithoff,

I am not familiar with your data and I confess that I am not that familiar with the details of ppmlhdfe, so it is difficult to know what is going on. My guess is that you are not including the fixed effects you think you are including. I suggest you use a subsample of your data and estimate both by ppmlhdfe and by ppml with the dummies, and check that you get the same results.

Best wishes,

Joao