Current Account Estimation Technique in Panel Daraset

[Naizen]

Hi everybody,

I am going to estimate Current Account Balance over GDP througout a panel dataset including 12 European Countries on a 10 years long time series from 2002 to 2011. If I got it well, data should be particularly problematic since time series are autocorrelated, there should be cross-country interferences and eteroskedasticity either. Some of the determinants purposed by literature seem to be correlated each other either
[population growth, dependency ratio (inactive population over active population), gdp, gdp growth, investment ratio, long term interest rates, initial net foreign asset, oil balance, fiscal balance and real effective exchange rates]

So my question is: what would you use to deal with all of these problems at a time?

(I wish not to influence anyone with this but I was thinking of using a Prais-Winstein regression (PCSE) or a FGLS with AR(1) disturbance what do you think?)


Thank you very much for help