regression with time and country dummy variables

#1
Hi guys,

I need some help. I have data for 20 countries and 5 fiscal years. I've read that in order to control for country and year-effects I need to introduce dummy variables for each country and each year. I did that but I don't know what I should use: regression or GLM. I don't know what is the difference between them. In GLM there are fixed and random effects. DO you know the difference? Should I put my dummies into fixed or random?

Thanks
 
#2
Hi Zofia,

When you run a simple OLS regression with dummy variables for each country and each year, the coefficients that you obtain on your independent variables are the same as in TSCSREG with fixed effects (two types of fixed effects, one for counties and one for years).

The reason is simply that the dummies that you introduce are the fixed effects but, aditionnally, when you run OLS regression with these dummies you also estimate the coefficient on these dummies. TSCSREG uses a Within estimator to eliminate fixed effects.

So I suggest that you use TSCSREG.

Etienne