mardi 16 avril 2019

Stationarity test in STATA using Panel Data

Hope someone is able to solve my problem!

Firstly, I am working with panel data of 44 countries over a 13 year period and analyzing the impact of foreign aid on these countries.

Secondly, I have roughly 8 different variables that I am wanting to run a regression for: GDP per capita, life expectancy, school enrollment, production sector aid, etc.

Thirdly, I want to essentially test the independent variables (possibly also DV?) for their stationarity using unit root tests. From what I learned from using help in Stata and Stata blogs, internet in general, some literature, etc... is that I should run an Augmented Dickey Fuller test to check for stationarity, but I am unsure what adding a trend/lag/intercept will do when testing for stationarity. If someone can explain this to me, that's be great!

Lastly, my plan is afterwards to run either a Fixed Effect or Random Effect regression. My understanding is that you can run a Hausman test first to determine which is more relevant. Grateful for any advice or assistance! Thanks!

Aucun commentaire:

Enregistrer un commentaire