outreg2 issue with Stata version 15.1?

[Jeff Grogger]

Last week I ran a program with the line below and it executed as expected:

outreg2 using $opath/variance_decomp_schooling_xsection79_28only.txt , long noast replace ///
   adds(nhh, `nhh',  n, `n',  RHO, `RHO' ,seRHO, `seRHO' )  ctitle(white, xsection obs)

Today I updated to Stata version 15.1. When I ran the program, I got this:

. outreg2 using $opath/variance_decomp_schooling_xsection79_28only.txt , long noast replace ///
>    adds(nhh, `nhh',  n, `n',  RHO, `RHO' ,seRHO, `seRHO' )  ctitle(white, xsection obs)
invalid syntax
r(198);

Some experimentation revealed that the problem lies with the "long" option, since deleting it eliminated the error:

. outreg2 using $opath/variance_decomp_schooling_xsection79_28only.txt ,  noast replace ///
>    adds(nhh, `nhh',  n, `n',  RHO, `RHO' ,seRHO, `seRHO' )  ctitle(white, xsection obs)
dir : seeout

The problem is that, without the "long" option, the output becomes very difficult to read (it's from the "mixed" command). I tried to update outreg2, but got a message that it was already up to date. I realize that outreg2 is externally supported, but I was hoping someone else might have come across this (a search for "outreg2" in the forum didn't turn up anything relevant). Short of that, if someone could tell me how to contact the relevant people at ssc, that would be helpful.

[Nick Cox]

The only person to approach is the program author, Roy Wada, who is not a member here.

The "people at SSC" is just one person, Kit Baum, who doesn't maintain the code for any programs but his own.

Roy hasn't updated this program since 2014. He alone can speak for whatever is ongoing or planned.

[Jeff Grogger]

Thanks for the quick response. Would be a shame to lose that functionality...

[Nick Cox]

I would check out programs that are being maintained.

[William Lisowski]

Just to point out what you may already have discovered, at the bottom of the help outreg2 output the Author section gives an email address for Roy Wada, as does the output from search outreg2.

[Jeff Grogger]

Thanks. No excuse but to say that it's a long way from the top of the documentation to the bottom!

[Victor Jarosiewicz]

Problem with outreg2 and Stata 15:

I just moved universities and have Stata 15.1/MP at the new one and still have access to Stata 15.0/SE at the old one. I just discovered that the following code works just fine inside a do-file on the old one (Stata 15.0/SE) but not the new one (Stata 15.1/MP):

reg investment cashflow lagtobinsq laglnassets ceoid* cfoid* otherid* i.year i.gvkey, noomitted vsquish noemptycells vce(cluster gvkey)
gen inv_sample=e(sample)

/* why does outreg2 crash with this data and this regression and do-file but works if entered manually??? */

outreg2 using "`outregdir'/`outregtable'", ///
keep(*ceoid* *cfoid* *otherid*) auto(4) bdec(3) sdec(3) ///
ctitle("INVESTMENT W/ CEO_CFO_OTHER") nocons nose ///
adds("Adj R2 ", e(r2_a), "Firms", e(N_clust)) ///
addnote("Std. Err. adjusted for `e(N_clust)' clusters in gvkey", ///
"Run at $S_TIME $S_DATE", "Using data from ","$S_FN", ///
"Using do-file: ", "`dofilepath'/`dofilename'" ) ///
excel replace

What is even more puzzling is that while the code crashes when executing as a do-file (no error code, it just shuts down the Stata application completely), if I only run the top two lines (the regression and the gen line) as a do-file, but then input the outreg2 command (either defining the local macros in the command line and then running the outreg2 command, or substituting the file names manually), then it also runs just fine on either 15.0/SE or 15.1/MP platform.

I have a lot of regressions to run so I'm trying to run everything via do-files. I lose access to the prior institution's Stata 15.0/SE at the end of the month, and need to be able to rerun these in the future.

I have noted the comment above that outreg2 is an ado file not maintained by Stata, so I'm sending this message also to Roy Wada next, but the fact that the same command runs on 15.0 but not 15.1, and runs outside of a do file but not inside of one makes me also wonder if that's a platform glitch as well as something in the ado file.

This is my first post (long time listener, first time caller, as the old saying goes) so I apologize for any info you find extraneous.

[Hua Peng (StataCorp)]

Victor Jarosiewicz Please make sure your State is fully updated by

update all

If problem persists, please email tech support [email protected] with the dataset and the do-file which cause the crash.

[Victor Jarosiewicz]

That did it! That was too easy - I just got this license for Stata days ago so I didn't think to update it again. Mea culpa