Dear Stata-list users,
I'm running a switching regression model with a
structural selection equation. The model is as
follows:
d* = a(y1 - y2) + xb + e (1)
d = 1 if d* > 0
y1 = z1a1 + e1 (2)
y2 = z2a2 + e2 (3)
For illustrative purpose, suppose y's are income for
migrants vs non-migrants.
To estimate the model, (2) and (3) are substituted
into (1) to give a reduced-form selection equation.
Then the model is estimated using Heckman's 2-step
procedure. Finally, predicted values of y1 and y2 are
plugged back into (1) to obtain estimates of the
parameters of the structural selection equation.
Here's my question: In the final step, since predicted
y's are used, do I need to account for this in
calculating the standard errors of parameter
estimates, as I must do while estimating the income
equations? (I've seen papers that adjusted for the
income equations but not the structural selection
equation.) If yes, how shall I do it? By, say, the
Murphy-Topel procedure?
Hope I've made myself clear. Looking forward to your
advice. Thx.
Wei Wang
Dept. of Economics
University of Pittsburgh
____________________________________________________________________________________
The fish are biting.
Get more visitors on your site using Yahoo! Search Marketing.
http://searchmarketing.yahoo.com/arp/sponsoredsearch_v2.php
*
* 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/
