Hello Statalist members,

I'm currently working on my bachelor thesis and therefore I'm analyzing the trade diverting effects of antidumping duties, which the EU implemented on several Solarproducts from China in 2013 (I use the harmonized System and there are ten HS Numbers affected by those duties).
In my dataset I work with a cluster of countries (6 countries) which is the aggregation of the trade flows for a specific HS number in a year.

During my bachelor (business administration) I only learned the "normal" OLS Regression.
I know from literature, that the OLS Regression isn't really useful for using it with the gravity equation introduced by Tinbergen.

Since I cannot upload my .dta here, you will find it in the following DropBox link (perhaps someone is willing to help me )
https://www.dropbox.com/s/9nxfempz0j...taset.dta?dl=0
So my advisor told me to also look for a regression with fixed effects and for the difference in differences approach by controlling for other ten HSnumbers, which weren't affected by any duties.

First of all I really don't know if my dataset is correct. For example if I have to put all the HSnumbers into one variable like I did, or to make ten HS variables which will be 1 if the trade flow belongs to the HSnumber. If I would do the latter, my number of Observation would decrease, which I thought would be bad.
However my next problem is to regress with fixed effects.
I know, that I have to work with the xt commands and of course, first I have to define my panel variable.
But what is the panel variable in my case?
I think that only the HSnumber and the year would be a appropriate way.
In the following there are my assumptions:

egen hs1=group( HS)

xtset hs1 year, yearly

xtreg ant_Solarimport ad logGDP_reporter logGDP_partner, fe

The coefficients are the same as in the following normal OLS Regression
reg ant_Solarimport ad logGDP_reporter logGDP_partner

After that I also tried to do a Hausman Test to explain why I will use the fixed effects model in my Thesis, but the result isn't really helpful for me as you can see:
egen hs1=group( HS)

xtset hs1 year, yearly

xtreg ant_Solarimport ad logGDP_reporter logGDP_partner, fe

estimate store fe

xtreg ant_Solarimport ad logGDP_reporter logGDP_partner, re -> also same results as the xtreg, fe and the normal reg command

estimate store re

Hausman fe re
The Prob>chi2 = 1.0000, which was not expected in my opinion.

So long Story Short, I know it is a huge request.
But I really don't know how to continue with my work, also because my advisor isn't really cooperative.

If I should make something more clear about my data please ask!

Thank you very very much for your time.
I truly hope that someone could help me, since I'm despairing somehow.

Kind regards from germany
Felix Flegel