st: Marginal effects with outreg or estout using the new margins command

[Danielle Li <danielle-li@kellogg.northwestern.edu>]

Dear all,

I'd like to output marginal effects from a fixed effects regression.
I have tried 4 approaches involving the outreg2, estout, mfx, and
margins commands and none of them work (described below).

I would be really grateful if anyone has any ideas.  Ideally, I'd to
report both the coefficients and the marginal effects in the same
output table, but I am also happy to just output the marginal effects
and coefficients into separate tables and combine later.

Thanks so much,
Danielle



1.  Using margins with outreg2

areg y x [aweight=weight], cluster(is) absorb(is)
margins, eyex(x) atmeans
outreg2 x using filename, se nolabel bdec(3) rdec(3)  nocons excel
replace margin

This command runs without error but when I look at my outreg file,
there are no coefficients reported (e.g. the space where the result is
is just blank)

2.  Using mfx instead of margins

areg y x [aweight=weight], cluster(is) absorb(is)
mfx, eyex predict
outreg2 x using filename, se nolabel bdec(3) rdec(3) excel replace mfx

This fails because I get this error using mfx

mfx, eyex
default predict() is unsuitable for marginal-effect calculation
r(119)

Note: any predict option I explicitly try results in the same error.

3.  Using estout with margins

eststo raw: areg y x [aweight=weight], cluster(is) absorb(is)
eststo mfx: margins, eyex(x) atmeans
estout raw mfx using filename, replace

This runs without error, but eststo raw and eststo mfx end up saving
the same output even though, when I run the margins, command, stata
successfully reports elasticities that are different from the
coefficients in the areg step. So I get two columns with the same
information.

4.  All of the above with xtreg instead of areg. I get the same error
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/