Re: st: Exporting tables of output from non-ereturn commands using outreg2

[John Luke Gallup <jlgallup@pdx.edu>]

Alex,

I think you want to make a table out of the test statistics and critical values of the -dfgls- test, which are stored in r(results) after the command runs.  

It is possible to trick -outreg2- or -outreg- to display tables like this if you can post the results with -ereturn post- as if they were estimation results.  However you need both a "b" vector _and_ a "V" matrix.  You could convert the r(results) matrix into a rowvector "b", and create a dummy variance matrix "V", make the row and column names of your "V" matrix the same as the column names of "b" and post it with -ereturn post b V-.  However this is hard to work with because the results are now in a single row, and you need to suppress reporting the fake V results.

An easier solution is to use the command -frmttable-, distributed with the -outreg- command, which takes a Stata matrix and generates a table, as in the following code:

webuse lutkepohl2
dfgls ln_inv
frmttable using dfgls, statmat(r(results))

You can include the other -dfgls- r() statistics in the -frmttable- table with the -addrows- option.

John


On Jul 19, 2012, at 7:12 AM, Alex Letko wrote:

> I was wondering if there is an easy way to use outreg2 (or another
> results-exporting program), to export results of the dfgls command.  I'm a
> relatively new user, but I'll do my best to explain.
> 
> From the manual for outreg2: "outreg2 can report any regression output,
> provided it conforms to the usual convention for ereturn list."
> 
> I realize that dfgls does not save anything to the ereturn list, only to
> r().  I thought I was getting around this by using the eststo command to
> save the dfgls results to a file and then specify that file in the outreg2
> command, like so:
> 
> eststo: dfgls lnHH, maxlag(30)
> outreg2 [est5] using results1.rft, stats(tstat) replace
> 
> I get a table that gives me the DF stat for the 1st lag, but I need to show
> all the lags.  The normal outreg2 specifications don't seem to work, since
> dfgls is not a normal regression (there is no way to specify giving me
> more lags).
> 
> Basically, all I want is a standard output table for my unit root tests,
> showing DF stats and the significance levels for each lag.  Should I not be
> using outreg2 for this? Is there an easier program? (I tried with estout
> but it does not recognize the stored file at all so I can't use that)
> 
> Thanks for any help,
> 
> Alex
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/