I am estimating the following:
proc panel data = data;
id gvkey rankyear;
model y = cera_pl_vol cera_mi_vol revi_pl_vol revi_mi_vol
/ranone noint hccme = 3 cluster;
run;
quit;
I want to check the Hausman-Statistic to infer wether random effects are appropriate or not. I estimate this for a set of different panels.
In some cases I get a m-value. In some cases I do not get a m-value. Why is that so? How can I appropriately check wether random effects are appropriate (random effects for gvkey)?
Examples are attached below: