Page 1 of 1

Fixed Effect Model and weights

Posted: Sat Mar 09, 2019 4:54 am
by tifrei
Hello,

Im working on a research paper and have run into a problem.
My data contains 1556 observations from 120 different companies from 2004S1 to 2018 S2.

In order to evaluate which regression to use for my data i run the Hausman test which suggested a Fixed Effect Models should be used.
After that i ran the wald test to check for hetroscedasticity. The data shows hetroscedasticity.

Therefore i figured out that i should use fixed effect and cross-section weights. But unfortunalety the following error appears : "Positive or non-negative argument to function expected in computation of group weight (variance). " and i dont know what to do?

Any suggestions or help?