Switching from panel regression to "fama Macbeth" 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 would like  to run the same regression in  "fama Macbeth" method.

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

Thanks,

Lior

[ets_kps]

there probably are more elegant ways to do this but here is a quick find on the web.  Have fun. http://kelley.iu.edu/nstoffma/fe.html

[lior]

Hi,

Thanks for the link.

Actually I've already looked at this link before  but as I don't understand the code its hard for me to use it...

could you please  transform the following regression:

proc panel data=invest.table_regressionfinal ;

id GVKEY DATADATE;

model r_tmw=eps_deflate deltaeps_deflate/fixone ;

run;

into the "fama Macbeth" code (which include Newey-West adjustment for standard errors)

which is presented at the link as:

proc sort data=pe; by variable; run;

%let lags=3;

ods output parameterestimates=nw;

ods listing close;

proc model data=pe;

by variable;

instruments / intonly;

estimate=a;

fit estimate / gmm kernel=(bart,%eval(&lags+1),0) vardef=n; run;

quit;

ods listing;

proc print data=nw; id variable;

var estimate--df; format estimate stderr 7.4;

run;

More over, I havn't found in the web the implication of this method, meaning what is the different in using

"fama Macbeth" method comper to a regular regression?

Thanks,

Lior