Re: st: compute standard errors with delta method of decomposed (in two parts) mfx from heteroskedastic probit model

[Maarten Buis <maartenlbuis@gmail.com>]

On Fri, Jan 13, 2012 at 3:32 PM, Luca Fumarco wrote:
> I am implementing a statistical decomposition to size the discrimination in the labour market, at a very early stage.
> For that, I am using a technique presented in Neumark (August 2011, page 19/20) -this is the most recent version of the working paper that I have found-.
> It is simply the decomposition of the marginal effect of ethnicity/race in the heteroskedastic probit model. Whilst it is simple to compute the two parts, I do not know how to compute the standard errors (he used the dealt-method)
>
> -> How can I compute the s.e. of those two parts?
> -> Do you know whether a command for the authomatic implementation of such decomposition already exists on Stata?

In order to answer that question we need to know what you did, which
is probably described in that working paper, but given that reference
it is impossible for us to find it, and thus it becomes very hard to
answer your question. Please see the Statalist FAQ (a link to it is at
the bottom of every post on Statalist) and comments from Cox (2011)
and Buis (2011)(*).

A general comment that may help is to do the computation using
-nlcom-, see -help nlcom-, as that will use the delta method for you.

-- Maarten

(*) A joke borrowed from Cox (2012).


--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------
*
*   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/