difference in coefficient

[chourouk bouguerra]

My model to regress :

R_(t+1)=α₁ RD_t+ α₂ DD_t+ u_(t+1)

with RD_t DD_t are dummy variables (if DD_t =1, RD_t= 0 and if DD_t =0, RD_t= 1)

H₀ : α₁ - α₂= 0
I regress this model and I obtained α_{1 and} α₂

My question is how can I test the signification of the difference of α_{1 and} α₂
I mean like the column ''diff'' in the picture.

[Clyde Schechter]

-help lincom-

[chourouk bouguerra]

thank you it helped me a lot.
Another question if you dont mind , for the p-value obtained from the newey west 1987 , how can i find it in stata , i just add Vce (robust) to my reg command,
but i find p value too huge compare to this table.

reg VWRTBL α1 α2, noconstant vce(robust)
lincom α1- α2
( 1) α1 - α2 = 0

[Alfonso Sánchez-Peñalver]

Several comments.

You don't describe your model specification, and/or the nature of your data. It is going to be very hard for anyone to figure out what the issue is unless you give much more detail. And even then, there may not even be an issue, but we cannot figure that out either.

Second, if you cite any work, please provide the reference.

Third, and last, vce(robust) in regress is only robust to heteroskedasticity, not to correlation. The table that is shown in the pic in #1 seems to deal with panel data. Are your data in panels? Because if so, you should consider using xtreg, and there

vce(robust) is also robust to correlation within the panels, because it is equivalent to vce(cluster panelvar). Of course this relates to the first point.

[Rich Goldstein]

see

help newey

[chourouk bouguerra]

Dear Alfonso Sanchez-Penalver,
I describe my modele in my first post and II mentioned that independent variables are dummy variables, and I wrote my regression model.
My dependent variable the returns of CRSP from 1972 to 2018 .
I apologize if I didn't specify the source correctly, but yesterday was my first day of using stata and it was my first publication here.
the reference of this work is, the presedential puzzle political cycles and stock market santa-clara and valkanov 2003.
Thank you.