panel regression versus pooled regression

[lior]

Hello,

I run the following panel regression:

proc panel data=invest.table_regressionfinal ;

id GVKEY DATADATE;

   model r_tmw=eps_deflate deltaeps_deflate/fixone ;

run;

now I want to run the same regression in "pooled" method.

Anyone happen to know how  to formulate the pooled procedure and what are the implications of  switching from panel regression to pooled regression ?

thanks,

Lior

[ets_kps · Accepted solution]

Hi Lior,

This link explains the "pooled" regression model.SAS/ETS(R) 13.2 User's Guide

In short, pooled regression is just another way of saying that you are assuming no id or time fixed effects.  In essence, you are "pooling" the data together and running one big regression and assuming each observation, wrt both time and id are independent.  Probably not a realistic assumption.

For your model, you would write...

proc panel data=invest.table_regressionfinal ;

id GVKEY DATADATE;

   model r_tmw=eps_deflate deltaeps_deflate/ pooled neweywest ;

run;

You likely would want to do some sort of SE correction using NEWEYWEST or similar on the model statement.

Hope this helps. -Ken

[lior]

Hi Ken,

Thank you very much. it is very helpfull.

Do you also farmiliar with fama macbeth regression? (meaning how to formulate the procesdure and what are the implications of switching to this regression)?

Thanks,

Lior