Fixed effects Poisson for trade exports but not gravity

[junlei luo]

Dear all,

I am performing a country-specific analysis to estimate exports using ppmlhdfe with the option absorb(year industry) and confused about how to argue PPML is appropriate (there are zero exports). As the literature on Poisson for international trade seems to always take on a cross country analysis, I'm not sure if it can be directly applied to my case. In particular, I would like to ask:

1. Can all the arguments for using Poisson for gravity models on trade flows be applied to a country-specific model describing aggregate industry exports (no destinations) with time and industry fixed effects?
2. Since there is no info on destinations, this is a two-way fixed effect model (industry and time only) but the fixed effects are not increasing in the same proportions as they would with importer and exporter fixed effects - does this matter?
3. Is it even possible to correctly specify the conditional mean of exports without a full gravity model, and if not does that mean Poisson is not an option?

Any advise would be greatly appreciated!

[Joao Santos Silva]

Dear junlei luo,

I do not think that I fully understand what you are trying to do, but if your dependent variable has a lower bound at zero and no upper bound, some form of Poisson regression is at least a good starting point.

Best wishes,

Joao