Thanks to Kit Baum, outreg2 is available from SSC.
outreg2 is based on the original outreg by John John Luke Gallup.
I have replaced the internal logic, making it compatible with any regression
output within Stata. This claim should hold true, provided the regression
output conforms to the usual convention for the eclass returns. (xtmixed,
for example, deviates from the established convention by returning the
number of group, e(N_g), as a 1x1 matrix, thus making it incompatible with
the original outreg. estimates_table cannot report it, either. You get a
blank. outreg2 contains a work-around for it).
I have written outreg2 with the view that the formation of regression table
is an iterative part of research process. Many regression tables are formed
that do not see the light of publication. The formation of regression table
should be fast and easy, thus lowering the hurdle for its use (this point
has been mentioned in different forms by Kit B and Steve S). Automation that
requires too much fiddling is no automation at all.
Among other things, outreg2 will automatically format the number of reported
digits and will accept the saved estimates, if you must have one. Two
companion commands, seeout and shellout, provide immediate access to the
formed tables (shellout works only for Windows XP/NT). outreg2 will convert
the tab-delimited ASCII table into LaTeX (per Kit B's help), MS Word, and MS
The codes are rather long and it is likely that some errors exist. They
should be reported to me in private, unless it is something that is better
discussed on the Statalist.