Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: RE: binary quantile regression


From   "Steven Lehrer" <steven_lehrer@hotmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: RE: binary quantile regression
Date   Tue, 03 Oct 2006 21:36:41 +0000

Binary regression quantiles is a semiparametric discrete choice estimator that imposes fewer asumpotion on the underlying data than other semiparametric estimators (i.e Klein and Spady or Ichimura). The cost is a slower rate of convergence. Basically you can view Manski's maximum score estimator as a median estimator, and Kordas (2004) outlines the benefits of using Manski's (1975, 1985) binary regression quantiles to provide consistent estimates of the conditional probability of some outcome given an individual's covariates at different points of the distribution. While Limdep has a algorithm to do maximum score it appears to perform fairly poorly. In fact, if you plug the solution in the Limdep manual as a starting value in the algorithm available on Kordas's website (with the same data) you will find a new solution which does a better job at maximizing the objective function. In addition the S-Plus code can do other quantiles. The S-Plus code calls some Fortran subroutines and from my own experience is fairly easy to use. My hunch is that programming BRQ in Stata would not just be time consuming but the program would take alot longer to converge on a solution then the S-Plus code.

Hope this helps,
Steve


From: "Feiveson, Alan H. (JSC-SK311)" <alan.h.feiveson@nasa.gov>
Reply-To: statalist@hsphsun2.harvard.edu
To: <statalist@hsphsun2.harvard.edu>
Subject: st: RE: binary quantile regression
Date: Tue, 3 Oct 2006 09:44:43 -0500

I wonder if the reason there was no response is that it is not clear
what is meant by "binary quantile regression". If you have a binary
response variable, it only takes on two values and hence does not have
continuous quantiles that can be modeled via regression. Perhaps that is
not what was meant by "binary quantile regresion". Please clarify.

Al Feiveson

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of
Paul.Zorn@pc.gc.ca
Sent: Tuesday, October 03, 2006 7:50 AM
To: statalist@hsphsun2.harvard.edu
Subject: st: binary quantile regression

Below is a question posted Jan. 25, 2004 for which I can not find a
response. Does anyone have any information on binary quantile regression
in Stata? Thank you.

Paul Zorn.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I was wondering if any of you are doing work using binary quantile
regression in Stata. Here are two links to papers with examples on
credit scoring and labor force participation:
http://www.econ.upenn.edu/~kordas/research/creditscore.pdf
http://www.econ.upenn.edu/~kordas/research/sbrq.pdf
The author provides the code in S-Plus but I could not find anything on
this model in Stata.
Thanks, Jose Pagan.

*
*   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/

*
*   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/
_________________________________________________________________
Donít waste time standing in lineótry shopping online. Visit Sympatico / MSN Shopping today! http://shopping.sympatico.msn.ca

*
* 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/




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