Help with bachelor's degree

[Wilhelm Wiklund]

Hello

I am studying the effects of aging population on GDP/capita growth.

GDP/capita growth = Over65 + Under15 + Life Expectancy growth + Average Education years growth + Trade openness.

I have data of all OECD countries from 1995 to 2021

The Over 65 variable is the proportion of people over 65 relative to the working age (in %) .

I have tried using fixed effects to run my regressions

When I use the code

reg BNPCG Over65 Under15 Trade LifeExpG EducationG i.year

I receive a significant result on over65, same with

xtset Country_n year
xtreg BNPCG Over65 Under15 Trade LifeExpG EducationG, fe

However when I try to combine both time and place fixed effects

xtset Country_n year
xtreg BNPCG Over65 Under15 Trade LifeExpG EducationG i.year,fe

My significance shoots from 1% to 96%

Is this accurate or am I missing something?

[Carlo Lazzaro]

Wilhelm:
I do not understand what you mean by the following statement

My significance shoots from 1% to 96%

.
That said, you two-way fixed effect model shows a decent/good R-sq within, but, with such a long T, you should use cluster-robust standard errors.
In addition, I would also test wether quadratic terms (see -fvvarlist- notation) for -LifeExpG- and/or -EdicationG- make sense in your regression.