Dear Joao Santos Silva Tom Zylkin ,

Currently, I'm doing an investigation about the effects of Free Trade Agreements on Peruvian trade (imports and exports) in the period 2000-2018. For that purpose, I'm using the gravity model as a benchmark and the PPML as the econometric technique. However, some problems raised with my model specification. In particular, I have problems with the variable LnDistance since it is the same for each pair of countries every year (distance between countries does not change): I have a singular matrix. So, I have some questions.

1) Is correct to use the PPML technique with a dynamic panel? Is there a stata comand for that?
2) If I include fixed effects for each exporter-importer pair, can I overcome the problem of the singular matrix (i.e. the fixed effect absorb the variable LnDistance)? Is it correct?