Sumon has problems using -oaxaca- after -selmlog- (see
http://www.crest.fr/pageperso/lmi/gurgand/selmlog.htm).
The coefficients vector returned by the -selmlog-
command contains coefficients associated with variables
that do not exist in the data. -oaxaca- breaks while
trying to compute the means of these variables.
The solution to the problem is to add such variables
back to the data before executing -oaxaca-. The added
variables should contain all zeros (to ensure that
they have no effect on the decomposition results).
Example:
. sysuse auto, clear
. gen byte replow = rep78<4
. gen price2 = price if foreign==0
. reg price2 weight
. selmlog price2 weight if replow==1, sel(foreign = turn)
. est sto replow
. selmlog price2 weight if replow==0, sel(foreign = turn)
. est sto rephigh
. gen byte _m0=0
. gen byte _m1=0
. oaxaca replow rephigh
Hope this helps,
ben
> -----Original Message-----
> From: [email protected] [mailto:owner-
> [email protected]] On Behalf Of Sumon Bhaumik
> Sent: Saturday, January 07, 2006 11:47 AM
> To: [email protected]
> Subject: Decomposition of selmlog estimates
>
> Hi,
>
> The Oaxaca command can now be used to decompose differences between
the
> LHS varriable of two groups of economic agents following both OLS and
> Heckman selection models. But it still cannot be used for
decomposition if
> the selection model is of the Bourguignon et al type, i.e., if the
command
> if "selmlog" instead of "Heckman". Is there any way to use Oaxaca or a
> similar command after selmlog?
>
> Regards,
>
> Sumon
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
