Hello!

I am having a problem trying to perform a two sample t test on the outcome of two regressions. It would be amazing if you could help me out.

I run a regression between portfolio returns and market returns with various event dummies (eg. d_e1_1, ..... d_e19_1; for events 1 - 19)
this regression I first run for "good" portfolio and after for "bad" portfolio, then I want to figure out if the coefficients for the events significantly differ for the two portfolio's.

The regression output looks the following:

Source | SS df MS Number of obs = 747
-------------+---------------------------------- F(20, 726) = 178.11
Model | .034618922 20 .001730946 Prob > F = 0.0000
Residual | .007055371 726 9.7181e-06 R-squared = 0.8307
-------------+---------------------------------- Adj R-squared = 0.8260
Total | .041674293 746 .000055864 Root MSE = .00312

-----------------------------------------------------------------------------------------
meanret | Coef. Std. Err. t P>|t| [95% Conf. Interval]
------------------------+----------------------------------------------------------------
msciworld_excess_return | .9031561 .0154274 58.54 0.000 .8728685 .9334437
d_e1_1 | -3.27e-06 .0018037 -0.00 0.999 -.0035443 .0035378
d_e2_1 | .0011861 .0024208 0.49 0.624 -.0035665 .0059387
d_e3_1 | -.0039503 .0022075 -1.79 0.074 -.0082842 .0003835
d_e4_1 | .0034561 .0018038 1.92 0.056 -.0000851 .0069973
d_e5_1 | -.0003422 .0018056 -0.19 0.850 -.0038871 .0032026
d_e6_1 | -.0023766 .0018057 -1.32 0.189 -.0059216 .0011684
d_e7_1 | -.0007311 .001807 -0.40 0.686 -.0042788 .0028165
d_e8_1 | -.0009108 .001804 -0.50 0.614 -.0044525 .0026308
d_e9_1 | .0012886 .0018046 0.71 0.475 -.0022542 .0048315
d_e10_1 | -.0023667 .0022078 -1.07 0.284 -.0067011 .0019678
d_e11_1 | -.0006082 .0018063 -0.34 0.736 -.0041545 .002938
d_e12_1 | .0032565 .0024225 1.34 0.179 -.0014994 .0080124
d_e13_1 | .0060298 .0018061 3.34 0.001 .0024839 .0095757
d_e14_1 | .0005106 .0020484 0.25 0.803 -.0035109 .0045321
d_e15_1 | -.0020765 .0029162 -0.71 0.477 -.0078017 .0036487
d_e16_1 | .0048814 .0029402 1.66 0.097 -.000891 .0106538
d_e17_1 | -.0006468 .0022051 -0.29 0.769 -.004976 .0036824
d_e18_1 | .0023479 .0019477 1.21 0.228 -.0014759 .0061716
d_e19_1 | .0017832 .0018038 0.99 0.323 -.0017581 .0053245
_cons | .0001487 .0001179 1.26 0.208 -.0000827 .0003801
-----------------------------------------------------------------------------------------



How can i store this information for two regressions (good and bad) and then retrieve the data to perform a two sample t test


A second issue, I also would like to test the differences between the sum of all the event coefficients (and the standard errors) in the sample.

A third concern, I would like to be able to "play" around with the events in the sample; eg. now there might be 19 events, but it should also work if i try to include less events (eg. only 6).

What would be the best way to do this? I would really appreciate the help!