How to export the results of multiple linear regressions into Excel by sorted group?

[Jessica Wen]

I'm doing mutiple linear regressions, regressing daily stock returns on three factors. I have 1680 stocks, and for each stock, I need to conduct 57 monthly regressions. I first use the "sort" command to sort daily excess returns into month groups and then run regressions. However, when exporting the results to excel, I can only get the results for all stocks in 1 month instead of all 57 months.
How can I export the results for all stocks in all 57 months at one time? I think the problem is with the outreg2 command~
I\ve pasted part of the data below. m, s, v are the 3 independent variables, and y1 represents the daily excess return of stock i. The leftmost column is the month each excess return belongs to. I really need help! Thanks a lot!
My current commands are as follows:
forvalues i=1(1)1680{
sort B, stable
by B:reg y`i' m s v
outreg2 using 1.xls
}

	m	s	v	y1	y2	y3	y4
2010-04	0.014403	-0.00101	-0.00035	0.008555	0.024145	0.00345	0.018391
2010-04	0.004004	-0.00221	-0.00269	0.001217	0.004046	-0.00619	-6.5E-05
2010-04	0.000375	0.004833	-0.00419	-0.00263	-0.01747	0.008745	-0.01324
2010-04	-0.01927	-0.00373	0.006913	0.001424	-0.00264	-0.01364	-0.00813
2010-04	-0.00967	-0.01977	0.002564	0.01927	0.005089	-0.04961	-0.02649
2010-05	-0.01115	0.00139	-0.01732	-0.0273	-0.04494	-0.04736	-0.03973
2010-05	0.015091	0.007471	-0.00742	-0.00507	0.005303	0.033369	0.036891
2010-05	-0.03947	0.004469	-0.01851	-0.07544	-0.04145	-0.02163	-0.04409
2010-05	-0.01899	0.001662	-0.01243	-0.00767	0.005505	-0.00207	-0.02419
2010-05	-0.00292	-0.00442	0.016185	0.009244	0.011015	-0.01011	0.015665
2010-05	-0.02382	-0.0042	0.0026	-0.05215	-0.00692	-0.00311	-0.02219
2010-05	-0.00563	-0.01185	0.016632	0.036568	0.012348	-0.02957	0.013509
2010-05	0.024402	-0.00497	-0.00608	0.031409	0.004021	-6.6E-05	0.004398

[Christos Makridis]

Hmm, I use esttab instead of outreg, and will look into it further, but here's one sure way of doing it use the matrix commands.

forval i=.. {
quietly reg y s v m if id==`i'
matrix b_s`i'=_b[s] if id==`i'
matrix b_v`i'=_b[v] if id==`i'
matrix b_m`i'=_b[m] if id==`i'
}

and then you can put all those in a matrix however you'd like, such as

matrix coefs=(b_s1,b_s2,b_s3,... \ b_v1, b_v2,...)
then you can see the matrix with "matrix list coefs" and export it using outtable to any format, excel included.

[Oded Mcdossi]

You my also find the statsby command very useful. Type help statsby for more details and follow the examples.