Country specific effects in a panel regression

[youno]

Dear moderators and others,

I am conducting my master thesis about the determinants of reserve currencies. I have 5 countries (currencies) and several variables. I stacked the variables in a way that each variable contains the information of the 5 countries.

Inflation country A
inflation country B
...
GDP country A
GDP country B

How do I eliminate the country specific effects. My supervisor told me to use 'fixed' cross section effects. The results make no sense. When I use 'fixed' time period effect, the results make sense and look more or less the same as when I use no effects at all. What is the correct way to eliminate the specific effects if data is stacked like this?

Thank you very much in advance!

Mark