Hi, I am currently working on my thesis. I have a dataset of approximetely 120 reporting and partner countries fråm the year 1996 to 2017. I find my results rather peculiar and was wondering if I could get some input. I am running a baseline equation that is

Code:
reg lnexport lndist contig comlang_off comcol fta_wto EUtrade Brexit EUBrexit export_* import_* year_*, robust cluster(distw)
Array

Now, i find the distance abnormaly high? Shouldnt it be around -1? I also find it peculiar that "EUtrade", which is if both the reporting and partner country is an EU country, is negative and significantly negative. Is there something i can change in my code to make it better?

I know OLS isn't perfect for my dataset. I will also run a PPML. When I look at other papers and they run an OLS, the lndist seems better.

Glad to get some input.

Sincerely,
John