st: Question on Heteroscedasticity Problems with an IV Tobit model

[Maarten Buis <maartenlbuis@gmail.com>]

--- Nicholas Dadzie wrote me privately
> My aim is to estimate the effect of a variable X on Y which
> is censored. From other empirical work read, the suggestion
> is to use Tobit's model and the IV Tobit if there is an issue of
> endogeneity. The IV Tobit is my preferred model , using
> Newey's two-step minimum chi-square estimator, but further
> reading shows that  Tobit models are susceptible to
> homoscedasticity assumptions.
>
> My question is that , after dealing with the endogeneity in the X
> variable (or in the IV Tobit framework) how do I test or check for
> heteroscedasticity and then correct for it?
>
> I read about the CLAD estimator (Censored least absolute
> deviations model), but this fails to converge for my model.
> Is there a way that i can estimate the IV Tobit model and have
> heteroscedastic robust standard errors for inference?

Such questions need to be sent to Statalist not privately to its
individual members. For the reasons why this is so see:
<http://www.stata.com/support/faqs/resources/statalist-faq/#private>

As far as I know the answer to your question is no. The
(homoscedastic) variance feeds into the mean equation of the Tobit
model, so you cannot neatly separate the two. But there are others on
this list who are much more knowledgeable on this issue.

-- Maarten

---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany

http://www.maartenbuis.nl
---------------------------------
*
*   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/