Greetings, I´ve done my best to answers these questions myself but I need some help. The following is within a Gravity Model of Trade framework. I´ve attached my .dta and .do.
1.
Do country-pair fixed effects mean that you generate a dummy which takes the value of unity for each country pair across the time period in a panel?
1.1.
Would this then render all other time-invariant explanatory variables useless and also make the , fe option for regression in stata unnecessary?
2.
If so, what should I do when the Hausman test suggests FE, but litterature suggests using country-pair fixed effects to limit endogeneity, whilst I would also like to keep my time-invariant explanatory variables since they combined with the ,re stata option leads to a much greater r^2 than , fe plus omitting my time-invariants do?
Also, any general help and guidence is appreciated.
Thank you for your time!
Best regards, Alfred
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing