Hello everyone,
I am currently working on my bachelor thesis and I am writing here to get some insight on how to do my estimation analysis properly.

I am analyzing the effect of an outbreak of an epidemic on a country's trade and cooperation with all other countries - more specifically the outbreak of ebola in Guinea, Sierra Leone and Liberia.

I am estimating the data with the PPML estimation technique and want to do a difference-in-differences analysis. I want to investigate how the trade, from a country infected with ebola, is affected by the epidemic. Therefore I want to investigate if the epidemic has created a new trend in the export to other countries. The dataset I created consists of three countries infected with ebola: Guinea, Sierra Leone and Liberia, and three countries without: Benin, Burkina Faso and Ghana.
The dataset consists of counterpart trade export for each of these countries, year, GDP, distance, contiguity, common language, colony, common currency, gatt_o(1 if origin country is a member of WTO, 0 if not), gatt_d(same but for destination country), fta_hmr(if the origin and destination country is in a free trade agreement), ebola_o, ebola_d(1 if infected with ebola), ebola_both(if both countries are affected).

I have some questions I hope can be answered here and also if you have any advice for me, I would gladly receive them

1) I want to create some time dummies, so I can observe the effect. How do I generate these? And should I divide the dataset into 'Ebola countries' and 'no ebola countries' and then estimate the results to see if there is an effect, or can I combine them in one dataset? Will it generate a good estimate then?

2) In PPML, we have to log all the continuous variables and this is what happens when I try to log: What do I do wrong?

. ppml log(export) log(gdp_o) log(gdp_d) log(distw) contig comlang_off colony comcur gatt_o gatt_d fta_hmr ebola_o ebola_d ebola_both

factor-variable and time-series operators not allowed
r(101);

3) Regarding trade endogeneity - I read that if I include pair fixed effects to account for endogeneity, then all my time invariant variables will be dropped because the pair fixed effects already accounts for them. Do I understand it correctly? And what could be the disadvantage of dropping the time invariant variables?

Your help is greatly appreciated. Thank you very much