Variables omitted due to multicollinearity- is it about the type of variable?

[Alba Alvarez]

Helle everyone,

I'm having some difficulties with a regression. When using xtreg with fixed effects two of my variables are omitted due to multicollinearity.

My panel data analysis encompasses 12 countries and the time period is 2003 to 2017. The variables I have are:

Dependent variable: export2: export value

Independent variables: (7-12 are just governance indicators).

1. culdist: cultural distance
2. geod: geographical distance
3. fta: free trade agreements
4. gdppc2: GDP per capita
5. inflation
6. exchrate: exchange rate
7. account:
8. polinest
9. goveff
10. regqual
11. rol
12. coco

When I run the fixed effect analysis, the cultural distance and the geographical distance are omitted due to collinearity.
The first thing I thought is that both variables are the same for each country over the years.

Question 1: Is it possible that that is the reason of the multicollinearity?
Question 2: if so, how do I fix that problem so those variables are not omitted?

*Handling missing values
replace exchrate=.a if exchrate == 999

* Exploratory plots
encode cntry, gen(country)
xtset country year
xtline export2
xtline export2, overlay

bysort country: egen export2_mean=mean(export2)
twoway scatter export2 country, msymbol(circle_hollow) || connected export2_mean country, ///
msymbol(diamond) || , xlabel(1 "A" 2 "B" 3 "C" 4 "D" 5 "E" 6 "F" 7 "G" 8 "H" 9 "I" 10 "J" 11 "K" 12 "L")

bysort year: egen export2_mean1=mean(export2)
twoway scatter export2 year, msymbol(circle_hollow) || connected export2_mean1 year,
msymbol(diamond) || , xlabel(2003(1)2017)

*FE regression
xtreg export2 culdist geod fta account polinest goveff regqual rol coco gdppc2 ///
inflation exchrate, fe
estimates store fixed

Thank you so much.

Alba

[Carlo Lazzaro]

Alba:
welcome to this forum.
1) Only you can know whether those predictors are time-invariant or not.
2) it they actually are time-invariant, there's nothing you can do to halt the -fe- machinery to wipe them out-