-cmp- from ssc appears to be the solution for all "unusual" 2SLS problems. Unfortunately, it seems not to work with count nor panel data. Also, behavior of -qvf- (from SJ) is unknown to me (but I am sure it doesn't work with panels). The only Stata commands for sample selection and treatment effects which I know having a count Y1 don't work with panels and require a dummy or multinomial Y2.
Turning to the literature, I don't remember if Mullahy (1997) works with panels or cross-sections (it is on the Review of Economics and Statistics, sorry for not providing full reference).
Sorry for providing negative news,
Nicola
At 02.33 27/10/2007 -0400, "Nguyen Cong Minh" wrote:
>Hi,
>
>I am working on a model in panel data context
>
>Y1(i,t) = alpha + b1*Y2(i,t) + b2*Z(i,t) + c(i) + error(i,t)
>
>where
>- - Y1 is a count data (0, 1, 2...) or fractional data (0 <= Y1 <= 1)
>- - Y2 is a count endogenous explanatory variable
>- - Z are other exogenous variables
>- - c is unobserved individual effect
>- - error is disturbance term, it could be normal or not.
>
>I have been looking in the literature to find the methods but found nothing
>yet. Does anyone know something related to this model from literature or
>Stata? The only paper I know is Terza (1998), Journal of Econometrics. Is
>there any Stata package for this kind of model yet? Please let me know if
>you know anything related to this problem.
>
>Ref:
>
>Terza, 1998, Estimating count data models with endogenous switching: sample
>selection and endogenous treatment effects. Journal of Econometrics, 84,
>129-154.
>
>Thank so much,
>
>Minh Nguyen
>PhD student, Economics
*
* 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/
