Re: st: RE: Cumulative Distribution Function

[Alison Wong <aliwong@ucalgary.ca>]

Hi:
Here is more information about my problem.  In a block sense, I am 
trying to calculate the Inverse Mill Ratio (IMR) to be use to calculate 
some form of generalized residual.  I read this method from Francis 
Vella's article:

"A Simple Estimator for Simultaneous Models with Censored Regressors", 
International Economic Review, Volume 34, Issue 2 (May, 1993), 441-457. 
In his article he suggested if you have a censored endogenous variable 
in a simultaneous equation system, than you could use a two step 
procedure (similar to the one Heckman presents) to get consistent 
estimates.  The first step is to regress the reduced form of the 
censored endogenous variable using tobit.  By that you should get 
estimates of coefficients (beta) and standard errors (sigma), then you 
could use these estimates to calculate the generalized residuals. The 
problem is in his generalized residual equation, he has the cumulative 
distribution function (cdf) and probability density function (pdf) in it 
and presented in the form similar to the Inverse Mill Ratio. This is the 
part that I can't figure out how to calculate, because I can't find 
anything in Stata that would calculate the Inverse Mill Ratio, of 
course, it calculate the IMR in the two step Heckman command but I don't 
want to use the Heckman command, so is there any other ways to obtain 
the IMR?

I hope I provide enough information this time.  If anyone is interested 
in this paper pls. let me know and I could send you a copy of this 
paper.  I would really appreciate if anyone who had read this article to 
tell me how to calculate the generalized residuals using Stata. 
Thank you!

Alison


FEIVESON, ALAN H. (AL) (JSC-SD) (NASA) wrote:


> Alison - 
> I'm afraid your question is not specific enough. What model are you running
> to get estimates of your coefficients? What random variable do you want to
> get the cdf and/or pdf for? What distribution does it have under the assumed
> model? Do you want to evaluate this cdf/pdf as a function prescribed by your
> model over a range of values, or do you want to compute an empirical cdf or
> pdf from data?
>
>
> Al Feiveson
>
>
> -----Original Message-----
> From: Alison Wong [mailto:aliwong@ucalgary.ca]
> Sent: Thursday, September 12, 2002 12:01 PM
> To: statalist@hsphsun2.harvard.edu
> Subject: st: Cumulative Distribution Function
>
>
> Hi:
>
> I have a problem of getting the cumulative distribution function (cdf) 
> and probability density function (pdf).  In particular I need to get the 
> cdf and pdf evaluated at the estimated coefficients (beta) and standard 
> errors (sigma).  I realize there are the "cumul" & "kdensity" command in 
> stata, but those only allows a single variable, but I have more than one 
> variable.
>
> Does anyone have a better idea of how to get cdf or pdf?
>
> Thanks.
>
> Alison Wong
>
> *
> *   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/

*
*   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/