Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

st: RE: how to use margins to evaulate an estimated function at two different sets of values for covariates and obtain the standard error of a function of the difference between the two


From   Murat Genc <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   st: RE: how to use margins to evaulate an estimated function at two different sets of values for covariates and obtain the standard error of a function of the difference between the two
Date   Mon, 8 Jul 2013 09:03:33 +0000

Dear Statalisters,

I'm using Stata 12.  I have estimated a system of 24 equations. Let's say that the equations are denoted by s`i'.  I am interested in predicting and obtaining the standard errors of the proportionate change in  
	(s`i')*food_exp)/price_f`i'
when some covariates are changed (and where food_exp and price_f are some variables in the data set).  

I calculated the initial predictions by 
	forvalue i=1/24 {
	est restore surebootall
	margins,expression((xb(s`i')*food_exp)/price_f`i') noesample post
	est store initial`i'
	} 

and the new predictions by 
	forvalue i=1/24 {
	est restore surebootall
	margins,expression((xb(s`i')*food_exp)/price_f`i') noesample post
	est store new`i'
	}
Using the new values of the covariates.  

So, what I need is the standard error of  (new`i'- initial`i')/ initial`i'.  This is where I get stuck, because I cannot use two estimates on the same command (such as nlcom).  I thought of using a single margins command with at() option to specify the new values for the covariates, but at() option expects a number for the covariate.  That makes it impossible, because these values are all different for every observation in the sample and I cannot possible enter them one by one.  I tried creating scalars and using something like at(var1=scalar1), but that does not work since scalar1 is not a number.

Does anyone have a suggestion?  

Many thanks in advance.

Murat


                                          Dr M Genç 
Department of Economics                Research Fellow                
University of Otago            and     Institute for the Study of Labor (IZA)
P O Box 56                                 Bonn, Germany
Dunedin 9054, New Zealand 
_______________________________________________ 
         E-Mail: [email protected]    
         Phone: +64 3 479 8644                                  
         Fax: +64 3 479 8174                                   
FFFFFFF Web: <www.otago.ac.nz/economics>      




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index