Re: Re: st: comparing coefficients using suest (same sample, same DV)

[Christopher Baum <kit.baum@bc.edu>]

<>
On Aug 14, 2012, at 2:33 AM, Tirthankar wrote:

> sysuse auto, clear
> gmm (price - {xb: mpg foreign}-{b0}) /// regression 1
> 	(price - {xc: mpg foreign rep78}-{c0}), /// regression 2
> 	winitial(identity) ///
> 	instruments(1: mpg foreign) /// regressors exogenous
> 	instruments(2: mpg foreign rep78) ///
> 	wmatrix(unadjusted, independent) /// moment conditions independent
> 
> // test
> test [xb_mpg]_cons = [xc_mpg]_cons

This is indeed a more elegant way to do this, but note that it gets a bit messed up a bit because of the missing cases in rep78
that affect the second model but not the first. If applied to my prior example,

sysuse auto, clear
gmm (price - {xb: mpg}-{b0}) /// regression 1
	(price - {xc: mpg trunk}-{c0}), /// regression 2
	winitial(identity) ///
	instruments(1: mpg) /// regressors exogenous
	instruments(2: mpg trunk) ///
	wmatrix(unadjusted, independent) /// moment conditions independent

// test
test [xb_mpg]_cons = [xc_mpg]_cons

It yields roughly the same p-value as does my earlier posting, keeping in mind that GMM uses z-statistics rather than t-statistics.

Kit


Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                             An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.html
  An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html


*
*   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/