Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: -gmm- and -outreg2-


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: -gmm- and -outreg2-
Date   Fri, 17 Aug 2012 11:21:15 -0400

Joseph Monte <hmjc66@gmail.com>
The log of E(y|X) is not the same as E(log y|X). -ivpois- assumes
E(y|X)=exp(Xb) or log E(y|X) = Xb
but your -ivreg2- call assumes
E(log Y|X)=Xb
which makes no sense when Y can take on the value zero,
and generically gives different estimates.
See also
http://blog.stata.com/2011/08/22/use-poisson-rather-than-regress-tell-a-friend/
http://www.stata.com/meeting/boston10/boston10_nichols.pdf

-ivpois- is only obsolete because you can use official -gmm- to get
the same answer; -ivpois- can still be used if you prefer the output.
If it is helpful to people, I could modify -ivpois- to use -gmm- and
return estimates in a friendlier way when run on version 11 or higher.

On Fri, Aug 17, 2012 at 8:42 AM, Joseph Monte <hmjc66@gmail.com> wrote:
> Thanks, Scott. I need to read up on -gmm- as my question (resulting
> from ignorance) pertained more to why the output using -gmm- and
> -ivreg2- were different rather than the horizontal display of the
> output using -gmm-.
>
> Joe
>
> On Wed, Aug 15, 2012 at 6:38 PM, Scott Merryman
> <scott.merryman@gmail.com> wrote:
>> Perhaps I am not following, but in the first example -outreg2- (-ssc
>> desc outreg2-)  is outputting the results from -gmm- and the in second
>> example it is outputting the results from -ivreg2- (-ssc desc
>> ivreg2-).  Given that there are 17 equations in the gmm estimated
>> model and -outreg2-, by default, outputs each equation in a separate
>> column makes the display different than the output from -ivreg2-.
>> Perhaps you want -onecol- option on -outreg2-?
>>
>> Scott
>>
>>
>> On Wed, Aug 15, 2012 at 9:54 AM, Joseph Monte <hmjc66@gmail.com> wrote:
>>> Richard,
>>>
>>> Here is an example:-
>>>
>>> use http://fmwww.bc.edu/ec-p/data/wooldridge/card, clear
>>> loc x "exper* smsa* south mar black reg662-reg669"
>>> ivpois wage `x', endog(educ) exog(nearc2 nearc4)
>>> gmm (wage*exp(-{xb:`x' educ}-{b0})-1), inst(`x' nearc2 nearc4)
>>> outreg2 using myfile, replace excel
>>>
>>> Is this how it is supposed to display? This output display is
>>> different from the one below.
>>>
>>> use http://fmwww.bc.edu/ec-p/data/wooldridge/card, clear
>>> loc x "exper* smsa* south mar black reg662-reg669"
>>> ivpois wage `x', endog(educ) exog(nearc2 nearc4)
>>> gmm (wage*exp(-{xb:`x' educ}-{b0})-1), inst(`x' nearc2 nearc4)
>>> estat overid
>>> ivreg2 lw `x' (educ=nearc2 nearc4)
>>> outreg2 using myfile, replace excel
>>>
>>> Thanks,
>>>
>>> Joe
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index