Event study - Beta prediction

[Barbara Heerkens]

Hi all,

I'm a bit stuck with predicting the beta's of all my firms in the sample with the index.

I entered the following (I did not create a group-id since the company_id was already in the form of _j for all returns belonging to the same firm)
gen predicted_return=.
forvalues i=1/78{
l _j if _j==`i' & dif==0
reg returns indexreturn if _j==`i' & estimation_window==1
predict p if _j==`i'
replace predicted_return = p if _j==`i' & event_window==1
drop p
}

Now I got a really nice outcome, except for the Beta of the first company, _j=1.
The Beta is, according to the outcome, exactly 1, which is probably not the case.

| _j |
|----|
13892. | 1 |
+----+

Source | SS df MS Number of obs = 128
-------------+---------------------------------- F(1, 126) = .
Model | .011142044 1 .011142044 Prob > F = .
Residual | 0 126 0 R-squared = 1.0000
-------------+---------------------------------- Adj R-squared = 1.0000
Total | .011142044 127 .000087733 Root MSE = 0

------------------------------------------------------------------------------
returns | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
indexreturn | 1 . . . . .
_cons | 0 (omitted)
------------------------------------------------------------------------------
(option xb assumed; fitted values)
(16,016 missing values generated)
(40 real changes made)


Does anyone know how I fix this?

Thank you!!