Problem of classical assumption test
Posted: Sun Oct 30, 2016 7:06 am
by bayu_sutikno
Sir, permit me to asking you about problem Problem of classical assumption test. My panel data has both heteroscedasticity and autocorrelation problem, what's method that I can remove them at once ??
thanks a lot Sir for your help and goodness.
Re: Problem of classical assumption test
Posted: Sun Oct 30, 2016 4:25 pm
by jmagomez
Dear Bayu_sutikno,
There is no magic method to solve these problems.
You will have to work over autocorrelation or heteroscedasticity and, probably, will improve the other one.
You can add lags of the dependent variable, trying to improve autocorrelation.
You can transform variables, using log transformation, trying to improve heteroscedasticity.
You will have to work a little bit.
Regards.