Fixed years effect in panel data regression

[Thomas Kok]

Hello everyone,

I am quite new to stata and I am trying to run a panel data regression to the effect of inward FDI on economic growth (GDP per capita). I include country and year effects to account for fixed effects in the model. Since the country effect is included with the xtset of the data, I included dummy variables for the year effects. However, my stata software omits the first and last year of these dummy variables and I am not really sure how it works and what steps I must take in stata to solve this problem. I took logarithms for a few variables due to violations of normality assumptions and a lag variable for fdi. I also added the file with data to this post.

[Carlo Lazzaro]

Thomas:
1) the best way to go -fe- with a continuous regressand is -xtreg,fe-;
2) given the previous point, the best way to include year fixed effect is to have a unique -timevar-;
3) -year1- is omitted to save you from the dummy trap;
4) the last year is omitted due to collinearity with other predictors;
5) logging your variables due to the violation of normality is not mandatory, as normality is a weak requirement for residual distribution only.