I am currently running FE vs RE via a Hausman test for panel data.


I am then taking 3 subsamples of this dataset: e.g. Large banks, medium sized banks and small banks just for reference/ robust
Dependent Variable: Liquidity Ratio
Independent Variable: Size, nim, roa, loan loss reserve etc....






I have found that the full dataset, large banks and small banks successfully reject the null hypotheses and lean towards the FE test model outcome.

However, the medium sized banks fails to reject the Hausman test null.


1. Should I be doing subsample Hausman tests in the first place?
2. Is this an issue?

Many Thanks,

V