PPML with unbalanced panel data, gaps in data, and data asymmetries

[Dan Eran]

I wish to use ppml to estimate a structural gravity model for the years 2000-2016, for ~ 130 countries.
I have some data constraints, and I would like to know if it is plausible to still use ppml under these circumstances.
Specifically, I wish to know:

1. Can one use the ppml estimator when the data is unbalanced?

2. Can one use the ppml estimator when there are gaps in the data? e.g., I have trade data between South Africa and Peru for the years 200-2003, 2007, 2010-2016, and trade data between Israel and Taiwan between 200-2005 and 2008-2016.

3. Can one use the ppml estimator when I have data for just one country in a pair? e.g., I have data on exports from the UK the Poland but not vice-versa.

Many thanks,
Dan

[Joao Santos Silva]

You can use PPML in all those cases, but the quality of the results obviously depends on the quality of your data (this applies to all estimators, not just PPML).

Best wishes,

Joao

[Dan Eran]

Thank you very much (: