country/regional fixed effects

[Wendy Brown]

Hi I have a question regarding country fixed effects. I am quite new to this subject, so apologies for the silly question!

I have a model that try to consider what determines the adoption of a specific technology, my independent variables are fdi and some local variables (Patents, export, DVA, etc). I have 34 countries for 12 years. I am running OLS and IV. with time fixed effects. Is it really necessary to have country fixed effects if I control for GDP or population and regional areas? I divided my countries in economic and geographic regions and I am running my models with these three dummies. Is it ok or do I have to insert country fixed effect? Thanks

[Phil Bromiley]

Welcome to Statalist. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata outout, and sample data using dataex.

While this may differ with academic area, the general prescription would be that you need to include country fixed effects and many would also recommend year dummies (which can be added to xtreg by including i.year among the rhs variables. Using fixed or random effects is almost automatic with panel data. The exception (which might lead to random effect or the between estimator) is if stable panel-level variables that you include in the model matter substantively.