Re: st: comparing coefficients accross models

[Traci Schlesinger <tracias@Princeton.EDU>]

Hi Partha!

yes, this is very helpful.  thanks for the suggestions.

best
traci

Partha Deb wrote:
> Traci,
>
> Your substantive question got my attention when you first posted, and 
> I'll admit I didn't read your post, nor subsequent helpful posts 
> carefully, so with due apologies to those who've made suggestions, 
> etc. I'll suggest that the -suest- type test you had in mind initially 
> seemed perfectly appropriate to me.
>
> You have 2 equations.
>
> y1 = b0 + b1 * x + u1
> y2 = a0 + a1 * x + u2
>
> You can think of these as a SUR, but if your regressors are the same 
> in both equations (as I've written it out), there is no advantage to 
> using an SUR procedure.  But, it's perfectly reasonable to test the 
> hypothesis: H0: b1=a1 using a Wald test.  That's what -suest- does.  
> Perhaps newey2 doesn't e(b) and V(b) from each model, which is what 
> you need for the test (and -suest-), but ivreg2 should.
>
> Hope this helps.
>
> Partha
*
*   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/